Effect of electronic health (eHealth) on quality of life in women with breast cancer: A systematic review and meta‐analysis of randomized controlled trials

Abstract Background Women with breast cancer and improved survival face some specific quality of life (QOL) issues. Electronic health (eHealth) is a useful tool aiming to enhance health services. However, evidence remains controversial about the effect of eHealth on QOL in women with breast cancer. Another unstudied factor is the effect on specific QOL functional domains. Therefore, we undertook a meta‐analysis about whether eHealth could improve the overall and specific functional domains of QOL in women with breast cancer. Methods PubMed, Cochrane Library, EMBASE, and Web of Science were searched to identify appropriate randomized clinical trials from database inception to March 23, 2022. The standard mean difference (SMD) served as the effect size and the DerSimonian–Laird random effects model was constructed for meta‐analysis. Subgroup analyses were conducted according to different participant, intervention, and assessment scale characteristics. Results We initially identified 1954 articles excluding duplicates and ultimately included 13 of them involving 1448 patients. The meta‐analysis revealed that the eHealth group had significantly higher QOL (SMD 0.27, 95% confidence interval [95% CI] 0.13–0.40, p < 0.0001) than the usual care group. Additionally, although not statistically significant, eHealth tended to improve the physical (SMD 2.91, 95% CI −1.18 to 6.99, p = 0.16), cognitive (0.20 [−0.04, 0.43], p = 0.10), social (0.24 [−0.00, 0.49], p = 0.05), role (0.11 [0.10, 0.32], p = 0.32), and emotional (0.18 [0.08, 0.44], p = 0.18) domains of QOL. Overall, consistent benefits were observed in both the subgroup and pooled estimates. Conclusions eHealth is superior to usual care in women with breast cancer for improved QOL. Implications for clinical practice should be discussed based on subgroup analysis results. Further confirmation is needed for the effect of different eHealth patterns on specific domains of QOL, which would help improve specific health issues of the target population.


| INTRODUCTION
Breast cancer is the most common malignancy among women worldwide. 1 The global mortality-to-incidence rate was about 36% in 2000 and decreased to 28% in 2020, which means women with breast cancer (including patients and survivors [defined as no cancer recurrence]) live longer than before due to the implementation of mammographybased screening and improved treatments. 1,2 Quality of life (QOL) is of particular importance to breast cancer among women with a longer survival time. However, they face some specific QOL issues that make their lives different from those of healthy people, including intimacy issues, 4 loss of womanhood, 5 and body image distortion, along with the disruptive effects of treatments. 6 Therefore, we need to take the QOL of women with breast cancer into consideration, instead of focusing only on their survival time.
Electronic health (eHealth) is a useful tool aiming to enhance health services and collect patient information through the internet and related technologies (e.g., websites, software, digital gaming, etc.), and helps to tackle barriers including distance, time, cost, and lack of health providers. 3 For example, mobile applications (apps) were found to help solve certain health problems without increasing the risk of exposure to the 2019 novel coronavirus disease (COVID-19) during the pandemic. 4 In addition, Mobile apps are well-accepted by women with breast cancer. 5 Since the evidence remains controversial about the effect of eHealth on QOL in women with breast cancer, [6][7][8][9] it is important to conduct a systematic review and metaanalysis to conclude. Another unstudied factor is the effect of eHealth on specific functional domains of QOL, including physical, cognitive, social, role, and emotional domains. [10][11][12][13][14] Hence, our study aimed to conduct a metaanalysis to explore the effects of eHealth interventions on the overall and five functional domains of QOL.

| Inclusion and exclusion criteria
According to the PICOs (population, intervention, comparison, and outcome), the following inclusion criteria were established: (1) population: women currently or previously diagnosed with breast cancers, regardless of whether completion of medical treatment (e.g., surgery, adjuvant chemotherapy, and radiotherapy); (2) intervention and comparison: eHealth intervention and usual care. eHealth intervention should be healthcare delivered through the internet and related technologies (e.g., websites, software, etc.). Usual care referred to basic healthcare services, including traditional health education (e.g., paper-based instruments) and routine nursing care, etc.; (3) outcome: primary outcome being health-related QOL. The assessment instruments and follow-up duration were not restricted; (4) study design: randomized clinical trials (RCTs); (5) studies being published in English.
The exclusion criteria were (1) studies including other cancers, (2) significant baseline differences between groups, (3) a review, retrospective study, or case report, (4) not divided into eHealth group and control group, and (5) insufficient data in the article.

