Cardiology - (2018) 1e9

Canadian Journal of

Review

Breaking Barriers: Mobile Health Interventions forCardiovascular Disease

Harry Klimis, MBBS,a,b,c Jay Thakkar, MBBS, MD (INT MEDICINE), FRACP,a,b,c and

Clara K. Chow, MBBS, FRACP, PhDa,b,c

aUniversity of Sydney, Sydney, New South Wales, AustraliabDepartment of Cardiology, Westmead Hospital, Sydney, New South Wales, Australia

cThe George Institute for Global Health, Sydney, New South Wales, Australia

ABSTRACTCardiovascular disease (CVD) is a leading global cause of death andmorbidity and prevention needs to be strengthened to tackle this.Mobile health (mHealth) might present a novel and effective solutionin CVD prevention, and interest in mHealth has grown dramaticallysince the advent of the smartphone. In this review, we discuss mHealthinterventions that target multiple cardiovascular risk factors simulta-neously in the context of primary as well as secondary prevention.There is some evidence that mHealth interventions improve a range ofindividual CVD risk factors, but a relative paucity of evidence onmHealth interventions improving multiple CVD risk factors simulta-neously. The existing data suggest mHealth programs improve overallCVD risk, at least in the short term. Interpretation of the evidence isdifficult in the context of poor methodology and mHealth modalitiesoften being a part of large complex interventions. In this review weidentify a number of unanswered questions including: which mode ofmHealth (or combination of interventions) would be most effective,what is the durability of intervention effects, and what degree ofpersonalization and interactivity is required.

Received for publication November 8, 2017. Accepted February 11, 2018.

Corresponding author: Dr Harry Klimis, Westmead Hospital CardiologyDepartment, PO Box 533, Wentworthville, New South Wales 2145,Australia. Tel.þ61-2-8890-3125.

E-mail: harry.klimis@sydney.edu.auSee page 8 for disclosure information.

https://doi.org/10.1016/j.cjca.2018.02.0120828-282X/� 2018 Canadian Cardiovascular Society. Published by Elsevier Inc. A

R�ESUM�ELa maladie cardiovasculaire (MCV) est une des principales causes demortalit�e et de morbidit�e dans le monde et il est urgent de mettrel’accent sur sa pr�evention pour s’attaquer à ce problème. Les appli-cations mobiles en sant�e pourraient constituer une solution innovanteet efficace pour la pr�evention de la MCV, et l’int�erêt pour ce type detechnologie a connu un essor fulgurant depuis les premiers t�el�ephonesintelligents. Dans cette revue, nous examinons les applications mo-biles en sant�e qui ciblent simultan�ement plusieurs facteurs de risquecardiovasculaire dans le contexte de la pr�evention tant primaire quesecondaire. Les donn�ees probantes dont on dispose indiquant que lesinterventions utilisant des applications mobiles en sant�e ont pour effetd’am�eliorer plusieurs facteurs de risque de MCV individuels sont rela-tivement plus nombreuses que celles attestant une am�eliorationsimultan�ee des facteurs de risque de MCV multiples. Les donn�eesexistantes laissent croire que les programmes d’applications mobilesen sant�e diminuent le risque de MCV global, du moins à court terme.L’interpr�etation des donn�ees probantes est difficile dans le contexted’une m�ethodologie inad�equate et du fait que les modalit�es utilisantdes applications mobiles en sant�e sont souvent int�egr�ees à des in-terventions complexes à grande �echelle. Dans cette revue, nousrecensons plusieurs questions encore sans r�eponse, y compris lessuivantes : quel serait le mode d’utilisation le plus efficace des ap-plications mobiles en sant�e (ou d’une combinaison d’interventions),quelle est la durabilit�e des effets de l’intervention, et quel est le degr�ede personnalisation et d’interactivit�e requis.

1

Cardiovascular disease (CVD) is the leading cause of deathand disability worldwide, with coronary heart disease (CHD)being the largest contributor.1,2 In 2012, 31% of all globaldeaths were due to CVD, with 80% occurring in low- and

middle-income countries. Multiple nonmodifiable (age, sex,ethnicity, family history) and modifiable (lifestyle andbehavioural) risk factors contribute to the lifetime risk ofCVD. Modifiable risk factors such as smoking, dyslipidemia,hypertension, abdominal obesity, physical inactivity, psycho-social factors, and diet account for most of the global risk formyocardial infarction (MI).3 Thus, effective prevention pro-grams aimed at a combination of these modifiable CVD riskfactors in primary as well as secondary prevention settings arelikely to yield best results.

Secondary prevention of CHD, including pharmaco-therapy and lifestyle modification, has been shown to reduce

ll rights reserved.

Figure 1. A text-message based prevention program delivered on a smartphone (left) and a standard cellular phone (right). The TEXT ME programsends regular text messages to encourage participants to lead healthier lifestyles. Photograph courtesy of Clara Chow. Reproduced with permissionfrom Clara Chow on behalf of TEXT ME.

2 Canadian Journal of CardiologyVolume - 2018

morbidity, mortality, and improve functional status andquality of life.4 However, despite strong evidence, there issubstantial underutilization of secondary prevention measuresin patients with CHD.5,6 For example, the European Actionon Secondary and Primary Prevention by Intervention toReduce Events (EUROASPIRE IV) study7 showed that mostpatients who had a coronary event did not meet guidelinerecommendations on CHD risk factor control. The reasonsfor the relatively poor implementation of secondary preven-tion are multiple, and include health system-, community-,and individual level-factors.8 Changing lifestyle behavioursand maintaining medication adherence is difficult for manyindividuals even if they know that they have CHD, and in theprimary prevention setting it is even more challenging toengender enough motivation to aggressively manage riskfactors.

