The efficacy of mobile health in alleviating risk factors related to the occurrence and development of coronary heart disease: A systematic review and meta‐analysis

Abstract The association between the efficacy of mobile health and the occurrence and development of coronary heart disease (CHD) is still unclear. Mobile health can alleviate the risk factors for CHD. PubMed, EMbase, Web of Science, The Cochrane Library, CNKI, WanFang, and VIP databases were searched from inception through May 28, 2020. Randomized controlled trials of the effect of mobile health in alleviating the risk factors of CHD's occurrence and development were included. Risks of bias were assessed by two independent reviewers by using the RevMan 5.3, GRADEpro, and RoB2.0 to generate findings. Meta‐analyses were performed to investigate the effects of mobile health on risk factors for CHD. Subgroup analyses were conducted. Sixteen randomized controlled trials, including 3898 patients with CHD, were included. Meta‐analysis results showed that mobile health can reduce BMI (mean difference [MD] = − 1.24, 95% CI = − 2.02 to − 0.45, p < .05), waist circumference (MD = − 4.40, 95% CI = − 4.72 to − 4.08, p < .00001), total cholesterol (TC) level (MD = − 0.43, 95% CI = − 0.64 to − 0.22, p < 0.00001), low‐density lipoprotein cholesterol (LDL‐C) level (MD = − 0.31, 95% CI = − 0.48 to − 0.15, p < .05), diastolic blood pressure (MD = − 2.01, 95% CI = − 3.40 to − 0.623, p < .05), and depression (MD = − 8.32, 95% CI = − 12.83 to − 3.81, p < .05) and increase high‐density lipoprotein cholesterol level (MD = 0.16, 95% CI = 0.01 to 0.32, p < .05) with statistically significant differences. The results of subgroup analyses indicated that the simple mobile health intervention has more remarkable advantages in reducing BMI, TC, LDL‐C, and systolic blood pressure than the complex mobile health intervention. Mobile health can alleviate the risk factors for CHD and has a certain effect on the prevention and recovery of CHD. Simple mobile health has a remarkable advantage. Limited by the quantity and quality of included studies, future research enrolling high‐quality studies should be taken to verify the above conclusions.


| INTRODUCTION
Coronary heart disease (CHD) refers to coronary arterial stenosis or obstruction caused by progressive coronary atherosclerotic lesions and the resulting myocardial ischemia or necrosis. 1,2 CHD is the main component of global cardiovascular disease. 3 Given its high prevalence and mortality rate, CHD has become a notable public health concern. 4 Studies have shown 5 that atherosclerosis is an inflammatory disease, and sickness (such as high blood pressure and abnormal lipoprotein content) are risk factors for the occurrence and promotion of inflammation. According to the 2020 American Heart Association report, 6 patients with CHD are at a high risk of recurring coronary events. The main causes of recurrence are hypertension, hypercholesterolemia, dyslipidemia, and obesity or overweight. 7 Current CHD prevention guidelines give high priority to the intensive control of CHD risk factors. 8 At present, clinicians guide patients to self-monitor risk factors through health education and follow-up. However, these efforts have limited success due to the lack of family medical equipment, medical knowledge deficit, and weak self-supervision and management. The mobile health (mhealth) can overcome these limitations and can remotely monitor and guide patients' home rehabilitation through smartphone applications, wearable devices, and text messages. The mhealth improves the patients' lifestyle and quality of life and assists them in meeting their individual needs. 9 This study has used the AMSTAR 2.0 10 for systematic review/meta-analysis, reported data in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, analyzed the influence of mhealth on CHD risk factors, and provided evidence of higher methodological and reporting quality for the effect of mhealth on alleviating CHD risk factors.

| Search strategy
This study is a systematic review and meta-analysis performed in accordance with the PRISMA guidelines. 11 The search was conducted by two independent reviewers. Relevant randomized controlled trials

| Selection criteria
Studies were included in accordance with the following eligibility criteria: (1) Designed as a RCT or a cluster RCT; (2)

| Data extraction and endpoints
Two researchers screened the literature independently and performed data extraction and cross-checking. Disagreements were resolved through discussion or negotiation with a third reviewer. Titles and abstracts were read first to exclude evidently irrelevant literatures.
The full texts of the remaining articles were reviewed for final inclusion. If necessary, the original research authors were contacted by email or phone to obtain undetermined but important information.
The following data were extracted: (1) Basic information of the included research (e.g., title, author, journal, and year), (2

| Data analysis and synthesis
The RevMan5.3 software was used for statistical analysis. Measurement data used the mean difference (MD) as the effect analysis statistics, and each effect size provided 95% CI. The heterogeneity among the results of the included studies was analyzed using the χ2 test (test level: α = 0.1), and the degree of heterogeneity was quantitatively judged in combination with I 2 . The fixed-effects model was used for analysis if no statistical heterogeneity among the results of each study was observed. If a statistical heterogeneity was observed among the results of each study, the source of the heterogeneity was further analyzed. After the exclusion of the influence of evident clinical heterogeneity, the random-effects model was used for analysis.
The level of meta-analysis was set to α = 0.05. Evident clinical heterogeneity was treated using the subgroup analysis.

| Study quality assessments
The included studies' risk of bias was evaluated independently and cross-checked by two investigators. Assessments were performed using the GRADEpro and the RCT bias risk assessment tool ROB2.0, which was recommended by the Cochrane manual. 12 3 | RESULTS A total of 3198 articles were retrieved from seven databases, and 1000 articles were duplicates. By reading the title, abstract, and keywords, 2148 articles were screened for inappropriate intervention measures, population, and non-RCTs. The remaining 50 full-text articles were assessed independently against the eligibility criteria by two researchers. A total of 34 articles were excluded. Five studies were unable to obtain the full text due to language. The researcher tried to contact the author but got no response. The primary outcome indicators of 21 articles did not meet the inclusion criteria. Two articles had participants that did not match. The study types of six articles did not match.