| Data extraction
Two authors independently reviewed the identified articles. The following baseline and study characteristics were extracted from each included publication: publication information, participant characteristics, intervention method and duration, and QOL assessment scales (Tables 1 and 2).

| Risk of bias assessment
The risk of bias in the clinical trials included in our metaanalysis was assessed according to the recommendations of the Cochrane Handbook of Systematic Reviews of Interventions (http://handb ook.cochr ane.org.), including the following domains: selection bias (random sequence generation and concealment of allocation), performance bias (blinding of participants and personnel), detection bias (blinding of outcome assessment), attrition bias (incomplete outcome data), and reporting bias (selective outcome reporting).

| Quality assessment
The methodological quality of studies in this meta-analysis was evaluated by the Jadad Scale, which scores from 0 to 7 (0-3: low quality; 4-7: high quality). With this tool, each study was assessed in four separate categories: randomization, concealment of allocation, double blinding, and withdrawals and dropouts. 15 The Jadad Scale scores of the included studies ranged from 3 to 5 (Table 1).

| Statistical analysis
Primary outcomes were measured by the difference in QOL (including overall and five functional domains of QOL) between eHealth and control groups. The overall QOL was measured by the total score of each scale included in our studies shown in Table 1; the functional outcomes were measured by the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30), which includes scales measuring physical, role, emotional, cognitive, and social domains. Higher scores of overall QOL or functional domains represent better QOL or functioning. Anxiety, as measured by the State-Trait Anxiety Inventory (STAI), the Hospital Anxiety and Depression Scale (HADS), and the Brief Symptom Inventory (BSI), was a secondary outcome of our study; depression was another secondary outcome measured by the Beck Depression Inventory, the BSI, and the HADS. For scales measuring anxiety and depression, higher scores mean worse issues. We used the standard mean difference (SMD) and a corresponding 95% confidence interval (95% CI) between the eHealth and control groups as the effect size. Given the clinical and methodological quality heterogeneity across trials, the DerSimonian-Laird random effects model was constructed for pooled estimates.
Between-study heterogeneity was quantified by the I 2 statistic and interpreted qualitatively as low (25%-50%), moderate (50%-75%), and high (75%-100%). 16 To identify the sources of heterogeneity, the subgroup, sensitivity, and meta-regression analyses were performed. The meta-regression analysis also investigated if any covariate moderate treatment effect sizes. Covariates for subgroup and meta-regression analyses included health status (patients vs. survivors), medical treatment (undergoing vs. not undergoing), income level (high-income vs. uppermiddle-income countries), whether communication with health providers (yes vs. no), whether mobile-based (yes vs. no), follow-up period (≤3 months vs. 3-12 months), QOL assessment scales (breast cancer-specific vs. cancerspecific vs. general), study quality (high [Jadad scale score >3] vs. low [Jadad scale score ≤3]) and the publication year (within last 10 years vs. earlier years). Sensitivity analyses were conducted by omitting each study in turn. Publication bias was assessed through funnel plots. Meta-regression analysis was performed by Stata/SE 16.0 (StataCorp LLC) while other analyses were by Review Manager 5.4 (The Cochrane Collaboration).
To explore the potential source of heterogeneity, we conducted subgroup, meta-regression, and sensitivity analyses. In subgroup analyses shown in Table 3 showed numerically more effective influence of eHealth on QOL than those from the counterpart subgroups; but the differences between subgroups were not statistically significant.
Meta-regression analyses revealed that the effect sizes did not statistically significantly vary according to participant characteristics, types of eHealth intervention, follow-up period, assessment scale, study quality, and the publication year (all p > 0.30, Table S1). In sensitivity analyses, the pooled estimates were not significantly altered when any one research was omitted in turn, with the SMD ranging from 0.21 (95% CI 0.11-0.32) to 0.29 (0.15-0.43). Therefore, the pooled estimate for QOL is robust in the current study.

| Different functional domains of QOL
Our study assessed the effects of eHealth on five functional domains of QOL, including physical, cognitive, social, role, and emotional domains, as measured by the functional subscales of the EORTC QLQ-C30.

| Publication bias
The potential publication bias of QOL studies was performed, as shown in the funnel plot in (Figure 5). The results showed that a publication bias might exist in the research.

