Types of studies

Randomised controlled trials (RCTs), quasi‐randomised controlled trials (QRCTs), controlled before and after studies (CBAs), and interrupted time series (ITS) with at least three time points before and after the intervention.

We consider including QRCT, CBA and ITS designs, as preliminary literature searching indicates a scarcity of RCTs of text messaging interventions. However, if a sufficient number of well‐conducted RCTs is identified (upon which we can draw reliable conclusions), we will exclude other study designs during the review stage. If all study designs are included, we will perform a sensitivity analysis to assess whether the results change by including non‐RCT designs.

Types of participants

We will consider all study participants regardless of age, gender and ethnicity, as well as all types and stages of diseases. We will include studies in all settings, i.e. primary care settings (services of primary health care), outpatient settings (outpatient clinics), community settings (public health services) and hospital settings. We will not exclude studies according to the type of healthcare provider (e.g. nurse, doctor, allied staff).

Types of interventions

We will include interventions that use Short Messaging Service (SMS or 'text messaging) or Multimedia Message Service (MMS) messaging in self management of long‐term conditions, i.e. often slowly‐progressing conditions that affect people's lives over a long period of time. Relevant interventions include those that provide disease‐related information to patients, support self monitoring of illness, improve adherence to treatment or medications or both, or offer a channel for peer‐to‐peer networking and support through SMS or MMS messaging. The messaging needs to be between a healthcare provider (either in person or automated) or a 'treatment buddy' (e.g. a lay health worker or peer supporter) and a patient, regardless of who sends the first SMS.

We will make comparisons between outcomes of mobile phone messaging and no intervention, as well as other modes of communication such as face‐to‐face, postal letters, calls to landline or mobile telephones, e‐mail or via electronic health records; and if applicable, automated versus personal text messaging.

Types of outcome measures

A number of processes and outcomes may be affected by interventions that aim to enhance and/or facilitate the communication between patients and/or carers, and healthcare providers (individuals or institutions) using mobile phone messaging. We will seek and extract all relevant outcomes which are likely to include some or all of the following categories:

1. Health outcomes as a result of the intervention, including physiological measures, e.g. blood pressure; clinical assessments; biomarker values; self reporting of symptom resolution or quality of life;

2. Patients' evaluation of the intervention, including satisfaction, readiness to use, timeliness, availability and/or convenience;

3. Healthcare providers' evaluation of the intervention, including satisfaction, readiness to use, timeliness and/or convenience;

4. Capacity to self manage long‐term conditions, including lifestyle modification, understanding of disease, impact on independence and responsibility, self esteem and/or creation of a supportive environment;

5. Costs and cost‐effectiveness of intervention;

6. Patient and/or healthcare providers' perceptions of safety.

We will also seek outcomes related to potential harms or adverse effects of the intervention, such as misreading or misinterpretation of data, transmission of inaccurate data, loss of verbal and non‐verbal communication cues, issues of privacy and disclosure, or failure or delay in the message delivery.

We will also record data related to the medical setting and clinical context of patients to assess any changes in outcomes within that context.

Selection of studies

All review authors will be involved in the selection process. Two review authors will independently assess the potential relevance of all titles and abstracts identified from the electronic searches. We will retrieve full text copies of all articles judged to be potentially relevant from the titles and abstracts. At least two review authors will then independently assess these articles for inclusion. A meeting of all review authors will check the final list of included and excluded studies, and any doubts or disagreements about particular studies will be resolved by discussion. Next, we will undertake a manual follow‐up of references from key publications. Where the description of the intervention is not sufficiently detailed to allow the review authors to judge whether it meets the review's inclusion criteria, the trial authors will be contacted and, where possible, more detailed descriptions and/or materials will then be assessed.

Data extraction and management

We will extract the following data from the included studies, using a data extraction form modified from the Cochrane Consumers and Communication Review Group's data extraction template:

1. General information: title, authors, source, publication status, date published, language, review author information, date reviewed.

2. Study methods: aims of intervention, aim of study, study design, methods of participant recruitment, inclusion/exclusion criteria, informed consent and ethical approval, funding.

3. Risk of bias: data to be extracted depends on the study design (see 'Assessment of risk of bias in included studies').

4. Patients: description, geographic location, setting, number, age, gender, ethnicity, socioeconomic status distribution. If relevant: principal health problem or diagnosis, stage of illness, treatment received.

5. Providers: description, geographic location, setting, age, gender.

6. Interventions: description including technical specifications on SMS and handset provider, duration of intervention, purpose of intervention, initiator of intervention, message content, details of control/usual or routine care, co‐interventions.

7. Outcomes: principal and secondary outcomes as specified above, methods of assessing outcomes, follow up for non‐respondents, timing of outcome assessment, adverse events.

8. Results: for outcomes and times of assessment, control and intervention groups if applicable.

Two review authors independently will extract full descriptions of the interventions onto a standard form. The standard forms will then be sent to other review authors who will check these descriptive data. Any discrepancies between the two review authors' data extraction sheets will be discussed by the data extractors and resolved in a joint meeting with other review authors. If data are missing, we will attempt to contact the authors of the studies to obtain the information.

Assessment of risk of bias in included studies

We will assess and report on the risk of bias of included studies in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008) which recommends the explicit reporting of sequence generation, allocation concealment, blinding of participants, providers and outcome assessors, incomplete outcome data, selective outcome reporting and other sources of bias for RCTs. We will assess the risk of bias of cluster RCTs, quasi‐randomised controlled trials, CBA and ITS studies using a modified version of the risk of bias tool, and in accordance with the guidelines of the Cochrane Consumers and Communication Review Group (Ryan 2007).

