Mobile phone‐based interventions for improving adherence to medication prescribed for the primary prevention of cardiovascular disease in adults

Melissa J Palmer; Kazuyo Machiyama; Susannah Woodd; Anasztazia Gubijev; Sharmani Barnard; Sophie Russell; Pablo Perel; Caroline Free

doi:10.1002/14651858.CD012675.pub3

Mobile phone‐based interventions for improving adherence to medication prescribed for the primary prevention of cardiovascular disease in adults

Authors' declarations of interest

Version published: 26 March 2021 Version history

https://doi.org/10.1002/14651858.CD012675.pub3

Collapse all Expand all

Abstract

available in

Background

Cardiovascular disease (CVD) is a major cause of disability and mortality globally. Premature fatal and non‐fatal CVD is considered to be largely preventable through the control of risk factors by lifestyle modifications and preventive medication. Lipid‐lowering and antihypertensive drug therapies for primary prevention are cost‐effective in reducing CVD morbidity and mortality among high‐risk people and are recommended by international guidelines. However, adherence to medication prescribed for the prevention of CVD can be poor. Approximately 9% of CVD cases in the EU are attributed to poor adherence to vascular medications. Low‐cost, scalable interventions to improve adherence to medications for the primary prevention of CVD have potential to reduce morbidity, mortality and healthcare costs associated with CVD.

Objectives

To establish the effectiveness of interventions delivered by mobile phone to improve adherence to medication prescribed for the primary prevention of CVD in adults.

Search methods

We searched CENTRAL, MEDLINE, Embase, and two other databases on 7 January 2020. We also searched two clinical trials registers on 5 February 2020. We searched reference lists of relevant papers. We applied no language or date restrictions.

Selection criteria

We included randomised controlled trials investigating interventions delivered wholly or partly by mobile phones to improve adherence to cardiovascular medications prescribed for the primary prevention of CVD. We only included trials with a minimum of one‐year follow‐up in order that the outcome measures related to longer‐term, sustained medication adherence behaviours and outcomes. Eligible comparators were usual care or control groups receiving no mobile phone‐delivered component of the intervention.

Data collection and analysis

We used standard methodological procedures recommended by Cochrane. The main outcomes of interest were objective measures of medication adherence (blood pressure (BP) and cholesterol), CVD events, and adverse events. We contacted study authors for further information when this was not reported.

Main results

We included 14 trials with 25,633 randomised participants. Participants were recruited from community‐based primary and tertiary care or outpatient clinics. The interventions varied widely from those delivered solely through short messaging service (SMS) to those involving a combination of modes of delivery, such as SMS in addition to healthcare worker training, face‐to‐face counselling, electronic pillboxes, written materials, and home blood pressure monitors. Some interventions only targeted medication adherence, while others additionally targeted lifestyle changes such as diet and exercise. Due to heterogeneity in the nature and delivery of the interventions and study populations, we reported most results narratively, with the exception of two trials which were similar enough to meaningfully pool in meta‐analyses.

The body of evidence for the effect of mobile phone‐based interventions on objective outcomes of adherence (BP and cholesterol) was of low certainty, due to most trials being at high risk of bias, and inconsistency in outcome effects. Two trials were at low risk of bias.

Among five trials (total study enrolment: 5441 participants) recording low‐density lipoprotein cholesterol (LDL‐C), two studies found evidence for a small beneficial intervention effect on reducing LDL‐C (−5.30 mg/dL, 95% confidence interval (CI) −8.30 to −2.30; and −9.20 mg/dL, 95% CI −17.70 to −0.70). The other three studies found results varying from a small reduction (−7.7 mg/dL) to a small increase in LDL‐C (0.77 mg/dL). All of which had wide confidence intervals that included no effect.

Across 13 studies (25,166 participants) measuring systolic blood pressure, effect estimates ranged from a large reduction (MD −12.45 mmHg, 95% CI −15.02 to −9.88) to a small increase (MD 2.80 mmHg, 95% CI 0.30 to 5.30). We found a similar range of effect estimates for diastolic BP, ranging from −12.23 mmHg (95% CI −14.03 to −10.43) to 1.64 mmHg (95% CI −0.55 to 3.83) (11 trials, 19,716 participants). Four trials showed intervention benefits for systolic and diastolic BP with confidence intervals excluding no effect, and among these were all three of the trials evaluating self‐monitoring of blood pressure with mobile phone‐based telemedicine. The fourth trial included SMS and provider support (with additional varied features). Seven studies (19,185 participants) reported 'controlled' BP as an outcome, and intervention effect estimates varied from negligible effects (odds ratio (OR) 1.01, 95% CI 0.76 to 1.34) to large improvements in BP control (OR 2.41, 95% CI: 1.57 to 3.68). The three trials of clinician training or decision support combined with SMS (with additional varied features) had confidence intervals encompassing benefits and harms, with point estimates close to zero. Pooled analyses of the two trials of interventions solely delivered through SMS were indicative of little or no beneficial intervention effect on systolic BP (MD −1.55 mmHg, 95% CI −3.36 to 0.25; I²= 0%) and small increases in controlled BP (OR 1.32, 95% CI 1.06 to 1.65; I²= 0%).

Based on four studies (12,439 participants), there was very low‐certainty evidence (downgraded twice for imprecision and once for risk of bias) relating to the intervention effect on combined (fatal and non‐fatal) CVD events.

Two studies (2535 participants) provided low‐certainty evidence for the effect of the intervention on cognitive outcomes, with little or no difference between trial arms for perceived quality of care and satisfaction with treatment.

There was moderate‐certainty evidence (downgraded due to risk of bias) that the interventions did not cause harm, based on six studies (8285 participants). Three studies reported no adverse events attributable to the intervention. One study reported no difference between groups in experience of adverse effects of statins, and that no participants reported intervention‐related adverse events. One study stated that potential side effects were similar between groups. One study reported a similar number of deaths in each arm, but did not provide further information relating to potential adverse events.

Authors' conclusions

There is low‐certainty evidence on the effects of mobile phone‐delivered interventions to increase adherence to medication prescribed for the primary prevention of CVD. Trials of BP self‐monitoring with mobile‐phone telemedicine support reported modest benefits. One trial at low risk of bias reported modest reductions in LDL cholesterol but no benefits for BP. There is moderate‐certainty evidence that these interventions do not result in harm. Further trials of these interventions are warranted.

PICOs

Population

Intervention

Comparison

Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Plain language summary

available in

Interventions delivered by mobile phone to help people adhere to medication to prevent heart and circulatory disease

Review question

We reviewed the evidence on the effect of interventions delivered by mobile phone to help people in taking their medication to prevent cardiovascular disease (for example, heart attacks and strokes).

Background

Around 17.6 million people die from cardiovascular disease every year. Medications can help to prevent cardiovascular disease, but many people who have been given these medications do not take them as often or as consistently as recommended. This means that the medication will not work as well as it could to prevent cardiovascular disease. Interventions delivered through mobile phones, for example, prompting by text messaging, may be a low‐cost way to help people to take their medication as recommended.

Study characteristics

The evidence is up to date to January 2020. We found 14 studies that tested interventions delivered at least partly by mobile phone, which followed up participants for at least 12 months.

Key results

We were not able to combine the results of most of the trials, because the interventions were very different. Two studies were at low risk of bias and 12 were at high risk of bias. The effects of the interventions were inconsistent across studies, and so we are not confident about their findings. Self‐monitoring of blood pressure plus telemedicine support by mobile phone may improve blood pressure control, but we are not confident about the findings due to trials being at risk of bias. Interventions delivered by text message alone may have little or no effect on blood pressure control. Interventions which included text messages and clinician training or clinician decision support (with or without additional features) may have little or no effect on blood pressure or cholesterol. The effects of the interventions which included text messages and provider support (with or without other features) were inconsistent across studies, and so we are not confident about their findings. We are uncertain about the effects of apps held by the patient or apps with additional provider support. Some interventions delivered by mobile phone may help people to take their medication, but the benefits are small or modest. Some trials found that the interventions did not have any beneficial effect. There was no evidence to suggest that these types of interventions caused harm.

Authors' conclusions

Implications for practice

Our results are based on 14 trials, of which two were considered to be at low risk of bias.

One trial at low risk of bias reported a reduction in low‐density lipoprotein cholesterol (LDL‐C) of 5.3 mg/dl (Choudhry 2018). The Cholesterol Treatment Trialists' Collaboration estimates that for each 1 mmol/L (38.67 mg/dL) reduction in LDL‐C there is a consistent 20% relative risk reduction for major vascular events, regardless of baseline risk (CTT 2012). So this equates to a 2.7 % relative risk reduction in major cardiovascular events. The other four trials measuring LDL‐C as an outcome reported intervention effects ranging from a 9.2 mg/dL reduction to a 0.77 mg/dL increase, meaning that even the larger of these effects would have a small or modest impact on clinical outcomes.

No trials at low risk of bias reported reductions in blood pressure (BP). All three of the trials of interventions involving a home blood‐pressure monitoring system alongside telemedicine support via mobile phone reported reductions in systolic BP of −6.6, −7.1,and 4.7 mmHg respectively, and of diastolic BP of −5.4, −3.9 and −1.3 respectively (He 2017; Logan 2012; McManus 2018). Differences in the effects may be due to differences in the control group (standard care or self‐monitoring without mobile phone‐based telemedicine support), differences in the content and media for delivering the telemedicine support, chance or bias. A 10 mg drop in systolic BP or a 5 mg drop in diastolic BP gives 22% fewer coronary events at one year and 41% fewer strokes, so these effects are clinically important (Collins 1990; Law 2009). Those considering implementing similar interventions should take into account the risks of bias in these trials and whether additional costs are incurred. Liu 2015 also reported clinically important reductions in systolic and diastolic BP of −12 mmHg associated with an intervention including SMS with a computerised CVD risk evaluation and face‐to‐face counselling. The trials of interventions delivered by SMS alone or interventions delivering SMS to patients plus clinician training / decision support, reported little or no benefit on mean BP. The delivery of mobile phone‐based interventions is inexpensive once systems are set up and previous analyses of such interventions in other fields have demonstrated cost effectiveness (Guerriero 2013; Lester 2010). If interventions were shown to be cost‐effective the modest benefits achieved at low cost would be important if achieved across whole populations.

While many of the trials included components of the intervention delivered to healthcare providers as well as patients or delivered by other more resource‐intensive means to patients, such as face‐to‐face counselling sessions with healthcare workers in addition to text messaging, only two out of seven of these reported clinically and statistically significant intervention benefits.

Implications for research

One trial at low risk of bias reported benefits on LDL cholesterol. It remains unclear why this intervention reported benefits whilst others with some apparently similar components did not. Given the heterogeneity of the interventions, future research should evaluate the effect of this intervention in other settings to provide greater certainty about the effects.

Trials involving self‐monitoring and telemedicine support reported benefits but were at risk of bias. Further trials addressing methodological or reporting limitations are therefore needed. Trial results suggested that different means of providing the telemedicine support component may have different effects, which could usefully be evaluate by future trials.

The two interventions delivered by SMS alone were developed with input from users (Bobrow 2016; Tobe 2019). The intervention by Bobrow 2016 by targeted many of the barriers to adherence, which might be addressed using SMS. Nonetheless, the modest or absent benefit on mean BP and BP control is consistent with results of adherence interventions delivered by SMS for secondary prevention of CVD, HIV medication and diabetes (Adler 2017; Anglada‐Martinez 2015; Farmer 2016). Adherence is influenced by a wide range of service and social factors, in addition to the individual‐level factors like knowledge, motivation and skills (DiMatteo 2004; Julius 2009; Kardas 2013; Nieuwlaat 2014; Pound 2005; Vermeire 2001). Future adherence interventions should build on existing knowledge by considering the content of adherence intervention shown to be effective previously and by considering the broad range of factors influencing adherence that may be amenable to change. Future trials should target people most at risk of poor adherence and should exclude those known to be adherent.

For the indirect measures of adherence, the largest effect estimates related to those outcome measures reliant on participants' self‐report. Where possible, future trials should prioritise including measures of adherence which are less subject to bias resulting from unblinded participants.

Finally, given the heterogeneity that exists between behaviour‐change interventions, we believe that further high‐quality adequately‐powered trials of particular interventions would provide higher‐quality evidence relating to the effectiveness, compared with evidence based on attempts to pool multiple smaller, lower‐quality and potentially heterogeneous interventions and trials.

