Leveraging Telehealth for the Management of Breast Cancer: A Systematic Review

Background: Breast cancer affects 2.3 million women and kills 685,000 globally, making it the most prevalent cancer. The telemedicine modality has been used to treat the symptoms associated with breast cancer recovery. Objectives: To analyze the effectiveness of telemedicine to help women recover from the treatment-associated effects and promote overall recovery from breast cancer. Methods: Four databases were queried for published literature from the last 10 years. The systematic literature review was conducted in accordance with the Kruse Protocol and reported in accordance with PRISMA 2020. Results: Five interventions were identified in the literature, with the most dominant being eHealth and mHealth. The other interventions were telephone, video teleconference, and a combination of eHealth and mHealth. There were positive effects of these telemedicine interventions in 88% of the studies analyzed. Telemedicine is shown to positively affect physical and mental health, sleep outcomes, quality of life, and body image. The largest barriers to the adoption of telemedicine interventions are training, cost, workflow, time of providers, and low reimbursement. Conclusion: Telemedicine offers promise to both providers and breast cancer survivors to improve the physical and mental health detriments of both cancer and its associated treatments. It also helps women develop healthy habits to reduce the risk of reoccurrence.


Rationale
Breast cancer is a disease, originating in the breast, in which breast cells grow out of control [1]. The incidence of breast cancer is extensive. In 2020, for example, over 2.3 million women were diagnosed with this condition, and this resulted in 685,000 deaths globally. The 5-year prevalence was estimated at 7.8 million women, which establishes it as the world's most prevalent cancer [2]. Breast cancer treatment is effective when caught early. Treatment often includes surgical removal, radiation therapy, and medication, but all of these treatments come at a physical and emotional cost to the survivor. Providers have sought new and innovative means to help women through the treatment process and the aftermath of the emotional devastation it brings. Telemedicine offers some interventions.
Telemedicine is defined as healing at a distance through the use of information and communications technologies (ICT) [3]. Telemedicine takes on many forms, but in general, it provides clinical support and overcomes geographical boundaries to improve health outcomes through ICT. Although many distinguish between telehealth and telemedicine, the World Health Organization does not distinguish between them, therefore, telehealth and telemedicine will be used interchangeably in this study. One form of telemedicine is mHealth and eHealth, or mobile-based health and computer-based health, respectively. These take the form of mobile apps, text messages through short message service (SMS), telephonic calls, websites, and computer programs. Many eHealth interventions can now be accessed on mobile devices, therefore the lines between the modalities have become blurred.
Several forms of telehealth have been used for the last several years in the area of oncology, and specifically breast cancer. mHealth apps have shown effectiveness in improving mood, symptom interference, self-efficacy, self-esteem, and emotional functioning [4]. mHealth apps provide education and improve health literacy [5,6]. They improve medication adherence and help women with coping strategies [7,8]. Overall, mHealth apps have shown positive effects on the perception of physical benefits, psychological factors such as motivation, social factors such as group practice, and organizational factors including preplanning physical activity [9]. The paucity of evidence for clinical efficacy begs additional research. This is the justification for this study.
In 2021, a systematic review was published examining mHealth interventions' ability to improve the quality of life for cancer patients. They identified 25 articles over a period of 10 years. They found the most common issues addressed by mHealth were physical activity, mindfulness, and stress management. Overall, mHealth had a positive effect on patients [10].
In 2022, a scoping review was published that examined mHealth's ability to increase screening rates among Hispanic communities. Ten articles were selected out of an original result of 597 from a search that spanned ten years. The reviewers reported mHealth was effective at providing education and increasing health literacy [6].

Objectives
The purpose of this review is to analyze the effectiveness of telehealth interventions to manage breast cancer care and recovery.

Eligibility Criteria
To be included in the group of articles for analysis, studies had to be published in the English language in the last 10 years in peer-reviewed, academic journals, and used human adult females as subjects. To avoid confounding results, other reviews were excluded. Systematic reviews summarize the findings of previous results (from a set number of years). Including a systematic review from 2022 in the analysis, for instance, would include results from articles also analyzed separately. This would double count instances of findings, which would confound the results.

Information Sources
Four databases were queried: The U.S. Library of Medicine's PubMed (MEDLINE), the Cumulative Index of Nursing and Allied Health Literature (CINAHL), Web of Science, and Embase's Science Direct. These databases were searched on 2 August 2022. We also performed a journal-specific search of Healthcare.

Search Strategy
We used the U.S. Library of Medicine's Medical Subject Headings (MeSH) to create a Boolean search string to combine key terms into an exhaustive search: (mHealth OR telemedicine OR "mobile apps") AND ("breast cancer" AND "treatment"). The same search string was used in all databases, and as much as possible, we used the same filters in each database. MEDLINE was excluded from all databases except PubMed since PubMed includes the MEDLINE database. This action helped eliminate duplicates.