Mobile health (mHealth) technology has gained increasinginterest as a means to improve the delivery of CVD preven-tion in a scalable and affordable way. It has particular potentialto assist lifestyle modification and improve medical adherence.The Global Observatory for eHealth defines mHealth as“Medical and public health practices supported by mobiledevices, such as mobile phones, patient monitoring devices,personal digital assistants (PDAs) and other wireless devices.”9

It is expected that by 2018, 50% of the global population willown a smartphone (mobile phones with a computer operatingsystem), and total mobile phone ownership in low-incomecountries will almost reach ubiquitous levels.10,11 Due tothe widespread global penetration of mobile technology,mHealth approaches might support the exchange of healthinformation, improve community-based health assessment by

reducing access barriers, and support improved lifestylechoices.9,12

Another particular potential of mHealth is its ability toaddress multiple risk factors. Although many traditionalapproaches have focused on targeting single risk factors,13

targeting multiple risk factors in those at high global risk aremore likely to be cost effective than treatment decisions on thebasis of individual risk factor targets.14 Targeting multiplecardiovascular risk factors is an important principle of CHDprevention.15

The aim of the current report is to review the evidence onmHealth interventions for simultaneous multiple risk factorreduction in primary as well as secondary prevention settings.

Modes of mHealth DeliverySmartphone applications (apps) and text messaging (short

messaging service or SMS) are 2 popular means of deliveringmHealth interventions (Figs. 1 and 2).

Smartphone apps

Smartphone usage has been increasing, with 39% of adultsglobally using a smartphone in 2012 to 61% in 201416 andthis has powered the growth in apps. Health-related apps arepopular, and easily accessible via the iTunes store or GooglePlay. Many are free, although some incur a cost to downloador purchase additional features. In April 2015, there were12,991 apps in the iTunes store and 1420 apps in Google Playtargeting multiple cardiovascular risk factors.12 These appsmight use different functions of the phone such as cameras,global positioning system, as well as sensors/accelerometers.

Figure 2. An app-based secondary prevention program targeting individuals who have known ischaemic heart disease. The “My Heart, My Life” appfrom the Australian Heart Foundation includes educational videos on acute coronary syndromes, and a personal health tracker. Reproduced withpermission from the National Heart Foundation of Australia.

Klimis et al. 3mHealth Interventions for Heart Disease

Many apps track and monitor cardiovascular risk factors, forexample, blood pressure logs, quit smoking apps, and diet andactivity logs such as MyFitnessPal (MyFitnessPal, Inc; www.myfitnesspal.com). In addition to smartphone ownership,running app-based interventions generally require an internetand data plan, which results in greater mobile phone runningcosts. The complexity of apps is an attractive feature, but canlimit utility for the less technology-literate user.11 From atechnical perspective, apps need to be specifically developedfor the operating system type (eg, Android, Apple) and requireregular maintenance and updates.

The utility of apps to improve health outcomes has beenpoorly evaluated. There is limited research evidence for theireffectiveness in modifying objective measures of health, andapp development and provision are largely unregulated, withFood and Drug Administration regulation being limited tothose apps that transform the smartphone into a medicaldevice.17,18 In addition many of these apps are not accessibleor not well designed for use by the elderly population. Therate of use of smartphone apps is lowest in elderly individuals,who are at the highest risk of CVD.19

Text messaging

Total mobile phone (smartphone and standard) sub-scriptions have increased, with an estimated 96.8subscriptions per 100 persons worldwide in 2015.16 In

contrast to apps, text messages are universally available on allmobile phones. There is some research evidence supportingthe utility of SMS-based approaches; most have focused onsingle risk factors in younger populations and have shortperiods of follow-up. The effectiveness of SMS-based in-terventions has been reported with smoking,20 weight loss,21

physical activity,22 diabetes,23 and hypertension.24,25 Thereare also some examples of large-scale implementation of text-message interventions particularly for smoking cessation.20

Text-message interventions are relatively inexpensive,simple to use, not reliant on internet/Wi-Fi access, and allowalmost instantaneous communication. SMS can be custom-ized and delivered via a computer algorithm, which has beenshown to be more effective than generic (not customized) textmessages.12 The cost of text messages can be a barrier indeveloping countries and literacy can also be a barrier.Importantly, the text messages offer greater simplicity becausetexts just arrive to the phone, obviating the need to pull in-formation as is the usual means of interacting with an app.

Primary Prevention Strategies in mHealthTargeting Multiple Risk Factors

There have been only a few randomized controlled trials(RCTs) that have assessed the feasibility and effectiveness ofmHealth approaches targeting multiple cardiovascular risk

Table 1. Randomized controlled trials (RCTs) using mHealth primary prevention programs with a simultaneous multiple risk factor strategy

Study Participants Intervention Outcome

Text message-based interventionsLiu et al.26 N ¼ 589

Chinese workers without CHDMean age 60.57 � 8.97 years58.2% Male sex

12-month durationSMS and follow-up phone calls

with frequency dependent onCVD risk (between once permonth to once per week)

An initial 15-minute face-to-facecounselling session at baseline

Computerized CVD evaluation

Measured at 12 months10-year CVD risk for intervention group not significantly different at

1 year (�1.05%; P ¼ 0.096). CVD risk change betweenintervention and control�2.83% (P ¼ 0.001)