Summary of findings

Open in table viewer

Summary of findings 1. Mobile phone interventions compared to usual care for improving adherence to medication prescribed for primary prevention of cardiovascular disease

Outcomes	Impact	№ of participants^f (studies)	Certainty of the evidence (GRADE)
Mobile phone interventions compared to usual care for improving adherence to medication prescribed for primary prevention of cardiovascular disease
Patient or population: people prescribed medication for primary prevention of cardiovascular disease Setting: community and healthcare settings Intervention: mobile phone‐based interventions Comparison: usual care, passive text messages, or 'enhanced' usual care
Objective measure of medication adherence: Cholesterol (low‐density lipoprotein) follow‐up: range 1 – 2 years	2 studies found evidence of a small beneficial intervention effect on reducing LDL‐C (−9.20 mg/dL, and 5.3 mg/dL), and 3 studies found results varying from a small reduction (−7.7 mg/dL) to a small increase in LDL‐C (0.77 mg/dL), all of which had wide confidence intervals that included no effect.	5,441 (5 RCTs)	⊕⊕⊝⊝ Low^a,b
Objective measure of medication adherence: Blood pressure follow‐up: range 1 – 2 years	Systolic BP: 9 of 13 studies found lower systolic blood pressure with mobile‐phone interventions, although only 4 of these reductions in systolic blood pressure had confidence intervals excluding no effect. Across the 13 studies, effect estimates varied greatly, from those showing a large reduction (−12.45 mmHg) to those reporting a small increase (+2.80 mmHg) in systolic blood pressure. Meta‐analysis of 2 trials evaluating an intervention targeting adherence to blood pressure medication delivered solely by SMS messaging provided a pooled MD of −1.55 mmHg, 95% CI −3.36 to 0.25.	25,166 (13 RCTs)	⊕⊕⊝⊝ Low^a,b
	Diastolic BP: 8 of 11 studies found lower diastolic blood pressure with mobile‐phone interventions, but in 4 of these the confidence intervals included no effect. Across the 11 studies, effect estimates varied widely from those showing a large reduction (−12.23 mmHg) to those showing a small increase (+1.64 mmHg) in diastolic blood pressure.	19,716 (11 RCTs)
	Controlled BP: 7 studies reported 'controlled' blood pressure as an outcome, of which six reported increased blood pressure control with mobile phone interventions, although in only one of these studies did the confidence interval exclude no effect. Effect estimates varied from negligible (OR 1.01) to large improvements in blood pressure control (OR 2.41). Meta‐analysis of 2 trials evaluating an intervention targeting adherence to blood‐pressure medication delivered solely by SMS messaging indicated a modest beneficial intervention effect: pooled OR of 1.32, 95% CI 1.06 to 1.65.	19,185 (7 RCTs)
Combined CVD events	1 trial reported on deaths due to CVD, and 3 recorded non‐fatal CVD events. For 3 studies the effect estimate was in the direction of harm, and for the 4th it was in the direction of intervention benefit. However, the number of events in each trial was low and all effect estimates had wide 95% confidence intervals encompassing no effect.	12,439 (4 RCTs)	⊕⊝⊝⊝ Very low^{c, d}
Adverse events follow‐up: range 1 – 2 years	3 studies reported that there were no adverse events attributable to the intervention. 1 reported that there was no difference between groups in adverse effects of statins, and that no participants reported intervention‐related adverse events. 1 study reported that potential side effects were similar between groups. 1 study reported a similar number of deaths in the intervention and control arms, but did not provide further information relating to potential adverse events.	8285 (6 RCTs)	⊕⊕⊕⊝ Moderate^b
Cognitive outcome: satisfaction with treatment follow‐up: mean 1 year	1 study measured satisfaction with treatment, and found no evidence of a difference between intervention and control arms. 1 study reported on perceived quality of care, with little difference observed between the 2 groups.	2535 (2 RCT)	⊕⊕⊝⊝ Low^d,e
LDL‐C: low‐density lipoprotein cholesterol; BP: blood pressure; RCT: randomised controlled trial
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.
^aDowngraded one level for inconsistency: trial results included large variations in the degree to which the outcome was affected. ^bDowngraded one level for risk of bias: most trials at unclear risk of bias for multiple domains. ^c Downgraded two levels for imprecision: very few events and wide confidence intervals encompassing intervention benefit and harm. ^dDowngraded one level for risk of bias: trials at unclear or high risk of bias for several domains. ^eDowngraded one level for indirectness: based on two trials, one conducted in public sector clinic in Cape Town, South Africa, and one in community health settings in India. ^f Total study enrolment.

Background

Description of the condition

Cardiovascular disease (CVD) is a major cause of disability and mortality throughout the world (Roth 2018; WHO 2011; WHO 2016), with an estimated 17.8 million people dying from CVDs in 2017, accounting for almost a third of all global deaths (Roth 2018). However, premature fatal and non‐fatal CVD is considered to be largely preventable through the control of risk factors (WHO 2011).

Primary prevention of CVD refers to actions taken to reduce the incidence of clinical events due to coronary heart disease (CHD), cerebrovascular disease and peripheral vascular disease, among people with risk factors who have not yet developed clinically‐manifest CVD (WHO 2007). Primary prevention of CVD consists of lifestyle modifications (e.g. smoking cessation, increasing physical activity) and drug therapy (Piepoli 2016).

Lipid‐lowering and antihypertensive drug therapies for primary prevention are cost‐effective in reducing CVD morbidity and mortality among high‐risk people and are recommended by international guidelines (Piepoli 2016; WHO 2007). Recommendations relating to the use of antiplatelet drugs for primary prevention vary. The European Society of Cardiology (ESC) states that aspirin cannot be recommended in primary prevention due to its increased risk of major bleeding (Piepoli 2016); however, the US Preventive Services Task Force (USPSTF) recommends the use of aspirin when the 10‐year risk of CVD events reaches such a level that the benefits of aspirin, in terms of CVD events prevented, outweigh the potential harm of increased gastrointestinal haemorrhage (USPSTF 2014).

Adherence to long‐term medication is not ideal and results in costs in both health and economic terms (Piepoli 2016). Meta‐analyses have estimated rates of adherence to cardiovascular medications ranging from 50% to 60% (Chowdhury 2013; Naderi 2012), and there is some evidence that adherence is lower for primary prevention (Naderi 2012).

One study of health records of over 430,000 people in UK general practices found that 47% of people prescribed statins for primary prevention discontinued treatment (indicated by a greater than 90‐day gap between prescriptions). Among these people, 72% then restarted treatment (Vinogradova 2016). One study of Finnish healthcare registers found that 53% of women prescribed statin therapy for primary prevention were adherent (defined as exceeding 80% of the prescribed regimen) (Lavikainen 2016). It has been estimated that approximately 9% of cases of CVDs in the EU could be attributed to poor adherence to vascular medications (Chowdhury 2013). Improving adherence to medications for the primary prevention of CVD would help to maximise the clinical benefits for the wider population (WHO 2003). There is therefore considerable scope for increasing adherence to prescribed medicine, and thereby reducing morbidity, mortality and healthcare costs.

Description of the intervention

Mobile phone ownership is almost universal in high‐income countries and estimated to have reached over 90% in low‐ and middle‐income countries (ICT 2016). Given the broad reach of mobile phones and the potential for automation of delivery, interventions delivered by mobile phone are a potentially cost‐effective strategy to improve medication adherence. A range of media can be delivered through mobile phones, including text messages, picture messages, interactive‐voice response, telephone calls and, with increasing ownership of smart phones with Internet capabilities (ICT 2016), mobile applications.

How the intervention might work

A wide range of factors have been shown to be associated with medication non‐adherence (DiMatteo 2004; Julius 2009; Kardas 2013; Pound 2005; Vermeire 2001; WHO 2003). Mobile phone‐based interventions have the potential to target a number of these factors. For example, lack of adherence resulting from lack of information about the benefits of medication, lack of information about how they work and how to take them, misconceptions about medication adverse effects, complex or unclear advice or poor recall of information provided in consultations may be addressed through text messages providing short and simply‐worded snippets of information (Julius 2009; Kardas 2013; Pound 2005; Vermeire 2001). Experiences of adverse effects can be targeted through mobile phone‐delivered interventions by providing information about medication and facilitating a link to a healthcare professional for people experiencing problems with their medication. Lack of social support has also been linked to poor medication adherence; previous qualitative research found that the receipt of text message‐based intervention provided social support (Douglas 2013). Mobile phone‐delivered interventions can be designed to target psychological factors such as lack of motivation and low self‐efficacy (Free 2016).

Existing interventions targeting adherence to CVD medication have employed mobile technologies to: deliver medication reminders (Park 2014a); encourage self‐monitoring of medication intake (Park 2014a); encourage habit formation relating to medication‐taking behaviours (Bobrow 2014); provide information (Bobrow 2014; Park 2014a); and facilitate links to healthcare services where required (Bobrow 2014; Piette 2012).

Systematic reviews assessing the effect of mobile health (mhealth) interventions on medication adherence for a range of conditions, including HIV, non‐communicable diseases and prevention of transplant rejection have reported significant improvements (Anglada‐Martinez 2015; Park 2014b). An RCT has found mobile phone messaging to be effective in improving contraceptive use (Smith 2015). Few adverse effects of mobile phone‐based interventions have been reported; potential, but rare, adverse events may include road traffic accidents (Caird 2014).

Why it is important to do this review

Systematic reviews evaluating the effect of mhealth interventions have reported promising but inconclusive results about improved medication adherence, including adherence to medication for secondary prevention of heart disease (Adler 2017; Anglada‐Martinez 2015; Park 2014b). However, no systematic review has specifically examined the effect of mobile phone‐based interventions on adherence to medications for the primary prevention of CVD. Mobile phone‐based interventions are of particular interest, given their low cost and potential for widespread delivery.

Objectives

To establish the effectiveness of interventions delivered by mobile phone to improve adherence to medication prescribed for the primary prevention of CVD in adults.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs) of parallel‐group design that randomised by participant or by cluster. We did not include cross‐over trials, as this design would be inappropriate for assessing effects on cardiovascular events or mortality, due to the irreversible nature of these events. We only included trials with a minimum of one‐year follow‐up in order that the outcome measures relate to longer‐term, sustained medication adherence behaviours and outcomes. We included studies published as full text and as abstract only, and unpublished data.

Types of participants

We included adults (aged 18 years and over) who have been prescribed medication for the primary prevention of CVD. As this review focuses on the primary prevention of CVD, we only included studies involving participants who had not had a prior CVD event, defined as: a previous myocardial infarction, stroke, revascularisation procedure (coronary artery bypass grafting or percutaneous coronary intervention), people with angina, and people with angiographically‐defined CHD. Where we identified trials that included a subset of eligible participants, we contacted the authors to request data for only those participants of interest. When we were unable to access these data, we applied a cut‐off which included only trials in which at least 75% of participants met the criteria for primary prevention.

Types of interventions

We included trials of interventions delivered wholly or partly by mobile phone to improve adherence to cardiovascular medications prescribed for the primary prevention of CVD. We included interventions targeting adherence to antihypertensive drugs (thiazide‐like diuretic, angiotensin‐converting enzyme inhibitor, calcium channel blocker, beta‐blocker); lipid‐lowering drugs (statins); and antiplatelet drugs (low‐dose aspirin, non‐aspirin antiplatelet drugs). We only included trials targeting adherence to at least one of these medications. We also included trials of interventions that targeted medication adherence alongside other lifestyle modifications.

Intervention

Any mobile phone‐specific delivery mechanism, including short messaging service (SMS), multimedia messaging (MMS), applications (apps) and Interactive Voice Response. We included interventions employing a mix of delivery mechanisms of which at least one was mobile phone‐based, for example, interventions delivered by mobile phones in combination with traditional methods such as face‐to‐face communication and links to other types of support (e.g. healthcare support worker, telephone calls, Internet pages).

Comparator

Usual care and active controls, where the control group intervention had no component delivered by a mobile phone‐specific delivery mechanism.

Types of outcome measures

Reporting one or more of the outcomes listed here in the trial was not an inclusion criterion for the review.

Where outcomes (primary or secondary) were measured at multiple time points, we extracted data for the final point of measurement.

