Rehabilitación de la baja visión para mejorar la calidad de vida de adultos con problemas de visión
Resumen
Antecedentes
La rehabilitación de la baja visión tiene por objeto optimizar el uso de la visión residual después de una pérdida grave de la visión, pero también tiene por objeto enseñar habilidades para mejorar el funcionamiento visual en la vida cotidiana. Otros objetivos son ayudar a las personas a adaptarse a la pérdida permanente de la visión y mejorar el funcionamiento psicosocial. Estas habilidades promueven la independencia y la participación activa en la sociedad. La rehabilitación de la baja visión debería mejorar la calidad de vida (CdV) de las personas con deterioro visual.
Objetivos
Evaluar la efectividad de las intervenciones de rehabilitación de la baja visión en cuanto a la calidad (CdV) de vida relacionada con la salud (CdVRS), la calidad de vida relacionada con la visión (CdVRV) o el funcionamiento visual y otros resultados estrechamente relacionados con los pacientes en adultos con deterioro de la visión.
Métodos de búsqueda
Se realizaron búsquedas en las bases de datos electrónicas y los registros de ensayos pertinentes hasta el 18 de septiembre de 2019.
Criterios de selección
Se incluyeron ensayos controlados aleatorizados (ECA) que investigaban la CdVRS, la CdVRV y los resultados relacionados de los adultos con una deficiencia visual irreversible (criterios de la Organización Mundial de la Salud). Se incluyeron estudios que compararon las intervenciones de rehabilitación con el control activo o inactivo.
Obtención y análisis de los datos
Se utilizaron los métodos estándar previstos por Cochrane. La certeza de la evidencia se evaluó mediante los criterios GRADE.
Resultados principales
Se incluyeron 44 estudios (73 informes) realizados en América del Norte, Australia, Europa y Asia. Teniendo en cuenta la diversidad clínica de las intervenciones de rehabilitación de la baja visión, los estudios se clasificaron en cuatro grupos de tipos de intervención relacionados (y por comparador): (1) terapias psicológicas y/o programas de grupo, (2) métodos para mejorar la visión, (3) programas de rehabilitación multidisciplinar, (4) otros programas. Los comparadores fueron la no atención o la lista de espera como grupo de control inactivo, la atención habitual u otro grupo de control activo. Los participantes incluidos en los estudios informados eran principalmente adultos mayores con discapacidad visual o ceguera, a menudo como resultado de la degeneración macular relacionada con la edad (DMS). Los lugares de estudio solían ser hospitales o servicios de rehabilitación de baja visión. En la mayoría de los estudios se midieron los efectos a corto plazo (seis meses o menos). No todos los estudios informaron sobre la financiación, pero los que lo hicieron fueron apoyados por financiadores públicos o sin fines de lucro (N = 31), excepto dos estudios.
En comparación con los comparadores inactivos, se encontró evidencia de muy baja certeza de que no hay efectos beneficiosos sobre la CdVRS que se estimaron de manera imprecisa para las terapias psicológicas y/o los programas grupales (DME 0,26; IC del 95%: ‐0,28 a 0,80; participantes = 183; estudios = 1) y una estimación imprecisa que sugiere poco o ningún efecto de los programas de rehabilitación multidisciplinaria (DME ‐0,08; IC del 95%: ‐0,37 a 0,21; participantes = 183; estudios = 2; I2 = 0%); no hubo datos disponibles para los métodos para mejorar la visión u otros programas. En cuanto a la CdVRV, se encontró evidencia de baja o muy baja certeza del beneficio impreciso estimado con terapias psicológicas y/o programas grupales (DME ‐0,23; IC del 95%: ‐0,53 a 0,08; estudios = 2; I2 = 24%) y métodos para mejorar la visión (DME ‐0,19; IC del 95%: ‐0,54 a 0,15; participantes = 262; estudios = 5; I2 = 34%). Dos estudios que utilizaron programas de rehabilitación multidisciplinar mostraron resultados beneficiosos pero inconsistentes, de los cuales uno de los estudios, que tenía un bajo riesgo de sesgo y utilizó rehabilitación intensiva, registró un efecto muy grande y significativo (DME: ‐1,64; IC del 95%: ‐2,05 a ‐1,24), y el otro un efecto pequeño e incierto (DME ‐0,42; 95%: ‐0,90 a 0,07).
En comparación con los comparadores activos, se encontró evidencia de muy baja certeza de efectos beneficiosos pequeños o nulos sobre la CdVRS que se estimaron de manera imprecisa con terapias psicológicas y/o programas grupales, sin incluir ninguna diferencia (DME ‐0,09; IC del 95%: ‐0,39 a 0,20; participantes = 600; estudios = 4; I2 = 67%). También se encontró evidencia de muy baja certeza de efectos beneficiosos pequeños o nulos con los métodos para mejorar la visión, que se estimaron de manera imprecisa (DME ‐0,09; IC del 95%: ‐0,28 a 0,09; participantes = 443; estudios = 2; I2 = 0%) y programas de rehabilitación multidisciplinar (DME ‐0,10; IC del 95%: ‐0,31 a 0,12; participantes = 375; estudios = 2; I2 = 0%). Con respecto a la CdVRV, se encontró evidencia de baja certeza de efectos beneficiosos pequeños o nulos que se estimaron de manera imprecisa con terapias psicológicas y/o programas grupales (DME ‐0,11; IC del 95%: ‐0,24 a 0,01; participantes = 1245; estudios = 7; I2 = 19%) y evidencia de certeza moderada de efectos pequeños con métodos para mejorar la visión (DME ‐0,24; IC del 95%: ‐0,40 a ‐0,08; participantes = 660; estudios = 7; I2 = 16%). No se encontró ningún beneficio adicional con los programas de rehabilitación multidisciplinar (DME 0,01; IC del 95%: ‐0,18 a 0,20; participantes = 464; estudios = 3; I2 = 0%; evidencia de baja certeza).
Entre los resultados secundarios, se encontró evidencia de muy baja certeza de un beneficio significativo y grande, pero impreciso, sobre la autoeficacia o la autoestima para las terapias psicológicas y/o los programas grupales versus la lista de espera o la ausencia de atención (DME ‐0,85; IC del 95%: ‐1,48 a ‐0,22; participantes = 456; estudios = 5; I2 = 91%). Además, se encontró evidencia de muy baja certeza de un beneficio significativo y grande estimado sobre la depresión para las terapias psicológicas y/o programas de grupo versus lista de espera o sin atención (DME ‐1,23, IC del 95%: ‐2,18 a ‐0,28; participantes = 456; estudios = 5; I2 = 94%), y evidencia de certeza moderada de un beneficio pequeño versus la atención habitual (DME ‐0,14; IC del 95%: ‐0,25 a ‐0,04; participantes = 1334; estudios = 9; I2 = 0%). En los pocos estudios en los que se informaron eventos adversos (graves), éstos no parecían estar relacionados con la rehabilitación de la baja visión.
Conclusiones de los autores
En esta revisión Cochrane, no se encontró evidencia de beneficio de diversos tipos de intervenciones de rehabilitación de baja visión en la CdVRS. Se encontró evidencia de baja y moderada certeza, respectivamente, de un pequeño beneficio en la CdVRV en estudios que comparaban terapias psicológicas o métodos para mejorar la visión con comparadores activos.
El tipo de rehabilitación varió entre los estudios, incluso dentro de los grupos de intervención, pero se detectaron beneficios incluso si se comparan con los grupos de control activo. Se realizaron estudios en adultos con discapacidad visual, principalmente de edad avanzada, que vivían en países de altos ingresos y que a menudo padecían de DMS. La mayoría de los estudios incluidos sobre la rehabilitación de la baja visión tuvieron un corto seguimiento,
A pesar de estas limitaciones, la dirección consistente de los efectos en esta revisión hacia el beneficio justifica la realización de nuevas actividades de investigación de mejor calidad metodológica, incluidos los efectos de mantenimiento más prolongados y los costos de varios tipos de rehabilitación de la baja visión. También es necesario investigar los mecanismos de funcionamiento de los componentes de las intervenciones de rehabilitación en diferentes contextos, incluidos los países de bajos ingresos.
PICO
Resumen en términos sencillos
Rehabilitación de la baja visión para mejorar la calidad de vida de adultos con problemas de visión
¿Cuál es el objetivo de la revisión?
El objetivo de esta revisión fue averiguar si la rehabilitación de la baja visión puede mejorar la calidad de vida (CdV) de las personas con problemas de visión. Los autores de la revisión Cochrane recopilaron y analizaron todos los estudios relevantes para responder a esta pregunta y encontraron 44 estudios.
Mensajes clave
La rehabilitación de la baja visión no parece tener un impacto importante en la calidad de vida relacionada con la salud, sin embargo la evidencia es de muy baja certeza. Existe evidencia de baja certeza de que algunas intervenciones de rehabilitación para la baja visión, en particular las terapias psicológicas y los métodos para mejorar la visión, pueden mejorar la calidad de vida relacionada con la visión en personas con pérdida de la misma en comparación con la atención habitual.
¿Qué se estudió en la revisión?
Una persona con discapacidad visual tiene problemas con su vista. Si la pérdida de visión no puede corregirse con gafas o lentes de contacto, o tratarse de otra manera, entonces la rehabilitación de la visión baja puede ayudar a mejorar la calidad de vida de las personas con problemas de visión.
Hay diferentes tipos de rehabilitación de la baja visión y estos incluyen:
• Terapias psicológicas y programas de grupo para ayudar a las personas a adaptarse a la pérdida permanente de la visión y mejorar el bienestar.
• Métodos para mejorar la visión, como la enseñanza del uso de dispositivos de aumento u otras tecnologías, o la enseñanza de técnicas para mejorar el uso de la visión residual en la vida cotidiana.
• Programas de rehabilitación multidisciplinar, que pueden incluir el uso de dispositivos de aumento y terapias psicológicas, así como otros servicios, incluso en los hogares de los pacientes.
• Otros tipos menos comunes de servicios de rehabilitación como entrenamiento de equilibrio o programas de seguridad en el hogar.
¿Cuáles son los principales resultados de la revisión?
Los autores de la revisión Cochrane identificaron 44 estudios de rehabilitación de baja visión y CdV.
La mayoría de esos estudios se realizaron en los servicios de rehabilitación de la visión de países de altos ingresos. Muchos de los participantes en estos estudios eran mayores y tenían un diagnóstico de degeneración macular. En los estudios, se examinaron las terapias psicológicas y los programas de grupo, los métodos para mejorar la visión y los programas de rehabilitación multidisciplinar. En estos estudios, las personas con discapacidad visual rellenaron cuestionarios sobre su salud general, su visión u otros aspectos del bienestar.
En comparación con las personas con pérdida de visión que no recibieron ninguna rehabilitación de baja visión:
• Las personas con pérdida de visión que reciben terapias psicológicas y/o programas de grupo pueden experimentar:
⇒ ninguna mejora en la calidad de vida relacionada con la salud (muy baja certeza);
⇒ alguna mejora en la calidad de vida relacionada con la visión (baja certeza).
• Las personas con pérdida de visión que aprenden métodos para mejorar la visión pueden experimentar:
⇒ alguna mejora en la calidad de vida relacionada con la visión (muy baja certeza).
⇒ (no se disponía de datos sobre la calidad de vida relacionada con la salud).
• Las personas con pérdida de visión que participan en un programa de rehabilitación multidisciplinar pueden experimentar:
⇒ poca o ninguna mejora en la calidad de vida relacionada con la salud (muy baja certeza);
⇒ cierta mejora en la calidad de vida relacionada con la visión, en particular si se utiliza un programa intensivo (muy baja certeza).
En comparación con las personas con pérdida de visión que recibieron la atención habitual:
• Las personas con pérdida de visión que reciben terapias psicológicas y/o programas de grupo pueden experimentar:
⇒ poca o ninguna mejora en la CdV relacionada con la salud (muy baja certeza) o la CdV relacionada con la visión (baja certeza).
• Las personas con pérdida de visión que aprenden métodos para mejorar la visión pueden experimentar:
⇒ poca o ninguna mejora en la calidad de vida relacionada con la salud (muy baja certeza);
⇒ alguna mejora en la QOL relacionada con la visión (certeza moderada).
• Las personas con pérdida de visión que participan en un programa de rehabilitación multidisciplinar pueden experimentar:
⇒ poca o ninguna mejora en la CdV relacionada con la salud (muy baja certeza) o la CdV relacionada con la visión (baja certeza).
Había alguna evidencia de que las terapias psicológicas tenían un efecto positivo en la autoestima (certeza muy baja) y en la depresión (evidencia de certeza moderada)
No había evidencia que sugiriera algún daño (efectos adversos) de la rehabilitación, pero los datos eran limitados.
¿Cuál es el grado de actualización de la revisión?
Los autores de la revisión Cochrane buscaron estudios publicados hasta el 18 de septiembre de 2019.
Authors' conclusions
Summary of findings
| Psychological therapies and/or group programmes compared to waiting list or no care for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as problem‐solving treatment (PST), self‐management programme Comparison: passive control group such as a waiting list | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with EQ‐5D | 54 Short‐term maintenance effect | SMD 0.26 SD worse | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with NEI‐VFQ and NEI‐VFQ‐7 | 285 Short‐term direct or maintenance effect | SMD‐0.23 SD better | ⊕⊕⊝⊝ |
| Adverse events | 285 Short‐term direct or maintenance effect | Brody 2002: No data available Nollet 2016: 2 AEs and 10 SAEs not related to rehabilitation | ⊕⊕⊝⊝ |
| EQ‐5D: EuroQol 5 Dimensions; NEI‐VFQ: National Eye Institute Visual Functioning Questionnaire; SAEs: serious adverse events; VFQ: Visual Functioning Questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded 1 level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (small sample size of n = 54, and wide confidence interval crossing the line of no effect) Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Methods of enhancing vision compared to waiting list or no care for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as immediate low vision assessment, provision of magnifying aids and training, low vision outpatient service, customised prism glasses Comparison: passive control group such as a waiting list, delayed low vision assessment, low vision examination and no intervention, placebo prisms | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life | No data were available for this outcome | ||
| Vision‐related quality of life measured with questionnaires: NEI‐VFQ‐ 25, VA‐LV‐VFQ48, Activity Inventory, IVI | 262 Short‐term direct or maintenance effect | SMD ‐0.19 SDs (better) | ⊕⊝⊝⊝ |
| Adverse events | 262 Short‐term direct or maintenance effect | Kaltenegger 2019 reported no adverse events. No data available for the other studies | not applicable |
| NEI‐VFQ: National Eye Institute Visual Functioning Questionnaire; VA‐LV‐VFQ: Veterans Affairs Low Vision Visual Functioning Questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias in 3 out of 4 studies) and two levels due to very serious imprecision (sample size of n = 237, and wide confidence interval crossing the line of no effect) Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Multidisciplinary rehabilitation compared to waiting list or no care for better quality of life in visually impaired adults | ||||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as multidisciplinary low vision rehabilitation plus home visit, multidisciplinary low vision programme Comparison: passive control group such as a waiting list | ||||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence | Explanation |
| Health‐related quality of life measured with EQ‐5D, SF‐36 | 183 | SMD ‐0.08 SD (better) Short‐term direct or maintenance effect | ⊕⊝⊝⊝ | |
| Vision‐related quality of life measured with NEI‐VFQ, VFQ 48 questionnaire | 193 | See comment Short‐term direct or maintenance effect | ⊕⊝⊝⊝ | Both studies beneficial, but large effect in a large trial delivering intensive rehabilitation (Stelmack 2008: SMD: ‐1.64, 95%CI ‐2.05 to ‐1.24) and less benefit in the other (Acton 2016: SMD ‐0.42, 95%: ‐0.90 to 0.07), P = 0.0001 for inconsistency |
| Adverse events | 193 | Acton 2016: 19 AEs probably unrelated to treatment Stelmack 2008: no (S)AEs related to the study. Short‐term direct or maintenance effect | ⊕⊕⊝⊝ | |
| EQ‐5D: EuroQol 5 Dimensions; SAEs: serious adverse events; SF‐36: Short Form‐36 item Health Survey; VA‐LV‐VFQ: Veterans Affairs Low Vision Visual Functioning Questionnaire | ||||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | ||||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 183, and wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias), two levels due to very serious imprecision (sample size of N = 193, and wide confidence interval crossing the line of no effect) and one level due to heterogeneity (both studies beneficial, but important and significant effect in a large trial delivering intensive rehabilitation and less benefit in the other) 3Downgraded one level due to study limitations (unclear risk of performance bias) and no reporting of adverse events in 1 study Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | ||||
| Psychological therapies and/or group programmes compared to active control for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as a health education programme, self‐management programme with usual rehabilitation care, stepped‐care (including cognitive behavioural therapy, problem‐solving treatment and/or referral to general practitioner), expressive writing course, problem‐solving treatment, behavioural activation Comparison: active control group such as an individual low vision programme, usual low vision rehabilitation, usual care by low vision service or other care providers, neutral writing exercise, supportive therapy | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with EQ‐5D, SF‐36 and a general health measure | 600 Long and short‐term maintenance effect | SMD ‐0.09 SDs better | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with NEI‐VFQ‐17 and 25, IVI, LVQOL 18, vision=specific mental health questionnaire | 1245 Long and short‐term maintenance effect | SMD ‐0.11 SDs better | ⊕⊕⊝⊝ |
| Adverse events | 1453 Long and short‐term maintenance effect | No data available for Eklund 2008, Girdler 2010, Bryan 2014, Rees 2015, Rovner 2007, Rovner 2013, Rovner 2014, Tey 2019. Van der Aa 2015: no (S)AEs related to the study. | ⊕⊝⊝⊝ |
| EQ‐5D: EuroQol‐5 Dimensions; NEI‐VFQ: National Eye Institute Visual Functioning Questionnaire; IVI: Impact of Vision Impairment profile, LVQOL: low vision quality of life questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 473, and wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias), one level due to imprecision (large sample size of N = 1118, but confidence interval including the line of no effect) 3Downgraded one level due to study limitations (unclear risk of performance bias) and no reporting of adverse events in 7 studies Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Methods of enhancing vision compared to active control for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low vision services Intervention: low vision rehabilitation such as CCTV training sessions from a low vision therapist, home‐based low vision rehabilitation, low vision devices with instruction, usual comprehensive vision rehab and access to desk top video magnifier, CCTV and training, low vision support service, nonportable and portable electronic devices Comparison: active control group such as CCTV instructions from supplier, clinic‐based low vision rehabilitation, low vision devices without instruction, usual comprehensive vision rehab without access to desk top video magnifier, eccentric viewing training, placebo support by a nurse, nonportable devices only | |||
| Outcomes | № of participants | Standardized Mean Difference (SMD) with Low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with EQ‐5D, SF‐36 | 443 Short‐term maintenance effect | SMD ‐0.09 SD (better) | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with questionnaires: VISQOL, LVQOL subscales, VA‐LV‐VFQ‐48, VFQ‐25, Activity Inventory, IVI | 660 Short‐term direct or maintenance effect | SMD ‐0.24 SD (better) | ⊕⊕⊕⊝ |
| Adverse events | 660 Short‐term direct or maintenance effect | Burggraaff 2012: no AEs. Stelmack 2017: 10 AEs not treatment‐related. No data available for Draper 2016, Jackson 2017, Leat 2017, Pearce 2011, Taylor 2017. | ⊕⊝⊝⊝ |
| EQ‐5D: EuroQol 5 Dimensions, SF‐36: Short Form 36‐item Health Survey, VISQOL: vision‐related quality of life, VFQ: Visual Functioning Questionnaire; VA‐LV‐VFQ: Veterans Affairs Low Vision Visual Functioning Questionnaire; IVI: Impact of Vision Impairment profile, LVQOL: low vision quality of life questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 443, but wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias and other domains at unclear or high risk of bias in some studies); no downgrade for imprecision (large sample size of N = 660, and confidence interval consistent with small effects) 3Downgraded one level due to study limitations (unclear risk of performance bias) and no reporting of adverse events in 7 studies Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Multidisciplinary rehabilitation compared to active control for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low vision services Intervention: low vision rehabilitation such as pooled community + centre‐ and community‐based low vision service delivery, family rehabilitation intervention where family is present at all stages, enhanced low vision rehabilitation including home visits Comparison: active control group such as community placebo home visits, individual rehabilitation intervention with no family present, conventional low vision rehabilitation and control home visits from a community worker with no rehabilitation, conventional clinic‐based low vision rehabilitation including placebo home visits | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with WHO‐QOL, SF‐36 | 375 Long‐term maintenance effect | SMD ‐0.10 SD (better) | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with IVI, FAQ, VCM1 | 464 Short‐term direct or long‐term maintenance effect | SMD 0.01 SD (same) | ⊕⊕⊝⊝ |
| Adverse events | 464 Short‐term direct or long‐term maintenance effect | No data available | not applicable |
| SF‐36: Short Form‐36 item Health Survey; WHO‐QOL: World Health Organization Quality of Life; IVI: Impact of Vision Impairment profile; VCM1: Vision‐related quality of life Core Measure 1, FAQ: Functional Assessment Questionnaire, observer‐rated functional visual performance test | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 443, and wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias), one levels due to imprecision (large sample size of N = 660, but confidence interval including moderate or small effects) Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
Background
From a systematic review and meta‐analysis on global prevalence data, it has been estimated that approximately 216.6 million people have moderate to severe visual impairment (< 6/18) and that 36 million people are blind (Bourne 2017). Visual impairment is especially problematic in developing countries where approximately 80% of all visually impaired persons live. Vision loss mainly affects older people; 82% of those who are blind and 65% of those with mild to severe vision loss are 50 years or older (Pascolini 2012). Worldwide, the leading causes of visual impairment are uncorrected refractive error and cataract (Bourne 2017; Pascolini 2012). There certainly is a large regional variation, e.g. the leading conditions for visual impairment in high‐income countries are uncorrected refractive error, cataract, macular degeneration, diabetic retinopathy and glaucoma (Bourne 2014).