Statistically significant change in intervention vs controls frombaseline to 1 year in SBP (�5.55 vs 6.89 mmHg; P < 0.001),DBP (�6.61 vs 5.62 mm Hg; P < 0.001), TC (�0.36 vs�0.10mmol/L; P¼ 0.005), fasting BSL (�0.31 vs 0.02mmol/L;P < 0.001), BMI (�0.57 vs 0.29; P < 0.001), waist-hip ratio(�0.02 vs 0.01; P < 0.001)

Anand et al.27 N ¼ 343South Asian participants in Canada

without CHDMean age 50.6 � 11.4 years51.9% Male sex

6-month durationE-mail every 2 weeks with either

weekly e-mail or SMS(participant choice)

Measured at 12 monthsNo significant change in MI risk score (blood pressure, waist to hip

ratio, HbA1c, ApoB:ApoA) between intervention and controlgroups (�0.27; 95% CI, �1.12 to 0.58; P ¼ 0.53) afteradjusting for baseline scores

Smartphone-based interventionsZhang et al.28 N ¼ 80

Singaporean (76.3% Chinese)workers without CHD

65% Female sexAged 21-40 years

4-week durationSmartphone application,

20-minute briefing session,daily SMS

Measured at 4 weeksSignificantly better awareness of CHD as a leading cause of death

(c2 ¼ 6.486; P ¼ 0.039), better overall CHD knowledge level(t ¼ 3.171; P ¼ 0.002), and better behaviour concerning bloodcholesterol control (c2 ¼ 4.54; P ¼ 0.033) in interventiongroup compared with controls

ApoB:ApoA, ratio of apolipoprotein B to apolipoprotein A; BMI, body mass index; BSL, blood sugar level; CHD, coronary heart disease; CI, confidenceinterval; CVD, cardiovascular disease; DBP, diastolic blood pressure; HbA1c, glycosylated hemoglobin; mHealth, mobile health; MI, myocardial infarction; SBP,systolic blood pressure; SMS, short messaging service; TC, total cholesterol.

4 Canadian Journal of CardiologyVolume - 2018

factors to improve cardiovascular risk in a primary preventioncohort (Table 1).

Text message-based interventions

The largest of these was conducted in China and was anRCT of a 12-month intervention comprising text messagingand phone calls performed among 589 Chinese workers(allocated to receive an annual medical examination) withoutCHD aged between 45 and 75 years.26 The interventiongroup received personalized text messages targeting lifestylemeasures to reduce CVD risk (frequency determinedaccording to risk assessment) in addition to a computerizedCVD risk evaluation, an initial 15-minute face-to-face car-diovascular risk counselling session, follow-up phone calls,educational handbook, and a medical examination. An elec-tronic health prescription software was developed to prescribethe program to the intervention group. The software calcu-lated overall 10-year CVD risk, informed participants ofunhealthy practices, and provided an individualized inter-vention plan aimed to improve their risk factors. The controlgroup received the annual medical examination and reportdetailing their results. At 12 months, there was a significantlylower mean 10-year CVD risk score in the intervention armcompared with the control arm (adjusted difference, �2.83%;P ¼ 0.001). However, there was no significant reduction inthe mean CVD risk at 1 year compared with baseline for theintervention group. This suggests the mean differencebetween groups was driven by the relative increase in car-diovascular risk in the control group (1.77% [95% CI, 0.62-2.92] increase at 1 year compared with baseline). Despite this,the components of the 10-year risk of CVD (systolic bloodpressure [SBP], total cholesterol, body mass index [BMI])were all significantly lower at 12 months compared withbaseline in the intervention group. This might suggest the

study was underpowered to detect differences in overall10-year CVD risk, despite showing significant differences inthe individual components of the risk score. The difference at12 months compared with baseline in intervention vs controlswere also significantly different in SBP (�5.55 vs þ6.89 mmHg; P < 0.001), diastolic blood pressure (�6.61 vs þ5.62mm Hg; P < 0.001), fasting blood glucose (�0.31 vs þ0.02mmol/L; P ¼ 0.001), total cholesterol (�0.36 vs �0.10mmol/L; P ¼ 0.005), waist-to-hip ratio (�0.02 vs þ0.01;P < 0.001), and BMI (�0.57 vs þ0.29; P < 0.001).26

In contrast, Anand et al.27 published the results of a single-blind, community-based, RCT that recruited Asian adultsliving in Canada without CHD. The intervention in thisstudy was not a text-message only based intervention butrather a study examining the role of electronic messaging.Participants were provided with a choice of their preferredintervention (e-mail or SMS). The messages (either e-mail ortext) were of 2 types: messages that supported confidence inbehaviour change, and health advice focused on diet andphysical activity. Less than 8% chose SMS. At the 12-monthfollow-up, there was no significant difference in overall MIrisk score between the intervention and control groups. Theinvestigators attributed the negative results to the need toactively access e-mails rather than having text messagesdirectly delivered to the mobile phones.

Smartphone-based interventions

We could identify no RCTs assessing the effect of apps ontotal cardiovascular risk in a primary prevention cohort onreviewing the current literature. However, Zhang et al.28

recently developed a smartphone-based CHD preventionprogram called Care4Heart and conducted an RCT of 80adults without CHD to assess CHD awareness and lifestylebehaviour change. The intervention comprised a 4-week

Klimis et al. 5mHealth Interventions for Heart Disease

program involving a mobile app consisting of 4 learningmodules covering common signs and symptoms, risk factoreducation, and healthy lifestyle practices including smokingcessation and stress management. Users could calculate theirBMI, daily caloric intake, and assess their 10-year CHD risk.In addition, participants would have a 20-minute briefingsession and daily text messages with CHD prevention advice.The control group was provided with Web site addresses toperuse at their leisure. At the end of the 4-week program,participants in the intervention group had significantly betteroverall CHD knowledge. However, there were no significantdifferences in lifestyle behaviours between the groups, exceptwith improved behaviours related to blood cholesterol controlin the intervention compared with control arm.28 Unfortu-nately, assessing the difference in total CHD risk scores at theend of the program was not part of this trial.