Selection Process
Search results were filtered and abstracts were screened in accordance with the Kruse Protocol [11] and reported in accordance with PRISMA 2020 [12]. The Kruse Protocol was written to demonstrate the veracity of using the systematic literature review in higher education, but it outlines a proven methodology that has been published over 50 times in high-quality journals [11]. The PRISMA 2020 standard provides a systematic methodology to ensure standardized fields are reported for all systematic reviews and meta-analyses. Abstracts were screened by at least two reviewers.

Data Collection Process
An Excel spreadsheet, standardized in the Kruse Protocol, was utilized as a data extraction tool, collecting additional data at each step of the process. Three consensus meetings were held to identify articles for analysis, perform a narrative or thematic analysis, and perform additional analysis on the results to identify trends [11,13]. Abstracts were screened and studies were analyzed by at least two reviewers throughout the process.

Data Items
We collected the following fields of data: research database source, year of publication, authors, title of study, journal, study participants, experimental intervention, results compared to the control, medical outcomes, study design, study sample size, observations of bias, effect size (Cohen's d), sensitivity, specificity, and F1 (when reported), country of origin, statistics used, patient satisfaction, effectiveness, barriers to adoption, strength of evidence, and quality of evidence.

Study Risk of Bias Assessment
Each reviewer noted observations of bias (e.g., selection bias), and we assessed the quality of each study using the Johns Hopkins Nursing Evidence Based Practice tool (JHNEBP) [14]. These observations were recorded because they affect how to interpret the results, and because bias can limit external validity [15].

Effect Measures
Summary measures were not standardized because we accepted mixed methods and qualitative studies. Measures of effect were summarized in tables for those studies in which it was reported.

Synthesis Methods
Once data extraction was completed, a thematic analysis was performed to make sense of the data. [13] Themes were tabulated and summarized. Results across studies were analyzed for additional inferences and to identify heterogeneity.

Reporting Bias Assessment
We identified the strength and quality of evidence in accordance with the JHNEBP to provide us with an assessment of the applicability of the cumulative evidence and the limit of external validity.

Additional Analyses and Certainty Assessment
We performed a narrative/thematic analysis of the observations to convert them into themes, or common threads between articles. This helped us make sense of the data. We calculated the frequency of occurrence and reported them in affinity matrices. The frequency provided the probability of occurrence in the group of articles analyzed, and it provided confidence in the data analyzed.

Statistical Analysis
Measures of effect were collected during the data extraction process. Where possible, each effect was translated into an effect size equivalent to Cohen's d [16]. These measures were converted into a weighted average effect size by using the sample size for the weight. Figure 1 illustrates the study selection process with four databases. A kappa statistic was calculated to estimate the level of agreement between reviewers, (k = 0.92, almost perfect agreement) [17,18]. Results from four research databases presented 2021 results. Duplicates and those outside the date range were removed from screening. Using database filters, 1399 records were screened for full text, human subjects, English language, peerreviewed, and academic journals. Anything except peer-reviewed, published work was excluded along with other systematic literature reviews and meta-analyses. The remaining 68 records were assessed for eligibility. Protocols, editorials, and studies that would not address the objective statement were removed. The remaining group for analysis was 33.

Risk of Bias in and across Studies
Reviewers used the JHNEBP quality assessment tool to identify the strength and quality of evidence. Due to the strong methodologies chosen for review, the JHNEBP tool identified 76% of the articles as Strength I, which means the methodologies were experimental or RCTs (studies had control groups and used randomization). Only 2 studies were identified as Strength II, reserved for quasi-experimental studies. The rest were Strength III, which were a combination of non-experimental, qualitative, observational, prepost, or mixed methods. Additionally, the JHNEBP tool identified the quality of evidence based on sample size and consistency of evidence. Our group of articles chosen for analysis was 90% (30/33) Quality Q, and only 9% (3/33) were quality B.

Results of Individual Studies
Following the Kruse Protocol, reviewers independently extracted data and recorded observations about each study on a standardized Excel spreadsheet. As part of a thematic analysis, observations that occurred more than once were identified as themes [13]. These themes are tabulated in Table 2. Multiple observations of a similar nature are listed multiple times for studies, but an observation-to-theme match can be found in Appendices A and B. In 29/33 (88%) studies analyzed, an improvement in at least one area was noted. Additional observations collected in the data extraction step (sample size, bias, effect size, country of origin, statistics used, patient satisfaction, and the strength and quality of evidence from the JHNEBP tool) can be found in Appendix C. Effect sizes were only reported for 22 of the 33 studies (67%). The weighted average effect size was 0.21 (small).

