Types of studies

We will include randomised controlled trials (RCTs) including cluster‐ and quasi‐randomised trials. We will exclude cross‐over trials and will include studies reported in full text, those published as abstract only, and unpublished data.

Types of participants

We will include both adults (aged 18 years and over) and children (under 18 years) with a diagnosis of asthma, as per international or national guidelines, or whose condition was diagnosed by a healthcare professional and are currently prescribed maintenance asthma treatment (via any administration route), given alone or in combination with other controller therapies. We will include interventions that target parents or carers who are involved in managing maintenance asthma medication for any participant. We will exclude interventions that are targeted at healthcare professionals, as the review relates only to digital interventions targeted at patients.

We will exclude participants with the following co‐morbidities/characteristics.

1. Other respiratory comorbidities such as chronic obstructive pulmonary disease (COPD) or bronchiectasis.

We will include studies in which only a subset of participants meets the inclusion criteria (asthma diagnosis, prescribed maintenance treatment, or managing maintenance treatment for a participant diagnosed with asthma) if we can obtain disaggregated data.

Types of interventions

We will include studies comparing any interventions with a primary or secondary aim of improving adherence to maintenance asthma treatment (alone or in combination) that uses:

1. a digital component to deliver the intervention versus non‐digital delivery of the same adherence intervention; or

2. a digital component to deliver an intervention versus usual care. Usual care is defined as standard asthma care as per evidence‐based guidelines or standard care in the study setting.

Included digital interventions may be completely self‐delivered or may include an 'in‐person' or 'human' element whereby a healthcare professional or a trained peer is involved in the intervention. This can occur at the point of invitation to participate (e.g. introduction of the digital intervention and/or training of the patient to use the digital intervention) or on an ongoing basis (e.g. discussion of data from the digital intervention at regular consultations, use of remote adherence monitoring and feedback to the patient). The interventions may be delivered completely virtually (i.e. completely digital with no 'in‐person' element) or may include some face‐to‐face aspect (i.e. has an 'in‐person' element); delivery can be provided to individuals (e.g. with mobile apps or electronic monitoring) or to groups (e.g. online forums or computer games), and the intervention may be delivered on a one‐off or ongoing basis.

We will include the following cointerventions, provided they are not part of the randomised treatment.

1. Cointerventions for which more than one type of digital media is used.

2. Other cointerventions that are used in asthma management

When interventions have been described in insufficient detail to determine how the digital intervention was used, we will contact the authors of identified studies to obtain further information. In the case of non‐response after initial contact, we will follow up with study authors twice (once every two weeks), if required. If we receive no response after three contacts, we will exclude these studies from the review.

Types of outcome measures

Electronic searches

We will seek assistance from the Cochrane Airways Information Specialist to identify studies from the Cochrane Airways Trials Register, which is maintained by the Information Specialist for the Group.

The Cochrane Airways Trials Register contains studies identified from several sources.

1. Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL), through the Cochrane Register of Studies Online (crso.cochrane.org).

2. Weekly searches of MEDLINE Ovid SP (1946 to date).

3. Weekly searches of Embase Ovid SP (1974 to date).

4. Monthly searches of PsycINFO Ovid SP (1967 to date).

5. Monthly searches of Cumulative Index to Nursing and Allied Health Literature (CINAHL) EBSCO (1937 to date).

6. Monthly searches of Allied and Complementary Medicine (AMED) EBSCO.

7. Handsearches of the proceedings of major respiratory conferences.

Studies contained in the Trials Register are identified through search strategies based on the scope of Cochrane Airways. We will include details of these strategies, as well as a list of handsearched conference proceedings, in Appendix 1. See Appendix 2 for search terms used to identify studies for this review. These terms have been guided by previous Cochrane reviews such as the Normansell 2017 review (which identifies asthma adherence reviews, although we will not be restricting to inhaled corticosteroids) and the Marcano Belisario 2013 review (which focused on smartphone and tablet apps, although we will not be restricting the review to only these two digital media).

We will search the following trials registries.

1. US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov).

2. World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch).

We will search the reference lists of all primary studies and review articles for additional references.

We will search for studies from the year 2000, as technologies existing before this time are unlikely to be representative of contemporary technologies that support health apps – this is in line with the Cochrane smartphone app review by Marcano Belisario et al (Marcano Belisario 2013).

We will apply no language restrictions.

Searching other resources

We will check the reference lists of all primary studies and review articles for additional references. We will search relevant manufacturers' websites for study information.

We will search for errata or retractions from included studies published in full text on PubMed and will report within the review the date this was done.