| Quality appraisal
This article uses the ROB2.0 to evaluate the literature. The risk of bias is shown in Supplementary Figures 1 and 2. Three trials 15,17,19 (18.75%) were rated to have low risk of bias for the randomization process. Nine trials 14,15,[21][22][23][24][25][26]28 (56.25%) were rated to have low risk of bias for deviations from intended interventions. All trials were rated to F I G U R E 1 Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram have low risk of bias for missing outcome data, measurement of the outcome, and selection of the reported result. One trial 15 (6.25%) was rated to have low risk of bias for overall bias. In general, all studies have low or medium risk of bias due to insufficient reporting on the randomization or the allocation process. In most studies, blinding was not implemented and applicable for subjects, but the lack of blinding was unlikely to have influenced the assessment of primary outcome indicators.
Funnel plots were applied to evaluate the publication bias of the study, which included more than 10 articles. In this study, the SBP was incorporated into the literature to make the funnel chart. Results showed partial asymmetry, suggesting the possibility of publication bias ( Figure 2).

| Meta-analysis for outcome measures
In the included studies, a meta-analysis of three risk factors, namely, obesity, cholesterol, and blood pressure and emotions, was performed. The data evaluated in each field were summarized as follows.

| Waist circumference
Two RCTs 15,20 with 764 patients were included. The results of the fixedeffects model analysis showed that the waist circumference of the

| BMI
The heterogeneity test after merging still showed a large heterogeneity. Thus, the random-effects model was used for analysis. Four stud-

| Low-density lipoprotein cholesterol
Three studies 15,21,24 were included in the simple mhealth group and showed that the simple mhealth intervention was significantly better than the control in reducing the LDL-C level (MD = − 0.40, 95% CI = − 0.63 to − 0.18, p = .0004). Four studies 16,17,27,28 were included in the complex mhealth group and indicated that the complex mhealth intervention could not significantly reduce the LDL-C level (MD = − 0.24, 95%CI = − 0.57 to 0.08, p = .14; Supplementary Figure 15).

| DISCUSSION
Through the GRADE classification, this study has obtained evidence of different quality levels and demonstrated that the mhealth has a certain effect on alleviating the risk factors that cause the occurrence and development of CHD. In terms of obesity, high-quality evidence indicates that the mhealth can narrow the waist circumference of patients with CHD and play a remarkable role in controlling obesity. Moderate-quality evidence indicates that the mhealth reduces the BMI of patients with CHD. Lowquality evidence suggests that the mhealth has no significant effect on the hip circumference. In terms of blood pressure, low-quality evidence suggests that the mhealth can decrease the DBP but has no significant effect on the SBP. In terms of serum lipid, moderate-quality evidence shows that the mhealth significantly reduces total cholesterol and LDL-C levels and increases the HDL-C level. In terms of emotions, low-quality evidence shows that the mhealth can relieve depression in patients with CHD but has no effect on anxiety.
Obesity is an independent risk factor for the occurrence and development of cardiovascular diseases. 29 Compared with nonobese people, obese people have faster development of coronary atherosclerosis, and the incidence of cardiovascular disease differs due to fat distribution. 30 A previous study 31 shows that obesity-related indicators from the intervention group are significantly reduced after ≥24 weeks. In this study, the short-term mhealth intervention can reduce the patient's BMI and waist circumference but does not significantly improve the hip circumference, which may be related to the difference in the patient's fat distribution. A reasonable intake of diet can help patients control their body. Studies have confirmed that lowsalt, low-fat, and healthy eating habits can significantly reduce the incidence of CHD by 20%-33%, delay the progression of atherosclerosis, reduce cardiovascular and metabolic risk factors, and improve the quality of life. 32,33 Dyslipidemia is one of the prominent risk factors for CHD. 34 Increased LDL-C level and decreased HDL-C level may induce CHD. 35,36 The increase in LDL-C level causes arterial intimal injury, which causes fibrocyte-producing proliferative reactions and eventually develops into atherosclerosis. 37 The reduction in HDL-C level affects peripheral tissues, thereby affecting the antiatherosclerosis effect. 38 In this study, after the intervention of the mhealth, total cholesterol and LDL-C levels decrease significantly, and HDL-C level increases. These results are similar to the results of Dale, 39 indicating that the mhealth treatment is important to alleviate the occurrence and development of CHD. However, study 40 has shown that, compared with changes in cholesterol, 10 year ASCVD risk can better reflect the close relationship between blood lipids and the incidence of coronary heart disease.
There is no research related to 10 year ASCVD risk in the literature included in this study, and it can be combined with mhealth research in the future.
A large global prospective study shows that blood pressure levels are positively correlated with the incidence of CHD. 41

| Limitation
Some shortcomings remain in this study. First, the great heterogeneity in the type of intervention ranging from SMS to mobile application and wearables can be a great contributor to the variation of outcomes among studies. Furthermore, the remission of CHD risk factors is closely related to the intervention duration. However, the compliance with each intervention may be a confounder in long-term studies.
Considering that the intervention duration of the included studies is relatively short, the subgroup analysis cannot be conducted on the basis of the duration factor to explore long-term mhealth intervention.
At the same time, CHD-related emotional indicators are rarely included in the literature, whereas the quality of life, anxiety, and depression are also important factors that influence the occurrence and the recurrence of CHD. Thus, subsequent data collection for this aspect should be strengthened. The RCTs included in this study generally lack standard allocation concealment and blinding because they are not applicable. Although it was not downgraded in this study, further research is needed to strengthen the methods of randomization, allocation, and blinding to improve the quality of evidence.