Vision loss is one of the leading causes of disability in older people (Brody 2001). Besides physical dysfunction (Bookwala 2011; Hayman 2007; Kempen 2012), limitations in daily life activities (Bookwala 2011; Hayman 2007), depression and anxiety (Bookwala 2011; Kempen 2012; Van der Aa 2015), vision loss is associated with a decreased life satisfaction (Bookwala 2011; Brody 2001) and quality of life (QOL) (Stelmack 2012; Van Nispen 2009). In addition, visual impairment affects many social and work‐related aspects of people's lives, including computer work (Bruijning 2013) and driving (Massof 2007).
Part of the World Health Organization's (WHO) global action plan 2014 to 2019 is to assess and monitor the magnitude and causes of visual impairment and to invest in programmes to eliminate avoidable blindness, which is also strongly recommended by the European initiatives (e.g. EC Vision). This plan also urges countries to evaluate their programmes and services as part of their overall health care system. The need of governments and their partners to invest in supporting those with irreversible visual impairment to overcome the barriers that they face in daily life is stressed as well (WHO 2013). This is also reflected in the new standards for low vision rehabilitation that have been initiated by the WHO (WHO 2017). During the last few decades, the outcome of rehabilitation has been measured mainly by performance‐based measures, such as reading speed, walking distance, functional performance of daily activities and other performance‐based measures. General or disease‐specific QOL questionnaires and other patient‐reported outcome measures have become increasingly important in the assessment of quality of health care, including rehabilitation (Amtmann 2011).
Description of the condition
Vision impairment includes low vision as well as blindness. There are different definitions for visual impairment given in the literature. We adopted the WHO criteria, because these are the most widely used. The WHO defines low vision as best corrected visual acuity in the better eye < 3/18 but ≥ 3/60, and/or visual field < 30º around the central point of fixation. Blindness is defined as best corrected visual acuity in the better eye < 3/60, and/or visual field < 10º around the central point of fixation (WHO 2007). Furthermore, we defined irreversible vision impairment as vision impairment with a duration of at least six months and/or chronic diseases such as diabetic retinopathy, age‐related macular degeneration or glaucoma as causes of the irreversible vision impairment.
Description of the intervention
Low vision rehabilitation for adults usually is a professional service which aims to optimise residual vision, but also to teach visually impaired people skills in order to improve (visual) functioning in daily life. Other aims may be to help patients with adaptation to vision loss or to improve psychosocial functioning. This may lead to more independence and active participation in society. Low vision rehabilitation should ultimately improve quality of life (QOL) of visually impaired patients.
Low vision rehabilitation is not available everywhere and, when present, is organised differently in nearly every country. Some countries may have multidisciplinary in‐ or outpatient centres, where occupational therapists, optometrists, low vision specialists, clinical physicists, psychologists, social workers, mobility and orientation trainers and computer trainers work closely together. Other countries have a single service system, where prescription of optical aids is done by one organisation and, for example, social work is provided by another. Some countries have their outpatient services linked to ophthalmology departments, e.g. in academic hospitals, where others have not. Individual or group sessions with social workers or psychologists seem to be increasingly common, as are the home environment assessments and training sessions in the use of optical or other aids (e.g. canes) and low vision software. Training in leisure time or vocational activities are important aspects of rehabilitation as well. Depending on agreements between organisations or policies in different countries, low vision rehabilitation services are either commercial, non‐profit or charity organisations. In some countries, low vision aids and services are reimbursed by health insurance, exceptional medical expenses acts or veterans affairs (European Network for Vision Impairment Training Education & Research: ENVITER; Lim 2014; Owsley 2009).
Why it is important to do this review
There have been several studies in the field of low vision rehabilitation, focusing on objective tasks or specific measures of functional ability such as reading speed (Virgili 2018) or other performance‐based measures. Although these measures are important and necessary to assess functioning, they do not capture all facets of the individual state (Scott 1999). Comprehensive patient‐reported outcome measures such as HRQOL or disease‐specific QOL have been introduced because of the growing interest of governments and health insurance companies in these outcome measures as parameters for quality of care (Amtmann 2011; Massof 2001; Stelmack 2001). In the field of low vision, there has been an increased attention on the theoretical constructs of VRQOL and visual functioning as important outcomes of rehabilitation. These constructs are measured by self‐report questionnaires, which are increasingly analysed with modern psychometric techniques, i.e. item response theory models (Pesudovs 2007). A comprehensive literature review of Binns and colleagues has shown that the evidence of vision rehabilitation services is not very clear with respect to HRQOL or VRQOL (Binns 2012). The authors did not specifically assess methodological quality and included observational studies as well. Skelton and colleagues performed a Cochrane Review to assess effectiveness of environmental and behavioural interventions for visually impaired older adults (60+) and found no RCTs (Skelton 2013). Hence, to date there has not been a systematic review that assesses the effectiveness, including certainty of evidence, of the various existing rehabilitation interventions using HRQOL, VRQOL or closely related patient‐reported outcomes of visually impaired adults aged 18 years or over. There may be a potential overlap, however, with the secondary outcomes of a previous Cochrane Review on orientation interventions to improve mobility in people with visual impairment (Virgili 2010). Studies on low‐vision rehabilitation for children were summarised in a Cochrane Review by Barker 2015 and a review by Elsman 2019, which found limited evidence for developmental and rehabilitation interventions on the participation and QOL of children with a visual impairment.
In order to develop better understanding of the most effective healthcare interventions for visually impaired people, it is necessary to draw on all forms of relevant scientific evidence. An essential step in this process is to collect and analyse the evidence from quantitative and comparative studies. The global action plan for 2014 to 2019 of the WHO stresses the importance of assessing effectiveness of eye health and rehabilitation services (WHO 2013) as are the new standards for services (WHO 2017). This review will describe the state‐of‐the‐art regarding evidence for low vision rehabilitation interventions up to September 2019. We aim to quantitatively synthesise the entire available literature on HRQOL, VRQOL and other closely related patient‐reported outcomes as an effect of rehabilitation for visually impaired adults. The secondary patient‐reported outcomes include the physical and functional, psychological and social domains so as to capture all health and well‐being outcomes available.
Objectives
To assess the effectiveness of low vision rehabilitation interventions on health‐related QOL (HRQOL), vision‐related QOL (VRQOL) or visual functioning and other closely related patient‐reported outcomes in visually impaired adults.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials (RCTs). Non‐RCTs, quasi‐RCTs or observational studies were excluded.
Types of participants
We included studies in which the effect of low vision rehabilitation among adults (>= 18 years) of either gender, with a vision impairment according to the WHO 2007 definition. Included studies had to be about vision impairment of irreversible nature, which was defined as a duration of at least six months. This means there was a focus on chronic eye diseases such as diabetic retinopathy, age‐related macular degeneration or glaucoma. Studies that included participants of all ages were only included if separate data on adult participants were reported.
Types of interventions
We included RCTs that compared one or more rehabilitation interventions with waiting lists or no care, or, usual or other care. We excluded studies focusing on the following interventions or devices:
-
neuro‐rehabilitation interventions (e.g. eye‐movement after stroke, such as the Cochrane Review by Pollock 2011);
-
interventions to improve visual field loss after brain damage;
-
medical interventions (e.g. cataract surgery, retinal or telescope implants such as the review by Maghami 2014);
-
preferences of low vision aid designs (e.g. different brands or designs of head mounted magnification devices or canes).
Types of outcome measures
We included studies that measured HRQOL, VRQOL as two primary outcomes or related patient‐reported outcomes as secondary outcomes at any follow‐up time point after the intervention ended, such as physical and functional measures, psychological measures and/or social measures. RCTs with measured outcomes not related in any way to the outcomes of interest were considered ineligible, however, some descriptive information and main outcomes of these RCTs were reported in an additional table (Table 1).
| Author year (country) | Study design, follow‐up, sample size, % dropout | Sample: mean age, % female, vision impairment | Setting and study type | Outcome measures | Intervention and control groups | Outcome of study |
| Ballemans 2012 (Netherlands) | 2‐arm RCT, 8 weeks, N = 68, 10% | 76 years, 60%, AMD, glaucoma, other conditions | Low vision rehabilitation service IV other programmes | Process evaluation: usefulness and acceptability of intervention from participants’ and trainers’ perspectives | I: standardised orientation and mobility training C: regular orientation and mobility training (without protocol) | Both the standardised and regular training showed to be useful and acceptable from both participants’ and trainers’ perspectives |
| (New Zealand) | 4‐arm RCT, 1 year, N = 391, 8% | 84 years, 68%, AMD, cataract, DRP, eye surgery | Home‐based IV other programmes | Number of falls and injuries and costs of implementation of the intervention | I: home safety assessment and modification programme by OT I: exercise programme by PT + vitamin D I: intervention 1 + 2 C: social visits | The home safety programme reduced falls and was more cost‐effective than the exercise programme. The exercise group was not effective in reducing falls. |
| (China) | 2‐arm RCT, 16 weeks, N = 40, 45% | 84 years, vision impairment | Residential care homes IV other programmes | Knee proprioception, concentric isokinetic knee strength, sensory organisation | I: Tai Chi (Yang style) C: music percussion activity | After Tai Chi, significant improvement was seen in knee proprioception and visual and vestibular ratios. Tai Chi can improve balance control. |
| (United States) | 3‐arm RCT, Immediate, N = 38, 0% | 28 years, 47% ocular related blindness | Laboratory IV other programmes | Performance of virtual navigation in virtual target building, physical navigation and drop‐off tasks in actual target building | I: video game for full exploration of virtual environment I: directed navigation by sighted facilitator and specific paths in game C: video game but not corresponding with target virtual environment | Highly interactive and immersive exploration of the virtual environment engages a blind user to develop skills akin to positive near transfer of learning. |
| (Canada) | 2‐arm RCT, 7 weeks, N = 32, unknown | 83 years, partially sighted, legally blind | Day centre for the visually impaired IV other programmes | Balance tests with open and closed eyes, perceived balance | I: balance training C: sitting stretching exercises | Perceived confidence in balance increased after training, however, objectively no improvement was found. |
| (United States) | 2‐arm RCT, 5 weeks N = 228, 9% | 75 years, 58% N = 29 visually impaired, N = 199 normally sighted | Laboratory, home‐based II methods to enhance vision | Useful field of view performance | I: speed of processing training C: social and computer contact | Speed of processing training effectively improved useful field of view performance. |
| (United States) | 3‐arm RCT, 3‐6 months, N = 25 | 47 years, 48% various (hereditary) retinal diseases | Laboratory‐based training and on‐road driving II methods to enhance vision | Clinical vision tests, functional orientation and mobility tasks, driving skills, psychophysical measures | I: bioptic telescopes and training I: lenses and training C: lenses without training | Significant improvement in visual skills using a bioptic telescope, including driving‐related skills |
AMD: age‐related macular degeneration
C: control group
DRP: diabetic retinopathy
I: intervention group
OT: occupational therapy
PT: physical therapy
RCT: randomised controlled trial
Primary outcomes
We evaluated studies that assessed QOL using validated one‐dimensional or multidimensional questionnaires. Examples of generic HRQOL questionnaires are the Short Form‐36 (SF‐36) and the Euroqol‐5 dimensions (EQ‐5D). Examples of VRQOL and visual functioning questionnaires are the VF‐14, the National Eye Institute Visual Functioning Questionnaire (NEI‐VFQ) with different item lengths and the Low Vision Quality of Life questionnaire (LVQOL). HRQOL questionnaires are valuable for policy makers to compare effects of different healthcare interventions and to make informed decisions on how to spend resources. However, we are aware that HRQOL questionnaires are found to be less sensitive to change than disease‐specific outcome measures. Therefore, we believe interpretations should be made using both generic and disease‐specific outcome measures; we did not place a hierarchy in the sense that one outcome would be more important than the other.
Self‐reported questionnaires filled out by the individual, a relative or independent rater were considered eligible for inclusion. Additionally, the questionnaires that were used in different studies had to be validated in terms of reliability (the extent to which a test effectively measures the underlying concept correctly) and validity (the extent to which a test measures the underlying concept). Even though it adds to heterogeneity, different end points regarding timing were allowed and considered in the meta‐analysis:
-
short‐term: outcomes up to six months after baseline (direct effect within one month after intervention, or maintenance effects, one month to six months after intervention);
-
long‐term: outcomes more than six months after baseline (direct effect within one month after intervention, or maintenance effects, one month or longer after intervention).
Secondary outcomes
Secondary outcomes were patient‐reported outcomes closely related to QOL concerning health and well‐being, such as:
-
physical and functional measures (e.g. activities of daily living, mobility and orientation, reading);
-
psychological measures (e.g. depression, mood, anxiety, adaptation to vision loss, self‐esteem);
-
social measures (e.g. loneliness or independence).
The timing of the secondary outcome assessment was similar to that of the primary outcomes.
Adverse outcomes
If reported, all adverse outcomes in RCTs were mentioned in the review, which could be serious adverse events or adverse events, related or unrelated to the trial or intervention studied.
Search methods for identification of studies
Electronic searches
The Cochrane Eyes and Vision Information Specialist conducted systematic searches in the following electronic databases for RCTs and controlled clinical trials. There were no restrictions by language or year of publication. The date of the search was 18 September 2019.
-
Cochrane Central Register of Controlled Trials (CENTRAL; 2019, Issue 9) (which contains the Cochrane Eyes and Vision Trials Register) in the Cochrane Library (searched 18 September 2019) (Appendix 1).
-
MEDLINE Ovid (1946 to 18 September 2019) (Appendix 2).
-
Embase Ovid (1980 to 18 September 2019) (Appendix 3).
-
Cumulative Index to Nursing and Allied Health Literature (CINAHL) EBSCO (1937 to 18 September 2019) (Appendix 4).
-
PsycINFO Ovid (1806 to 18 September 2019) (Appendix 5).
-
International Standard Research Clinical Trial Number (ISRCTN) registry (www.isrctn.com/editAdvancedSearch; searched 18 September 2019) (Appendix 6).