Selection of studies

Two review authors (AC, ADS) will independently screen the titles and abstracts of the search results and will code them as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We will retrieve the full‐text study reports of all potentially eligible studies, and two review authors (VW, CC) will independently screen them for inclusion, while recording the reasons for exclusion of ineligible studies. We will resolve any disagreement through discussion, or, if required, we will consult a third person/review author (LH). We will identify and exclude duplicates and will collate multiple reports of the same study, so that each study, rather than each report, is the unit of interest in the review. We will record the selection process in sufficient detail to complete a PRISMA flow diagram and 'Characteristics of excluded studies' table (Moher 2009).

Data extraction and management

We will use Rayyan (Ouzzani 2016) to screen titles and abstracts of identified studies based on the aforementioned inclusion criteria. We will then obtain full‐text study reports/publications of included or potentially relevant studies, and two review authors will independently screen these to identify studies for inclusion and to identify and record reasons for exclusion of ineligible studies. We will resolve disagreements through discussion or, if not resolved, by consultation with the review team. We will exclude any duplicates and multiple reports of the same study.

We will use Covidence to extract study characteristics and outcome data. We will develop a data collection form to extract data and will pilot this form on at least one study in the review. Two review authors (from the following: AC, ADS, VW, LH, and CC) will extract the following study characteristics from included studies in duplicate.

1. Methods: date of study, study design and method of randomisation, length of follow‐up, total study duration, details of any 'run‐in' period, number of study centres and locations, study setting (healthcare setting and country), study withdrawals (study dropout and intervention dropout). We will attempt to distinguish between study versus intervention dropouts to better understand attrition behaviour, if possible, as per an earlier review (Sohanpal 2012)).

2. Participants: N (baseline and upon completion), mean age, age range, gender, severity of asthma, baseline lung function, smoking history, inclusion criteria, exclusion criteria, and differences between groups at baseline.

3. Interventions: intervention details, type of intervention (theory‐based vs non‐theory‐based), details of intervention provider, intervention target (primary and secondary), types of digital components used (technologies used), number of digital components, number of intervention sessions, interactivity with patient (i.e. a two‐way flow of information between the digital component and the patient), adherence feedback, concomitant medications, and excluded medications.

4. Comparison: details of comparison group.

5. Outcomes: primary and secondary outcomes specified and collected; methods of assessment of outcomes and time points reported.

6. Notes: funding of trial and notable conflicts of interest of trial authors.

Two review authors (from the following: AC, ADS, VW, LH, and CC) will independently extract outcome data from included studies. We will note in the 'Characteristics of included studies' table if outcome data were not reported in a usable way. We will resolve disagreements by consensus or by involving a third person/review author (RH). One review author (AC) will transfer data into the Review Manager file (RevMan 2014). We will double‐check that data are entered correctly by comparing data presented in the systematic review against study reports. A second review author (VW) will spot‐check study characteristics for accuracy against the study report.

Assessment of risk of bias in included studies

Two review authors (from the following: AC, VW, ADS, and LH) will assess risk of bias independently for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will resolve any disagreements by discussion or by involving another review author (CC). We will assess risk of bias according to the following domains.

1. Random sequence generation.

2. Allocation concealment.

3. Blinding of participants and personnel.

4. Blinding of outcome assessment.

5. Incomplete outcome data.

6. Selective outcome reporting.

7. Other bias.

We will judge each potential source of bias as high, low, or unclear and will provide a quote from the study report together with a justification for our judgement in the 'Risk of bias' table. We will summarise risk of bias judgements across different studies for each of the domains listed. We will consider blinding separately for different key outcomes when necessary (e.g. for unblinded outcome assessment, risk of bias for all‐cause mortality may be very different than for a patient‐reported pain scale). When information on risk of bias relates to unpublished data or correspondence with a trialist, we will note this in the 'Risk of bias' table.

When considering treatment effects, we will take into account the risk of bias for studies that contribute to that outcome.

Measures of treatment effect

For data reported as rates or times‐to‐events (e.g. exacerbations), we will analyse time‐to‐event or rate ratios. Reported rate ratios can be transformed into log‐rate ratios and analysed via a random‐effects model and by generic inverse variance (GIV).