There is some nonrandomised evidence. For example,Widmer et al.29 examined the effect of using a digital healthintervention (DHI) offered to a multicentre primary preven-tion cohort of 30,974 participants. The DHI was an onlineand smartphone-based portal, which contained educationalmaterial, and gave the user tasks to improve health includingtracking and logging activities. It also included reminders tocomplete tasks in the form of text messages and e-mails. Mostparticipants had no (14,173) or very low (defined as < 12 peryear; 12,260) log-ins, and 3360 had monthly, 651 weekly,and 530 semiweekly log-ins. Changes in waist circumference,weight, BMI, blood pressure, lipids, and blood glucose wereassessed at 1 year. Significant improvements in weight losswere seen with increasing DHI usage with the greatest changebeing in the semiweekly group compared with weekly (�3.39� 1.06 pounds; P ¼ 0.0013). Older, female, Hispanic in-dividuals with a greater number of risk factors were morelikely to adhere to the DHI.29

It is hard to draw conclusions at this stage from thisliterature, because of the few studies and varied interventions.The mixing of multiple modalities (eg, SMS and e-mail), andthe considerable heterogeneity between studies in frequency ofdelivery and level of interaction are all challenges to synthesisof the evidence. Additionally, mHealth methods have gener-ally been evaluated in selected settings where individuals arelikely to be highly motivated, and thus the results might notbe applicable across various demographic populations. Weneed more information on how patients interact with theseinterventions to identify how they work, which might assist inoptimizing mHealth interventions in these populations.

Secondary Prevention Strategies in mHealthTargeting Multiple Risk Factors

There have been a number of mHealth interventions thataim to deliver secondary prevention to patients with CVD andthat target multiple risk factors simultaneously (Table 2).

Text message-based interventionsThe Tobacco, Exercise, and Diet Messages (TEXT ME)

trial30 was a single-centre RCT of 710 patients with CHDthat examined the effect of a semipersonalized text message-based intervention compared with usual care on objectivemeasures of cardiovascular risk factors at 6 months. The

intervention comprised semipersonalized (customized tobaseline risk factors) text messages (4 per week) over a 6-month period, which provided advice, motivation, and sup-port on diet, smoking cessation if relevant, physical activity,and general cardiovascular information. The text messageswere 1-way, that is a response was not encouraged and par-ticipants were told that responses would not be answered. Theprimary outcome of low-density lipoprotein cholesterol wassignificantly lower in the intervention compared with theusual care group and similarly the intervention group achievedlower SBP, lower BMI, higher rates of smoking cessation, anda greater increase in physical activity at the 6-month finalfollow-up (Table 2).30

The Text4Heart trial31 performed in a New Zealandcohort was a smaller RCT (n ¼ 123), which also addressedmultiple behaviour changes (smoking cessation, physical ac-tivity, healthy diet, and nonharmful alcohol use) in those withestablished CHD. The trial lasted 24 weeks, and the inter-vention group had received 1 message per day (7 per week),which decreased to 5 messages each week from week 13, andhad access to a supporting Web site and a pedometer. Thestudy identified positive effects of adherence to healthy life-style behaviours at 3 months but not at 6 months. This raisesthe question about the durability of mHealth interventions aswell as the optimal frequency of text messages.

The TEXT ME and Text4Heart trials were both non- orminimally-interactive respectively, and messages semi-personalized (eg, only smokers would receive smoking cessa-tion messages). In contrast, Blasco et al.32 conducted a12-month single-blind RCT of 203 patients who had a pre-vious acute coronary syndrome and applied a Web-basedtelemonitoring service with personalized text messages onthe basis of cardiovascular data sent via mobile phone. Spe-cifically, the intervention group would send blood pressure,heart rate, weight (weekly), and glucose and lipids (monthly)results through their mobile phones following a structuredquestionnaire. A cardiologist would then access the data andsend individualized text messages with recommendationsabout CVD prevention on the basis of the data received. Atthe 12-month follow-up, participants in the interventiongroup were significantly (relative risk, 1.4; 95% confidenceinterval, 1.1-1.7) more likely to have an improved cardio-vascular risk profile than controls (69.6% vs 50.5%; P ¼0.01). Furthermore, more patients achieved control for bloodpressure and glycosylated hemoglobin in the interventiongroup compared with controls, with no difference in smokingcessation or low-density lipoprotein cholesterol.32

Smartphone-based interventions

Effects on total cardiovascular risk or simultaneous multi-ple cardiovascular risk factor reduction have been less prom-ising with other modalities of mHealth. Johnston et al.33

examined 174 ticagrelor-treated MI patients in a multicentreRCT. The intervention was an interactive patient support tool(app) that included 4 main modules in which patients wereencouraged to actively enter data on physical activity, weight,smoking habits, and a medication adherence e-diary. Addi-tionally, patients could enter data on blood pressure, choles-terol, and blood glucose levels. There was no significant

Table 2. Randomized controlled trials using mHealth secondary prevention programs with a simultaneous multiple risk factor strategy