Primary outcomes

Objective measures of adherence to treatment (low‐density lipoprotein cholesterol (LDL‐C), total cholesterol (TC) and high‐density lipoprotein cholesterol (HDL‐C), for the effect of statins; blood pressure for antihypertensive drugs; heart rate for the effect of atenolol; urinary 11‐dehydrothromboxane B2 for the antiplatelet effects of aspirin).
Combined CVD events (fatal or non‐fatal events).
Adverse effects including self‐reported road traffic accidents.

Secondary outcomes

Indirect measures of adherence to treatment (self‐report, tablet counts, medication event monitoring systems, pharmacy prescription data).
Fatal cardiovascular events.
Non‐fatal cardiovascular events (CHD, stroke).
Health‐related quality of life assessed using validated instruments (e.g. 36‐Item Short Form Health Survey (SF‐36), EQ‐5D).
Cognitive outcomes (any measures of: satisfaction with treatment, medication‐taking self‐efficacy, autonomy related to medication, attitudes (e.g. concerns about medicine adverse effects)).
Costs.

We also reported on the following process measures: extent of intervention received (e.g. number of text messages received, measures of use of allocated mobile application) and acceptability of intervention.

Search methods for identification of studies

Electronic searches

We identified trials through systematic searches of the following bibliographic databases on 7 January 2020:

Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (Issue 1 of 12, 2020);
Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, MEDLINE Daily and MEDLINE (Ovid, 1946 to 6 January 2020);
Embase (Ovid, 1980 to 2020 week 1);
CINAHL Plus (EBSCOhost, 1937 to 7 January 2020);
Conference Proceedings Citation Index‐Science (CPCI‐S) on Web of Science (Clarivate Analytics, 1990 to 7 January 2020).

The search strategies are presented in Appendix 1. The Cochrane sensitivity‐precision maximising RCT filter was applied to MEDLINE (Ovid) and adaptations of it to the other databases, except CENTRAL (Lefebvre 2011).

We carried out a search of www.ClinicalTrials.gov and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) Search Portal (apps.who.int/trialsearch/) for ongoing or unpublished trials on 5 February 2020.

We imposed no restriction by date or language of publication.

We did not perform a separate search for adverse effects of mobile phone‐based interventions targeting medication adherence. We considered adverse effects described in included studies only.

Due to the disruptions as a result of the Covid‐19 pandemic, it was not possible to publish this review within 12 months of the search being conducted. We therefore repeated the search strategy on 8 January 2021, and screened these results, with potentially eligible studies added to 'Studies awaiting classification'.

Searching other resources

We checked the reference lists of all included studies and reviewed relevant articles for additional references. We also examined relevant retraction statements and errata for included studies.

Data collection and analysis

Selection of studies

Two review authors independently screened the titles and abstracts of all identified potential studies to decide whether to retrieve the full text (eligible or potentially eligible/unclear studies) or to discard the study. Two review authors independently screened the retrieved full texts to identify studies for inclusion and identify and record reasons for exclusion of the ineligible studies in the Characteristics of excluded studies table. We resolved any disagreements though discussion, and where necessary, a third review author arbitrated. We excluded any duplicates. We collated multiple reports of the same RCT into a single entry. We completed a PRISMA flow diagram (Liberati 2009).

Data extraction and management

We used a standardised, prepiloted form to extract data from the included studies for assessment of study quality and evidence synthesis. We contacted chief investigators for additional information where necessary. We extracted the following information.

Methods: study design; total duration of study; study setting and date of study.
Participants: number randomised; number lost to follow‐up/withdrawn; number analysed; mean age; age range; gender; proportion meeting criteria of 'primary prevention'; and inclusion criteria and exclusion criteria.
Interventions: intervention; comparison; concomitant medications; excluded medications; intervention delivery mechanism (text messages/MMS/mobile application/combined); how intervention was developed; if intervention was personalised; and frequency and duration of intervention receipt.
Outcomes: primary and secondary outcomes specified and collected; adverse effects; and time points reported.
Notes: funding for trial and notable conflicts of interest of trial authors.

Two review authors independently extracted data and resolved any differences by returning to the original study reports and discussion with a third review author where necessary. One review author transferred data into Review Manager 5 (Review Manager 2020). To ensure that there were no errors in data entry, another review author checked that the data entered into Review Manager 5 were consistent with those in the data extraction form.

Assessment of risk of bias in included studies

Two review authors independently assessed the risks of bias for each study using the criteria detailed in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). For each of the following domains, we graded the potential bias as high, low or unclear.

Random sequence generation.
Allocation concealment.
Blinding of participants and personnel.
Blinding of outcome assessment (objective and subjective self‐reported outcomes assessed separately).
Incomplete outcome data.
Selective outcome reporting.
Other biases (selective cluster recruitment for cluster RCTS).

We resolved disagreements by discussion. Where necessary, we consulted a third review author to arbitrate. We constructed a 'Risk of bias' table including justifications for our judgements. Where information relating to the risk of bias came from unpublished data or correspondence with an author, we noted this. We summarised the 'Risk of bias' judgements across different studies for each of the domains listed. When considering treatment effects, we accounted for the risk of bias for the studies that contributed to that outcome.

For other potential sources of bias, we assessed evidence for selective cluster recruitment for the included cluster RCTS. We assessed the blinding of outcome assessment domains separately for objectively‐measured outcomes, and self‐reported subjective outcomes. Given the nature of the interventions included in this review, it is likely that blinding of participants and personnel would be impossible, as would blinding of self‐reported outcome assessment, so we expected trials to be categorised at high risk of bias for both of these domains. For the overall study assessment, we categorised a trial as being at low risk of bias if it was rated as low risk in all the domains listed above (with the exception of blinding of participants and personnel/self‐reported outcome assessment). Trials that were at high or unclear risk of bias for any of the domains (except blinding of participants and personnel/self‐reported outcome assessment) were categorised as being at high risk of bias.

Assessment of bias in conducting the systematic review

We conducted the review according to the published protocol and report any deviations from it in the Differences between protocol and review section (Palmer 2017).

Measures of treatment effect

We analysed dichotomous outcome data as odds ratios (ORs) with 95% confidence intervals (CIs). We analysed continuous outcome data as mean differences (MDs) with 95% CIs, or if a continuous outcome had been measured in multiple ways, as a standardised mean difference (SMD) with 95% CIs.

Unit of analysis issues

Due to the heterogeneity of the included studies' intervention content and delivery mechanisms, most of the results in the review are described narratively. However, in doing this we entered data in to RevMan to construct forest plots to aid data presentation without pooling. For cluster‐randomised trials, we extracted the effect estimates adjusted for clustering where these were reported. Where cluster‐randomised trials presented results which did not account for clustering, we recalculated effect estimates based on the 'effective sample size' derived from intracluster coefficients (ICCs) for CVD‐related outcomes by Singh 2015 and Lee 2020 (specific ICCs used are noted in the footnotes of the corresponding data and analyses). For the two meta‐analyses conducted in this review, one of the contributing trials had two eligible intervention arms (Bobrow 2016), and we therefore halved the number in the control group to avoid double‐counting. We excluded intervention arms not appropriate for this review.

Dealing with missing data

We contacted investigators to obtain further information where necessary (e.g. when the study included a mixed population of participants who met the criteria for primary prevention and participants who met the criteria for secondary prevention, and when only a subset of participants had been prescribed CVD preventive medication). We also planned to contact investigators or study sponsors to obtain missing data (e.g. when a study was available as abstract only). We planned that where this was not possible, and the missing data were considered a potential source of serious bias, we would conduct a sensitivity analysis to explore the impact of including such studies in the overall assessment of results.

Assessment of heterogeneity

With the exception of two trials, both of which used a text messaging‐based intervention to target blood pressure medication adherence and reported blood pressure outcomes (Bobrow 2016; Tobe 2019), we considered the included trials to be too methodologically heterogeneous to pool the data in a meta‐analysis. We therefore describe most results narratively. We planned to use the I² statistic to measure heterogeneity across the trials for the analysis of each outcome. In constructing the narrative forest plots for those outcomes reported by multiple studies, we calculated the I² statistic and reported this. Had we considered the trials methodologically similar enough to pool, and had we identified moderate to substantial heterogeneity (an I² statistic between 30% and 100%), we would have reported it and examined possible causes according to our prespecified subgroup analyses, subject to having a sufficient number of studies.

Assessment of reporting biases

We planned that if the results from more than 10 trials could be pooled, we would use a funnel plot to explore possible small‐study biases for the primary outcomes. However, we were able to pool only a maximum of two studies.

Data synthesis

We planned to carry out meta‐analyses only if it was meaningful to do so (i.e. if the interventions, participants and outcome measures were similar enough for pooling to make sense). Two trials were considered similar enough to pool results (Bobrow 2016; Tobe 2019); this was on the basis that both studies evaluated interventions targeting blood pressure medication adherence using text messages only, and recorded blood pressure outcomes. We did not undertake meta‐analyses for the rest of the included studies, as they were too heterogeneous in their content and delivery of their interventions. We present the effect estimates for outcomes reported by multiple studies in illustrative forest plots (without pooling); it should be noted that in transferring effect estimates from papers into Review Manager 5 using the generic inverse variance method, some CIs differed from those reported in the original paper by a decimal place, and where clustering has been accounted for using an external ICC (as described above), the CIs differ from those in the original report.

For the two studies pooled in meta‐analyses, we used a fixed‐effect model. In the presence of heterogeneity (an I² statistic in excess of 30%), we planned to examine whether this heterogeneity could be explained through our prespecified subgroup analyses. If these analyses accounted for the heterogeneity, we would only present the subgroup pooled effect estimates. If these subgroup analyses did not explain the heterogeneity, we would present results narratively. We intended to use fixed‐effect meta‐analysis and apply a conservative I² threshold to identify heterogeneity in this review to avoid overweighting smaller studies. This is because we consider that the heterogeneity observed in these behaviour‐change trials will primarily be a result of differences in the content of the interventions and differences in risk of bias.

Subgroup analysis and investigation of heterogeneity

We had planned to conduct the following subgroup analyses for the primary outcome of adherence to treatment if there had been sufficient studies to pool in meta‐analyses:

income region (by World Bank income group) (World Bank 2017);
how text messages were developed (i.e. theory‐based, incorporating user views and based on evidence relating to factors influencing behaviour‐targeted versus other);
delivery mechanisms (i.e. mobile phone messaging only, mobile applications only, combined mobile phone messaging and application, combined application and other).

Due to the limited number of studies, we were unable to conduct subgroup analyses. Should more trials become available for future updates of this review, we will re‐examine the planned subgroup analyses.

Sensitivity analysis

We planned to carry out a sensitivity analysis by only including studies with low risk of bias. As we only carried out a meta‐analysis of two studies, we did not conduct a sensitivity analysis.

Summary of findings and assessment of the certainty of the evidence

We created a 'Summary of findings' table of narrative results for the following outcomes: objective measures of adherence to treatment (cholesterol and blood pressure), combined CVD events (fatal and non‐fatal events), adverse events and cognitive outcomes. We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of the body of evidence as it related to the studies that contributed data for each outcome. We used methods and recommendations described in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2020) using GRADEpro software (GRADEpro GTD 2015). We justified decisions to downgrade the quality of studies using footnotes and made comments to aid readers' understanding of the review where necessary.

Results

Description of studies

Results of the search

From the previous version of this review there were four included studies and six ongoing studies. The new search of the databases retrieved 4206 records, and the search of the clinical trial registers retrieved an additional 53 records. After de‐duplication, we screened 2898 title and abstract records and excluded 2763 records. We assessed 135 full texts and excluded 85 references (78 studies). Combining the previous review and the updated search resulted in 13 ongoing studies (15 references), 19 studies (21 references) awaiting classification, and 14 studies (33 references) were eligible for inclusion. The flow diagram of search results is shown in Figure 1.

Figure 1

Study flow diagram.

Screening of the search that was repeated on 8 January 2021 resulted in a further 18 studies (19 references) awaiting classification.

Included studies

The Characteristics of included studies table presents details of the design, methods, participants, intervention, comparison and outcome measures for the studies included in this review. We identified 14 studies for inclusion, which were relatively heterogeneous, with particular variation in the nature (content and delivery) of the intervention, and the population.

Participants

The sample sizes of included studies ranged from 59 (Morillo‐Verdugo 2018) to 9642 (Peiris 2019), with a total of 25,633 randomised participants across all 14 included studies.