Results of Syntheses, Additional Analysis and Certainty of Evidence
Thematic analysis was performed on all studies. Themes and additional observations were summarized into affinity matrices. Results are sorted by frequency. Frequency is reflected not to imply importance, but only to identify the probability a theme or observation was found in the group of studies analyzed. Table 3 summarizes the results of the studies compared with a control group. For non-experimental studies, the "no control group" leads the results. This is done to avoid confounding the results. facilitators observed. Thirteen themes and four individual observations were identified by the reviewers for a total of 111 occurrences in the literature. The theme most often observed was "improved mental health", which occurred 16/111 (14%) occurrences [19,23,34,36,39,40,46,49,50]. This theme combined observations of anxiety, distress, fear of reoccurrence, depression, optimism, self-efficacy, and self-actualization. Sleep outcome was the next most frequently identified theme. It occurred 12/111 (11%) of the occurrences [20,22,28,30,47]. This theme included the following observations: sleep disturbance, insomnia, sleep efficiency, cognitive function, fatigue, and cancer fatigue. The next theme is an improved quality of life, which appeared in 9/111 (8%) of the occurrences [22,27,28,33,35,37,39,44,45]. Two themes appeared in 7/111 (6%) of the occurrences: improved body image [22,31,38,43,45] and improved physical health [27,31,34,44,49]. The body image theme was comprised of the following observations: waist circumference, fat mass, and weight. Two themes were identified in 6/111 (5%) of the occurrences: less numbness, pain, or swelling [22,27,48], and no statistical differences between the intervention and control groups [23,24,26,33,38,41]. Next was less nausea or vomiting [27,34,44]. This occurred in 5/111 (5%) of the observations. Although nausea and vomiting are highly correlated, they are not synonymous, so reviewers chose to report them separately, but they appeared together in two studies. Two themes appeared in 3/111 (3%) of the occurrences: improved global health/return to baseline functioning [22,35,43] and improved social support, and questions were answered by providers [21,29,50]. Two themes occurred in 2/111 (2%) of the occurrences: improved arm symptoms/upper limb functionality [37,48], and the app provided education and answered questions [32,42]. There were four observations that could not be fit into themes: improved exercise, improved medication adherence, improved fasting plasma glucose, and the complexity of the tool (app) takes more time for users to process [25,26,45,46]. Table 3. Results of studies, compared to control group.
The findings of this systematic literature review are congruent with that of Buneviciene et al. [10]. The intervention of mHealth and eHealth addressed the quality of life of patients in the areas of physical activity, mindfulness, and stress management. This review found multiple instances of improvements in mental health, physical health, sleep outcomes, and quality of life. Our findings are also consistent with Watanabe et al., in that eHealth and mHealth augmented medical education and health literacy [6]. eHealth and mHealth offer several possible interventions that show promise as a treatment modality of care, however the clinical efficacy of this modality shows mixed results. The difference in results could be due to a difference of methodology or a difference of measurement. While older patients do not often prefer eHealth and mHealth interventions, many other patients do prefer this modality. Even when the results of using the eHealth and mHealth modalities of care show equivalent, but not statistically greater efficacy, offering the modality may meet the preference of the patient. These issues should be addressed in future research considerations.
Future research should examine the reasons for the lack of significant results in some of the studies. Standardization of methodology and measurement should yield consistent results. The results reported in this review were inconsistent. This systematic review focused on breast cancer. Future reviews should examine other types of cancer, then a review of reviews should be conducted for all cancer. The results did not seem to follow any particular intervention. This means it could have been a bias in the sample. Many examples of both sample bias and selection bias were observed, which affect the external and internal validity, respectively.
The results of this review should give practitioners confidence that telehealth can provide viable interventions to help their patients assuage the effects of breast cancer recovery and chemotherapy. The results from the studies analyzed in this review demonstrate healthy habits, less nausea, lost weight, more strength, and an increase in personal confidence. Policy makers should explore other reimbursement mechanisms to ensure the extra time and money these interventions require is reimbursed.

Limitations
No study is without its limitations, and this literature review is no different. Only four databases were queried over 10 years for published works. A broader scope of databases, years, and sources of literature, such as grey literature, may have identified additional interventions and results. However, the reviewers chose these databases due to their wide availability, 10 years because telemedicine is a rapidly growing field, and published literature to ensure a peer review. Within the studies analyzed were multiple examples of selection and sample bias, which affect the internal and external validity, respectively.

Conclusions
Telehealth offers promise to help breast cancer survivors cope with the side effects of treatment, the mental anguish that shakes confidence, and the physical ailments that accompany chemotherapy. Several exercise applications show promise educating and helping survivors establish healthy habits to lower the risk of reoccurrence. The most significant barrier is training followed by cost, but these are not significant barriers to overcome.

Informed Consent Statement:
This review is conducted in accordance with the Kruse Protocol for writing a systematic review. It is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA). This review has been registered with PROSPERO: CRD42022347417.
Data Availability Statement: Data from this study can be obtained by asking the lead author.

Conflicts of Interest:
The authors declare no conflict of interest.