Study Participants Intervention Outcome

Text message-based interventionsChow et al.30 N ¼ 710 Australians with

proven CHDMean age 58 � 9.2 years82% Male sex

6-month durationSMS (4 per week)

Measured at 6 monthsSignificantly lower LDL-C (�5 mg/dL; 95% CI, �9 to 0; P ¼0.04), SBP

(�7.6 mmHg; 95% CI, �9.8 to �5.4; P < 0.001), BMI (�1.3; 95% CI,�1.6 to �0.9; P < 0.001) in intervention group compared with controls

Significantly increased physical activity in intervention group compared withcontrols (293.4 MET minutes per week; 95% CI, 102.0-484.8; P ¼ 0.003)

Significantly reduced levels of smoking in the intervention group vs controls(RR, 0.61; 95% CI, 0.48-0.76; P < 0.001)

Participants in the intervention group were more likely to achieve control of 4or more of 5 risk factors (LDL-C < 77 mg/dL; BP < 140/90 mm Hg,regular physical activity, smoking cessation, BMI < 25 (RR, 2.80; 95%CI, 1.95-4.02; P < 0.001)

Pfaeffli Daleet al.31

N ¼ 123 New Zealanderswith proven CHD

Mean age 59.5 � 11.1years

81.3% Male sex

24-week durationDaily SMS and supporting

Web site.

Measured at 3 and 6 monthsAt 3 months participants in the intervention group were more likely to self-

report adherence to healthy lifestyle behaviours compared with controls(AOR, 2.55; 95% CI, 1.12-5.84; P ¼ 0.03). This did not stay significant at6 months (AOR, 1.93; 95% CI, 0.83-4.53; P ¼ 0.13)

Significantly greater self-reported medication adherence scores in theintervention group compared with controls at 6 months (mean difference,0.58; 95% CI, 0.19-0.97; P ¼ 0.004)

Blasco et al.32 N ¼ 203 SpanishACS survivorsMean age 60.6 � 11.5)

years in the interventiongroup and 61 � 12.2years in the controlgroup

81.4% Male sex in theintervention group, and79.2% male sex in thecontrol group

12-month durationBidirectional personalized SMS

Measured at 12 monthsIntervention group more likely (RR, 1.4; 95% CI, 1.1-1.7) to have improved

cardiovascular risk profile compared with controls (69.6% vs 50.5%;P ¼ 0.01)

Intervention group compared with controls more likely to achieve target levelsof BP (62.1% vs 42.9%; P ¼ 0.012) and HbA1c (86.4% vs 54.2%;P ¼ 0.018). BMI was lower in intervention group compared withcontrols (�0.77 vs 0.29; P ¼ 0.005). There were no differences insmoking cessation or LDL-C

Smartphone-based interventionsJohnston et al.33 N¼174 ticagrelor-treated

Swedish MI patientsMean age 58 years81% Male sex

6-month durationSmartphone app (interactive

patient support tool app)

Measured at 6 monthsSignificantly greater adherence in intervention group compared with controls

(nonadherence score 16.6 vs 22.8; P ¼ 0.025)No significant difference in degree of smoking cessation, physical activity,

quality of life scoresPatient satisfaction higher in intervention vs control group (system usability

score 87.3 vs 78.1; P ¼ 0.001)Varnfield et al.34 N ¼ 120 Australians

post-MIMean age 55.7 � 10.4

years vs 55.5 � 9.6 yearsbetween groups

82% vs 85% male sexbetween groups

6-week duration6-month self-management

phase in which participantsfrom both groups wereencouraged to continuelifestyle changes

Smartphone app, SMS, andpreinstalled smartphoneaudio-visual files; 1-hourinitial face-to-face trainingsession; weekly scheduledtelephone conversations.Control group was standardCR.

Measured at 6 weeks and 6 monthsSignificantly higher uptake (80% vs 62%), adherence (94% vs 68%), and

completion (80% vs 47%) in intervention vs traditional CR group(all P < 0.05)

Both groups showed improvements in 6MWT (intervention 60 m [P < 0.00]vs 47 m in controls [P ¼ 0.001]) at 6 weeks, which was maintained at 6months. No significant difference between groups (P ¼ 0.4)

At 6 weeks the intervention group showed significant weight reduction (0.97%weight change; 95% CI, �1.8 to �0.1; P ¼ 0.02) and significantimprovements in mental health (K10: median, 14.6 [IQR, 13.4-16.0] to12.6 [IQR, 11.5-13.8]; P ¼ 0.001) and quality of life scores (EQ5D index:median, 0.84; IQR, 0.8-0.9 to 0.92; IQR, 0.9-1.0; P < 0.001). Thebetween group difference for the EQ5D index was significant at 6 weeks,but not the other listed outcomes

Between group differences for changes in 6MWT, LDL-C, HDL-C, EQ5Dindex, or K10 were not significant at 6 months

Widmer et al.35 N ¼ 80 Americans post-PCI for ACS

Mean age 63.6 � 10.9years vs 62.5 � 10.7years between groups

85% Male sex vs 78%between groups

3-month durationDHI consisting of online and

smartphone-based programin addition to standard CRprogram; 30-minute trainingsession for the DHI. Controlarm was standard CR

Measured outcomes at 90 days and emergency visits/rehospitalizations at180 days

Intervention group had greater weight loss compared with the control group(�5.1 � 6.5 kg vs �0.8 � 3.8 kg; P ¼ 0.02)

No significant difference in other metrics assessed (BP, lipids, glucose, HbA1c,exercise capacity, medication adherence)

No significant change in emergency visits/rehospitalizations (8.1% vs 26.6%;P ¼ 0.054).