Liu 2015 specified that participants must have had "no known cardiovascular disease" as an inclusion criterion, and therefore included 100% of participants meeting the criteria for primary prevention. Morillo‐Verdugo 2018 was among participants living with HIV over 35 years of age, and "receiving active ART with at least 1 drug prescribed for the treatment of hypertension, dyslipidaemia, angina pectoris, cardiovascular prophylaxis, or type 2 diabetes"; 100% of participants met the criteria for primary prevention (personal communication with author).The other included studies had a mix of participants: in five studies, at least 90% of participants met the criteria of primary prevention (Choudhry 2018; Márquez Contreras 2019; McManus 2018; Párraga‐Martínez 2017; Prabhakaran 2019); four studies included at least 78% primary‐prevention participants (Bobrow 2016; He 2017; Logan 2012; Peiris 2019); and in one study approximately 65% of participants met the criteria for primary prevention, but results were reported separately for participants according to whether or not they had previously had a CVD event (Gulayin 2019). Saleh 2018 did not report the proportion of participants meeting the criteria for primary prevention, but stated that participants had to be registered at the primary healthcare centres "as diabetics or hypertensive and aged 40 years or more". We sought clarification from study authors on the proportion who had not had prior CVD events, but had not received this information at the time of submission. Tobe 2019 did not report the proportion of primary prevention participants (personal communication stated that this was not recorded), but inclusion criteria included age 18 years or more, uncontrolled hypertension (≥ 140/90 mmHg or ≥ 130/80 mmHg for diabetics), and stable on current dose of antihypertensive (if treated) for at least eight weeks.

There was heterogeneity between trials in the proportion of participants who were taking medication for the primary prevention of CVD. For four studies, having been prescribed medication for CVD prevention (anti‐hypertensives or lipid‐lowering medication) was an inclusion criterion, and therefore 100% of participants were receiving anti‐hypertensive medication at baseline (Bobrow 2016; Choudhry 2018; Márquez Contreras 2019; McManus 2018). Logan 2012 included at least 89.1% of participants prescribed medication (hypertensive drugs or lipid‐lowering drugs or aspirin, or a combination of these); almost 85% of participants in He 2017 were using antihypertensive medications at baseline; and Párraga‐Martínez 2017 stated that 68.1% of their sample had been prescribed lipid‐lowering medication (but did not mention other types of CVD prevention drugs). Based on communication with trial authors, at least 58% of participants in Prabhakaran 2019 were taking hypertensive or lipid‐lowering medication. Almost half of participants in Peiris 2019 were taking at least one anti‐hypertensive, lipid‐lowering or anti‐platelet medication. In Morillo‐Verdugo 2018, 28% of participants were prescribed lipid‐lowering medication, with additional prescriptions for other types of CVD drugs (data provided by drug class, so proportion of participants taking at least one CVD‐related drug not available); and Tobe 2019 had 28% of participants on anti‐hypertensive medications at baseline (personal communication). In Gulayin 2019, no participants were on CVD medication at baseline, with current statin use being an exclusion criterion, but the intervention sought to get participants on medication for hyperlipidaemia where indicated, and improve adherence to this medication. Liu 2015 did not report the proportion of participants prescribed medication, but explicitly stated that the intervention targeted adherence to medication among those on treatment. Similarly, Saleh 2018 did not report this proportion, but stated that the intervention targeted medication compliance.

The mean age of participants varied from approximately 49 years (Saleh 2018) to 67 years (McManus 2018). The proportion of women in the trial samples ranged from 9.4% (Morillo‐Verdugo 2018) to 72% (Bobrow 2016).

Settings

All studies recruited from healthcare settings, whether from outpatients attending clinics or outreach/home visits co‐ordinated by local health providers. Three studies were conducted in Spain: one recruited through four primary care centres in Huelva (Southern Spain) (Márquez Contreras 2019); one recruited participants living with HIV who had moderate or high cardiovascular risk from five tertiary hospitals (Morillo‐Verdugo 2018); and one recruited participants from primary care clinics in three health districts of three Spanish autonomous communities (Párraga‐Martínez 2017). Two trials were carried out in India: one recruited through community health centres in Haryana (North India) and Karnataka (South India) (Prabhakaran 2019), and one from primary health centres in Andhra Pradesh (South India) (Peiris 2019). Two trials recruited through primary healthcare centres in Argentina (Gulayin 2019; He 2017). Two studies were based in Canada: Logan 2012 recruited from the offices or clinics of physicians practising in metropolitan Toronto; and Tobe 2019 recruited through community healthcare provision from Canada’s First Nations communities living on six reserves in Northern Ontario, Quebec and New Brunswick. Bobrow 2016 recruited from an outpatient chronic disease service in a single, large, public sector clinic in Cape Town, South Africa. Liu 2015 recruited from a health management centre in a hospital in Guangzhou, China. Choudhry 2018 recruited from primary care practice sites of a large multi‐specialty group practice in Massachusetts, USA. McManus 2018 recruited from general practices in the UK. Finally, Saleh 2018 recruited from primary healthcare centres located in rural areas and Palestinian refugee camps across Lebanon.

Interventions

The content and delivery of the interventions varied across studies. In most of the trials, the interventions involved general health education, for example, targeting behaviours such as lifestyle modifications including healthy diet and physical activity, alongside messaging focusing on medication adherence for those prescribed CVD medication (Gulayin 2019; He 2017; Liu 2015; Márquez Contreras 2019; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Prabhakaran 2019; Peiris 2019; Saleh 2018). The interventions evaluated by Bobrow 2016 and Choudhry 2018 were specifically designed to focus primarily on medication adherence, with only a few references to other lifestyle modifications such as diet and physical exercise. Similarly, Tobe 2019 focused on the importance of blood pressure control and the rationale for medical therapy to encourage adherence. Two trials examined interventions which primarily consisted of blood pressure telemonitoring; one included feedback (via their smartphone) and could review their readings (Logan 2012), and the other did not provide such feedback but relied on participants to send their readings via SMS (McManus 2018). These mobile telemonitoring interventions were considered to implicitly target adherence to treatment as well as other health behaviours important for the control of blood pressure.

Two studies delivered the intervention solely through educational and motivational mobile‐phone text messages about hypertension and its medical therapy (Bobrow 2016; Tobe 2019), and one through a mobile application only, which allowed participants to record personal data, recommended BP levels as objectives, record the doctor's advice about the prescribed treatment, set reminder alarms, set a calendar of appointments or events, and record the results of the BP measurement (Márquez Contreras 2019). Three studies evaluated interventions involving text messages to participants, alongside additional components involving healthcare workers, such as a face‐to‐face counselling session (Liu 2015); pharmacotherapeutic follow‐up and an individual motivational interview (led by pharmacists) to work towards the achievement of pharmacotherapeutic objectives (Morillo‐Verdugo 2018), and additional training for healthcare providers focusing on clinical guidelines and provider‐patient communication strategies (Saleh 2018). Two studies examined interventions involving training for healthcare providers in clinical guidelines, a mobile/tablet‐based decision support system for clinical staff, and educational and motivational text messages directly to patients (Gulayin 2019; Prabhakaran 2019), and one study combined a tablet‐based decision support system for clinical staff with interactive voice response messaging delivered to participants (Peiris 2019). Choudhry 2018 evaluated an intervention involving text message (as reminders and motivational support for adherence), pillboxes, mailed personalised progress reports, and an individually‐tailored telephone consultation conducted by a staff clinical pharmacist. The intervention examined by He 2017 involved regular home visits from community health workers to providing education and counselling on lifestyle modification, home BP monitoring, and medication adherence skills; provision of an automatic home blood‐pressure monitor and seven‐day pill organiser; and individualised text messages to promote lifestyle changes and reinforce medication adherence. Párraga‐Martínez 2017 involved a combination of text messages, written information, and self‐completion cards for participants to record adherence to recommendations. Logan 2012 evaluated an intervention which involved the provision of a home blood‐pressure monitoring device and feedback to participants' smartphones, alongside an automated fax providing detailed information on the participants' status to their physicians on the day before their next scheduled appointment. In McManus 2018, intervention participants were provided with a home blood‐pressure monitoring device, and trained to send readings via a simple free SMS text‐based telemonitoring service, which alerted participants to contact their healthcare providers in response to very high or very low readings or if their average blood pressure was above target, reminded them if insufficient readings were transmitted, and presented readings to clinicians via a web interface.

Most studies had a control group that received usual care (Choudhry 2018; He 2017; Liu 2015; Márquez Contreras 2019; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Peiris 2019; Saleh 2018). The control group in Logan 2012 received the same home blood‐pressure monitoring equipment as the intervention group and a booklet containing information on the measurement of blood pressure, treatment of hypertension and goals of therapy. The control group in Bobrow 2016 received written information about hypertension and healthy living, and only received text messages that were sent to all trial participants, which primarily related to trial participation. The control group in Tobe 2019 received 'passive' text messages including healthy lifestyle and diet advice, with none of the content relating to blood pressure control. McManus 2018 had three arms: the telemonitoring arm (as described above), one arm which received usual care, and another group which was instructed to self‐monitor blood pressure and at the end of each week record their readings on paper and send them to their healthcare provider. For the purposes of this review, we extracted data from the usual‐care arm and excluded the self‐monitoring arm. In Prabhakaran 2019, participants in the control group received 'enhanced' usual care, whereby training on clinical guidelines was provided to healthcare providers; charts on the management of conditions were displayed prominently at the outpatient clinics; and nurses provided a lifestyle advice pamphlet to each participant. Gulayin 2019 reported that the control group received usual care, but clinics were also provided with educational flyers and written material for display.

Outcomes

All studies reported at least one objective measure related to medication adherence. Thirteen studies measured blood pressure (Bobrow 2016; Choudhry 2018; He 2017; Liu 2015; Logan 2012; Márquez Contreras 2019; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Prabhakaran 2019; Peiris 2019; Saleh 2018; Tobe 2019), and six studies measured cholesterol levels (at least one of the following: LDL‐C, HDL‐C, TC) (Choudhry 2018; Gulayin 2019; Liu 2015; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Prabhakaran 2019). Three studies provided outcome data relating to CVD events (McManus 2018; Peiris 2019; Tobe 2019) and one study reported on CVD‐related deaths (Bobrow 2016).

Six studies explicitly reported adverse events, including adverse medication effects of statins, intervention‐related adverse events and deaths (Bobrow 2016; He 2017; McManus 2018; Párraga‐Martínez 2017; Prabhakaran 2019; Tobe 2019).

Nine studies reported indirect measures of adherence to treatment (our secondary outcomes). Three studies included outcome data on self‐reported adherence measured using the Morisky‐Green scale (for lipid‐lowering medication (Gulayin 2019; Párraga‐Martínez 2017) and for anti‐hypertensive medication (He 2017)). Bobrow 2016 included self‐reported adherence to medication measured using a visual analogue scale, in addition to a measure of 'proportion of days of medication covered' (defined as the proportion of participants with 80% or more days covered with blood pressure‐lowering medication from prescribing and dispensing data routinely recorded in the clinical record, pharmacy record and Chronic Dispensing Unit record). Choudhry 2018 reported medication adherence for hyperlipidaemia and hypertension, assessed using prescription claims data and measured as the mean proportion of days covered (PDC) over the 12 months after randomisation. Márquez Contreras 2019 reported the percentage of participants who took antihypertensive drugs correctly on 80% to 100% of days, measured using a Medication Event Monitoring System (MEMS). Prabhakaran 2019 recorded the proportion of participants who reported adherence to their anti‐hypertensive drugs in the last seven days before the endline assessment. Morillo‐Verdugo 2018 recorded the proportion of participants adherent to 'concomitant medication' measured with "the Morisky‐Green questionnaire and pharmacy dispensing records". McManus 2018 recorded self‐reported adherence using the 'Medication Adherence Rating Scale', but the study report lacks detail on how this was scored.

Three trials also included a measure of quality of life (measured with the EuroQol Group 5‐Dimension Self‐Report Questionnaire) (Bobrow 2016; McManus 2018; Peiris 2019).

Two trials reported on cognitive outcomes: Bobrow 2016 compared satisfaction with treatment between the trial arms, and Prabhakaran 2019 asked participants about their perceived quality of care.