We will analyse continuous data as standardised mean differences (SMDs) using a random‐effects model and 95% confidence intervals (CIs). We will use mean differences (MDs) rather than SMDs however if measures are reported on the same scale, but we anticipate that most studies will use a variety of methods of measurement, in which case we will use the SMD. We will use the standard deviation (SD) of final (rather than baseline) measurements in the analysis. Although adherence can be presented as dichotomous or continuous, adherence generally is best considered as a continuous variable by nature (to avoid loss of valuable information and use of arbitrary cutoffs), which may be later dichotomised, depending on the adherence measurement method used (Saberi 2011). Therefore, we will treat adherence as continuous data in this review, and this will increase the power to detect a difference. We can adjust dichotomous data by applying a logit transformation and producing an SMD, which can be analysed via GIV. When available, we can use change from baseline scores.

We will undertake meta‐analyses only when this is meaningful, that is, when treatments, participants, and the underlying clinical question are similar enough for pooling to make sense, for example, studies using a similar method of digital intervention. We will describe skewed data narratively (e.g. as medians and interquartile ranges for each group).

When a single study reports multiple trial arms, we will include only the relevant arms. If two comparisons (e.g. intervention A vs control and intervention B vs control) are combined in the same meta‐analysis, we will combine the active arms or will halve the control group to avoid double‐counting.

If adjusted analyses are available (ANCOVA), we will use these as a preference in our meta‐analyses. If both changes from baseline and endpoint scores are available for continuous data, we will use change from baseline unless we note low correlation between measurements in individuals. In addition, we will not combine change from baseline and endpoint scores in analyses using the SMD. If a study reports outcomes at multiple time points, we will use the measure taken at the last follow‐up. We will use the SDs of final (rather than baseline) measurements in the analysis.

We will use intention‐to‐treat (ITT) or 'full analysis set' analyses when they are reported (i.e. those in which data have been imputed for participants who were randomly assigned but did not complete the study) in preference to available case or per‐protocol analyses, if both are reported.

Unit of analysis issues

For dichotomous outcomes, we will use participants, rather than events, as the unit of analysis (i.e. number of children admitted to hospital rather than number of admissions per child). However, if rate ratios are reported in a study (e.g. for exacerbations), we will analyse them on this basis. We will meta‐analyse data from cluster‐RCTs only if available data have been adjusted (or can be adjusted) to account for the clustering. In keeping with recommendations from the Cochrane Handbook for Systemaitc Reviews of Interventions, we will adjust cluster‐randomised data by inflating standard errors using a design effect (DE) calculated with an intracluster correlation coefficient (ICC).

Dealing with missing data

We will contact investigators or study sponsors to verify key study characteristics and to obtain missing numerical outcome data when possible (e.g. when a study is identified as an abstract only). When this is not possible and the missing data are thought to introduce serious bias, we will exclude the study.

Assessment of heterogeneity

We will use the Chi² test of homogeneity and the I² statistic to measure heterogeneity among the studies included in each analysis. If we identify substantial heterogeneity, we will report this and will explore the possible causes by performing prespecified subgroup analysis. Higgins et al suggests using an I² value of 75% and over to indicate high heterogeneity (Higgins 2003).

Assessment of reporting biases

If we are able to pool more than 10 studies, we will create and examine a funnel plot to explore possible small‐study and publication biases using Egger's t‐test.

Data synthesis

We will use a random‐effects model and will perform a sensitivity analysis using a fixed‐effect model.

Subgroup analysis and investigation of heterogeneity

We plan to carry out the following subgroup analyses.

1. Interventions that have used only one digital component versus interventions with multiple (more than one) digital component.

2. Different types of digital interventions (i.e. online vs computer‐based vs electronic monitoring devices).

3. Interventions with an interactive component versus non‐interactive interventions.

4. Digital interventions involving adherence feedback versus interventions that do not.

5. Theory‐based versus non‐theory‐based digital interventions.

6. Interventions for ICS versus non‐ICS therapies.

7. Primary versus secondary care setting (defined in terms of where participants were recruited for the study).

8. Interventions with an 'in‐person' component versus interventions that are fully digital and self‐delivered.

We will use the primary outcomes in the subgroup analyses.

1. Adherence to maintenance medication via any objective or validated subjective measure of adherence.

2. Asthma control via any validated self‐report instrument.

3. Exacerbations requiring at least oral corticosteroid treatment.

We will use the formal test for subgroup interactions available in Review Manager (RevMan 2014).

Sensitivity analysis

We plan to carry out the following sensitivity analyses while removing these items from primary outcome analyses.

1. Unpublished data.

2. Trials with high risk of selection bias.

3. Trials via subjective adherence outcome measurement methods.

4. Quasi‐randomised trials.

5. Non‐English studies.

6. Commercially funded studies.

We will compare results from a fixed‐effect model versus results from a random‐effects model.

For cluster‐randomised trials, we will run the main analyses using more and less conservative estimates of the ICC.