ACS, acute coronary syndrome; AOR, adjusted odds ratio; app, application; BMI, body mass index; BP, blood pressure; CHD, coronary heart disease;CI, confidence interval; CR, cardiac rehabilitation; DHI, digital health intervention; EQ5D, quality of life score; HbA1c, glycosylated hemoglobin; HDL-C, high-density lipoprotein cholesterol; IQR, interquartile range; K10, Kessler psychological distress scale; LDL-C, low-density lipoprotein cholesterol; MET, metabolicequivalent; mHealth, mobile health; MI, myocardial infarction; 6MWT, 6-minute walk test; PCI, percutaneous coronary intervention; RR, relative risk; SBP,systolic blood pressure; SMS, short messaging service.

6 Canadian Journal of CardiologyVolume - 2018

Klimis et al. 7mHealth Interventions for Heart Disease

change in multiple cardiovascular risk factors (smoking,physical activity) and quality of life at 6 months.33

Varnfield et al.34 published the results of an unblindedRCT that recruited 120 Australian patients who had a pre-vious MI to investigate the effect of a smartphone-based homeservice delivery (Care Assessment Platform) on cardiac reha-bilitation. The intervention group received a 6-week programinvolving an app for health and exercise monitoring, SMS,and audio-visual files. The app included a step counter andexercise diary. The control group participated in a traditionalcentre-based cardiac rehabilitation program. After the 6-weekprogram, participants were encouraged to maintain lifestylechanges. At the 6-month follow-up, there was a significantlyhigher uptake (80% vs 62%), adherence (94% vs 68%), andcompletion (80% vs 47%) of the rehabilitation program inthe intervention vs control group (all P < 0.05). There wereequivalent improvements in 6-minute walk test and healthydiet in both groups at 6 weeks, which was maintained at 6months. In addition, in the intervention group at 6 weeksthere was significantly improved psychological and quality oflife scores, and small but significant improvements in weightand waist circumference, not maintained at 6 months. Theauthors concluded that smartphone-based cardiac rehabilita-tion programs were used more and were at least as effective ascentre-based methods.34 Similarly, Widmer et al.35 recentlypublished a small RCT (N ¼ 80) using combined DHI and atraditional cardiac rehabilitation program over 3 months andcompared this with traditional rehabilitation. Primary out-comes were cardiovascular-related emergency visits and reho-spitalizations, and secondary outcomes were changes incardiovascular risk factor parameters. Whereas at 90 days,compared with baseline, there was a significant difference inweight loss (�5.1 � 6.5 kg vs �0.8 � 3.8 kg; P ¼ 0.02),there were no statistically significant differences in other pa-rameters including blood pressure, glycemic control, physicalactivity, and medication adherence. Additionally, althoughthere was a trend toward reduced hospital visits and admis-sions at 180 days, this did not reach statistical significance(8.1% vs 26.6%; P ¼ 0.054).35 This might be because thestudy was underpowered to detect significant differences inthe primary outcome of cardiovascular-related emergencyvisits and rehospitalizations.

The studies presented generally found nil or small im-provements in cardiovascular risk factors, but the small trialsizes mean they are unlikely to be powered to detect smallimprovements in cardiovascular risk factors. The theme of anumber of them also suggests the potential for reducing healthservice utilization and hospitalization, but further research isrequired to clarify this. If mHealth interventions were able toreach large numbers of people, small improvements might beimpactful on a population level. Few studies discuss orexamine the extent to which the interventions they proposewould be efficacious if implemented into practice, and thisshould be addressed in future studies.

Optimizing Medication AdherenceMedication adherence is a challenge in chronic disease

management, and mHealth strategies have also been proposedas having potential in addressing this.36,37 In a recent sys-tematic review that included 107 articles targeting adherence

for chronic diseases, most (40.2%) had used SMS-predominant interventions, and 23% used app-predominantmethods.38 Of the 6 cardiovascular RCTs that assessed theeffectiveness of mHealth on adherence outcomes (predomi-nantly SMS-based), 5 showed a significantly improvedmedication adherence rate in the intervention vs controlgroups. Published after this review was the multicentre RCTpublished by Johnston et al., previously referenced.33 Thisstudy in addition assessed drug adherence via an interactivepatient support app (described previously) in post-MI pa-tients. They showed that self-reported medication adherencewas higher in the intervention vs control arm at the 6-monthfollow-up (nonadherence score [lower score indicates betteradherence]: 16.6 vs 22.8; P ¼ 0.025).33

More recently, a meta-analysis was published that assessedSMS-based interventions on medication adherence in chronicdisease including CVD.39 In this study of 2742 patients across16 RCTs, participants who received SMS-based interventionswere more than twice as likely to have adequate medicationadherence compared with controls (odds ratio, 2.11; 95%confidence interval, 1.52-2.93; P < 0.001).39 The medianperiod of intervention was 12 weeks. Similarly, increasedadherence to pharmacologic and nonpharmacologic therapywas shown in a recent systematic review including 27 studies(5165 patients) that compared mHealth (all modalities) withstandard care.40

The studies presented have shown promise. However,most are predominantly text-message based, and of shortduration. More studies are needed to assess the more appro-priate medium for delivery (ie, text-message based, app-based,or combination), and long-term studies are required beforetranslation into current health care delivery.

Limitations of the Current Evidence and FutureImplications

It is apparent that mHealth strategies for CVD preventionhave the potential to improve outcomes. The body of evi-dence is generally supportive of this, however, there are still asignificant number of unanswered questions about how tooptimize interventions, how to maintain engagement and ef-ficacy, and information on how to quantify the effect size ofthese interventions is lacking.