Four trials reported data relating to our process measures, including satisfaction with the intervention (Párraga‐Martínez 2017), adherence to the intervention home blood‐pressure monitoring schedule (Logan 2012), proportion responding to messages (Bobrow 2016), and proportion opting to receive intervention text messages (Choudhry 2018).

Two trials reported on the costs associated with the intervention (He 2017; Prabhakaran 2019)

Further information requested

Of the trials which included participants who had or had not been prescribed CVD prevention medication, four studies did not report the proportion prescribed medication (Liu 2015; Prabhakaran 2019; Saleh 2018; Tobe 2019). We contacted authors for this information, and received it for two of the trials (Prabhakaran 2019; Tobe 2019). We also contacted authors of three trials (Bobrow 2016; Morillo‐Verdugo 2018; Saleh 2018) for information relating to the proportion of participants who had previously experienced a CVD event and received this information for two trials (Bobrow 2016; Morillo‐Verdugo 2018).

Excluded studies

See Characteristics of excluded studies table for details of excluded studies which narrowly missed the inclusion criteria.

Ongoing studies

We identified 13 ongoing studies (see Characteristics of ongoing studies table).

Risk of bias in included studies

Details of the risk of bias assessments for each of the included studies are presented in the 'Risk of bias' tables in the Characteristics of included studies table, and in Figure 2

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Twelve studies reported adequate random sequence generation and were at low risk of bias for this domain (Bobrow 2016; Choudhry 2018; Gulayin 2019; He 2017; Liu 2015; Márquez Contreras 2019; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Peiris 2019; Prabhakaran 2019; Tobe 2019). Two studies did not provide sufficient information and were therefore at unclear risk of bias for random sequence generation (Logan 2012; Saleh 2018).

Eight studies described their allocation concealment adequately and were at low risk of bias in this domain (Bobrow 2016; Choudhry 2018; Gulayin 2019; Márquez Contreras 2019; Peiris 2019; Prabhakaran 2019; Saleh 2018; Tobe 2019). The other six studies did not provide sufficient information on their allocation procedures and were therefore at unclear risk of bias for allocation concealment (He 2017; Liu 2015; Logan 2012; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017).

Blinding

In all but one of the included studies, the nature of the interventions precluded blinding of participants, the exception being Tobe 2019 which compared 'active' messages with 'passive messages' including general health information, meaning it would not be obvious to participants whether they were in the intervention or control group. This study also blinded personnel, so was at low risk of bias in this domain. In 10 studies, personnel were not blinded to group assignment (Choudhry 2018; Gulayin 2019; He 2017; Liu 2015; Márquez Contreras 2019; McManus 2018; Morillo‐Verdugo 2018; Peiris 2019; Prabhakaran 2019; Saleh 2018) and were at high risk of bias this domain. The remaining three studies were at unclear risk of bias: two of them stated that personnel were blinded (Bobrow 2016; Párraga‐Martínez 2017), and one study was not clear whether personnel were blinded (Logan 2012).

For the blinding of objective outcome assessment domain, five studies provided sufficient detail relating to the blinding of outcome assessors and were at low risk of bias for this domain (Bobrow 2016; Choudhry 2018; Peiris 2019; Prabhakaran 2019; Tobe 2019). Nine studies did not provide adequate detail relating to whether or not outcome assessors were blinded, the nature of data collection, or the nature of data entry, or both, and so were judged as being at unclear risk of bias in this domain (Gulayin 2019; He 2017; Liu 2015; Logan 2012; Márquez Contreras 2019; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Saleh 2018).

Eight of the included studies reported self‐reported subjective outcomes for extraction. All were judged to be at high risk of bias for this domain, as the participants could not be blinded to their allocation, and therefore this may have resulted in biased self‐reported outcomes (Bobrow 2016; Gulayin 2019; He 2017; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Peiris 2019; Prabhakaran 2019). This domain did not apply to the six other studies, as they did not report subjective outcomes. We assessed these as low risk of bias to provide a complete risk of bias assessment (Choudhry 2018; Liu 2015; Logan 2012; Márquez Contreras 2019; Saleh 2018; Tobe 2019).

Incomplete outcome data

Most of the included studies had high rates of follow‐up (85% or greater) with no evidence of differential loss to follow‐up, and reported sensitivity analyses finding consistent results or used appropriate methods to impute missing data; we therefore judged them to be at low risk of bias for the incomplete outcome data domain (Bobrow 2016; Choudhry 2018; Gulayin 2019; He 2017; Logan 2012; Márquez Contreras 2019; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Peiris 2019; Prabhakaran 2019).

One study reported that 27.5% of participants did not attend for follow‐up, and that they differed from those who did attend for follow‐up based on several characteristics. The study also reported that these missing values were likely to have little impact on the primary outcome, based on sensitivity analyses. However, it is unclear whether this may have affected other outcomes, and so we rated this study at unclear risk of bias for this domain (Liu 2015). We judged a second study to be at unclear risk of bias for this domain, as 14% loss to follow‐up was reported, but no further information nor analyses relating to these participants were provided (Tobe 2019).

Finally, Saleh 2018 was considered to be at high risk of bias for incomplete outcome data, as 28% of participants in the intervention group had no outcome data compared to only 1% in the control group, and this discrepancy was not explained.

Selective reporting

Five studies were at low risk of bias for selective reporting on the basis that they reported all (or at least main) outcomes as specified in their published protocols, and the trials were registered prior to start of recruitment (Choudhry 2018; He 2017; McManus 2018; Peiris 2019; Tobe 2019). Gulayin 2019 was considered to be at high risk of bias for this domain as the "proportion of patients with moderate and high CVD risk who have reduced their LDL‐c by 30% and 50% respectively" was specified as an outcome in the protocol, but no results for this are reported.

The rest of the studies were at unclear risk of bias for selective reporting. Bobrow 2016 reported outcomes as planned in their protocol, with the exception of one outcome that was flagged in protocol, but not in the trial report (‘hypertension knowledge'). This trial began recruiting in June 2012, but details of the protocol were not registered until December 2013, and so we cannot be certain as to what was planned before the trial began. Three trials also appeared to have been registered after recruitment had begun, with no published protocol identified (Liu 2015; Logan 2012; Saleh 2018). We could find no protocol for Márquez Contreras 2019, and while the report stated the trial had been registered it was unclear whether this was prior to recruitment. We found no protocol nor trial registry entry for Morillo‐Verdugo 2018. Párraga‐Martínez 2017 reported all outcomes as planned in the protocol, with the exception of cardiovascular events occurring during the study period. This was considered an important outcome, but it was not clear whether this outcome was not reported because no events occurred. In Prabhakaran 2019 not all of the secondary outcomes (health‐related quality of life and costs) were reported, but authors stated that the cost‐effectiveness analysis would be conducted if the intervention showed a substantial effect, which it did not. The supplementary materials for this trial listed the adverse events to be recorded, but no mention of adverse events is made in the trial report, although it is not clear whether this is because none occurred.

Other potential sources of bias

For other potential sources of bias, we assessed evidence for selective cluster recruitment for the included cluster‐RCTS. Four studies were considered at low risk of bias for this domain, with little difference between arms in relevant baseline characteristics (Choudhry 2018; Gulayin 2019; Peiris 2019; Saleh 2018). Three studies were at unclear risk of bias, either due to some imbalances in baseline characteristics (but it was not clear whether these would affect conclusions drawn) (He 2017; Prabhakaran 2019) or because key baseline characteristics were not measured (Márquez Contreras 2019). The remaining trials were not parallel RCTs and so this domain was not applicable; we judged them to be at low risk, to provide a complete 'Risk of bias' assessment (Bobrow 2016; Liu 2015; Logan 2012; McManus 2018; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Tobe 2019).

Effects of interventions

See: Summary of findings 1 Mobile phone interventions compared to usual care for improving adherence to medication prescribed for primary prevention of cardiovascular disease

We did not pool results in a meta‐analysis for most of the trials, as we deemed the content and delivery mechanisms of the interventions, and the study populations, too heterogeneous to allow meaningful pooling. The exceptions to this were the Bobrow 2016 and Tobe 2019 trials, both of which delivered the intervention solely through text messages about hypertension and it's medical therapy to target adherence, and recorded blood pressure outcomes (systolic blood pressure, and 'controlled' blood pressure).

In generating the illustrative forest plots, we also checked heterogeneity statistically (I² greater than 85% for systolic blood pressure (SBP) and for diastolic blood pressure (DBP); I²= 55% for controlled blood pressure; I² = 39% for total cholesterol; I² = 24% for LDL‐C; I² = 0% for HDL‐C). Based on these findings, we gave further consideration to pooling results from the three studies which reported HDL‐C (Liu 2015; Morillo‐Verdugo 2018; Párraga‐Martínez 2017) and from the five studies which reported LDL‐C (Choudhry 2018; Gulayin 2019; Liu 2015; Morillo‐Verdugo 2018; Párraga‐Martínez 2017). However, we still considered the interventions too diverse to warrant meaningful pooling: specifically, one intervention included face‐to‐face counselling alongside text‐messaging (Liu 2015); one consisted of written information and text messages (Párraga‐Martínez 2017); one involved pharmacotherapeutic follow‐up and an individual motivational interview (led by pharmacists), in addition to text messages (Morillo‐Verdugo 2018); another involved training for healthcare providers in clinical guidelines, a mobile/tablet‐based decision support system for clinical staff, and educational and motivational text messages directly to patients (Gulayin 2019); and Choudhry 2018 evaluated an intervention involving text messages (as reminders and motivational support for adherence), pillboxes, mailed personalised progress reports, and an individually‐tailored telephone consultation conducted by a staff clinical pharmacist.

We present results narratively below, and in Analysis 1.1; Analysis 1.2; Analysis 1.3; Analysis 1.4; Analysis 1.6; Analysis 1.7.

Primary outcomes

Objective measures of adherence to treatment

Cholesterol

Five trials reported LDL‐C levels (Analysis 1.1), two of which showed reductions in LDL‐C with confidence intervals excluding no effect: MD −9.20 mg/dL, 95% CI −17.7 to −0.70 (Párraga‐Martínez 2017) and MD −5.30 mg/dL, 95% CI −8.30 to −2.30 (Choudhry 2018).

Two trials reported results consistent with no intervention effect on LDL‐C: MD −1.60 mg/dL, 95% CI −25.78 to 22.58 (Morillo‐Verdugo 2018) and MD 0.77 mg/dL, 95% CI −4.64 to 6.18 (Liu 2015) (note: we converted mmol/L cholesterol to mg/dL using a multiplier of 38.67, as recommended by Rugge 2011).

Gulayin 2019 reported results separately according to the baseline CVD risk of participants as follows: moderate CVD‐risk participants MD −3.60 mg/dL, 95% CI −13.5 to 6.03, and high CVD risk‐participants MD −7.70 mg/dL, 95% CI −25.8 to 10.40. Morillo‐Verdugo 2018 also reported the proportion of participants with 'controlled' LDL‐C (without specifying the cut‐off applied), with no appreciable difference between the intervention arm (64.3%) and the control arm (63.2%).

We judged the evidence relating to the intervention effect on LDL‐C to be of low certainty, due to all but one of the trials contributing to this comparison being at unclear risk of bias across multiple domains, and the inconsistency in effect estimates across studies (summary of findings Table 1).

Of the four trials recording total cholesterol (Analysis 1.2), two found evidence of intervention benefit: MD −10.05 mg/dL, 95% CI −17.01 to −3.09 (Liu 2015) and MD −9.70 mg/dL, 95% CI −19.10 to −0.30 (Párraga‐Martínez 2017).

The other two reported results consistent with no intervention effect: MD −4.70 mg/dL, 95% CI −26.45 to 17.05 (Morillo‐Verdugo 2018) and MD −1.80 mg/dL, 95% CI −6.30 to 2.70 (Prabhakaran 2019). Morillo‐Verdugo 2018 also reported the proportion of participants with 'controlled' total cholesterol as follows: intervention arm 64.3% and the control arm 62.5%.

None of the three trials found evidence for an adverse effect on HDL‐C (Analysis 1.3): MD 1.16 mg/dL, 95% CI −1.55 to 3.87 (Liu 2015); MD 0.10 mg/dL, 95% CI −2.60 to 2.80 (Párraga‐Martínez 2017); MD 1.50, 95% CI −6.11 to 9.11 (Morillo‐Verdugo 2018).