-
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 18 September 2019) (Appendix 7).
-
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp; searched 18 September 2019) (Appendix 8).
Searching other resources
We searched reference lists of relevant articles to find additional trials. We manually searched the Visual Impairment Research journal from January 1999 to the last issue in December 2006 and several related proceedings. In addition, we contacted authors of relevant trials to identify further published and/or unpublished reports. We manually searched the proceedings of the International Society for Low Vision Research and Rehabilitation (ISLRR) of their International Conferences on Low Vision held between 1999 to 2017 for relevant trials. There were no language or data restrictions in the searches.
Data collection and analysis
Selection of studies
Three review authors, working independently, assessed the titles and abstracts resulting from the electronic searches. Full copies of all seemingly relevant papers were obtained and assessed according to the 'Criteria for considering studies for this review'. We only assessed trials that met these criteria. Disagreements about whether a trial should be included were resolved by discussion and consensus. In cases where additional information was needed before deciding whether to include a trial, we obtained the full report.
Data extraction and management
Three review authors working independently extracted data using a data extraction form developed by Cochrane. Any discrepancies were resolved by discussion. We contacted investigators to obtain missing data if necessary. For most studies, data were obtained directly from the trial reports. If the source of data was from a personal communication with an investigator, we reported this in Notes in the Characteristics of included studies table.
Furthermore, since QOL and depression data are often skewed with more persons having favourable outcomes, we considered log‐transforming data. However, since the raw data were not available and skewness was difficult to assess with certainty from means and standard deviations (SDs) only, it was decided not to create precision that might not hold in reality.
Assessment of risk of bias in included studies
Two review authors working independently assessed risk of bias according to the methods set out in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). Six parameters were considered: random sequence generation (selection bias), allocation concealment (selection bias), blinding (masking) of participants and personnel (performance bias), masking of outcome assessment (detection bias), incomplete outcome data addressed (attrition bias) and selective outcome reporting (reporting bias). See Table 2 for more information on these parameters.
|
Any other sources of bias were reported under notes in the Characteristics of included studies table. |
Each parameter was graded as a low, high or unclear risk. Every study was assessed separately by two review authors. We resolved discrepancies between assessments by discussion or by consulting another review author. Finally, we planned to address reporting bias by using funnel plots if more than 10 studies could be pooled. Since fewer than 10 studies could be pooled per comparison, it was decided not to report publication bias.
Measures of treatment effect
Data analysis was conducted according to Section 4.8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). Questionnaire outcomes usually are continuous outcomes with different scaling for different questionnaires (e.g. summary scores for Likert‐type scales). This means that standardised mean differences (SMD) are indicated to synthesise data. The SMD is the difference in mean effects in the experimental and control groups divided by the pooled standard deviation of participants’ outcomes (see Chapter 9, Section 9.2.3.2 of the Cochrane Handbook). The value of the SMD thus depends on both the size of the effect (the difference between means) and the standard deviation of the outcomes (the inherent variability among participants). The SMD expresses the intervention effect in standard units rather than the original units of measurement. However, this is inevitable when comparing outcomes using different questionnaires. We extracted means, mean differences, standard deviations, standard errors and/or P‐values from the studies in order to calculate SMDs, which were summarised in forest plots and in the 'Summary of Findings' tables. SMDs were interpreted as (Cohen's) effect sizes, where 0.2 represents a small effect, 0.5 represents a moderate effect and 0.8 represents a large effect. If only one study was available in a comparison, forest plots showed SMDs whereas in the text, mean differences were also reported for the main outcomes.
Unit of analysis issues
Participants (rather than eyes) were the unit of analysis in this systematic review. Unit of analysis issues were studied using multiple outcomes or subscales to measure a single construct (e.g. HRQOL measured by a physical and a mental health summary score of the SF‐36, or depression measured by two depression questionnaires), or studies having multiple intervention or control groups. This was solved (post hoc decision) by choosing only one questionnaire when similar constructs were measured in one study, or by combining the means and SDs of subscales (e.g. SF‐36 mental and physical summary score) into one effect estimate. Although it can be considered unconventional to report one score of a multidimensional questionnaire where it is common to report summary scores of subscales, this overcomes the issue of presenting correlated data in the meta‐analysis.
Dealing with missing data
If data were missing or difficult to interpret from a paper, if feasible, our intention was to contact the authors for more information on whether they thought missing data was (completely) at random. Missing data were expected to be an issue when many respondents were lost to follow‐up, or if there were an unequal loss to follow‐up in the trial arms (attrition bias). During the data extraction, missing data were described in tables for every study, also with respect to baseline imbalances between respondents and those lost to follow‐up on the main outcomes (missingness not at random).
Assessment of heterogeneity
Before data synthesis, heterogeneity was assessed by examining the characteristics of each study separately and for every planned comparison with respect to clinical diversity (variability in participants, interventions, controls and outcomes) and methodological diversity (variability in study design and risk of bias). We planned to use forest plots to present study results. These included a Chi2 test for statistical heterogeneity and the I2 test which estimated the amount of heterogeneity between trials by describing the percentage of variation between studies.
In order to deal with heterogeneity in the interventions and comparators, we decided (post hoc) to conduct two sets of analyses dividing studies with inactive control groups, such as no care or a waiting list, from studies with active comparators. An active comparator was defined as 'usual care' by some authors, but we acknowledge that no standard exists to define usual care and we briefly describe its components for each comparison. An active comparator could also be some other type of programme, such as a lower dose of the studied intervention. Finally, we did not conduct an overall meta‐analysis for 'low vision rehabilitation', but presented them for each subgroup of analyses. Nonetheless, we believed it to be important to present different intervention groups in the same forest plot since this would allow formal and graphical heterogeneity assessment.
Data synthesis
In most cases, relevant data for the separate treatment arms could be obtained from the questionnaire outcomes (continuous data) in order to calculate SMDs. If data were not available, these studies were then excluded from our efforts to synthesise data. If two or more treatment arms were compared to a control group, combined data of treatment arms were entered in the forest plots or treatment arms were ignored if these were adding to clinical diversity. Combining means and SDs of study arms outcomes, or other reasons such as combining means and SDs of subscales of questionnaires, was performed in IBM SPSS version 22 and according to Chapter 7 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
As was recommended in Chapter 11 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017), results were summarised in the 'Summary of Findings' (SoF) table using GRADE, where a) imprecision due to few participants (< 400), b) imprecision due to wide confidence intervals including no effect, imprecision due to wide confidence intervals including no effect and appreciable harm and benefit, c) unclear and/or high risk of bias limitations, d) indirectness (e.g. indirect comparison), inconsistency of effect expressed by extensive heterogeneity of results (I2 > 60%) were considered to be potential reasons to downgrade the certainty of evidence. Footnotes were used to explain reasons to downgrade the certainty of evidence. In the SoF table, the intervention subgroups were listed by most important outcomes (HRQOL, VRQOL) and by comparator (inactive or active), as were the SMDs, 95% confidence intervals, number of studies and persons included in the analysis. The overall assumed risk of the control group could not be calculated, since the outcome measures used were often different which means that different units of measurement were used.
Since heterogeneity was expected to be found (I² > 50%), the meta‐analysis was based on the random‐effects model if there were enough studies within one comparison. We had planned to use the fixed‐effects model If there were fewer than three trials in a certain comparison. However, this is not possible when a single forest plot is produced for several subgroups. Thus, we always used random‐effects and we chose pragmatically to comment on individual studies when heterogeneity was apparent due to poor overlap of 95% CIs of two studies in a subgroup.
Dealing with multiple study arms
To facilitate interpretation of results, we chose to ignore certain study arms in the data analysis as a post hoc decision, taking into account intervention and comparison subgroup definitions. This was done for three studies. In Christy 2012, the arm in which only the centre‐based intervention was presented was ignored. Three other arms remained: the community intervention arm was combined with the centre plus community intervention arm and compared to the community intervention plus placebo home‐visits which served as the control group (other care). In Reeves 2004, the conventional low vision rehabilitation arm was also ignored, facilitating a comparison between the community‐based enhanced rehabilitation to a placebo community‐based control group (other care). Finally, in Smith 2005, the standard prism prescription arm was ignored, facilitating a comparison between the customised prism spectacles and sham prism spectacles ('no' care).
Subgroup analysis and investigation of heterogeneity
Some studies presented only short‐ or long‐term data, or, both short‐ and long‐term data as trial outcomes. Since long‐term outcomes were scarce and were considered a possible threat to homogeneity of the data, we originally planned to first analyse data using short‐term outcomes only and if long‐term data were available, to enter them in the analyses and report separately if differences with only short‐term outcomes were relevant. However, given the small number of studies in each subgroup and the limitation in data reporting, we analysed the data as available.
Sensitivity analysis
We originally planned sensitivity analyses by excluding studies from the analysis if they presented unclear and/or high risk of bias on at least four out of six domains on which risk of bias was assessed, or, in case of outliers. However, following this procedure, given the small number of studies in each subgroup, with unclear or high risk of bias in at least one domain, we decided not to conduct such analyses.
Results
Description of studies
Results of the search
The electronic searches identified a total of 15,891 records (Figure 1). The Cochrane Information Specialist removed 2699 duplicates and screened the remaining 13,192 reports, of which 4626 were not relevant to the scope of the review. We reviewed the remaining 8566 reports and discarded 8403 records as not relevant to the scope of the review. We obtained 163 full‐text reports for potential inclusion in the review. We included 73 reports of 44 studies, see Characteristics of included studies table for details. Seventy‐two reports of 65 studies were excluded, see Characteristics of excluded studies table for details.
Study flow diagram.
Eighteen reports of 15 study protocols or design papers were found with potentially relevant outcomes for a future update of the current review (see Characteristics of ongoing studies) However, the studies by NCT00971464 and Zijlstra 2009 can be considered unpublished as reported by the study authors, and NCT00545220, Rees 2013, and Rubin 2011 might be considered unpublished since the end date of the study must have passed over more than two years ago (pending personal contact with the study authors).
Included studies
Forty‐four studies were reported in 73 documents and were included in the review (i.e. some studies had multiple publications; see Characteristics of included studies), of which 22 had a published study protocol in a journal and/or trials register (mainly the more recent studies). Twenty‐seven studies described the main HRQOL (N = 3) or VRQOL (N = 18) outcomes or both main outcomes (N = 8). Of these, 27 described both main and secondary outcomes (patient‐reported outcomes closely related to QOL) and in an additional 11 studies, only secondary outcomes were described.
Compared to the original protocol (Langelaan 2008), the focus on mono‐ versus multidisciplinary rehabilitation was not feasible due to clinical diversity. Low vision aids training or psychological treatment would both be considered monodisciplinary rehabilitation services, since one type of professional would be involved in treatment of the participant. However, based on content of the service and the professionals involved, to combine these completely different types of monodisciplinary rehabilitation did not seem to be the best option. Therefore, as a post hoc decision, they were categorised into four groups according to the types of interventions assessed:
-
psychological therapies and/or group programmes (20 studies);
-
methods of enhancing vision (14 studies);
-
multidisciplinary rehabilitation programmes (7 studies);
-
other programmes (3 studies)
In addition, the type of comparator was divided into inactive or active comparators:
-
inactive: no care (4 studies) or waiting list (14 studies);
-
active: usual care (13 studies) or other care (13 studies).
Types of interventions
1. Psychological therapies or group programmes, or both
Twenty studies assessed the effect of psychological therapies or group programmes, or both. In five studies, self‐management interventions were investigated (Brody 1999; Brody 2002; Girdler 2010; Rees 2015; Tey 2019). Two studies described educational interventions (Eklund 2008; Goldstein 2007). Individual problem‐solving treatment (PST) was investigated by Rovner 2007 and again in 2013 (Rovner 2013). PST falls under the umbrella of cognitive‐behavioural therapies where the focus is on the development of personal coping strategies that aim to solve daily issues and problems and target changing unhelpful patterns in cognitions, such as thoughts, beliefs, and attitudes, but also focus on changing behaviours and emotional regulation. In Van der Aa 2015, PST was part of a stepped‐care intervention, which also included watchful waiting, cognitive‐behavioural therapy based self‐help and a referral to the general practitioner. Nollet 2016 also studied PST and compared it to a referral to the general practitioner. There were other treatments which can be considered psychological therapies or group programmes, or both, i.e. group‐based peer support by Bradley 2005, expressive writing to deal with traumatic experiences by Bryan 2014, perceptual training versus group or individual psychological counselling by Conrod 1998, a stress‐reducing intervention for people with glaucoma by Kaluza 1996, group‐based rational emotive therapy by Mozaffar Jalali 2014, behavioural activation versus supportive treatment by Rovner 2014, social competence training for visually impaired employees by Rumrill 1999, and psychological therapy by Sun 2012.
Self‐management programmes
-
Brody 1999 compared a self‐management group programme of seven to 10 participants focusing on behavioural skills training in six sessions of 2 hours with a waiting list control group.
-
In a new study, Brody 2002 again compared a self‐management group programme, however this time, the intervention was compared to two control conditions. The self‐management group intervention of 8 to 10 participants consisted of six 2‐hour sessions and was led by an experienced professional in public health and behavioural medicine. The intervention was composed of both cognitive and behavioural components and focused on health education and enhancement of problem‐solving skills. The first control condition was an audiotape intervention consisting of a series of 12 hours of audiotapes of health lectures on age‐related macular degeneration (AMD) and healthy ageing, which they could listen to during a period of six weeks. The second control group was a waiting list control group.
-
Girdler 2010 also investigated a self‐management intervention plus usual care which they compared to the effects of usual care. The group intervention with six to 10 participants was led by an occupational therapist and a social worker and lasted eight weeks (total 24 hours). The programme was structured in several steps including understanding vision loss, optimising residual vision and other senses, practicing problem‐solving skills, daily life and safety skills, and focusing on the future. Usual care was an individual case management model and multidisciplinary work including, for example, a visual assessment and a prescription of low vision aids, occupational therapy, social work, and orientation and mobility training.
-
Rees 2015 investigated a low vision self‐management group programme held once a week over eight weeks, with 3‐hour sessions facilitated by two low vision rehabilitation counsellors. The programme also included a guest speaker from a national advocacy organisation, sessions with an orthoptist, a low vision technology expert and an orientation and mobility instructor. Two important components of the self‐management programme were problem‐solving skills training and goal planning. The control condition consisted of usual care, which was provided by the low vision rehabilitation service with an initial assessment by a member of the multidisciplinary team, an optometric assessment and prescription of optical aids. If needed, further training was provided by the multidisciplinary team.
-
Tey 2019 studied a low vision self‐management programme which was held four times during three 5‐hour weekly group sessions with six to eight participants. The intervention group also received usual care which consisted of standard ophthalmologic care, low vision aid training and referral to occupational or mobility training at the participant's request. In the self‐management group, participants could choose a goal they wished to achieve. Furthermore, there was a focus on the process of learning new techniques to enhance activities of daily living, information was provided, experiences with low vision were explored and problem‐solving skills were taught including positive self‐talk and the use of other senses (touch and hearing). The group work was aimed at sharing and practicing activities of daily living, as well.
Educational programmes
-
Eklund 2008 developed and investigated a health education programme called 'Discovering new ways', with groups of four to six participants. The programme was for eight weeks, two hours per week, and was led by an occupational therapist. The programme's main focus was on providing information and skills training. The intervention was compared with an individual standard intervention programme for the target group at the low vision clinic. Prescription of low vision aids and information about lighting was provided. If requested, additional information about the disease could be received. This programme mainly consisted of one to two 1‐hour sessions at the clinic followed by telephone contact.
-
Goldstein 2007 developed an educational video to help people achieve self‐efficacy with regard to obtaining and using assistive devices and other rehabilitation resources. The educational video addressed educational, emotional and motivational needs associated with low vision. Participants were given two weeks to watch the video. The control group was a waiting list control group.
Problem‐solving treatment (PST)
-
Rovner 2007 compared PST to usual care provided by an ophthalmologist or other health care providers. With the intervention, problem‐solving skills were taught to participants in addition to usual care (similar to control group). Six PST sessions were provided by trained therapists at the participant's home, with a duration of 45 to 60 minutes, for a period of eight weeks.
-
In a new study, PST was compared to supportive therapy as a placebo intervention (Rovner 2013). The intervention consisted of workshops, review and discussion of the PST at the participant's home. The supportive therapy control group was a standardised psychological nondirective treatment in which the therapist investigated the impact of vision loss on their lives. In contrast with PST, there was no discussion of vision function goals, problem‐solving or low vision rehabilitative strategies.
-
PST was also part of the stepped‐care intervention which consisted of four steps and was studied and compared to usual care by Van der Aa 2015. The first step comprised of three months of watchful waiting, involving an active decision not to treat the condition but, instead, to intermittently reassess its status. Telephone calls were made at baseline and after three months of watchful waiting. Participants could contact the executive researcher by telephone during this period, if necessary. In the second step, guided self‐help was based on a written, digital, audio, and Braille version of a self‐help course based on cognitive behavioural therapy (with specific vision‐related examples and exercises). Guidance was provided by trained and supervised occupational therapists from outpatient low vision rehabilitation organisations. Two face‐to‐face contacts took place at the beginning of the intervention for about 60 minutes and one to three telephone calls. Participants followed the intervention at home. In the third step, trained and supervised social workers and psychologists from the low vision rehabilitation centres offered problem‐solving treatment with a maximum of seven face‐to‐face contacts of about 60 minutes each. Finally, if symptoms of depression and anxiety persisted, a fourth step was available which was a referral to the general practitioner to discuss further treatment, such as the use of drug treatment. Stepped‐care plus usual care was compared to usual care and included outpatient low vision rehabilitation care and/or care that was provided by other healthcare providers.