| DISCUSSION
To our knowledge, this is the first meta-analysis to demonstrate the effect of eHealth on overall and specific functional domains of QOL among women with breast cancer. The results revealed that the eHealth group was associated with statistically significant improvement in QOL compared with the usual care group, regardless of participant characteristics, features of the eHealth platforms, and assessment scales. Regarding the functional subscales of EORTC QLQ-C30, the physical, cognitive, role, and emotional domains of QOL showed no significant group effects. However, the eHealth intervention did experience an improvement tread in these functional domains. In addition, compared with usual care, eHealth did not significantly relieve anxiety and depression in our study. Based on our findings, eHealth is worthy of healthcare use for both breast cancer patients and survivors. For the intervention effect on QOL, our results are consistent with a previous meta-analysis which indicated that telehealth intervention was superior to usual care in breast cancer patients for improved QOL. 14 However, the previous meta-analysis covered non-eHealth interventions (e.g., telephone) and did not consider the specific domains of QOL. Our study novelly and specifically evaluated the effect of eHealth on five functional domains of EORTC QLQ-C30. Since the same scale (EORTC QLQ-C30) was used for functional evaluation, information bias would be avoided to some extent. The results indicated F I G U R E 1 Flow diagram of choosing the appropriate articles. QOL, quality of life; RCTs, randomized clinical trials. that eHealth intervention tended to improve these functional domains, although the improvement was not statistically significant. Further confirmation is required for the effect of eHealth on specific health issues.
Our study showed that eHealth did not positively affect anxiety, which was similar to the result of a study about a nurse-led telephone follow-up and educational group program after breast cancer treatment. 25 Although not statistically significant, eHealth tended to relieve depression in our study. The eHealth intervention included in the depression analysis were psychosocial counseling by web-based e-mail, 19 mobile games developed to improve self-management, 21 and mobile breast cancer e-support programs. 26 A study conducted by Akechi et al. 27 demonstrated that smartphone psychotherapy reduced the fear of cancer recurrence and depression among breast cancer survivors. Psychotherapy through eHealth may be a promising way to reduce psychological issues.
Considering the different characteristics of participants, intervention, and outcome assessment instruments in the included RCTs, a random effects model was constructed for meta-analysis, and subgroup analyses were conducted. The use of a random-effects model measured variability between trials and weighted each study's contribution within the pooled effect. Through subgroup analyses, we found that mobile-based eHealth intervention, breast-cancer-specific QOL scales, and publications within the last 10 years suggested associations with higher effect sizes, and the possible explanations are as follows: (1) mobile-based eHealth interventions are more portable and utilization efficient than non-mobile-based interventions. These results could also indicate the improvement of eHealth intervention patterns in the future; (2) breast cancer-specific scales are sensitive for the measurement of breast cancer-specific QOL issue, 28,29 and (3) studies published within the last 10 years included mobile-based interventions, while those published earlier only involved nonmobile-based interventions. Overall, consistent benefits were observed in both the subgroup and pooled estimates. Therefore, the pooled estimate for QOL is generally robust in the current study.
The results of our study are similar to those of some studies including participants with other cancers. [30][31][32] One clinical trial showed that tele-motivational interviewing was an effective and acceptable intervention for overweight participants with cancer (e.g., breast, prostate, uterine, and colorectal cancers, and lymphoma) to improve their QOL. 32 A meta-analysis revealed that telehealth interventions were effective and alternative methods for improving QOL among cancer survivors. 30 More evidence is needed to better introduce eHealth and other relevant telehealth interventions to benefit cancer patients.
To the best of our knowledge, this is the first metaanalysis to demonstrate the effect of a wide spectrum of eHealth patterns on overall QOL and specific QOL functional domains in women with breast cancer. In addition, this study considers the association between potential covariates (e.g., participant and intervention characteristics) and the effect sizes to verify the stability of the synthesized results. Therefore, our results give a comprehensive perspective on the effect of eHealth on women with breast cancer. This study also has some limitations. First, we only included articles published in English, so we might have   missed some pertinent studies. In addition, insufficient data were available to identify the associations between breast cancer stages and the effect of eHealth on QOL. Future meta-analyses could account for this diversity between studies to avoid ceiling and floor effects.

| CONCLUSIONS
eHealth offers a promising way to improve the overall QOL of women with breast cancer. Implications for clinical practice should be discussed based on subgroup analysis results. Further confirmation is needed for the effect of different eHealth intervention models on specific domains of QOL, which would help improve specific health issues of the target population.

ACKNOWLEDGMENTS
We acknowledge the contribution of all authors. We thank Mufan Shao at the University of Chicago for his valuable comments.

FUNDING INFORMATION Cancer Prevention and Control Special Fund from General
Electric Company, Cancer Foundation of China.