Blood pressure

Thirteen studies reported at least one blood pressure outcome. Overall, we judged the evidence about blood pressure to be of low certainty, due to substantial inconsistency between studies in the degree to which the outcomes were affected, and because most of the trials were at unclear risk of bias across multiple domains (summary of findings Table 1).

Of the 13 studies recording systolic blood pressure, nine showed a reduction in the intervention arm compared to the control (Analysis 1.4). For four of these studies, the reduction in systolic blood pressure had confidence intervals which excluded no effect, as follows: MD −6.60 mmHg, 95% CI −8.60 to −4.60 (He 2017); MD −12.45 mmHg, 95% CI −15.02 to −9.88 (Liu 2015); MD −7.10 mmHg, 95% CI −11.61 to −2.59 (Logan 2012); MD −4.70 mmHg, 95% CI −7.00 to ‐2.40 (McManus 2018).

For the eight studies in which the confidence interval included no effect, the mean difference ranged from −3.96 mmHg (Morillo‐Verdugo 2018) to 0.83 mmHg (Párraga‐Martínez 2017). Choudhry 2018 recorded an increase in systolic blood pressure in the intervention arm of 2.80 mmHg with a confidence interval excluding no effect: 95% CI 0.30 to 5.30. The meta‐analysis (Analysis 1.5) of the two trials which evaluated an intervention targeting adherence to blood pressure medication delivered solely by SMS messaging provided a pooled MD of −1.55 mmHg, 95% CI: −3.36 to 0.25, I² = 0% (Bobrow 2016; Tobe 2019).

Eleven trials recorded diastolic blood pressure (Analysis 1.6). Four found a reduction in the intervention arm with confidence intervals excluding no effect: MD −5.40 mmHg, 95% CI −6.80 to −4.00 (He 2017); MD −12.23 mmHg, 95% CI −14.03 to −10.43 (Liu 2015); MD −3.90 mmHg, 95% CI −6.45 to −1.35 (Logan 2012); MD −1.30 mmHg, 95% CI −2.50 to −0.10 (McManus 2018). The remaining seven trials reported mean differences in diastolic blood pressure ranging from −3.64 mmHg, 95% CI −9.03 to 1.75 (Márquez Contreras 2019) to 1.64 mmHg, 95% CI −0.55 to 3.83 (Párraga‐Martínez 2017).

Seven studies reported 'controlled' blood pressure as an outcome (Analysis 1.7). Estimates varied from negligible effects (OR 1.01, 95% CI 0.76 to 1.34) (Peiris 2019) to large improvements in blood pressure control (OR: 2.41, 95% CI 1.57 to 3.68) (He 2017), athough for all but one study (He 2017), confidence intervals encompassed no effect. The pooled analysis (Analysis 1.8) of two trials which evaluated an intervention targeting adherence to blood pressure medication delivered solely by SMS messaging indicated a modest beneficial intervention effect: OR 1.32, 95% CI 1.06 to 1.65, I² = 0% (Bobrow 2016; Tobe 2019).

Heart rate

No studies reported heart rate.

Urinary 11‐dehydrothromboxane B2

No studies reported urinary 11‐dehydrothromboxane B.

Combined cardiovascular disease event (fatal or non‐fatal events)

One trial reported on deaths due to CVD (Bobrow 2016), and three trials recorded non‐fatal CVD events (McManus 2018; Peiris 2019; Tobe 2019). For three studies the effect estimate was in the direction of harm (Bobrow 2016; McManus 2018; Peiris 2019), and for the fourth it was in the direction of intervention benefit (Tobe 2019). However, the number of events in each trial was low and all effect estimates had wide 95% confidence intervals, encompassing no effect (Analysis 1.9). For further detail see 'fatal CVD events' and 'non‐fatal CVD events' in Secondary outcomes.

Adverse effects

Based on six trials, we found moderate‐certainty evidence that the mobile phone‐based interventions under study did not lead to adverse events (summary of findings Table 1). The evidence was of moderate certainty, due to the studies being at unclear risk of bias across multiple domains. Bobrow 2016 (1372 participants) reported no adverse events attributable to the intervention. Párraga‐Martínez 2017 (304 participants) reported that there were no differences between groups in experiencing adverse effects of statins (intervention group: 7 events; control group: 10 events), and no participants reported intervention‐related adverse events. McManus 2018 reported that potential side‐effects were similar between the groups. He 2017 stated that no adverse events were reported. Prabhakaran 2019 provided a list of adverse events that would be recorded, although they are not reported on in the trial report, and it is unclear whether this is because none occurred. However, Prabhakaran 2019 did report the number of deaths by arm (34 deaths in the intervention group and 21 deaths in the control group), but not the causes of death. Tobe 2019 stated that there were no reports of hypotension, and there were two deaths in the intervention arm: one due to pre‐existing cancer and one in a car accident (as a passenger). The other trials did not report on adverse events (Choudhry 2018; Gulayin 2019; Liu 2015; Logan 2012; Márquez Contreras 2019; Morillo‐Verdugo 2018; Peiris 2019; Saleh 2018).

Secondary outcomes

Indirect measures of adherence to treatment

Included studies reported a variety of different measures relating to adherence to prescribed medication. An overview of the trial results for indirect measures of medication adherence is presented in Table 1.

Open in table viewer

Table 1. Indirect measures of adherence

Trial	Outcome measure	Comparison	Intervention	Number (intervention)	Control	Number (Control)	Narrative results
Bobrow 2016 (1‐year follow‐up)	Proportion of days covered by dispensed BP medicine (prescription data)	Information‐only SMS vs control	MD 83.3% (95% CI 69.3 to 91.7)	457	79.2% (95% CI 4.6 to 91.4)	458	Median difference 5.2, quartiles 1 ‐ 3: 1.5 to 8.9; P = 0.006
		Interactive SMS vs control	MD 83.3% (95% CI 66.7 to 91.7)	457	79.2% (95% CI 64.6 to 91.4)	458	Median difference 3.8; quartiles 1 ‐ 3: 0.03 to 7.6; P = 0.048
	Proportion of participants with proportion of days covered ≥ 80% (prescription data)	Information‐only SMS vs control	63%	457	49.4%	458	Adjusted OR 1.86, 95% CI 1.39 to 2.49; P < 0.001
		Interactive SMS vs control	60%	457	49.4%	458	Adjusted OR 1.60, 95% CI 1.20 to 2.16; P = 0.002
	Self‐reported medication adherence (score range 5 – 10)	Information‐only SMS vs control	10 (quartiles 1 ‐ 3: 9 to 10)	457	10 quartiles 1 ‐ 3: 9 to 10)	458	Median difference 0.04, 95% CI −0.1 to 0.2; P = 0.70
	Self‐reported medication adherence (score range 5 – 10)	Interactive SMS vs control	10 (quartiles 1‐3: 9 to 10)	457	10 (quartiles 1‐3: 9 to 10)	458	Median difference 0.02, 95% CI –0.2 to 0.2; P = 0.80
Párraga‐Martínez 2017 (2‐year follow‐up)	Proportion adherent to lipid‐lowering medication according to self‐reported medication adherence (measured using 'adapted Morisky‐Green test')	—	77.2%	Disaggregated not reported	64.1%	Disaggregated not reported	P = 0.029 220 in total, not reported by group
He 2017 (18‐month follow‐up)	High adherence to BP medication (Morisky score = 8)	—	66.1%	629	53.0%	542	Risk difference^a: 13.1%, 95% CI 7.0 to 19.2; P < 0.001
Gulayin 2019 (1‐year follow‐up)	Participants at moderate CVD risk: High adherence to lipid‐lowering medication (Morisky score = 8)	—	46.9%	58	50.1%	54	Risk difference^a −3.2, 95% CI −27.9 to 21.5); P = 0.7994
Gulayin 2019 (1‐year follow‐up)	Participants at high CVD risk: High adherence to lipid‐ lowering medication (Morisky score = 8)	—	30.3%	75	45.8%	58	Risk difference −15.5, 95% CI −42.6 to 11.6; P = 0.2616
Prabhakaran 2019 (1‐year follow‐up)	Self‐reported adherence to antihypertensive drug on all 7 days prior to endline assessment	—	81.1%	1027	57.9%	1119	Risk difference^b 23.1%, 95% CI 14.6 to 31.6%; P < 0.001
Choudhry 2018 (1‐year follow‐up)	Lipid‐lowering medication: mean proportion of days covered over the 12 months after randomisation (prescription data)	—	48.2	1467	44.1	1503	Mean difference^a 4.5, 95% CI 2.1 to 6.8 (P‐value not reported)
Choudhry 2018 (1‐year follow‐up)	BP medication: mean proportion of days covered over the 12 months after randomisation (prescription data)	—	42.7	529	35.9	486	Mean difference^a 8.5, 95% CI 5.4 to 11.7 (P‐value not reported)
Márquez Contreras 2019 (1‐year follow‐up)	Proportion taking BP medication correctly on 80% ‐ 100% of days (MEMS)	—	86.3%	73	62.7%	75	Risk difference^c 21.6%, 95% CI −1.2 to 44.5; P = 0.064
McManus 2018 (1‐year follow‐up)	Mean adherence score for BP medication (MARS questionnaire score) (unclear what the score range is as applied in this report)	—	24.0	~ 327 (exact n unclear)	23.9	~ 348 (exact n unclear)	Adjusted mean difference 0.02, 95% CI −0.20 to 0.25; P = 0.833
Morillo‐Verdugo 2018 (1‐year follow‐up)	Proportion adherent to 'concomitant medication'^d ‐ measured "with the Morisky‐Green questionnaire and pharmacy dispensing records [....] patients were considered adherent [....] if they obtained a positive score"	—	87.7%	29	58.3%	24	Risk difference 27.9%, 95% CI 5.5 to 51.3

CI: confidence interval; SMS: short messaging service
^aresult reported in paper accounts for clustering.
^bcalculated using effective sample size (ICC: 0.05, average cluster size: 83).
^ccalculated using effective sample size (ICC:0.05, average cluster size: 37).
^dvarious medications (see Characteristics of included studies for more detail).

Six studies reported on adherence to blood pressure‐lowering medication, of which five found evidence of intervention benefit. Bobrow 2016 (1372 participants) presented 12‐month outcome data for the median difference in the proportion of days covered by dispensed blood pressure medication, finding evidence for a modest benefit for both the information‐only text‐messaging intervention group (83.3% with intervention versus 79.2% with control; median difference 5.2, quartiles 1 ‐ 3: 1.5 to 8.9; P = 0.006), and the interactive text‐messaging group (83.3% with intervention versus 79.2% with control; median difference: 3.8, quartiles 1 ‐ 3: 0.03 to 7.6; P = 0.048), compared with the control group receiving usual care. There were similar results for the outcome of achieving 80% or more days covered (information‐only text‐messaging group versus control: OR 1.86, 95% CI 1.39 to 2.49; P < 0.001; interactive text‐messaging group versus control: OR 1.60, 95% CI 1.20 to 2.16; P = 0.002) (it is not clear how the underlying proportions compared, as the authors did not report the proportion achieving 80% or more days covered for the control group). However, there was no evidence of benefit for the outcome of self‐reported medication adherence (information‐only text‐messaging group versus control: median difference 0.04, quartiles 1 ‐ 3: −0.1 to 0.2; P = 0.70; interactive text‐messaging group versus control: median difference 0.02, quartiles 1 ‐ 3: −0.2 to 0.2, P = 0.80). He 2017 found evidence of intervention benefit, with 66.1% of the intervention group reporting high medication adherence (based on the Morisky‐Green test) compared with 53.0% of the control group at 1 year (P < 0.001). Similarly, Prabhakaran 2019 reported higher medication adherence at 1‐year follow‐up in the intervention group (81.1%) compared to the control group (57.9%) (P < 0.001) (based on the proportion who reported taking their drugs on all seven days prior to endline assessment). Márquez Contreras 2019 reported a greater proportion of participants in the intervention group taking their blood‐pressure medication correctly on 80% to 100% of days recorded via MEMS (86.3% versus 62.7%, P = 0.064). Based on the mean proportion of days covered over 12 months as indicated by prescription data, Choudhry 2018 reported modest beneficial intervention effect (MD 8.5, 95% CI 5.4 to 11.7). No evidence of intervention benefit was reported by McManus 2018, based on a mean self‐reported adherence score (MD 0.02, 95% CI −0.20 to 0.25).