-
Nollet 2016 also investigated PST and compared it to a referral to the general practitioner and to a waiting list condition. PST consisted of a brief manualised cognitive behavioural therapy based on the therapy used by Rovner 2007, however, this time the intervention included large‐print self‐help materials on depression and a list of vision‐related organisations. The optometrists providing the low vision assessment shared the participant’s treatment plan with the psychologists. Therapists who were rigorously trained, worked with the participants on an individual basis in their own home or at one of the research centres to teach them a seven‐step method for approaching and solving their problems. The referral condition consisted of a standardised letter sent to the participant’s physician within two weeks of randomisation. It informed the physician that his or her patient had screened positive for significant depressive symptoms and asked the physician to offer treatment according to National Institute for Health and Care Excellence (NICE) guidelines. Participants in the waiting list control arm received no intervention other than the 6‐week follow‐up low vision assessment.
Other psychological or group treatments, or both
-
Bradley 2005 studied group‐based peer support and information provision. Discussion groups were organised and six leaflets with information were distributed in six weekly sessions of 1.5 hours which were led by people who were experienced in living with macular degeneration. The control group was on a waiting list for six weeks.
-
Bryan 2014 investigated an expressive writing intervention and compared it to a non‐emotional topic writing task. The intervention group was asked to write about their most stressful experiences related to Stargardt’s eye disease and to reveal their deepest emotions. Participants wrote for 20 minutes on three separate days during a 1‐week period. The control group was instructed to write about their day completely factually and without emotions.
-
Conrod 1998 investigated the effect of perceptual training which consisted of five weekly one‐hour training sessions to enhance visual skills versus individual psychological counselling or group counselling with the same frequency.
-
Kaluza 1996 investigated the effect of a stress‐reducing intervention for people with glaucoma. The aim of the programme was to support patients' coping with glaucoma and to enable them to self‐regulate stress‐induced elevated intraocular pressure by teaching them relaxation techniques. Home assignments were part of the intervention as well. Eight weekly 90‐minute group sessions were conducted by an experienced clinical psychologist. The control group was a waiting list control group.
-
Kamga 2017 investigated the effect of a cognitive behavioural therapy‐based self‐care tool intervention (anti‐depressant skills workbook, mood monitoring tool, DVD on depression) plus up to three coaching 10‐minute phone calls by a trained former nurse. Controls were a waiting list group receiving the intervention after follow‐up with one phone call.
-
Mozaffar Jalali 2014 studied a group‐based rational emotive behavioural therapy which is a comprehensive, active‐directive psychotherapy which focuses on resolving emotional and behavioural problems and irrational beliefs. It was offered by therapists and compared to no treatment.
-
Rumrill 1999 studied the effects of a social competence training programme which aimed to increase accommodation request activity, self‐efficacy and knowledge of the Disabilities Act among employees with a visual impairment. Training duration was 16 weeks by a rehabilitation professional. The intervention was compared to a waiting list control group.
-
In contrast to earlier studies in which PST was investigated, Rovner 2014 now studied behavioural activation which is considered a functional analytic psychotherapy which focuses on targeting behaviours that might maintain or worsen depression. This therapy leaves the 'changing cognitions' part out, which can be quite difficult for older persons. Sessions were one hour in duration and were offered six times by occupational therapists, on top of low vision rehabilitation treatment. The control group received supportive therapy, offered in similar dose and intensity. Supportive therapy is a non‐directive, psychological treatment that provides emotional support and was used as a control for attention in this explanatory trial.
-
Sun 2012 investigated psychological therapy given to people with glaucoma while receiving clinical therapy from ophthalmologists versus clinical therapy only. Therapy was provided by psychodynamic psychiatrists for about six months. Ophthalmologists and nurses were also involved in psychological support. Ophthalmologists talked about prognosis, offered help to eliminate any fears, and to have a positive outlook on the disease. Specialist nurses provided psychological care during the period of surgery, but also offered telephone counselling and family follow‐up after discharge from the hospital. Participants were encouraged to listen to light music, and stroll in the woodlands or park to relax their minds. The control group received clinical treatment only, such as drug treatment and surgery, if needed.
2. Methods of enhancing vision
Fourteen studies assessed methods of enhancing vision. In six studies, low vision device training was assessed, i.e. CCTV‐training (Burggraaff 2012) and general low vision device training, including magnifiers or microscopes (Kaltenegger 2019; Mielke 2013; Pearce 2011; Scanlan 2004; Stelmack 2017). In five studies, prescription of low vision devices such as magnifiers and CCTVs was investigated sometimes with some form of training or self‐practice (Dunbar 2013; Jackson 2017; Leat 2017; Patodia 2017; Taylor 2017). The other studies focused on the effect of light adaptation (Brunnström 2004), vision strategies in reading training (Seiple 2011) and use of low vision devices, such as prism spectacles (Smith 2005). Control groups received delayed treatment (waiting list or cross‐over), placebo treatment or usual care.
Low vision device training
-
Burggraaff 2012 investigated the effect of a standardised CCTV training programme. The treatment group received usual instructions from the supplier when the CCTV was delivered at home combined with training sessions in the use of the device from a low vision therapist. Training sessions were once a week, with a duration of 60 minutes. Participants were trained until they had practiced with every assignment or until no further improvement in reading ability was to be expected. The control group received only usual instructions from the supplier who delivered the CCTV.
-
Kaltenegger 2019 investigated reading training with sequentially presented text (RSVP) on a laptop computer for home‐based training in addition to magnifying aids. The control group performed a placebo training (doing crossword puzzles) in addition to magnifying aids. Both groups performed the training exercises at home for 6 weeks, 5 days/week, 30 min/day.
-
Mielke 2013 investigated the effects in a pilot study of professional visual rehabilitation. The treatment group received magnifying visual aids and training. There was a waiting list control group.
-
Pearce 2011 investigated the effect of adding low vision device training on top of consultation with an optometrist for conventional low vision assessment. The intervention group received a one‐hour appointment with a low vision specialist to review handling of low vision devices and to discuss problems at home. The control group received a placebo treatment by a nurse, only measuring weight, height, vision and blood pressure in addition to the conventional low vision assessment.
-
Scanlan 2004 investigated a low vision service that made use of an extended period of education to assist clients with the use of low vision devices. The intervention group received an extended teaching programme in reading with microscopes (near vision telescopes), consisting of five one‐hour sessions at a low vision clinic. The control group received a traditional teaching session at the low vision clinic of one hour in reading with microscopes.
-
Stelmack 2017 studied low vision rehabilitation, where participants received basic low vision services plus one to three therapy sessions which included an instruction in eccentric viewing, the use of low vision devices, environmental modification, integration of low vision devices into the participant's lifestyle, and homework assignments to practice using low vision devices for everyday tasks. The control group received basic low vision services, where the optometrist dispensed low vision devices without therapy or homework assignments.
Prescription of low vision devices
-
Dunbar 2013 evaluated a low vision assessment, within two weeks of enrolment in a hospital‐based low vision centre versus a delayed assessment (waiting list). During the assessment, information was provided on aids and services and prescribed low vision aids were dispensed.
-
Jackson 2017 evaluated the immediate provision of a CCTV to usual comprehensive vision rehabilitation with optical aids of preference as compared to usual rehabilitation alone.
-
Leat 2017 compared the prescription of a CCTV associated with the request to practice everyday for 10 minutes twice for six weeks with a control group receiving eccentric viewing training and then home training with an observer giving feedback on accuracy.
-
Patodia 2017 evaluated a low vision outpatient treatment programme including examination, prescription of low vision devices for four‐week use to determine which would be most beneficial, and a single training session, i.e. binocular and monocular telescopes, telemicroscopes, magnifiers, portable CCTV and absorptive filters. Controls received a low vision examination, but no intervention.
-
Taylor 2017 evaluated the benefit of adding an electronic portable device on top of non‐electronic optical devices, compared to the use of non‐electronic optical devices in controls.
Other methods of vision enhancement
-
Brunnström 2004 evaluated the effect of additional light adaptation in the living room on top of basic light adjustments in other rooms such as the kitchen, bathroom and hall, which was the usual care received by the control group.
-
Seiple 2011 evaluated reading rehabilitation training exercises which were derived from the literature and from the laboratory’s curriculum. Three training groups were compared in a cross‐over design and also compared to a waiting list control group: 1) visual awareness and eccentric viewing; 2) control of reading eye movements; 3) reading practice with sequential presentation of lexical information.
-
Smith 2005 evaluated the effect of prism spectacles in people with AMD. Participants in the first treatment group received custom bilateral prisms which matched the participants' preferred power and base direction. Treatment group two received standard bilateral prisms (six prism diopters). The control group received placebo prisms, consisting of spectacles matched in weight and thickness to prism spectacles but without the prism.
3. Multidisciplinary rehabilitation programmes
Seven studies evaluated outpatient multidisciplinary programmes which sometimes included home visits by professionals.
-
Acton 2016 evaluated the effect of home visits from a visual rehabilitation officer who assessed areas such as emotional issues, low vision function, lighting, personal care, medication management, kitchen tasks, household chores, entitlements, orientation and mobility, and communication. The control group was on a waiting list receiving no care.
-
Christy 2012 evaluated different methods of multidisciplinary low vision service delivery. Group 1 received centre‐based service delivery, which took place in vision rehabilitation centres and included a clinical assessment and prescription of low vision devices. A qualified rehabilitation professional trained all individuals in independent living skills, based on their individual needs, for three consecutive days (four to six hours per day) and six to 12 follow‐up visits (two to five hours per visit) at the centre. Group 2 received community‐based service delivery by trained rehabilitation workers who provided rehabilitation at the participant’s home environment. Frequency was similar to group 1 and family members and neighbours were involved in training. Group 3 received centre‐ and community‐based service delivery. Individuals were trained in needs‐based independent living skills on a one‐to‐one basis for three consecutive days, four to six hours a day at the centre by a rehabilitation professional. Additional training was provided at the participant's home by a community‐based rehabilitation worker with a similar frequency and family members and neighbours were also involved in training. Finally, group 4 received centre‐based services with non‐interventional community visits. The low vision intervention was similar to the model of the centre‐based service delivery, however, no additional follow‐up training was provided at the rehabilitation centre. Home visits made by community workers were non‐interventional.
-
Draper 2016 compared two programmes delivering five visits over two to three months: a home‐based low vision rehabilitation with visit 1, 3, 5 in the clinic and visit 2 and 4 at home, including low vision aids prescription and training, and home adaptations. The second programme was clinic‐based low vision rehabilitation with all five visits in the clinic.
-
McCabe 2000 evaluated a model of outpatient vision rehabilitation in order to assess whether involving families in rehabilitation would result in better rehabilitation outcomes. All participants in the study received the “Infirmary's standard vision rehabilitation programme” which included training in the use of prescribed optical and non‐optical devices, instruction in adaptive techniques, and adjustment counselling. In contrast to controls, in the treatment arm where family was involved, time was spent educating family members about the ophthalmic condition and the rehabilitation process. Families also used simulators to understand the nature of the participant's vision loss. The social work interview included an exploration of the meaning of vision loss for the family unit and the ways the family members worked together to adapt to the loss. In the control group, individual rehabilitation interventions focused solely on the participant. Family members were excluded from all sessions.
-
Pankow 2004 tested the effectiveness of a vision rehabilitation programme which included orientation and mobility training, blind rehabilitation teaching, such as occupational therapy, and a low vision evaluation by optometrists and social workers, all depending upon the participants’ needs. The control group was on a waiting list and merely received education regarding ocular disease.
-
Reeves 2004 investigated the effectiveness of different strategies of low vision rehabilitation. Group 1 received conventional low vision rehabilitation (CLVR) by the hospital eye service, which included optometric services or other hospital‐based services, such as social work (usual care). Group 2 received enhanced low vision rehabilitation (ELVR), which was CLVR enhanced by home visits from a rehabilitation officer who provided low vision interventions. Participants received three visits which took place at approximately two weeks, four to eight weeks and at four to six months after the first hospital low vision assessment. The control groups received a placebo treatment, which was CLVR supplemented with home visits from a community care worker, who did not provide rehabilitation, but gave general advice and support unrelated to vision (CELVR).
-
Stelmack 2008 evaluated an outpatient low vision rehabilitation programme for veterans with moderate and severe vision loss due to macular disease (LOVIT). Treatment consisted of five weekly sessions (approximately two hours per session) at the low vision clinic to teach strategies for more effective use of remaining vision and use of low vision devices. In addition, one home visit was provided to teach environmental adaptations and to set up low vision devices so that participants could practice using them in their homes. Each participant was assigned five hours of homework per week after each therapy session to practice performing everyday tasks. The control group was a waiting list control group, which received additional bimonthly telephone calls from low vision therapists during the four months.
4. Other programmes
There were three other intervention studies that were considered eligible for inclusion in this review.
-
-
Gleeson 2014 studied the Alexander technique to improve balance in participants with vision loss. The Alexander Technique lesson was given by an accredited teacher 12 times for 30 minutes in 12 weeks in addition to usual care. The technique provides a cognitive construct for examining habitual responses to the familiar stimuli that precede voluntary movements. Everyday activities were practiced during the lessons such as movements between sitting and standing, getting to and from the floor, walking, climbing stairs and carrying things. The control group received usual care from Guide Dogs.
-
Pinninger 2013 investigated a tango dance programme for people with AMD with sessions of one and a half hours twice a week, for four weeks. The control group was a waiting list control group.
-
Waterman 2016 used three treatment arms; 1) home safety programme by occupational therapist visiting twice and making safety modifications plus one phone call; 2) home safety plus home exercise programme stressing strength and balance, at least three times every week for 30 minutes, walking twice per week by occupational therapists who visited two times, and visits three times and phone calls twice per week by peer mentors; 3) control: usual care from NHS which made three social visits, as well as two telephone calls by trained voluntary lay visitors.
-
Types of participants
Among the psychological therapies or group programmes, or both , the number of participants randomised in the trials ranged from 12 to 265. The most common cause of visual impairment was AMD in most studies. Between 50% and 100% of the participants were female. Eight studies were performed in the USA, three in Australia, two in Germany, two in the United Kingdom, two in Canada, one in the Netherlands and Belgium, one in Singapore, one in China, one in Iran, and one in Sweden. Except for Rumrill 1999, most studies focused on older adults recruited from hospitals or low vision services.
-
Bradley 2005 included 12 participants with AMD (United Kingdom, mean age 76 years);
-
Brody 1999 included 92 participants with AMD (USA, mean age 79 years); a separate study, Brody 2002, included 252 participants with AMD (mean age 81 years), however, for the Brody 2006 study citation, a subsample of 32 participants was chosen (Brody 2002);
-
Bryan 2014 randomised 81 adults with Stargardt's disease (juvenile AMD, USA, mean age 42 years);
-
Conrod 1998 randomised 49 adults with low vision (78% AMD, Canada, mean age 70 years) and included a group of 50 adults who were fully sighted as well (mean age 70 years);
-
Kamga 2017 randomised 80 participants of whom 55% had AMD and 45% diabetic retinopathy as a cause of low vision (Canada, mean age 76 years);
-
Eklund 2008 included 253 participants with AMD (Sweden, mean age 79 years); participants were allowed to have relatively good visual acuities;
-
Girdler 2010 included 77 participants with age‐related vision disorders (Australia, 79% AMD, mean age 79 years);
-
Goldstein 2007 included 156 low vision participants (eye disorders not reported, USA, median age 78 years);
-
Kaluza 1996 included 23 participants with primary angle glaucoma (visual acuity not reported, Germany, mean age 52 years);
-
Mozaffar Jalali 2014 randomised 60 participants with blindness (Iran, age 20 to 60 years);
-
Nollet 2016 randomised 85 participants with low vision (United Kingdom, mean age 70 years);
-
Rees 2015 randomised 153 participants, of which 70% had AMD (Australia, mean age 80 years);
-
Rovner 2007 included 206 AMD participants (USA, mean age 81 years); Rovner 2013 included 241 participants with AMD (neovascular 88%, mean age 82 years); and in their latest study, Rovner 2014 randomised 188 AMD participants (USA, mean age 84 years);
-
Rumrill 1999 included 48 participants with low vision (eye disorders not reported, USA, mean age 44 years);
-
Sun 2012 included 100 participants with glaucoma (China, mean age 62 years);
-
Tey 2019 randomised 165 participants with glaucoma, DRP, RP, pathological myopia and other eye conditions (Singapore, mean age 60 years);
-
Van der Aa 2015 randomised 265 participants with low vision of which 46% had AMD (Netherlands and Belgium, mean age 74 years).
Among the methods of enhancing vision studies, the number of participants randomised in the trials varied between 14 and 323. In most studies, the cause of low vision was AMD. Between 3% and 73% of the participants were female. Four studies were from the UK, three from the USA, three from Canada, one from Sweden, one from the Netherlands, and two from Germany. All studies focused on older adults.
-
Brunnström 2004 included 46 participants of whom 61% had dry and wet AMD; other specified diagnoses were glaucoma (11%) and retinitis pigmentosa (4%) (Sweden, mean age 76 years);
-
Burggraaff 2012 included 122 participants of whom 68% had AMD (the Netherlands, mean age 77 years);
-
Coleman 2006a included 131 participants, however, the majority had relatively good vision (VA (visual acuity) > 20/32). Participants with a VA < 20/40 had AMD or cataract, numbers not reported (USA, mean age 80 years);
-
Dunbar 2013 included 100 participants with diabetic retinopathy (UK, mean age 57 years);
-
Kaltenegger 2019 included 37 participants with AMD (Germany, median age 72 years);
-
Mielke 2013 included 20 participants with AMD (Germany, mean age 79 years);
-
Jackson 2017 included 37 participants of which 73% had AMD or juvenile MD (USA, mean age 72 years);
-
Leat 2017 included 14 participants with AMD (Canada, mean age 82 years);
-
Patodia 2017 included 16 participants with glaucoma (Canada, mean age not reported);
-
Taylor 2017 included 100 participants of which 61% had AMD and 39% other vision disorders (UK, mean age 71 years);
-
Pearce 2011 included 120 low vision participants (eye disorders not reported, UK, mean age 73 years);
-
Scanlan 2004 included 64 participants with AMD (Canada, mean age 81 years);
-
Seiple 2011 included 30 participants with AMD in the intervention groups (USA, median age 79 years) and 6 in the control group (mean 78.5 years);
-
Smith 2005 included 243 participants with AMD (UK, mean age 81 years);
-
Stelmack 2017 randomised 323 participants with various types of macular disease of which only 3% was female (USA, mean age 80 years).