However, there is already an overwhelming number ofcardiovascular mHealth options available to consumers. Manyof these available mHealth interventions have not undergonesignificant evaluation, and most apps have not gone throughFood and Drug Administration approval. Although directharm from an mHealth intervention might seem unlikely, alack of benefit is a harm to public health, and the delivering ofincorrect information potentially even more harmful. Imple-mentation of high-quality and effective mHealth interventionsneeds to be our goal. And although the background hasbecome increasingly complex with the range of freewareavailable to consumers, as clinicians and patients we still seekthose interventions that are most effective.

Thus, future studies are needed to better evaluate mHealthinterventions with respect to long-term outcomes, optimizingprograms (including program duration, frequency of textmessages, degree of complexity, degree of interactivity, anddegree of personalization), and safety, effectiveness, and cost-

8 Canadian Journal of CardiologyVolume - 2018

effectiveness. Furthermore, studies assessing mHealth pro-grams across different socioeconomic classes, cultural groups,and in different languages would be highly valuable indetermining the potential reach of these programs and how tocustomize and optimize them for wider delivery. Importantly,future mHealth producers should collaborate with cliniciansand regulatory agencies to ensure the interventions are safeand effective, and outcome measures standardized.

There are important limitations to consider in our review.Because of the paucity of RCTs evaluating mHealth onmultiple risk factor reduction, this is a narrative literaturereview and not a systematic review. Additionally, we have onlyincluded published studies.

ConclusionsmHealth interventions are a novel, exciting, and expanding

field in medicine that will potentially transform health caredelivery by improving access to treatments that wouldotherwise require frequent clinic/hospital visits. In particular,SMS-based interventions, because of their low cost andubiquity, could assist in achieving equity in delivery of car-diovascular prevention programs to developed as well asdeveloping nations. There is evidence supporting the utility ofmHealth programs, at least in the short-term. However, thereare many unknown variables including which mode ofmHealth (or combination of) would be most efficacious. Thecurrent challenge is ensuring patients have access to evidence-based choices that have the input of clinicians.

Funding SourcesClara K. Chow is supported by National Health and

Medical Research Council Career Development Award(APP1105447) co-funded by a Future Leader Fellowshipfrom the National Heart Foundation.

DisclosuresThe authors have no conflicts of interest to disclose.

References

1. World Health Organisation. Media Centre. Cardiovascular Diseases(CVDs). Available at: http://www.who.int/mediacentre/factsheets/fs317/en/. Accessed February 5, 2016.

2. Australian Institute of Health and Welfare. Monitoring acute coronarysyndrome using national hospital data: an information paper on trendsand issues. Cat. no. CVD 57 2011. Canberra: AIHW, 2011.

3. Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable riskfactors associated with myocardial infarction in 52 countries (theINTERHEART study): case-control study. Lancet 2004;364:937-52.

4. Smith SC Jr, Benjamin EJ, Bonow RO, et al. AHA/ACCF secondaryprevention and risk reduction therapy for patients with coronary andother atherosclerotic vascular disease: 2011 update: a guideline from theAmerican Heart Association and American College of CardiologyFoundation endorsed by the World Heart Federation and the PreventiveCardiovascular Nurses Association. J Am Coll Cardiol 2011;58:2432-46.

5. Bethell HJ, Evans JA, Turner SC, Lewin RJ. The rise and fall of cardiacrehabilitation in the United Kingdom since 1998. J Public Health (Oxf)2007;29:57-61.

6. Chew DP, Anderson FA, Avezum A, et al. Six-month survival benefitsassociated with clinical guideline recommendations in acute coronarysyndromes. Heart 2010;96:1201-6.

7. Kotseva K, Wood D, De Bacquer D, et al. EUROASPIRE IV: a Euro-pean Society of Cardiology survey on the lifestyle, risk factor and ther-apeutic management of coronary patients from 24 European countries.Eur J Prev Cardiol 2016;23:636-48.

8. Thakkar JB, Chow CK. Adherence to secondary prevention therapies inacute coronary syndrome. Med J Aust 2014;201:S106-9.

9. World Health Organisation. mHealth: new horizons for health throughmobile technologies: second global survey on eHealth, Vol 2016. Geneva,Switzerland: World Health Organization, 2011.

10. Fierce Wireless. Report: global smartphone penetration to jump 25% in2014. Available at: http://www.fiercewireless.com/wireless/report-global-smartphone-penetration-to-jump-25-2014-led-by-asia-pacific. AccessedJuly 20, 2017.

11. Bastawrous A, Armstrong MJ. Mobile health use in low- and high-income countries: an overview of the peer-reviewed literature. J R SocMed 2013;106:130-42.

12. Burke LE, Ma J, Azar KM, et al. Current science on consumer use ofmobile health for cardiovascular disease prevention: a scientific statementfrom the American Heart Association. Circulation 2015;132:1157-213.

13. Smith SC Jr. Current and future directions of cardiovascular risk pre-diction. Am J Cardiol 2006;97:28A-32A.

14. World Health Organisation. Global Action Plan for the Prevention andControl of NCDs 2013-2020. Available at: http://www.who.int/nmh/events/ncd_action_plan/en/. Accessed March 10, 2016.

15. Ndindjock R, Gedeon J, Mendis S, Paccaud F, Bovet P. Potential impactof single-risk-factor versus total risk management for the prevention ofcardiovascular events in Seychelles. Bull World Health Organ 2011;89:286-95.