Three studies recorded adherence to lipid‐lowering medication. Choudhry 2018 recorded the mean proportion of days covered for lipid‐lowering medication based on prescription data, with a slightly higher level of adherence evident in the intervention group (MD 4.5, 95% CI 2.1 to 6.8). Párraga‐Martínez 2017 found evidence of intervention benefit for the proportion of participants reporting adherence to lipid‐lowering therapy (measured using the Morisky‐Green test) at two years post‐randomisation (77.2% with intervention versus 64.1% with control; P = 0.029), whereas no beneficial effect was recorded for self‐reported adherence to lipid‐lowering therapy (again measured using the Morisky‐Green test) in Gulayin 2019 (moderate CVD‐risk participants: intervention 46.9%, control 50.1%, P = 0.799; high CVD‐risk participants: intervention 30.3%, control 45.8%, P = 0.262).

Finally, Morillo‐Verdugo 2018 reported adherence to 'concomitant medication' (which could have referred to various CVD‐related medication types ‐ see Characteristics of included studies for more detail), with higher adherence recorded in the intervention group (87.7%) compared with the control group (58.3%).

Fatal cardiovascular events

Bobrow 2016 (1372 participants) reported that two participants in the information‐only text‐messaging group died due to ischaemic heart disease, two participants in the interactive text‐messaging group died due to congestive cardiac failure, and there were no deaths in the control group known to be due to CVD. There were slightly more participants in the usual‐care arm who were lost to follow‐up due to 'lost contact' (14 participants), compared to the information SMS arm (seven participants), and the interactive SMS arm (seven participants). It is therefore possible that this differential lost to follow‐up due to lost contact could have underestimated deaths, including those due to CVD, in the usual‐care arm.

Non‐fatal cardiovascular events

McManus 2018 reported that cardiovascular events (new atrial fibrillation, angina, myocardial infarction, coronary artery bypass graft or angioplasty, stroke, peripheral vascular disease, or heart failure) were recorded in nine participants in the control group, and 11 in the intervention group. Tobe 2019 stated that one participant in the control group had a stroke, and one participant in the intervention group had a myocardial infarction. Peiris 2019 reported that 107/4348 participants in the intervention group reported a new CVD event, compared with 62/4294 in the control group, but the confidence interval was wide and encompassed benefit and harm (OR 1.42, 95% CI 0.78 to 2.62).

Health‐related quality of life assessed using validated instruments

Bobrow 2016 reported the median difference in quality of life as measured by the Euro‐Qol 5‐Dimension Index, finding no effect of the information‐only text messages (median difference 0.01, quartiles 1 ‐ 3: −0.01 to 0.02; P = 0.50) or the interactive text messages (median difference: 0.003, quartiles 1 ‐ 3: −0.02 to 0.02; P = 0.73) compared with the control group. McManus 2018 and Peiris 2019 also reported on quality of life with a mean difference of −0.03, 95% CI −0.06 to −0.001; P = 0.0384, and 0.02, 95% CI 0.00 to 0.04, P = 0.03, respectively.

Cognitive outcomes

Bobrow 2016 measured satisfaction with treatment and found no evidence of difference between intervention arms and control arm (information‐only text‐messaging group versus control: median difference 0, quartiles 1 ‐ 3: −0.3 to 0.3; P > 0.99; interactive text‐messaging group versus control: median difference 0, quartiles 1 ‐ 3: −0.3 to 0.3; P > 0.99). Prabhakaran 2019 reported on perceived quality of care, with little difference observed between the two groups (intervention: 96.6%, control: 95.0%).

Costs

Two studies provided information on costs. He 2017 reported the total cost per participant (intervention and healthcare costs) as follows: mean costs intervention arm: USD 178.6, control arm: USD 67.6 (MD USD 102.7, 95% CI 61.0 to 144.4). In relation to costs, Choudhry 2018 stated "clinical pharmacists spent a total of 985 hours conducting these calls, or 29 minutes per patient. Assuming a mean annual pharmacist salary of USD 120 000, this amounts to USD 30 per patient per year. Our intervention also had other components although their marginal costs were small".

Process measures

Four studies reported relevant process measures. Párraga‐Martínez 2017 recorded satisfaction with the intervention, finding that 90.8% (95% CI 85.9 to 95.7) of the 155 intervention‐group participants reported being satisfied or very satisfied with the intervention at two years' post‐randomisation. Logan 2012 recorded a 65.4% (standard deviation 30) adherence rate to the home blood‐pressure measurement schedule (taking a minimum of eight readings per week) in the intervention group. Bobrow 2016 reported that 50% of participants allocated to the interactive SMS intervention arm responded to messaging. Choudhry 2018 reported that of the 1069 intervention participants who received a telephone consultation, 194 (18.1%) opted in to receive text messages, and among all 2038 intervention participants, 1804 (88.5%) were sent quarterly progress reports.

Discussion

Summary of main results

This review provides low‐certainty evidence about the effects of adherence interventions delivered by mobile phone, with some trials reporting modest benefits and other no benefits. There was moderate‐certainty evidence that the interventions did not cause harm. In our review, we identified 14 trials, of which two were at low risk of bias (Choudhry 2018; Peiris 2019). The trials varied widely in the behaviours targeted, content and delivery mechanisms of the interventions, and the populations targeted. Due to these differences, we mostly summarised results narratively.

The evidence for the intervention effect on LDL cholesterol was of low certainty. Two of the five studies reporting LDL cholesterol as an outcome recorded evidence of intervention benefit, albeit of a modest size (Choudhry 2018; Párraga‐Martínez 2017).

The body of evidence relating to the effect of mobile phone‐based interventions on blood pressure was also of low certainty. Four of the 13 studies recording systolic blood pressure showed evidence of intervention benefit, with confidence intervals excluding no effect (He 2017; Liu 2015; Logan 2012; McManus 2018). The same four trials also demonstrated a reduction in diastolic blood pressure associated with the intervention (He 2017; Liu 2015; Logan 2012; McManus 2018). The direction of the point estimates was more consistently positive for the outcome of 'controlled' blood pressure, although the confidence intervals excluded no effect in only one trial (He 2017). Pooled analysis of two trials showed there was little or no benefit for systolic blood pressure for interventions delivered solely through educational and motivational text messages about hypertension and its medical therapy, although there was a modest increase in the proportion of participants with 'controlled' blood pressure (Bobrow 2016; Tobe 2019).

Nine studies reported indirect measures of medication adherence, of which seven reported evidence of intervention benefit (Bobrow 2016; Choudhry 2018; He 2017; Márquez Contreras 2019; Morillo‐Verdugo 2018; Párraga‐Martínez 2017; Prabhakaran 2019), ranging from a relatively small increase in adherence in the intervention arm based on prescription data (Bobrow 2016; Choudhry 2018) to the largest effect estimates (a 23.1% and a 27.9% absolute increase in adherence) recorded through self‐reported data (Morillo‐Verdugo 2018; Prabhakaran 2019).

Based on four studies, there was very low‐certainty evidence relating to the intervention effect on combined (fatal and non‐fatal) CVD events (Bobrow 2016; McManus 2018; Peiris 2019; Tobe 2019).

Overall completeness and applicability of evidence

The studies were conducted in a range of high, upper‐middle, and lower‐middle settings and some specifically targeted more disadvantaged settings within those countries, providing reasonable confidence in the applicability of results across settings. Given that one of our inclusion criteria was that trials have a minimum of one‐year follow‐up, we can be confident that our results are applicable to longer‐term, sustained medication adherence behaviours and outcomes. Few studies reported on fatal or non‐fatal cardiovascular events, meaning we were unable to establish whether the modest benefits observed in individual trials for cholesterol and blood pressure translated into such patient‐relevant outcomes. In the trials involving clinical decision support systems, whereby prescriptions and dosages may have been altered during the study period, we cannot be sure of the contribution of increased medication adherence to the reductions in cholesterol and blood pressure reported. The relative contribution of improved medication adherence is also uncertain in the trials which included a mix of participants who had and had not been prescribed CVD medication, and those which targeted lifestyle modifications alongside medication‐taking behaviour. Furthermore, in many of these trials, adherence to medication was a secondary rather than a primary outcome, meaning that these studies were not designed around the focus of this review.

Quality of the evidence

Using GRADE methodology, we assessed the certainty of the evidence for our narrative synthesis of objective outcomes of medication adherence (LDL‐C, SBP and DBP), cognitive outcomes and adverse events. The evidence was of low certainty across all outcomes, with the exception of adverse events, for which we rated the evidence as of moderate certainty. We downgraded the certainty of the evidence for objective outcomes of medication adherence by one level as a result of inconsistency in effect estimates which spanned both clinically‐meaningful improvements and null effects. We downgraded the certainty of the evidence for all five outcomes considered by one level because most of the included studies were at high risk of bias. Eleven of the studies were at unclear risk of bias for at least two of the domains, indicating inadequate reporting of the trial methods in these studies, which limited our ability to make clear judgements about the level of risk of bias. Finally, the evidence relating to the cognitive outcomes of satisfaction with treatment and perceived quality of care was also downgraded for indirectness, because this was based on two trials measuring different outcomes. Half of the trials in this review randomised by clusters rather than individuals, and not all measured indicators of adherence at baseline, so there was uncertainty about the extent of imbalanced relevant baseline characteristics in these trials.

Potential biases in the review process

Our inability to conduct a meta‐analysis for most outcomes means that this review cannot benefit from examining pooled effect estimates based on larger sample sizes than the individual trials. Furthermore, publication bias, whereby trials with positive findings are more likely to be published, may have biased the selection of included studies for this review. However, we tried to overcome this through searching clinical trial registries for prospectively‐registered trials. We decided to only include trials with a minimum of one‐year follow‐up in order that results were applicable to longer‐term sustained behaviour change in adherence, which would therefore be more important in improving health status. This means that we are unable to comment on the effectiveness of mobile phone‐based interventions for short‐term adherence to medication prescribed for the primary prevention of CVD.

Agreements and disagreements with other studies or reviews

Our findings of mixed evidence for the effects of mobile phone‐delivered interventions to increase adherence to medication prescribed for the primary prevention of CVD and no reported harms are consistent with those of a Cochrane Review examining the effectiveness of text‐messaging interventions to improve adherence to medication prescribed for the secondary prevention of CVD (Adler 2017). These findings are broadly consistent with systematic reviews of mhealth interventions to improve medication adherence across conditions, although these reviews included short‐term studies and non‐RCT designs, which are subject to bias (Anglada‐Martinez 2015; Park 2014b; Ng 2020). One systematic review examining RCTs of monitoring and messaging interventions targeting medication adherence for the management of type 2 diabetes found no evidence for an improvement in medication adherence in their pooled meta‐analyses of five trials (Farmer 2016). Our finding that pooled analyses of interventions delivered by text messaging alone indicated small benefits, some of which achieved statistical significance, is consistent with the findings from trials using SMS alone targeting adherence to HIV medication, which also report small benefits of borderline clinical and statistical significance (Da Costa 2012; Orrell 2015; Pop‐Eleches 2011; Sabin 2015). The three (out of four) studies reporting evidence of intervention benefit for lowering blood pressure with confidence intervals excluding no effect were the only studies which included the provision of home blood‐pressure monitoring systems in combination with mobile phones in the intervention. These positive intervention effects are consistent with the modest benefits of monitoring interventions in general (Carrasco 2008; Lim 2011; McKinstry 2013; Yoo 2009). The small or modest benefits reported may reflect the challenges involved in improving adherence, and overall inconclusive findings relating to adherence interventions in general, which have previously been noted in a Cochrane Review of all adherence interventions (Nieuwlaat 2014).

Figure 1

Study flow diagram.