Among the multidisciplinary rehabilitation programmes, the number of participants randomised in the trials ranged from 30 to over 300 adult participants. Also in the multidisciplinary intervention studies, AMD was often the most common cause of low vision. Studies reported between 2% and 66% females among their populations. There were four American studies, two from the UK and one from India. Except for the Indian study, all studies focused on (older) adults.
-
Acton 2016 included 71 participants of whom 58% had AMD (UK, mean age 75 years);
-
Christy 2012 included 436 participants of whom 347 were aged 16 years or older, AMD was the most common cause of low vision (32%, India);
-
Draper 2016 included 55 participants with low vision, eye disorders not reported (USA, mean age approximately 64 years);
-
McCabe 2000 included 97 participants of whom 64% had AMD (USA, mean age 69 years, median 76);
-
Pankow 2004 included 30 participants who most frequently had AMD (USA, mean age 78 years);
-
Reeves 2004 included 226 participants with AMD (UK, median age 81 years);
-
Stelmack 2008 included 126 participants (98% male) mostly with macular disease (USA, mean age 79 years).
There were three studies focusing on other programmes related to low vision rehabilitation:
-
Gleeson 2014 randomised 120 visually impaired participants of whom 64% were female (USA, mean age 77.5 years);
-
Pinninger 2013 included 17 female participants with AMD (Australia, mean age 79 years);
-
Waterman 2016 included 49 participants with various eye diseases of which 57% had AMD (UK, mean age 81 years).
Types of outcomes
Main outcome measures
There were several QOL outcome measures used. For thepsychological therapies or group programmes, or both, potential comparisons could be made on HRQOL between Bradley 2005 who used the 12‐item Well‐Being questionnaire, Brody 1999 with the Quality of Well‐Being scale which is a comprehensive measure of HRQOL that includes functional scales for mobility, physical activity and social activity, Eklund 2008 who used a general health measure, stating diseases and Girdler 2010 who used the well‐known Short Form‐36 (SF‐36). In addition, Nollet 2016, Tey 2019 and Van der Aa 2015 used the Euroqol‐5 dimensions (EQ‐5D). On VRQOL and visual functioning, potential comparisons could be made between Brody 2002, Bryan 2014, Nollet 2016, Rovner 2007, Rovner 2013 and Rovner 2014 who all used the National Eye Institute Visual Functioning Questionnaire of different lengths (NEI‐VFQ‐ 7, 17 and 25 items). In addition, Rovner 2013 and Rovner 2014 used the Activity Inventory, Bradley 2005 used the MacDQOL, Nollet 2016 used the LV‐VFQ‐48 for reading ability, Rees 2015 and Tey 2019 the Impact of Vision Impairment (IVI) and Van der Aa 2015 the Low Vision Quality of Life questionnaire (LVQOL).
For methods of enhancing vision , potential comparisons could be made on HRQOL between Brunnström 2004 who used the Psychological and General Well‐Being scale (PGWB) which included items on well‐being, vitality and depression, Burggraaff 2012, Stelmack 2017 and Taylor 2017 who used the EQ‐5D, Stelmack 2017 who also used the SF‐36, and Taylor 2017 the WHO‐5. On VRQOL and visual functioning, potential comparisons could be made between Burggraaff 2012 who used the LVQOL, Leat 2017, Mielke 2013, Scanlan 2004 and Smith 2005 who used the NEI‐VFQ or the NV‐VFQ‐15 or Taylor 2017 who also used the VISQOL to measure VRQOL. In addition, Pearce 2011 and Dunbar 2013 used the Activity Inventory, Burggraaff 2012 part of the Dutch ICF Activity Inventory, and Jackson 2017 the reading subscale of the Activity Inventory, however, they also used the IVI; Kaltenegger 2019; Patodia 2017, Seiple 2011 and Stelmack 2017 used the Veterans Affairs VFQ‐48.
For themultidisciplinary rehabilitation programmes , potential comparisons could be made on HRQOL between Acton 2016 who used the EQ‐5D, Christy 2012 who used the World Health Organisation Quality Of Life questionnaire (WHOQOL), and Reeves 2004 and Stelmack 2008 who both used the SF‐36. On VRQOL and visual functioning, potential comparisons could be made between Christy 2012 who used the Impact of Visual Impairment (IVI), Reeves 2004 who used the Vision‐Related Quality of Life Core Measure (VCM1), Acton 2016; Draper 2016 and Stelmack 2008 who used the Veterans Affairs Visual Functioning Questionnaire‐48 (VA‐VFQ‐48) and McCabe 2000 who used the Functional Assessment Questionnaire which measures visual function and overall well‐being.
For the other programmes , measurements of VRQOL were made by Pinninger 2013 who used the NEI‐VFQ, Gleeson 2014 who used the IVI emotional well‐being subscale and Waterman 2016 who used the VCM1 and also the SF‐12 for a measure of HRQOL.
Secondary outcome measures
Secondary outcomes were patient‐reported outcomes closely related to QOL, concerning concepts of health and well‐being. With respect to physical and functional measures, examples were the Functional Independence Measure for Blind Adults (FIMBA), included by Pankow 2004 as an assessment of the participant's ability to perform daily tasks, orientation and mobility skills independently; the Dependency in ADL levels by Eklund and colleagues (2008; seeEklund 2008); and the Melbourne Low Vision ADL (MLVA) which consisted of questions on the performance of specified typical activities of daily living (not) dependent on vision, used by Smith 2005. The Perceived Visual Ability Scale (PVAS) was used by Gleeson 2014 for measuring mobility difficulty for a range of situations. The Health and Impact Questionnaire, used by Brody 1999 and Brody 2002, is an instrument that obtains, for example, questions about medical history, current medical conditions, medications and the impact of AMD. Brunnström 2004 used an interview in which ADL‐type questions were posed. Finally, Taylor 2017 used the five timed instrumental activities of daily living questionnaire (5‐TIADL).
The questionnaire for measuring perceived security in performing daily activities was developed as a primary outcome for the purpose of evaluating the health education programme of Eklund 2008. It measures participants' perceived security in performing specified tasks. Girdler 2010 used the Activity Card Sort as a primary outcome to measure participation levels.
Psychological measures, for example, to measure depression, mood disorders and anxiety, were as follows: the centre for Epidemiological Studies – Depression scale (CES‐D, used by Bryan 2014, Burggraaff 2012; Mielke 2013; Stelmack 2008, Van der Aa 2015), the Geriatric Depression Scale (GDS; Girdler 2010; Gleeson 2014, Mielke 2013; Nollet 2016; Pinninger 2013; Leat 2017 which was used as a primary outcome measure in the study by Brody as well (seeBrody 2002), the Beck Depression Inventory‐II (also used by Nollet 2016), the Patient Health Questionnaire‐9 (PHQ‐9 used by Acton 2016; Kamga 2017; Rovner 2013 and Rovner 2014), the Montgomery–Åsberg Depression Rating Scale (MADRS used by Kaltenegger 2019), the Profile Of Mood States (POMS) for emotional distress, a psychological strain questionnaire (used by Brody 1999; Brody 2002 as a primary outcome measure), the Warwick Edinburgh Mental Well‐being Scale by Acton 2016, the Perceived Stress Scale (PSS) by Bryan 2014, the Life Orientation Test‐Revised (LOT‐R; used by Brody 2002) which assesses an optimistic versus a pessimistic life outlook and the Mood Episodes section and the Global Assessment of Function Scale of the SCID‐IV (Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM‐IV) used by Brody 2002. Rovner 2007 assessed depressive disorders with the Schedule for Affective Disorders and Schizofrenia and the Structured Interview Guide for the Hamilton Depression Rating Scale. Van der Aa 2015 also used a diagnostic interview, namely the Mini International Neuropsychiatric Interview which is based on the DSM‐IV. Jackson 2017; Mozaffar Jalali 2014, Rees 2015 and Tey 2019 measured depression, anxiety and stress symptoms with the DASS. The study of Van der Aa 2015 also focused on anxiety, which was measured with the Hospital Anxiety Depression Scale ‐ Anxiety subscale (HADS‐A); Kamga 2017 measured generalised anxiety disorder with the GAD‐7.
Adaptation to vision loss was measured with the Adaptation to age‐related Vision Loss questionnaire (AVL, used by Acton 2016; Burggraaff 2012; Christy 2012; Girdler 2010; Rees 2015; Van der Aa 2015) and the Nottingham Adjustment Scale (NAS, used by Pankow 2004 and Reeves 2004).
Self‐efficacy and self‐esteem were measured with the Macular Degeneration Self‐Efficacy Questionnaire (AMD‐SEQ) which evaluates the degree of self‐confidence in the individual’s ability to handle situations related to AMD (used by Brody 1999 and Brody 2002 and adapted and used by Girdler 2010 to reflect age‐related vision loss). Mozaffar Jalali 2014 used Eysenck's self‐esteem questionnaire. Girdler 2010, Rees 2015 and Tey 2019 used the Generalised Self‐Efficacy Scale (GSES) and Kamga 2017 and Pinninger 2013 a Self‐Esteem Scale (SES). Rumrill 1999 used the Accommodation Self‐efficacy measure and an Accomodation Activity scale to assess a participant’s self‐efficacy in the accommodation request process. Sheres Self‐Efficacy Scale was designed to measure general expectancies. Conrod 1998 used a ‘beliefs questionnaire’ and an ‘expectations questionnaire’ to assess functional effects and attitudes toward low vision and blindness and whether expectations about the future play a meaningful role in the adjustment process. Goldstein 2007 measured emotional responses (sadness, fear, confusion, peace) to low vision and self‐efficacy with regard to obtaining and using assistive devices. Kaluza 1996 measured psychological strain with the Kurzfragebogen zur Aktuellen Beanspruchung (KAB). Mozaffar Jalali 2014 used the Jones irrational beliefs questionnaire. Pinninger 2013 used the Satisfaction With Life scale (SWL) which measures global perceptions of well‐being and contentment with life.
Social functioning was measured with the Duke Social Support Index (DSSI) which measures satisfaction in terms of frequency, content and quality of support and social interaction with family and friends (Brody 2002). Acton 2016 used the University of California Los Angeles Loneliness Scale, Bryan 2014 used a the Medical Outcomes Study Social Support Scale and Brunnström 2004 used a one‐item measure of loneliness. Gleeson 2014 measured participation with Keele's assessment of participation (KAP) and socialisation which was defined as the number of contacts with others. Rumrill 1999 used the Work Experience Survey which measures career maintenance barriers in people with disabilities.
Length of follow‐up, direct and maintenance effects
Short‐term follow‐up included measurements directly or within one month after treatment, one month after baseline (Jackson 2017; Mozaffar Jalali 2014; Patodia 2017 and Pinninger 2013), six weeks after baseline (Bradley 2005 and Leat 2017, whereas for Kaltenegger 2019 this measurement was used in the study before cross‐over of the control group to treatment), eight weeks after baseline (Kaluza 1996 and Kamga 2017), eight to nine weeks after baseline (Conrod 1998), 10 weeks after baseline which was within four weeks after the intervention (Brody 1999), two to three months after baseline (Draper 2016), four months after baseline (Stelmack 2008 and Taylor 2017), 18 weeks after baseline (Seiple 2011), six months after baseline (Sun 2012 and Waterman 2016), and, for McCabe 2000, the follow‐up measurement seemed to occur directly after one to five intervention visits but the length of follow‐up was not reported.
There were also short‐term follow‐up measurements reported with maintenance effects of between one and six months: the maintenance period was approximately four weeks after intervention for the study by Girdler 2010 with follow‐up 12 weeks after baseline, four to six weeks (Pankow 2004, 12 to 14 weeks after baseline), six weeks (Bryan 2014, seven weeks after baseline), seven weeks (Scanlan 2004, 12 weeks after baseline), approximately two to two and a half months (Burggraaff 2012; Dunbar 2013; Mielke 2013 and Pearce 2011, three months after baseline; Rovner 2014 and Rumrill 1999, four months after baseline), three to three and a half months (Goldstein 2007, three months after baseline and Stelmack 2017, four months after baseline), approximately four months (Acton 2016; Nollet 2016; Rovner 2007 and Rovner 2013, six months after baseline), approximately four and a half months (Brody 2002, six months after baseline) and six months post‐intervention and baseline (Brunnström 2004 and Tey 2019).
Long‐term follow‐up consisted of maintenance effects of four to five months post‐intervention (Christy 2012, nine months after baseline), six months following treatment (Rees 2015, approximately eight months after baseline), six to nine months post‐intervention (Reeves 2004, 12 months after baseline), nine months post‐intervention (Gleeson 2014, 12 months after baseline) and 28 months after intervention (Eklund 2008, approximately 30 months after baseline). Finally, Van der Aa 2015 produced a maintenance effect of at least 12 months following intervention, which was 24 months after baseline and included an intervention period of between three and 12 months.
Excluded studies
Of the 163 potentially useful full papers that were screened, there were 72 reports of 65 studies that were subsequently excluded mainly because they were RCTs or studies with other designs on topics that were not the focus of this review (e.g. medical or vision restoration therapy), or because of study designs: nonrandomised controlled studies (NRCTs), observational studies without comparison groups, of which two studies were written in Japanese, literature reviews, and other reasons such as irrelevant population (children) or intervention or questionnaire development articles; for some of the excluded studies that might be of interest, see: Characteristics of excluded studies table.
Furthermore, some studies were excluded because they did not measure any of the prespecified review outcomes, but the interventions could be interpreted as relevant, therefore, an additional table is included, see Table 1. For example, the quasi‐RCT of Culham 2009 was excluded, because it was set up to compare reading speed and satisfaction with different head mounted devices. Although a Visual Functioning questionnaire was used as a secondary outcome, the questionnaire was modified and the design of the study was not set up to properly assess VRQOL. Moreover, the study has been evaluated in the Cochrane Review by Virgili 2018 in its main outcomes, including a 'Risk of bias' assessment. The study by La Grow 2004 was excluded because of the matching of the intervention and a contrast group; randomisation was not possible. Participants in the contrast group were recommended by ophthalmic and optometric practices, living in areas where low vision services were not available.
Ongoing studies
Ongoing studies
There were 14 study protocols found of ongoing randomised studies that could not be used in the current review, but should possibly be included in the future update.
Risk of bias in included studies
The results of the quality assessment are described for all parameters which are shown in the Characteristics of included studies and 'Risk of bias' tables and which are summarised in Figure 2 and Figure 3. Almost all studies lacked information to properly assess risk of bias for some of the criteria. However, from the 20 psychological therapies, or group programmes, or both (type I interventions), Rovner’s studies (Rovner 2007; Rovner 2013; Rovner 2014), but also Van der Aa 2015 performed well on most risk of bias criteria. Considering the 14 studies in which methods of enhancing vision (type II interventions) were described, the studies by Burggraaff 2012; Dunbar 2013; Smith 2005 and Stelmack 2017 performed well. With respect to the seven multidisciplinary rehabilitation programmes (type III interventions), Reeves 2004 and Stelmack 2008 performed best and also from the other programmes (type IV interventions) Gleeson 2014 performed very well.
Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study. Studies were categorized into: (I) psychological therapies and/or group programs; (II) methods of enhancing vision; (III) multidisciplinary rehabilitation programs.
Not all studies reported on funding, but those that did were supported by public or non‐profit funders (N = 31), except for two studies which received low vision aids by commercial producers (Jackson 2017; Pankow 2004).
Allocation
In most studies (N = 27), the random sequence generation was considered to be good, because random number tables, or computer random number generators or other low tech methods were used to randomise participants (e.g. numbers drawn from a hat). In one study, risk for selection bias was assessed as high; Rumrill 1999 used a matching procedure in which first pairs of two participants were made, after which randomisation took place where one of a pair would be in the intervention group and the other in the control group. In other studies, methods were unclear.
In more than half of the studies (N = 24), allocation concealment was not reported, however, in the other 20 studies, it was reported and considered to be appropriately done e.g. by central allocation methods or use of sealed opaque envelopes.
Blinding
In the field of low vision, most trials had a pragmatic nature in which masking of participants and personnel was not possible. Therefore, these studies were assessed as having an unclear risk (N = 40). Still, for one study masking was possible and was done properly (Smith 2005). The study investigators randomised three types of prism glasses, of which one placebo was made similar in weight. Participants and the investigator doing the analyses were not aware which glasses were prescribed. In two studies, however, masking could not have been performed, and there could have been contamination between trial interventions. I.e. personnel were trained in the different treatment options in the study of Conrod 1998, which means personnel could have conveyed elements of treatment to the control group. Another example is the study by Pankow 2004, where researchers were also trainers of the intervention and outcome assessors at the same time.
In more than half of the studies (N = 24), the risk of detection bias was assessed as low risk, because masking of assessors was reported and performed accurately. In some studies, participants were notified that they should not reveal their allocation concealment. To gain insight into possible detection bias, outcome assessors were sometimes asked to guess in which treatment arm the participant would be (e.g. Rovner 2013; Van der Aa 2015). In 13 studies, the risk of detection bias was unclear and in another seven studies, it was considered to be high. Goldstein 2007 and Pankow 2004 reported that outcome assessors were not masked to group assignment; Pinninger 2013 reported that volunteers assisted in completion of the pre‐ and post‐test questionnaires; for the treatment arm, this meant an assessment at the treatment location, whereas the control group was assessed by telephone. Rumrill 1999 and Scanlan 2004 reported that trainers were also outcome assessors, which may stimulate socially desirable answering and may influence unbiased reporting of participants.