16. International Telecommunications Union. Measuring the Information So-ciety Report 2015. Available at: http://www.itu.int/en/ITU-D/Statistics/Documents/publications/misr2015/MISR2015-w5.pdf. Accessed October5, 2017.

17. Neubeck L, Lowres N, Benjamin EJ, et al. The mobile revolutioneusingsmartphone apps to prevent cardiovascular disease. Nat Rev Cardiol2015;12:350-60.

18. Meng Y, Wong SS. Trends and features of top 100 grossing health andfitness iPhone apps [abstract 1028.5]. FASEB J 2014;28(suppl):1028.5.

19. Pew Research Center. Older adults and technology use. Available at: http://www.pewinternet.org/2014/04/03/older-adults-and-technology-use. AccessedJune 2, 2017.

20. Whittaker R, McRobbie H, Bullen C, Rodgers A, Gu Y. Mobile phone-based interventions for smoking cessation. Cochrane Database Syst Rev2016;4:CD006611.

21. Siopis G, Chey T, Allman-Farinelli M. A systematic review and meta-analysis of interventions for weight management using text messaging.J Hum Nutr Diet 2015;28(suppl 2):1-15.

22. Maddison R, Pfaeffli L, Whittaker R, et al. A mobile phone interventionincreases physical activity in people with cardiovascular disease: resultsfrom the HEART randomized controlled trial. Eur J Prev Cardiol2015;22:701-9.

23. Liang X, Wang Q, Yang X, et al. Effect of mobile phone intervention fordiabetes on glycaemic control: a meta-analysis. Diabet Med 2011;28:455-63.

Klimis et al. 9mHealth Interventions for Heart Disease

24. Logan AG, Irvine MJ, McIsaac WJ, et al. Effect of home blood pressuretelemonitoring with self-care support on uncontrolled systolic hyper-tension in diabetics. Hypertension 2012;60:51-7.

25. Bobrow K, Farmer AJ, Springer D, et al. Mobile phone text messages tosupport treatment adherence in adults with high blood pressure (SMS-Text Adherence Support [StAR]): a single-blind, randomized trial. Cir-culation 2016;133:592-600.

26. Liu Z, Chen S, Zhang G, Lin A. Mobile phone-based lifestyle inter-vention for reducing overall cardiovascular disease risk in Guangzhou,China: a pilot study. Int J Environ Res Public Health 2015;12:15993-6004.

27. Anand SS, Samaan Z, Middleton C, et al. A digital health intervention tolower cardiovascular risk: a randomized clinical trial. JAMA Cardiol2016;1:601-6.

28. Zhang H, Jiang Y, Nguyen HD, Poo DC, Wang W. The effect of asmartphone-based coronary heart disease prevention (SBCHDP) pro-gramme on awareness and knowledge of CHD, stress, and cardiac-relatedlifestyle behaviours among the working population in Singapore: a pilotrandomised controlled trial. Health Qual Life Outcomes 2017;15:49.

29. Widmer RJ, Allison TG, Keane B, et al. Workplace digital health isassociated with improved cardiovascular risk factors in a frequency-dependent fashion: a large prospective observational cohort study.PLoS One 2016;11:e0152657.

30. Chow CK, Redfern J, Hillis GS, et al. Effect of lifestyle-focused textmessaging on risk factor modification in patients with coronary heartdisease: a randomized clinical trial. JAMA 2015;314:1255-63.

31. Pfaeffli Dale L, Whittaker R, Jiang Y, et al. Text message and internetsupport for coronary heart disease self-management: results from theText4Heart randomized controlled trial. J Med Internet Res 2015;17:e237.

32. Blasco A, Carmona M, Fernandez-Lozano I, et al. Evaluation of a tele-medicine service for the secondary prevention of coronary artery disease.J Cardiopulm Rehabil Prev 2012;32:25-31.

33. Johnston N, Bodegard J, Jerstrom S, et al. Effects of interactive patientsmartphone support app on drug adherence and lifestyle changes inmyocardial infarction patients: a randomized study. Am Heart J2016;178:85-94.

34. Varnfield M, Karunanithi M, Lee CK, et al. Smartphone-based homecare model improved use of cardiac rehabilitation in postmyocardialinfarction patients: results from a randomised controlled trial. Heart2014;100:1770-9.

35. Widmer RJ, Allison TG, Lennon R, et al. Digital health interventionduring cardiac rehabilitation: a randomized controlled trial. Am Heart J2017;188:65-72.

36. Ho PM, Magid DJ, Shetterly SM, et al. Medication nonadherence isassociated with a broad range of adverse outcomes in patients with cor-onary artery disease. Am Heart J 2008;155:772-9.

37. World Health Organisation. Adherence to Long-term Therapies: Evi-dence for Action. Available at: http://www.who.int/chp/knowledge/publications/adherence_report/en/. Accessed October 2, 2017.

38. Hamine S, Gerth-Guyette E, Faulx D, Green BB, Ginsburg AS. Impactof mHealth chronic disease management on treatment adherence andpatient outcomes: a systematic review. J Med Internet Res 2015;17:e52.

39. Thakkar J, Kurup R, Laba TL, et al. Mobile telephone text messaging formedication adherence in chronic disease: a meta-analysis. JAMA InternMed 2016;176:340-9.

40. Gandhi S, Chen S, Hong L, et al. Effect of mobile health interventionson the secondary prevention of cardiovascular disease: systematic reviewand meta-analysis. Can J Cardiol 2017;33:219-31.