In all cases, two review authors will independently assess the risk of bias of included studies, with any disagreements resolved by discussion and consensus of the team. We will use a template to guide the assessment of risk of bias, and will judge each domain as 'yes' (indicating a low risk of bias), 'no' (indicating a high risk of bias) or 'unclear' (indicating an uncertain risk of bias). We will present all included studies by study type and risk of bias level.

We will contact study authors for additional information about the included studies, or for clarification of the study methods as required. We will present the results of the risk of bias assessment in tables, and incorporate the results of this assessment into the review through systematic narrative description and commentary about each of the domains, leading to an overall assessment the risk of bias of included studies and a judgement about the internal validity of the review's results.

Measurement of effect of the intervention

We will use risk ratios (RR) as effect measures for dichotomous outcomes and standardised mean differences (SMD) for continuous outcomes. We will report confidence intervals with all measures of effect.

Unit of analysis issues

We will take into account unit of analysis issues resulting from cluster‐randomised trials, repeated measurements and studies with more than two treatment groups. If applicable, the data will be analysed using the recommendations in Cochrane Collaboration Open Learning Material on issues related to the unit of analysis (Alderson 2002a).

Dealing with missing data

We will obtain relevant missing data from study authors, if feasible. We will carefully evaluate important numerical data such as screened, eligible and randomised patients as well as intention‐to‐treat and per‐protocol population. With incomplete outcome data (such as drop‐outs, misses to follow‐up and withdrawn study participants), we will assess and report the risk of bias as 'low' or 'high' as guided by the Cochrane Handbook (Higgins 2008) and identify the numbers as well as the reasons for incomplete data (requesting for additional information from the study authors if necessary). Issues of last‐observation‐carried‐forward (LOCF) will be critically appraised and compared to specifications of primary outcome parameters and power calculation.

Assessment of reporting biases

We will assess reporting biases statistically and using funnel plots in RevMan 5. We will assess selection bias, performance bias, attrition bias and detection bias in the included studies using the checklist provided in Ryan 2007.

Data synthesis

We will present a narrative overview of the findings, including tabular summaries of extracted data. We aim to structure the narrative primarily according to the intended purpose of the message. We will consider whether it is appropriate to combine the studies quantitatively once we have completed the search. The decision is likely to rest on the diversity of interventions and outcome measures used in the studies. Studies will be classified on the following issues:

• Study design: RCTs, QRCTs, CBAs, ITS;

• Outcome measures used, as described at Types of outcome measures.

If quantitative analysis is undertaken, the meta‐analysis will depend on the outcomes reported. For continuous data, where outcomes have been measured in a standard way across studies, we will report the SMD and confidence intervals (Alderson 2002b). For dichotomous data, when outcomes have been measured in a standard way, we will report the RR (Deeks 2001). In such cases, we will take a cautious approach to combining results, and will detail the rationale. We will conduct statistical analysis according to the guidelines in the Cochrane Handbook (Higgins 2008).

Assessment of heterogeneity

In the event of substantial clinical, methodological or statistical heterogeneity, we will not combine study results in meta‐analysis. We will identify heterogeneity by visual inspection of the forest plots, by using a standard Chi² test and a significance level of alpha = 0.1, in view of the low power of such tests. We will also examine heterogeneity with I², where I² values of 50% or more indicate a substantial level of heterogeneity (Higgins 2003). When heterogeneity is found, we will attempt to determine potential reasons for it by examining individual study characteristics and those of subgroups of the main body of evidence.

Subgroup analysis

If there are sufficient trial data, and if appropriate, we will analyse subgroups according to age, as mobile phone messaging usage and general healthcare utilisation patterns are different across age groups. We will distribute patients into three subgroups; 0 to 18, 18 to 55, over 55. The distribution was made on the basis of a survey by MORI in 2005 (MORI 2005). The use of text messaging by different age groups varied, but was significant and consistently high except in the over 55 age group where around only 40% of users had communicated by text messaging.

Sensitivity analysis

We will perform sensitivity analyses where appropriate in order to explore the influence of the following factors on effect size:

• excluding unpublished studies;

• taking account of risk of bias of included studies, as specified above;

• excluding any large studies to establish how they impact on the results;

• excluding studies using the following filters: criteria used for clinical diagnosis and eligibility for intervention, language of publication, source of funding (industry versus other), country;

• the length of the interval between delivery of the intervention and measurement of the effect.

We may also test the robustness of the results by repeating the analysis using different measures of effects size (risk difference, odds ratio etc.) and different statistical models (fixed effect and random effects models).

Consumer participation

The draft review will be circulated for peer review by consumers in the Cochrane Collaboration. We will examine whether consumers were involved in the design and implementation of each included study. We also aim to receive consumer feedback from organisations that may represent consumers/patients who would be affected by the interventions identified in our review.

Dissemination plans

We will aim to promote and disseminate the findings of our review through publication of the review in peer‐reviewed journal(s), by presenting the findings at conferences and scientific meetings and by giving talks on the topic of the review for broad audiences varying from public and policy‐makers to healthcare providers. We will also promote and disseminate the findings of the review through organisations representing consumers ‐ as described above.