Incomplete outcome data
In 24 studies, the risk of attrition bias was considered to be low, because follow‐up rates and compliance were similar in the groups, analyses were often based on the intention‐to‐treat principle and attrition was limited. For 10 studies, attrition bias was unclear, but another 10 studies seemed to have a high risk. In these studies, a high risk assessment was often due for multiple reasons. Per‐protocol or as‐treated analyses were performed in Brody 1999, Mielke 2013 and Scanlan 2004 and possibly in Conrod 1998 and Kaluza 1996. In addition, Bryan 2014, Conrod 1998, Jackson 2017, Kaltenegger 2019, Leat 2017 and Mielke 2013 reported a relatively high attrition rate. Finally, Scanlan 2004 replaced participants with deteriorating vision, which can be considered a design flaw. Although it is difficult to indicate to what magnitude, missingness in the data analysis may have influenced the means and SDs of single studies and subsequently the SMD of the pooled outcomes.
Selective reporting
In 15 studies, protocols or trial registrations revealed that predefined outcomes of interest were assessed. In 26 studies, it was unclear if all relevant outcomes were assessed and in four studies, a high risk was expected because e.g. the outcome measure was not completely reported or other durations of follow‐up were planned. Contacting the authors did not provide us with an answer.
Effects of interventions
See: Summary of findings for the main comparison Psychological therapies and/or group programmes compared to waiting list or no care for better quality of life in visually impaired adults; Summary of findings 2 Methods of enhancing vision compared to waiting list or no care for better quality of life in visually impaired adults; Summary of findings 3 Multidisciplinary rehabilitation compared to waiting list or no care for better quality of life in visually impaired adults; Summary of findings 4 Psychological therapies and/or group programmes compared to active control for better quality of life in visually impaired adults; Summary of findings 5 Methods of enhancing vision and/or group programmes compared to active control for better quality of life in visually impaired adults; Summary of findings 6 Multidisciplinary rehabilitation compared to active control for better quality of life in visually impaired adults
Low vision rehabilitation versus waiting list or no care
Fifteen studies provided data for this comparison. The effects on the primary outcomes are shown in Figure 4 and Figure 5.
Forest plot of comparison: 1 Low vision rehabilitation versus waiting list or no care, outcome: 1.1 Health‐related quality of life.
Forest plot of comparison: 1 Low vision rehabilitation versus waiting list or no care, outcome: 1.2 Vision‐related quality of life.
Primary outcomes
Health‐related quality of life
Three studies investigated the effect of a psychological or multidisciplinary interventions against an inactive comparator on the HRQOL outcome (Analysis 1.1).
Nollet 2016 (54 participants) investigated PST (group I) and compared it to a referral to the general practitioner; the short‐term maintenance effect was imprecise and did not suggest a benefit (standard mean difference (SMD) 0.26, 95% confidence interval (CI) ‐0.28 to 0.80; participants = 54; studies = 1). Two studies investigated multidisciplinary rehabilitation (group III). Stelmack 2008 (114 participants) evaluated an outpatient multidisciplinary low vision rehabilitation programme for veterans with moderate and severe vision loss due to macular disease (LOVIT) and Acton 2016 (67 participants) evaluated the effect of home visits from a visual rehabilitation officer. Both found effects (i.e. short‐term direct and maintenance, respectively) which favoured rehabilitation, but estimates were very imprecise and included no effect (SMD ‐0.08, 95% CI ‐0.37 to 0.21; participants = 183; studies = 2; I2 = 0%). This evidence was of very low‐certainty due to imprecision (‐2) and risk of bias (‐1) for both comparisons.
Vision‐related quality of life
Nine studies investigated the effect of psychological therapies, methods for enhancing vision and multidisciplinary interventions against an inactive comparator on the VRQOL outcome (Analysis 1.2).
Two studies investigated group I interventions (psychological therapies and/or group programmes), i.e. Brody 2002: short‐term direct effects of a self‐management group programme; Nollet 2016: short‐term maintenance effects of PST. We found low‐certainty evidence of small and imprecisely estimated benefit (SMD ‐0.23, 95% CI ‐0.53 to 0.08; studies = 2; I2 = 24%).
Five studies investigated group II interventions (methods of enhancing vision), i.e. Dunbar 2013: rehabilitation in a hospital‐based low vision centre; Kaltenegger 2019: reading (self‐)training at home; Mielke 2013: professional rehabilitation with prescription of magnifying visual aids and training; Patodia 2017: prescription of low vision devices and training; Smith 2005: prism spectacles in AMD participants. All effects were short‐term maintenance effects, except for Patodia 2017 which investigated short‐term direct effects. The meta‐analysis suggested a small benefit but effects were moderately heterogenous and imprecisely estimated including no benefit (SMD ‐0.19, 95% CI ‐0.54 to 0.15; participants = 262; studies = 5; I2 = 34%). This evidence was of very low‐certainty due to imprecision (‐2), and risk of bias (‐1).
Two studies investigated group III interventions (multidisciplinary rehabilitation programmes) and yielded heterogeneous short‐term maintenance effects. Stelmack 2008 (126 participants) found a large and significant benefit in the LOVIT study (SMD: ‐1.64, 95% CI ‐2.05 to ‐1.24), which was at low risk of bias and used an intensive rehabilitation programme. Acton 2016 (67 participants) evaluated the effect of home visits from a visual rehabilitation officer and found a benefit but estimates were imprecise and included no effect (SMD ‐0.42, 95%: ‐0.90 to 0.07); the study was at unclear risk of bias regarding at least allocation concealment and estimates were imprecise. We did not consider the pooled estimate of these two studies because their confidence intervals did not overlap (P < 0.0001 for inconsistency). The two studies had domains at unclear, but not high risk of bias.
Secondary outcomes
Physical aspect of quality of life
Two studies investigated the effect on activities of daily living as a 'physical aspect' of QOL (Analysis 1.3) and against an inactive comparator.
In group II, Smith 2005 (151 participants, short‐term maintenance effect) evaluated the effect of prism spectacles (methods of enhancing vision) in AMD participants and found no benefit (SMD ‐0.06; 95%CI ‐0.38 to 0.26). In group III, Pankow 2004 (30 participants, short‐term maintenance effect of multidisciplinary rehabilitation) delivered a vision rehabilitation programme which included orientation and mobility training, blind rehabilitation teaching, such as occupational therapy and a low vision evaluation by optometrists and social workers, all depending upon the participants’ needs and found no difference, but estimates were imprecise (SMD 0.07; 95%CI ‐0.64 to 0.79). The certainty of evidence was very low for both comparisons (‐2 for imprecision, ‐1 for risk of bias).
Psychological aspect of quality of life
Nine studies investigated the effect of the intervention groups on depressive symptoms (Analysis 1.4), five studies on self‐esteem (Analysis 1.5) and two studies on adaptation to vision loss Analysis 1.6 as a 'psychological aspect' of QOL and against an inactive comparator.
Depression
Five studies assessing group I psychological therapies and/or group programmes on depression (Brody 1999 and Brody 2002: short‐term direct and maintenance effects, respectively, of self‐management programmes; short‐term direct effects by Kaluza 1996: stress‐reducing intervention, and Mozaffar Jalali 2014: group‐based rational emotive behavioural therapy; and short‐term maintenance effect of PST by Nollet 2016) found a large and significant, but imprecise benefit (SMD ‐1.23, 95% CI ‐2.18 to ‐0.28; participants = 456; studies = 5; I2 = 94%); effects were heterogenous and very large in one outlying study (Mozaffar Jalali 2014). The certainty of this evidence was very low (‐1 for risk of bias, ‐2 for imprecision, ‐1 for inconsistency). However, after removing the outlier which was a study with mainly unclear risk of bias, the effect was moderate and heterogeneity decreased substantively due to reasons that could not be explained (SMD ‐0.46, 95% CI ‐0.66 to ‐0.26; participants = 396; studies = 4; I2 = 0%), thereby also increasing the certainty of evidence to moderate (‐1 for risk of bias, ‐1 for imprecision).
Kaltenegger 2019 compared a reading (self‐)training at home with doing crossword puzzles as a placebo, and Mielke 2013 compared magnifying aids and training with a waiting list (both studies short‐term maintenance effect of group II methods for enhancing vision) and found a large but imprecisely estimated benefit (SMD ‐0.86, 95% CI ‐1.50 to ‐0.23; participants = 44; studies = 2; I2 = 0%); very low‐certainty of evidence for imprecision (‐1) and risk of bias (‐1).
Acton 2016 and Stelmack 2008 investigated the short‐term maintenance and direct effects, respectively, of group III multidisciplinary rehabilitation and found a small benefit which was imprecisely estimated and included no benefit (SMD ‐0.16, 95% CI ‐0.44 to 0.13; participants = 193; studies = 2; I2 = 0%). The certainty of this evidence was very low (‐1 for risk of bias, ‐2 for imprecision).
Self‐esteem
Five studies investigated the effect on self‐esteem (Analysis 1.5), all assessing group I psychological therapies and/or group programmes (Brody 1999; Brody 2002: short‐term direct and maintenance effects, respectively, of self‐management programmes; Goldstein 2007: short‐term maintenance effect of an educational video to help participants achieve self‐efficacy; Mozaffar Jalali 2014: short‐term direct effect of a group‐based rational emotive behavioural therapy; Rumrill 1999: short‐term maintenance effect of a social competence training). Overall effects were all in the direction of benefit but very heterogenous, thus the benefit was significant, but imprecisely estimated (SMD ‐0.85, 95% CI ‐1.48 to ‐0.22; participants = 550; studies = 5; I2 = 91%). The certainty of this evidence was very low (‐1 for risk of bias, ‐1 for imprecision, ‐1 for inconsistency). We were unable to find reasons for such high heterogeneity, but pooled the data as effects were in the direction of benefit for all studies.
Adaptation to vision loss
Acton 2016 and Pankow 2004 investigated the short‐term maintenance effect of group III multidisciplinary rehabilitation interventions on adaptation to vision loss (Analysis 1.6) and found an effect in the direction of benefit but very imprecisely estimated and included no benefit (SMD ‐0.14, 95% CI ‐0.54 to 0.26; participants = 97; studies = 2; I2 = 0%). The certainty of this evidence was very low (‐1 for risk of bias, ‐2 for imprecision).
Social aspect of quality of life
Although the studies by Acton 2016, Brody 2002, and Rumrill 1999 used some measure to investigate effects of an intervention versus an inactive comparator on a social aspect of QOL, these were not feasible for entering in a meta‐analysis.
Low vision rehabilitation versus active control group
Twenty‐three studies provided data for this comparison, in which interventions of interest were compared to usual care or other interventions. The effects on the primary outcomes are shown in Figure 6 and Figure 7.
Forest plot of comparison: 2 Low vision rehabilitation versus active comparator, outcome: 2.1 Health‐related quality of life.
Forest plot of comparison: 2 Low vision rehabilitation versus active comparator, outcome: 2.2 Vision‐related quality of life.
Primary outcomes
Health‐related quality of life
Nine studies investigated the effect of psychological therapies, methods for enhancing vision, multidisciplinary interventions or other interventions against an active comparator on the HRQOL outcome (Analysis 2.1).
There were four studies in group I psychological interventions: (Eklund 2008: long‐term maintenance effect of an educational programme by an occupational therapist versus standard low vision clinic; Girdler 2010 and Tey 2019: short‐term maintenance effects of a self‐management programme plus usual care versus usual care; Van der Aa 2015: long‐term maintenance effect of a stepped‐care intervention including cognitive‐behavioural therapy‐based self‐help and/or PST plus usual care versus usual care consisting of outpatient low vision rehabilitation care or any other (medical) care needed). Overall effects suggested small benefits, but were heterogenous and included no benefit (SMD ‐0.09, 95% CI ‐0.39 to 0.20; participants = 600; studies = 4; I2 = 67%) and larger in Eklund 2008 and Girdler 2010. The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
Two studies investigated short‐term maintenance effects of group II methods for enhancing vision (Burggraaff 2012: standardised CCTV training programme versus usual instruction by the supplier; Stelmack 2017: additional sessions on eccentric viewing, use of low vision devices, environmental modification plus low vision care versus low vision care only). Overall effects were modest and imprecisely estimated, including no benefit (SMD ‐0.09, 95% CI ‐0.28 to 0.09; participants = 443; studies = 2; I2 = 0%). The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
Two studies investigated long‐term maintenance effects of group III multidisciplinary rehabilitation (Christy 2012; different methods of multidisciplinary low vision service delivery versus centre‐based non‐interventional community visits; Reeves 2004: home visits from a rehabilitation officer in addition to conventional low vision rehabilitation versus conventional low vision rehabilitation only). The results suggested a small or no benefit (SMD ‐0.10, 95% CI ‐0.31 to 0.12; participants = 375; studies = 2; I2 = 0%). The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
Waterman 2016 (group IV other programmes, 43 participants) investigated short‐term direct effects of a home safety programme by occupational therapist with usual care and found an imprecise effect which did not suggest a benefit (SMD ‐0.05, 95% CI ‐0.70 to 0.60; very low‐certainty evidence).
Vision‐related quality of life
Nineteen studies investigated the effect of psychological therapies, methods for enhancing vision, multidisciplinary interventions or other interventions against an active comparator on the VRQOL outcome (Analysis 2.2).
Seven studies investigated group I psychological therapies and/or group programmes where short‐term maintenance effects were studied by Bryan 2014: expressive writing intervention versus non‐emotional topic writing task; Rovner 2007: PST versus usual care by an ophthalmologist; Rovner 2013: PST versus standardised psychological nondirective treatment; Rovner 2014: behavioural activation versus standardised psychological nondirective treatment; and long‐term maintenance effects by Rees 2015: low vision self‐management group programme; Tey 2019: self‐management programme, and Van der Aa 2015: stepped‐care intervention including cognitive‐behavioural therapy‐based self‐help plus usual care versus usual care consisting of outpatient low vision rehabilitation care). Overall, a small benefit was found (SMD ‐0.11, 95% CI ‐0.24 to 0.01; participants = 1245; studies = 7; I2 = 19%). The certainty of this evidence was low due to risk of bias (‐1) and imprecision (‐1).
Seven studies investigated group II methods for enhancing vision (short‐term maintenance effects by Burggraaff 2012: of a standardised CCTV training programme versus usual instruction by the supplier; Stelmack 2017: additional sessions on eccentric viewing, use of low vision devices, environmental modification plus low vision care versus low vision care only, and Pearce 2011: low vision device training on top of consultation with an optometrist for conventional low vision assessment versus placebo treatment by a nurse; and short‐term direct effects by Draper 2016: home‐based versus clinical‐based rehabilitation; Jackson 2017: provision of CCTV in addition to optical aids versus optical aids only; Leat 2017: CCTV provision and training versus eccentric fixation training; and Taylor 2017: use of portable electronic devices on top of optical devices versus optical devices only). The meta‐analysis of these studies suggested small, but significant benefits (SMD ‐0.24, 95% CI ‐0.40 to ‐0.08; participants = 660; studies = 7; I2 = 3%). The certainty of this evidence was moderate due to risk of bias (‐1), whereas we did not downgrade for imprecision since the confidence interval was consistent with small effects..
Three studies investigated group III multidisciplinary rehabilitation (long‐term maintenance effects by Christy 2012: different methods of multidisciplinary rehabilitation versus centre‐based services with non‐interventional community visits; and Reeves 2004: home visits from a rehabilitation officer in addition to conventional low vision rehabilitation versus conventional low vision rehabilitation only; and short‐term direct effects by McCabe 2000: involvement of families in addition to individual outpatient vision rehabilitation versus individual rehabilitation only). Overall effects were not beneficial (SMD 0.01, 95% CI ‐0.18 to 0.20; participants = 464; studies = 3; I2 = 0%). The certainty of this evidence was low due to risk of bias (‐1) and imprecision (‐1).
Among group IV, other types of interventions, Gleeson 2014 (120 participants) as a long‐term maintenance effect compared the Alexander technique to improve balance in people with vision loss with usual Guide Dog care (charity foundation). Waterman 2016 investigated a short‐term direct effect of a home safety programme with and without exercise versus social visits. In both studies, the benefits found were very imprecisely estimated (very low‐certainty evidence).
Secondary outcomes
Physical aspect of quality of life
Three studies investigated the effect of psychological therapies and/or group programmes, or other interventions on activities of daily living as a 'physical aspect' of QOL (Analysis 2.3) against an active comparator.
Two studies used group I psychological therapies and/or group programmes (Eklund 2008; long‐term maintenance effect of a health education programme versus standard low vision programme: Girdler 2010: short‐term maintenance effect of a self‐management intervention plus usual care versus usual care which was individual case management model and multidisciplinary work). An overall significant moderate effect was recorded (SMD ‐0.39, 95% CI ‐0.67 to ‐0.12; participants = 208; studies = 2; I2 = 0%). The certainty of this evidence was low due to risk of bias (‐1) and imprecision (‐1).
Among other types of interventions, Gleeson 2014 (120 participants) investigated a long‐term maintenance effect of the Alexander technique to improve balance in people with vision loss with usual low vision care delivered by Guide Dogs (charity foundation) and found no benefits compared to low vision care.
Psychological aspect of quality of life
Several studies investigated the effect of the intervention groups on depressive symptoms (Analysis 2.4), on self‐efficacy or self‐esteem (Analysis 2.5) and on adaptation to vision loss (Analysis 2.6) as a 'psychological aspect' of QOL and against an active comparator.
Depression
Nine studies investigated group I psychological therapies and/or group programmes on the effects on depression (short‐term maintenance effects by Bryan 2014: expressive writing intervention versus non‐emotional topic writing task; Kamga 2017: short‐term direct effect of a cognitive behavioural therapy‐based self‐care tool versus usual care; Girdler 2010 and Tey 2019: self‐management programmes plus usual care versus usual care; Rovner 2007: PST versus usual care by an ophthalmologist; Rovner 2013: PST versus standardised psychological nondirective treatment; Rovner 2014: behavioural activation versus standardised psychological nondirective treatment; and long‐term maintenance effects of Rees 2015: low vision self‐management group programme; and Van der Aa 2015: stepped‐care plus usual care compared to usual care). Overall results were suggestive of small benefits that were imprecisely estimated (SMD ‐0.14, 95% CI ‐0.25 to ‐0.04; participants = 1334; studies = 9; I2 = 0%). The certainty of this evidence was moderate due to risk of bias (‐1).