Navigate to figure in ReviewOpen in new tab

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Navigate to figure in ReviewOpen in new tab

Analysis 1.1

Comparison 1: Mobile phone intervention versus control, Outcome 1: Change in low‐density lipoprotein cholesterol (mg/dL)

Navigate to figure in ReviewOpen in new tab

Analysis 1.2

Comparison 1: Mobile phone intervention versus control, Outcome 2: Change in total cholesterol (mg/dL)

Navigate to figure in ReviewOpen in new tab

Analysis 1.3

Comparison 1: Mobile phone intervention versus control, Outcome 3: Change in high‐density lipoprotein cholesterol (mg/dL)

Navigate to figure in ReviewOpen in new tab

Analysis 1.4

Comparison 1: Mobile phone intervention versus control, Outcome 4: Change in systolic blood pressure (mmHg)

Navigate to figure in ReviewOpen in new tab

Analysis 1.5

Comparison 1: Mobile phone intervention versus control, Outcome 5: Pooled change in systolic blood pressure (mmHg)

Navigate to figure in ReviewOpen in new tab

Analysis 1.6

Comparison 1: Mobile phone intervention versus control, Outcome 6: Change in diastolic blood pressure (mmHg)

Navigate to figure in ReviewOpen in new tab

Analysis 1.7

Comparison 1: Mobile phone intervention versus control, Outcome 7: Controlled blood pressure

Navigate to figure in ReviewOpen in new tab

Analysis 1.8

Comparison 1: Mobile phone intervention versus control, Outcome 8: Pooled controlled blood pressure

Navigate to figure in ReviewOpen in new tab

Analysis 1.9

Comparison 1: Mobile phone intervention versus control, Outcome 9: Combined fatal and non‐fatal CVD events

Navigate to figure in ReviewOpen in new tab

Summary of findings 1. Mobile phone interventions compared to usual care for improving adherence to medication prescribed for primary prevention of cardiovascular disease

Outcomes	Impact	№ of participants^f (studies)	Certainty of the evidence (GRADE)
Mobile phone interventions compared to usual care for improving adherence to medication prescribed for primary prevention of cardiovascular disease
Patient or population: people prescribed medication for primary prevention of cardiovascular disease Setting: community and healthcare settings Intervention: mobile phone‐based interventions Comparison: usual care, passive text messages, or 'enhanced' usual care
Objective measure of medication adherence: Cholesterol (low‐density lipoprotein) follow‐up: range 1 – 2 years	2 studies found evidence of a small beneficial intervention effect on reducing LDL‐C (−9.20 mg/dL, and 5.3 mg/dL), and 3 studies found results varying from a small reduction (−7.7 mg/dL) to a small increase in LDL‐C (0.77 mg/dL), all of which had wide confidence intervals that included no effect.	5,441 (5 RCTs)	⊕⊕⊝⊝ Low^a,b
Objective measure of medication adherence: Blood pressure follow‐up: range 1 – 2 years	Systolic BP: 9 of 13 studies found lower systolic blood pressure with mobile‐phone interventions, although only 4 of these reductions in systolic blood pressure had confidence intervals excluding no effect. Across the 13 studies, effect estimates varied greatly, from those showing a large reduction (−12.45 mmHg) to those reporting a small increase (+2.80 mmHg) in systolic blood pressure. Meta‐analysis of 2 trials evaluating an intervention targeting adherence to blood pressure medication delivered solely by SMS messaging provided a pooled MD of −1.55 mmHg, 95% CI −3.36 to 0.25.	25,166 (13 RCTs)	⊕⊕⊝⊝ Low^a,b
	Diastolic BP: 8 of 11 studies found lower diastolic blood pressure with mobile‐phone interventions, but in 4 of these the confidence intervals included no effect. Across the 11 studies, effect estimates varied widely from those showing a large reduction (−12.23 mmHg) to those showing a small increase (+1.64 mmHg) in diastolic blood pressure.	19,716 (11 RCTs)
	Controlled BP: 7 studies reported 'controlled' blood pressure as an outcome, of which six reported increased blood pressure control with mobile phone interventions, although in only one of these studies did the confidence interval exclude no effect. Effect estimates varied from negligible (OR 1.01) to large improvements in blood pressure control (OR 2.41). Meta‐analysis of 2 trials evaluating an intervention targeting adherence to blood‐pressure medication delivered solely by SMS messaging indicated a modest beneficial intervention effect: pooled OR of 1.32, 95% CI 1.06 to 1.65.	19,185 (7 RCTs)
Combined CVD events	1 trial reported on deaths due to CVD, and 3 recorded non‐fatal CVD events. For 3 studies the effect estimate was in the direction of harm, and for the 4th it was in the direction of intervention benefit. However, the number of events in each trial was low and all effect estimates had wide 95% confidence intervals encompassing no effect.	12,439 (4 RCTs)	⊕⊝⊝⊝ Very low^{c, d}
Adverse events follow‐up: range 1 – 2 years	3 studies reported that there were no adverse events attributable to the intervention. 1 reported that there was no difference between groups in adverse effects of statins, and that no participants reported intervention‐related adverse events. 1 study reported that potential side effects were similar between groups. 1 study reported a similar number of deaths in the intervention and control arms, but did not provide further information relating to potential adverse events.	8285 (6 RCTs)	⊕⊕⊕⊝ Moderate^b
Cognitive outcome: satisfaction with treatment follow‐up: mean 1 year	1 study measured satisfaction with treatment, and found no evidence of a difference between intervention and control arms. 1 study reported on perceived quality of care, with little difference observed between the 2 groups.	2535 (2 RCT)	⊕⊕⊝⊝ Low^d,e
LDL‐C: low‐density lipoprotein cholesterol; BP: blood pressure; RCT: randomised controlled trial
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.
^aDowngraded one level for inconsistency: trial results included large variations in the degree to which the outcome was affected. ^bDowngraded one level for risk of bias: most trials at unclear risk of bias for multiple domains. ^c Downgraded two levels for imprecision: very few events and wide confidence intervals encompassing intervention benefit and harm. ^dDowngraded one level for risk of bias: trials at unclear or high risk of bias for several domains. ^eDowngraded one level for indirectness: based on two trials, one conducted in public sector clinic in Cape Town, South Africa, and one in community health settings in India. ^f Total study enrolment.

Summary of findings 1. Mobile phone interventions compared to usual care for improving adherence to medication prescribed for primary prevention of cardiovascular disease

Navigate to table in Review

Table 1. Indirect measures of adherence

Trial	Outcome measure	Comparison	Intervention	Number (intervention)	Control	Number (Control)	Narrative results
Bobrow 2016 (1‐year follow‐up)	Proportion of days covered by dispensed BP medicine (prescription data)	Information‐only SMS vs control	MD 83.3% (95% CI 69.3 to 91.7)	457	79.2% (95% CI 4.6 to 91.4)	458	Median difference 5.2, quartiles 1 ‐ 3: 1.5 to 8.9; P = 0.006
		Interactive SMS vs control	MD 83.3% (95% CI 66.7 to 91.7)	457	79.2% (95% CI 64.6 to 91.4)	458	Median difference 3.8; quartiles 1 ‐ 3: 0.03 to 7.6; P = 0.048
	Proportion of participants with proportion of days covered ≥ 80% (prescription data)	Information‐only SMS vs control	63%	457	49.4%	458	Adjusted OR 1.86, 95% CI 1.39 to 2.49; P < 0.001
		Interactive SMS vs control	60%	457	49.4%	458	Adjusted OR 1.60, 95% CI 1.20 to 2.16; P = 0.002
	Self‐reported medication adherence (score range 5 – 10)	Information‐only SMS vs control	10 (quartiles 1 ‐ 3: 9 to 10)	457	10 quartiles 1 ‐ 3: 9 to 10)	458	Median difference 0.04, 95% CI −0.1 to 0.2; P = 0.70
	Self‐reported medication adherence (score range 5 – 10)	Interactive SMS vs control	10 (quartiles 1‐3: 9 to 10)	457	10 (quartiles 1‐3: 9 to 10)	458	Median difference 0.02, 95% CI –0.2 to 0.2; P = 0.80
Párraga‐Martínez 2017 (2‐year follow‐up)	Proportion adherent to lipid‐lowering medication according to self‐reported medication adherence (measured using 'adapted Morisky‐Green test')	—	77.2%	Disaggregated not reported	64.1%	Disaggregated not reported	P = 0.029 220 in total, not reported by group
He 2017 (18‐month follow‐up)	High adherence to BP medication (Morisky score = 8)	—	66.1%	629	53.0%	542	Risk difference^a: 13.1%, 95% CI 7.0 to 19.2; P < 0.001
Gulayin 2019 (1‐year follow‐up)	Participants at moderate CVD risk: High adherence to lipid‐lowering medication (Morisky score = 8)	—	46.9%	58	50.1%	54	Risk difference^a −3.2, 95% CI −27.9 to 21.5); P = 0.7994
Gulayin 2019 (1‐year follow‐up)	Participants at high CVD risk: High adherence to lipid‐ lowering medication (Morisky score = 8)	—	30.3%	75	45.8%	58	Risk difference −15.5, 95% CI −42.6 to 11.6; P = 0.2616
Prabhakaran 2019 (1‐year follow‐up)	Self‐reported adherence to antihypertensive drug on all 7 days prior to endline assessment	—	81.1%	1027	57.9%	1119	Risk difference^b 23.1%, 95% CI 14.6 to 31.6%; P < 0.001
Choudhry 2018 (1‐year follow‐up)	Lipid‐lowering medication: mean proportion of days covered over the 12 months after randomisation (prescription data)	—	48.2	1467	44.1	1503	Mean difference^a 4.5, 95% CI 2.1 to 6.8 (P‐value not reported)
Choudhry 2018 (1‐year follow‐up)	BP medication: mean proportion of days covered over the 12 months after randomisation (prescription data)	—	42.7	529	35.9	486	Mean difference^a 8.5, 95% CI 5.4 to 11.7 (P‐value not reported)
Márquez Contreras 2019 (1‐year follow‐up)	Proportion taking BP medication correctly on 80% ‐ 100% of days (MEMS)	—	86.3%	73	62.7%	75	Risk difference^c 21.6%, 95% CI −1.2 to 44.5; P = 0.064
McManus 2018 (1‐year follow‐up)	Mean adherence score for BP medication (MARS questionnaire score) (unclear what the score range is as applied in this report)	—	24.0	~ 327 (exact n unclear)	23.9	~ 348 (exact n unclear)	Adjusted mean difference 0.02, 95% CI −0.20 to 0.25; P = 0.833
Morillo‐Verdugo 2018 (1‐year follow‐up)	Proportion adherent to 'concomitant medication'^d ‐ measured "with the Morisky‐Green questionnaire and pharmacy dispensing records [....] patients were considered adherent [....] if they obtained a positive score"	—	87.7%	29	58.3%	24	Risk difference 27.9%, 95% CI 5.5 to 51.3
CI: confidence interval; SMS: short messaging service ^aresult reported in paper accounts for clustering. ^bcalculated using effective sample size (ICC: 0.05, average cluster size: 83). ^ccalculated using effective sample size (ICC:0.05, average cluster size: 37). ^dvarious medications (see Characteristics of included studies for more detail).

Table 1. Indirect measures of adherence

Navigate to table in Review

Comparison 1. Mobile phone intervention versus control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Change in low‐density lipoprotein cholesterol (mg/dL) Show forest plot	5		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.2 Change in total cholesterol (mg/dL) Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.3 Change in high‐density lipoprotein cholesterol (mg/dL) Show forest plot	3		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.4 Change in systolic blood pressure (mmHg) Show forest plot	13		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.5 Pooled change in systolic blood pressure (mmHg) Show forest plot	2	1494	Mean Difference (IV, Fixed, 95% CI)	‐1.55 [‐3.36, 0.25]

1.6 Change in diastolic blood pressure (mmHg) Show forest plot	11		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

1.7 Controlled blood pressure Show forest plot	7		Odds Ratio (IV, Fixed, 95% CI)	Totals not selected

1.8 Pooled controlled blood pressure Show forest plot	2	1494	Odds Ratio (IV, Fixed, 95% CI)	1.32 [1.06, 1.65]

1.9 Combined fatal and non‐fatal CVD events Show forest plot	4		Odds Ratio (IV, Fixed, 95% CI)	Totals not selected

Comparison 1. Mobile phone intervention versus control

Navigate to table in Review

Cochrane Review language

Website language

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

PICOs

PICOs

Population

Intervention

Comparison

Outcome

Plain language summary

Interventions delivered by mobile phone to help people adhere to medication to prevent heart and circulatory disease

Visual summary

Authors' conclusions

Implications for practice

Implications for research

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Intervention

Comparator

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Assessment of bias in conducting the systematic review

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

Results

Description of studies

Results of the search

Included studies

Participants

Settings

Interventions

Outcomes

Further information requested

Excluded studies

Ongoing studies

Risk of bias in included studies

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Primary outcomes

Objective measures of adherence to treatment

Cholesterol

Blood pressure

Heart rate

Urinary 11‐dehydrothromboxane B2