There were two studies in the group II methods for enhancing vision (short‐term maintenance effect by Burggraaff 2012: standardised CCTV training programme versus usual instruction by the supplier and a short‐term direct effect by Jackson 2017: provision of CCTV in addition to optical aids versus optical aids only). Results suggested some benefits but were imprecisely estimated. (SMD ‐0.22, 95% CI ‐0.59 to 0.15; participants = 162; studies = 3; I2 = 12%). The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
In group IV, other programmes, Gleeson 2014 (120 participants) compared the Alexander technique to improve balance in people with vision loss with usual Guide Dog care (charity foundation) and found no difference (very low‐certainty evidence).
Self‐efficacy or self‐esteem
Four studies investigated group I psychological therapies and/or group programmes on the effects on self‐efficacy or self‐esteem (Analysis 2.5): Girdler 2010 and Tey 2019: short‐term maintenance effects of self‐management interventions plus usual care versus usual care; Kamga 2017: short‐term direct effect of a cognitive behavioural therapy‐based self‐care tool versus usual care; Rees 2015: long‐term maintenance effects of a low vision self‐management group programme versus multidisciplinary team). Results suggested a small and imprecisely estimated benefit and included no benefit (SMD ‐0.06, 95% CI ‐0.26 to 0.15; participants = 427; studies = 4; I2 = 13%). The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
In the group III, multidisciplinary programmes, Reeves 2004 compared home visits from a rehabilitation officer in addition to conventional low vision rehabilitation versus conventional low vision rehabilitation as a long‐term maintenance effect and found no benefit (very low‐certainty evidence).
Adaptation to visual loss
Six studies investigated the effect of several interventions on adaptation to visual loss (Analysis 2.6) against an active comparator.
Three studies investigated group I psychological therapies and/or group programmes (Girdler 2010: short‐term maintenance effect of a self‐management intervention plus usual care versus usual care which was an individual case management model and multidisciplinary work; and long‐term maintenance effects by Rees 2015: low vision self‐management group programme versus multidisciplinary team; and Van der Aa 2015: stepped‐care plus usual care compared to usual care). The meta‐analysis suggested a small or no benefit on this dimension (SMD ‐0.11, 95% CI ‐0.28 to 0.07; participants = 495; studies = 3; I2 = 0%). The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
One study in group II, methods for enhancing vision group, which investigated short‐term maintenance effects (Burggraaff 2012: standardised CCTV training programme versus usual instruction by the supplier) found no benefit (very low‐certainty evidence).
Two studies in group III, multidisciplinary rehabilitation, both investigated long‐term maintenance effects (Christy 2012: different methods of multidisciplinary rehabilitation versus centre‐based services with non‐interventional community visits; Rees 2015: low vision self‐management group programme versus multidisciplinary team) and found no additional benefit (SMD ‐0.02, 95% CI ‐0.24 to 0.19; participants = 376; studies = 2; I2 = 0%). The certainty of this evidence was very low due to risk of bias (‐1) and imprecision (‐2).
Social aspect of quality of life
Although the studies by Brunnström 2004, Bryan 2014, and Gleeson 2014 used some measure to investigate effects of an intervention versus an active comparator on a social aspect of QOL, these were not feasible for entering in a meta‐analysis.
Adverse events and serious adverse events
Only a few studies reported data on adverse events (AEs) or serious adverse events (SAEs). Nollet 2016 reported two AEs of which one was related to pharmacological treatment of depression rather than to rehabilitation and 10 SAEs which were unrelated to study interventions. Acton 2016 reported 19 AEs that were probably unrelated to treatment, mainly worsening of depression. Stelmack 2017 reported 10 AEs that were not treatment‐related. Stelmack 2008, Kamga 2017, Waterman 2016, Burggraaff 2012 and Van der Aa 2015 reported no (S)AEs that were treatment‐related. For all other studies, no data were available.
Discussion
Summary of main results
In this review, we have taken a broad perspective to map and summarise the evidence from RCTs in which several types of rehabilitation interventions were evaluated and which aimed to improve QOL of adults with low vision. We adopted both HRQOL and VRQOL as primary outcomes, because general and disease‐specific measures are both used across medical specialities in order for policy makers to make informed decisions about resources.
The interpretation of the results of this review are complicated by the fact that low vision rehabilitation is not a standard process, as interventions are highly individually tailored and can vary in different settings, where a mixture of different optometric or therapeutic components are used. For this reason, we grouped the study interventions in four broad categories. Unlike several other Cochrane Reviews, which present results by comparison, in the main text we have presented effects on an outcome basis, which allows us to explore the consistency of the effects of several types of low vision rehabilitation intervention on each QOL, VRQOL or related outcome. In order to offer a different perspective, the following summary will be comparison‐based, that is, the effects for all outcomes will be described for each comparison.
Psychological therapies and/or group programmes compared to no care showed an effect towards improvement of VRQOL, but this estimate was not sufficiently precise to exclude small benefit and the evidence was of very low‐certainty. No effect was found in a single, small study on VRQOL. Instead, a large effect was found on depression and self‐efficacy or self‐esteem but these were imprecisely estimated, although significant, and the certainty of the evidence was very low. Psychological therapies and/or group programmes compared to usual care or active control showed a small benefit on HRQOL and VRQOL with imprecise estimates that may have included no benefit, with low‐certainty evidence. A small effect of moderate‐certainty was found on depression and low‐certainty evidence of some benefit was also found on activities of daily living. Effects on self‐efficacy or self‐esteem were still positive but very uncertain and with low‐certainty of evidence.
Methods for enhancing vision compared to no care showed moderate‐certainty evidence of small effects for VRQOL and very low‐certainty evidence of effects directed towards benefit for HRQOL and depression but estimates were imprecise. Methods for enhancing vision compared to usual or other care also showed effects towards benefit for HRQOL, VRQOL and depression, but estimates were very imprecise with low‐ or very low‐certainty of evidence.
Multidisciplinary rehabilitation compared to no care showed no effect on HRQOL, and two heterogeneous studies showed a large effect on VRQOL, where the larger trial at low risk of bias was using intensive treatment; the small trial showed a small effect on VRQOL. Imprecise estimates of benefit were detected for adaptation to vision loss and no effect on activities of daily living. All evidence was of low‐ or very low‐certainty. Multidisciplinary rehabilitation compared to usual or other care showed imprecisely estimated effects in the direction of benefit for HRQOL and no benefit for VRQOL, for self‐efficacy or self‐esteem and adaptation to vision loss. All evidence was of low‐ or very low‐certainty.
Three heterogeneous studies evaluated other programmes and readers should refer to each study for this evidence.
The summary made above has no direct interpretation due to the heterogeneity in the interventions, even within subgroups, but it suggests that small effects of several types of low vision rehabilitation may exist for different QOL and related outcomes, especially VRQOL and depression.
The effect size of low vision rehabilitation was small (below 0.5 standard deviation (SD) units and probably around 0.20 SD units) for most intervention types and outcome intersection, except for the effect of psychological therapies on depression. However, larger effects are seldom achieved using QOL tools as outcome measurement instruments, including when effective treatments are assessed. As an example, in participants with diabetic macular oedema, Bressler 2014 found a 2‐year difference in best‐corrected visual acuity (BCVA) increase by 9.7 (SD: 14.5) Early Treatment Diabetic Retinopathy Study (ETDRS) letters for monthly ranibizumab 0.5 mg compared to sham, which is an effect size of 0.67, while the increase in QOL (NEI‐VFQ 25 composite score) was 5.4 (SD: 13.5), corresponding to an effect size of 0.4, meaning that even an intensive and long‐term active pharmacological treatment that improves vision in one eye had a small or moderate impact on QOL, a figure that changed little whether the treatment was applied to the better or the worse eye. We think that small or moderate effects of low vision rehabilitation on VRQOL could be valuable and should be evaluated in further studies including their cost‐effectiveness. Moreover, the intensity of the intervention matters, since effects were recorded for several low vision rehabilitation interventions compared to both no care and usual or other active care. This may have been the reason why the largest benefit for multidisciplinary rehabilitation programmes (versus no care) was recorded in Stelmack 2008, the LOVIT study conducted in the USA, which used an intensive rehabilitation scheme.
Overall completeness and applicability of evidence
Our review offers a broad picture of evidence of RCTs on several types of low vision rehabilitation to improve QOL and related outcomes. Apart from the obvious challenge in discussing the applicability of this heterogeneous evidence in different settings, it must be noted that only a few studies investigated long‐term maintenance effects on QOL or related outcomes, so we do not know whether any benefit persists. Studies were generally focused on short‐term maintenance (within six months) or direct effects of interventions (within one month of intervention completion). Adverse events, or the lack of, were rarely reported in studies. In the few studies reporting on adverse events, these seemed to be related to associated treatments, e.g. antidepressants, rather than directly to the interventions studied. In addition, some interventions were not standard 'low vision rehabilitation' such as prism glasses in AMD (Smith 2005), or tango dance lessons (Pinninger 2013). All studies were conducted in North America; Europe or Australia and only four in Asia, but none in South America or Africa.
With respect to the larger effects of psychological therapies, or group programmes, or both, we only synthesised data on depressive symptomatology, self‐efficacy and self‐esteem and adaptation to vision loss for its close relatedness to the psychological aspect of QOL. The effect of preventive interventions on the incidence of depression and/or anxiety disorders in visually impaired (older) adults was disregarded, as these reflect serious medical conditions. Still, we believe that, since depression and anxiety are a problem in increasing numbers of older visually impaired adults, these preventive interventions may be considered another important development within low vision rehabilitation settings. Examples are interventions specifically developed to prevent depression and anxiety disorders or reduce symptoms (e.g. Nollet 2016; Rees 2015; Rovner 2014 and Van der Aa 2015).
Although most studies in the current review focused on psychological therapies and/or group interventions and showed a positive effect, this does not imply that these therapies are better than methods of enhancing vision or multidisciplinary care to improve QOL. It is possible, though, that methods to enhance vision target VRQOL more directly than psychological interventions which might target the psychological outcomes better, such as depression or self‐esteem.
Nonetheless, there was some evidence of potential benefit on several outcomes for both these interventions. Future evidence may clear up this issue thus research is needed to confirm that methods to enhance (residual) vision should be made accessible to all, taking into account both effectiveness and cost. Apart from that, multidisciplinary approaches, which might include psychological therapies or group programmes, or both, may further enhance QOL and related outcomes as was shown in several studies, even when having usual care as a comparator, which sometimes included rehabilitation services.
In addition, it could be observed in the current review that some low vision rehabilitation interventions aim to stimulate patient empowerment leading to increased self‐esteem in our vulnerable group of patients. This is encouraging and following important developments in healthcare (Samoocha 2010) should receive more attention in rehabilitation settings or future studies as well. More high‐quality studies are needed to prove the effect of low vision rehabilitation interventions in improving activities of daily living (e.g. occupational therapy) and to enhance adaptation to vision loss. Adaptation to vision loss may be a long process, however, it may be influenced by group interventions, peer support or psychological interventions aiming to increase insight into the acceptance of living with a disability. Whether these types of interventions are accessible to patients is unclear. Lessons learned from other fields in healthcare could be used as an inspiration to develop new treatments focusing on adaptation to visual disability.
Referring patients to low vision rehabilitation requires customisation of the various available interventions which should be tailored to the individual patient’s wishes and needs. Gaining the medical and functional history of a patient or assessing the patient's needs from his/her own perspective, for example, using the Activity Inventory (Massof 2005) are promising developments to efficiently refer patients to the right intervention. In turn, more prospective studies are needed to predict which patients benefit most from a specific service. The complexity and diversity of patients' needs may indicate a need to adapt any rehabilitation programme to individual patient profiles. Research should be conducted on prognostic factors to develop those patient profiles for specific rehabilitation programmes. Also, there is little evidence yet on dose‐response relations between the intensity of treatment needed and the outcomes studied. Of interest, a landmark study conducted by US Veterans, the LOVIT study (Stelmack 2008), adopted an intensive rehabilitation programme and yielded a very large benefit on VRQOL. If feasible, RCTs addressing differences between intensive versus less intensive models should be set up. In addition, there are only a few studies that report on cost‐effectiveness and/or cost‐utility of low vision rehabilitation interventions (e.g. Stroupe 2008; Eklund 2008, Van der Aa 2015 and Stelmack 2017).
A limitation of our findings is that included studies were mainly conducted on participants with AMD living in high‐income countries. Studies are needed in middle and low‐income countries if low vision services are available. Other ‘forgotten’ subgroups which should be separately addressed are young and working‐age adults. Only one and unfortunately rather low‐quality RCT was found addressing work‐related issues in living with vision loss. As the prevalence of visual impairment in working age adults and children is low, we encourage collaboration with other (inter)national research groups for organising properly powered trials providing more and stronger evidence for the effect of rehabilitation programmes. Apart from QOL in younger adults, these outcomes should focus on return to work. Other subgroups should receive more attention as well in the implementation and studies into effective interventions, e.g. people with multiple disabilities, such as intellectual disabilities or concurrent hearing disabilities. The latter have been studied in two Dutch RCTs by Vreeken 2013 and Roets‐Merken 2015 which studied the effect of dual‐sensory loss interventions in older adults. Finally, in clinical practice, training in the use of modern devices such as specific ‘visually impaired user‐friendly’ computer software, tablets and smartphones are increasingly offered. These interventions may have the potential to increase participation in society but trials are needed to evaluate their effectiveness and cost‐effectiveness.
User groups need to be empowered to be more vocal and create the demand for more and better research and the funding for it. with the aim of designing and conducting high‐quality RCTs in rehabilitation programmes that are up to date with respect to patients' needs.
Quality of the evidence
The certainty of the evidence that aimed to assess the effectiveness of low vision rehabilitation interventions on QOL or related outcomes was generally low or very low. Only the certainty of the effect of psychological therapies and/or group programmes versus usual care on depression was moderate.
Reasons for downgrading assessments using GRADE were mostly because of imprecision due to few participants in the comparisons with wide confidence intervals, often suggesting a benefit but including no effect. Moreover, many studies were at unclear risk of bias for allocation concealment and masking. Because rehabilitation is generally impossible to mask, one could argue whether downgrading the certainty of the evidence for this item is fair in this field. We also suggest that research should be conducted to disentangle true benefit from socially desirable responses. Therefore, in view of the subjectivity of QOL outcomes which might be prone to response bias, we encourage researchers to use at least masked outcome assessors and also to develop outcome measures which can be assessed without knowing the intervention status of the participant. Some of the more recent studies which can be found in this review should be used as examples in the field to improve the methodological quality of obtaining evidence of low vision rehabilitation interventions.
Potential biases in the review process
Although all relevant literature databases were searched which led to over 11,000 hits and additional searches were performed on reference lists of relevant studies, conference proceedings and journals, we cannot be certain that we included all available studies. Compared with the review of Binns 2012, the seven trials that were found in that study were also identified in our literature search. In contrast to Binns 2012, nonrandomised trials (NRCT) or observational studies were excluded in this review, however, some of the RCTs included in this study had considerable risk of bias. In a future version of this review, it may be worth including non‐randomised studies with a robust assessment of risk of bias to get a more complete picture of all the relevant evidence.
A potential limitation of the methodology in this review is represented by the use of the SMD, which is the most common pooling method used when instruments use different scales. Various alternatives are available to interpret a meta‐analysis using the SMD (Higgins 2017). We decided to adopt a simple rule of thumb to interpret the magnitude of the effect sizes: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. Using the data extracted for Brody 2002 as an example, this study found a total score difference of ‐0.95 (95%CI: ‐2.84, 0.94) between the self‐management programme and a waiting list using the NEI‐VFQ questionnaire, for which a clinically significant important difference is commonly believed to be five points. The corresponding SMD we found was ‐0.13 (‐0.40, 0.14), which is consistent with the interpretation of the raw NEI‐VFQ estimates of no difference both for the point estimate and its uncertainty.
Agreements and disagreements with other studies or reviews
Binns 2012 concluded from their comprehensive review, in which they included seven RCTs and over fifty observational studies, that rehabilitation services resulted in improved clinical and functional ability outcomes, but that the effects on mood, VRQOL and HRQOL were less clear. We found it difficult to compare our results to those of Binns 2012 due to different inclusion criteria. Binns 2012 was a broad purpose review and we referred to it for a thorough assessment of methodological issues in current low vision research, including a discussion of choice of QOL tools, and other issues regarding study design.
Study flow diagram.
Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study. Studies were categorized into: (I) psychological therapies and/or group programs; (II) methods of enhancing vision; (III) multidisciplinary rehabilitation programs.
Forest plot of comparison: 1 Low vision rehabilitation versus waiting list or no care, outcome: 1.1 Health‐related quality of life.
Forest plot of comparison: 1 Low vision rehabilitation versus waiting list or no care, outcome: 1.2 Vision‐related quality of life.
Forest plot of comparison: 2 Low vision rehabilitation versus active comparator, outcome: 2.1 Health‐related quality of life.
Forest plot of comparison: 2 Low vision rehabilitation versus active comparator, outcome: 2.2 Vision‐related quality of life.
Comparison 1 Low vision rehabilitation versus waiting list or no care, Outcome 1 Health‐related quality of life.
Comparison 1 Low vision rehabilitation versus waiting list or no care, Outcome 2 Vision‐related quality of life.
Comparison 1 Low vision rehabilitation versus waiting list or no care, Outcome 3 Activities of daily living (QOL physical aspect).
Comparison 1 Low vision rehabilitation versus waiting list or no care, Outcome 4 Depression (QOL: psychological aspect).
Comparison 1 Low vision rehabilitation versus waiting list or no care, Outcome 5 Self‐efficacy or self‐esteem (QOL: psychological aspect).
Comparison 1 Low vision rehabilitation versus waiting list or no care, Outcome 6 Adaptation to vision loss (QOL: psychological aspect).
Comparison 2 Low vision rehabilitation versus active comparator, Outcome 1 Health‐related quality of life.
Comparison 2 Low vision rehabilitation versus active comparator, Outcome 2 Vision‐related quality of life.
Comparison 2 Low vision rehabilitation versus active comparator, Outcome 3 Activities of daily living (QOL physical aspect).
Comparison 2 Low vision rehabilitation versus active comparator, Outcome 4 Depression (QOL: psychological aspect).
Comparison 2 Low vision rehabilitation versus active comparator, Outcome 5 Self‐efficacy or self‐esteem (QOL: psychological aspect).
Comparison 2 Low vision rehabilitation versus active comparator, Outcome 6 Adaptation to vision loss (QOL: psychological aspect).
| Psychological therapies and/or group programmes compared to waiting list or no care for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as problem‐solving treatment (PST), self‐management programme Comparison: passive control group such as a waiting list | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with EQ‐5D | 54 Short‐term maintenance effect | SMD 0.26 SD worse | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with NEI‐VFQ and NEI‐VFQ‐7 | 285 Short‐term direct or maintenance effect | SMD‐0.23 SD better | ⊕⊕⊝⊝ |
| Adverse events | 285 Short‐term direct or maintenance effect | Brody 2002: No data available Nollet 2016: 2 AEs and 10 SAEs not related to rehabilitation | ⊕⊕⊝⊝ |
| EQ‐5D: EuroQol 5 Dimensions; NEI‐VFQ: National Eye Institute Visual Functioning Questionnaire; SAEs: serious adverse events; VFQ: Visual Functioning Questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded 1 level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (small sample size of n = 54, and wide confidence interval crossing the line of no effect) Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Methods of enhancing vision compared to waiting list or no care for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as immediate low vision assessment, provision of magnifying aids and training, low vision outpatient service, customised prism glasses Comparison: passive control group such as a waiting list, delayed low vision assessment, low vision examination and no intervention, placebo prisms | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life | No data were available for this outcome | ||
| Vision‐related quality of life measured with questionnaires: NEI‐VFQ‐ 25, VA‐LV‐VFQ48, Activity Inventory, IVI | 262 Short‐term direct or maintenance effect | SMD ‐0.19 SDs (better) | ⊕⊝⊝⊝ |
| Adverse events | 262 Short‐term direct or maintenance effect | Kaltenegger 2019 reported no adverse events. No data available for the other studies | not applicable |
| NEI‐VFQ: National Eye Institute Visual Functioning Questionnaire; VA‐LV‐VFQ: Veterans Affairs Low Vision Visual Functioning Questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias in 3 out of 4 studies) and two levels due to very serious imprecision (sample size of n = 237, and wide confidence interval crossing the line of no effect) Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Multidisciplinary rehabilitation compared to waiting list or no care for better quality of life in visually impaired adults | ||||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as multidisciplinary low vision rehabilitation plus home visit, multidisciplinary low vision programme Comparison: passive control group such as a waiting list | ||||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence | Explanation |
| Health‐related quality of life measured with EQ‐5D, SF‐36 | 183 | SMD ‐0.08 SD (better) Short‐term direct or maintenance effect | ⊕⊝⊝⊝ | |
| Vision‐related quality of life measured with NEI‐VFQ, VFQ 48 questionnaire | 193 | See comment Short‐term direct or maintenance effect | ⊕⊝⊝⊝ | Both studies beneficial, but large effect in a large trial delivering intensive rehabilitation (Stelmack 2008: SMD: ‐1.64, 95%CI ‐2.05 to ‐1.24) and less benefit in the other (Acton 2016: SMD ‐0.42, 95%: ‐0.90 to 0.07), P = 0.0001 for inconsistency |
| Adverse events | 193 | Acton 2016: 19 AEs probably unrelated to treatment Stelmack 2008: no (S)AEs related to the study. Short‐term direct or maintenance effect | ⊕⊕⊝⊝ | |
| EQ‐5D: EuroQol 5 Dimensions; SAEs: serious adverse events; SF‐36: Short Form‐36 item Health Survey; VA‐LV‐VFQ: Veterans Affairs Low Vision Visual Functioning Questionnaire | ||||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | ||||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 183, and wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias), two levels due to very serious imprecision (sample size of N = 193, and wide confidence interval crossing the line of no effect) and one level due to heterogeneity (both studies beneficial, but important and significant effect in a large trial delivering intensive rehabilitation and less benefit in the other) 3Downgraded one level due to study limitations (unclear risk of performance bias) and no reporting of adverse events in 1 study Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | ||||
| Psychological therapies and/or group programmes compared to active control for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low‐vision services Intervention: low vision rehabilitation such as a health education programme, self‐management programme with usual rehabilitation care, stepped‐care (including cognitive behavioural therapy, problem‐solving treatment and/or referral to general practitioner), expressive writing course, problem‐solving treatment, behavioural activation Comparison: active control group such as an individual low vision programme, usual low vision rehabilitation, usual care by low vision service or other care providers, neutral writing exercise, supportive therapy | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with EQ‐5D, SF‐36 and a general health measure | 600 Long and short‐term maintenance effect | SMD ‐0.09 SDs better | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with NEI‐VFQ‐17 and 25, IVI, LVQOL 18, vision=specific mental health questionnaire | 1245 Long and short‐term maintenance effect | SMD ‐0.11 SDs better | ⊕⊕⊝⊝ |
| Adverse events | 1453 Long and short‐term maintenance effect | No data available for Eklund 2008, Girdler 2010, Bryan 2014, Rees 2015, Rovner 2007, Rovner 2013, Rovner 2014, Tey 2019. Van der Aa 2015: no (S)AEs related to the study. | ⊕⊝⊝⊝ |
| EQ‐5D: EuroQol‐5 Dimensions; NEI‐VFQ: National Eye Institute Visual Functioning Questionnaire; IVI: Impact of Vision Impairment profile, LVQOL: low vision quality of life questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 473, and wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias), one level due to imprecision (large sample size of N = 1118, but confidence interval including the line of no effect) 3Downgraded one level due to study limitations (unclear risk of performance bias) and no reporting of adverse events in 7 studies Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Methods of enhancing vision compared to active control for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low vision services Intervention: low vision rehabilitation such as CCTV training sessions from a low vision therapist, home‐based low vision rehabilitation, low vision devices with instruction, usual comprehensive vision rehab and access to desk top video magnifier, CCTV and training, low vision support service, nonportable and portable electronic devices Comparison: active control group such as CCTV instructions from supplier, clinic‐based low vision rehabilitation, low vision devices without instruction, usual comprehensive vision rehab without access to desk top video magnifier, eccentric viewing training, placebo support by a nurse, nonportable devices only | |||
| Outcomes | № of participants | Standardized Mean Difference (SMD) with Low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with EQ‐5D, SF‐36 | 443 Short‐term maintenance effect | SMD ‐0.09 SD (better) | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with questionnaires: VISQOL, LVQOL subscales, VA‐LV‐VFQ‐48, VFQ‐25, Activity Inventory, IVI | 660 Short‐term direct or maintenance effect | SMD ‐0.24 SD (better) | ⊕⊕⊕⊝ |
| Adverse events | 660 Short‐term direct or maintenance effect | Burggraaff 2012: no AEs. Stelmack 2017: 10 AEs not treatment‐related. No data available for Draper 2016, Jackson 2017, Leat 2017, Pearce 2011, Taylor 2017. | ⊕⊝⊝⊝ |
| EQ‐5D: EuroQol 5 Dimensions, SF‐36: Short Form 36‐item Health Survey, VISQOL: vision‐related quality of life, VFQ: Visual Functioning Questionnaire; VA‐LV‐VFQ: Veterans Affairs Low Vision Visual Functioning Questionnaire; IVI: Impact of Vision Impairment profile, LVQOL: low vision quality of life questionnaire | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 443, but wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias and other domains at unclear or high risk of bias in some studies); no downgrade for imprecision (large sample size of N = 660, and confidence interval consistent with small effects) 3Downgraded one level due to study limitations (unclear risk of performance bias) and no reporting of adverse events in 7 studies Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Multidisciplinary rehabilitation compared to active control for better quality of life in visually impaired adults | |||
| Patient or population: adults (>= 18 years) with an irreversible visual impairment Setting: low vision services Intervention: low vision rehabilitation such as pooled community + centre‐ and community‐based low vision service delivery, family rehabilitation intervention where family is present at all stages, enhanced low vision rehabilitation including home visits Comparison: active control group such as community placebo home visits, individual rehabilitation intervention with no family present, conventional low vision rehabilitation and control home visits from a community worker with no rehabilitation, conventional clinic‐based low vision rehabilitation including placebo home visits | |||
| Outcomes | № of participants | Standardised mean difference (SMD) with low vision rehabilitation | Certainty of the evidence |
| Health‐related quality of life measured with WHO‐QOL, SF‐36 | 375 Long‐term maintenance effect | SMD ‐0.10 SD (better) | ⊕⊝⊝⊝ |
| Vision‐related quality of life measured with IVI, FAQ, VCM1 | 464 Short‐term direct or long‐term maintenance effect | SMD 0.01 SD (same) | ⊕⊕⊝⊝ |
| Adverse events | 464 Short‐term direct or long‐term maintenance effect | No data available | not applicable |
| SF‐36: Short Form‐36 item Health Survey; WHO‐QOL: World Health Organization Quality of Life; IVI: Impact of Vision Impairment profile; VCM1: Vision‐related quality of life Core Measure 1, FAQ: Functional Assessment Questionnaire, observer‐rated functional visual performance test | |||
| GRADE Working Group grades of evidence High‐certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low‐certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. | |||
| 1Downgraded one level due to study limitations (unclear risk of performance bias) and two levels due to very serious imprecision (sample size of N = 443, and wide confidence interval crossing the line of no effect) 2Downgraded one level due to study limitations (unclear risk of performance bias), one levels due to imprecision (large sample size of N = 660, but confidence interval including moderate or small effects) Note: the following rule of thumb can be used to interpret the clinical magnitude of meta‐analysis results expressed as SMD: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. | |||
| Author year (country) | Study design, follow‐up, sample size, % dropout | Sample: mean age, % female, vision impairment | Setting and study type | Outcome measures | Intervention and control groups | Outcome of study |
| Ballemans 2012 (Netherlands) | 2‐arm RCT, 8 weeks, N = 68, 10% | 76 years, 60%, AMD, glaucoma, other conditions | Low vision rehabilitation service IV other programmes | Process evaluation: usefulness and acceptability of intervention from participants’ and trainers’ perspectives | I: standardised orientation and mobility training C: regular orientation and mobility training (without protocol) | Both the standardised and regular training showed to be useful and acceptable from both participants’ and trainers’ perspectives |
| (New Zealand) | 4‐arm RCT, 1 year, N = 391, 8% | 84 years, 68%, AMD, cataract, DRP, eye surgery | Home‐based IV other programmes | Number of falls and injuries and costs of implementation of the intervention | I: home safety assessment and modification programme by OT I: exercise programme by PT + vitamin D I: intervention 1 + 2 C: social visits | The home safety programme reduced falls and was more cost‐effective than the exercise programme. The exercise group was not effective in reducing falls. |
| (China) | 2‐arm RCT, 16 weeks, N = 40, 45% | 84 years, vision impairment | Residential care homes IV other programmes | Knee proprioception, concentric isokinetic knee strength, sensory organisation | I: Tai Chi (Yang style) C: music percussion activity | After Tai Chi, significant improvement was seen in knee proprioception and visual and vestibular ratios. Tai Chi can improve balance control. |
| (United States) | 3‐arm RCT, Immediate, N = 38, 0% | 28 years, 47% ocular related blindness | Laboratory IV other programmes | Performance of virtual navigation in virtual target building, physical navigation and drop‐off tasks in actual target building | I: video game for full exploration of virtual environment I: directed navigation by sighted facilitator and specific paths in game C: video game but not corresponding with target virtual environment | Highly interactive and immersive exploration of the virtual environment engages a blind user to develop skills akin to positive near transfer of learning. |
| (Canada) | 2‐arm RCT, 7 weeks, N = 32, unknown | 83 years, partially sighted, legally blind | Day centre for the visually impaired IV other programmes | Balance tests with open and closed eyes, perceived balance | I: balance training C: sitting stretching exercises | Perceived confidence in balance increased after training, however, objectively no improvement was found. |
| (United States) | 2‐arm RCT, 5 weeks N = 228, 9% | 75 years, 58% N = 29 visually impaired, N = 199 normally sighted | Laboratory, home‐based II methods to enhance vision | Useful field of view performance | I: speed of processing training C: social and computer contact | Speed of processing training effectively improved useful field of view performance. |
| (United States) | 3‐arm RCT, 3‐6 months, N = 25 | 47 years, 48% various (hereditary) retinal diseases | Laboratory‐based training and on‐road driving II methods to enhance vision | Clinical vision tests, functional orientation and mobility tasks, driving skills, psychophysical measures | I: bioptic telescopes and training I: lenses and training C: lenses without training | Significant improvement in visual skills using a bioptic telescope, including driving‐related skills |
| AMD: age‐related macular degeneration | ||||||
|
Any other sources of bias were reported under notes in the Characteristics of included studies table. |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Health‐related quality of life Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 I: psychological therapies and/or group programs vs waiting list or no care | 1 | 54 | Std. Mean Difference (IV, Random, 95% CI) | 0.26 [‐0.28, 0.80] |
| 1.2 III: multidisciplinary rehabilitation vs waiting list or no care | 2 | 183 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.08 [‐0.37, 0.21] |
| 2 Vision‐related quality of life Show forest plot | 9 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.1 I: psychological therapies and/or group programs vs waiting list or no care | 2 | 285 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.23 [‐0.53, 0.08] |
| 2.2 II: methods of enhancing vision vs waiting list or no care | 5 | 262 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.19 [‐0.54, 0.15] |
| 2.3 III: multidisciplinary rehabilitation vs waiting list or no care | 2 | 193 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.04 [‐2.24, 0.17] |
| 3 Activities of daily living (QOL physical aspect) Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 3.1 II: methods of enhancing vision vs waiting list or no care | 1 | Std. Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 III: multidisciplinary rehabilitation vs waiting list or no care | 1 | Std. Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Depression (QOL: psychological aspect) Show forest plot | 9 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.1 I: psychological therapies and/or group programs vs waiting list or no care | 5 | 456 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.23 [‐2.18, ‐0.28] |
| 4.2 II: methods of enhancing vision vs waiting list or no care | 2 | 44 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.86 [‐1.50, ‐0.23] |
| 4.3 III: multidisciplinary rehabilitation vs waiting list or no care | 2 | 193 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.16 [‐0.44, 0.13] |
| 5 Self‐efficacy or self‐esteem (QOL: psychological aspect) Show forest plot | 5 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 5.1 I: psychological therapies and/or group programs vs waiting list or no care | 5 | 550 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.85 [‐1.48, ‐0.22] |
| 6 Adaptation to vision loss (QOL: psychological aspect) Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.1 III: multidisciplinary rehabilitation vs waiting list or no care | 2 | 97 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.14 [‐0.54, 0.26] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Health‐related quality of life Show forest plot | 9 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 I: psychological therapies and/or group programs vs usual or other care | 4 | 600 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.39, 0.20] |
| 1.2 II: methods of enhancing vision vs usual or other care | 2 | 443 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.28, 0.09] |
| 1.3 III: multidisciplinary rehabilitation vs usual or other care | 2 | 375 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.10 [‐0.31, 0.12] |
| 1.4 IV: other programs vs usual or other care | 1 | 43 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.05 [‐0.70, 0.60] |
| 2 Vision‐related quality of life Show forest plot | 19 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.1 I: psychological therapies and/or group programs vs usual or other care | 7 | 1245 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.11 [‐0.24, 0.01] |
| 2.2 II: methods of enhancing vision vs usual or other care | 7 | 660 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.24 [‐0.40, ‐0.08] |
| 2.3 III: multidisciplinary rehabilitation vs usual or other care | 3 | 464 | Std. Mean Difference (IV, Random, 95% CI) | 0.01 [‐0.18, 0.20] |
| 2.4 IV: other programs vs usual or other care | 2 | 163 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.21 [‐0.53, 0.10] |
| 3 Activities of daily living (QOL physical aspect) Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 I: psychological therapies and/or group programs vs usual or other care | 2 | 208 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.39 [‐0.67, ‐0.12] |
| 3.2 IV: other programs vs usual or other care | 1 | 120 | Std. Mean Difference (IV, Random, 95% CI) | 0.11 [‐0.25, 0.47] |
| 4 Depression (QOL: psychological aspect) Show forest plot | 13 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.1 I: psychological therapies and/or group programs vs usual or other care | 9 | 1334 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.14 [‐0.25, ‐0.04] |
| 4.2 II: methods of enhancing vision vs usual or other care | 3 | 162 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.22 [‐0.59, 0.15] |
| 4.3 IV: other programs vs usual or other care | 1 | 120 | Std. Mean Difference (IV, Random, 95% CI) | 0.03 [‐0.33, 0.39] |
| 5 Self‐efficacy or self‐esteem (QOL: psychological aspect) Show forest plot | 5 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 5.1 I: psychological therapies and/or group programs vs usual or other care | 4 | 427 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.06 [‐0.26, 0.15] |
| 5.2 III: multidisciplinary rehabilitation vs usual or other care | 1 | 133 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.22 [‐0.56, 0.12] |
| 6 Adaptation to vision loss (QOL: psychological aspect) Show forest plot | 6 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.1 I: psychological therapies and/or group programs vs usual or other care | 3 | 495 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.11 [‐0.28, 0.07] |
| 6.2 II: methods of enhancing vision vs usual or other care | 1 | 122 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.30 [‐0.65, 0.06] |
| 6.3 III: multidisciplinary rehabilitation vs usual or other care | 2 | 376 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.02 [‐0.24, 0.19] |
