Use of the internet has increased considerably since the early 1990s. The World Bank reports that almost 44% of people across the globe used the internet in 2015, compared with 0.25% in 1993 [1]. This number is expected to increase to over 50% by 2019 [2]. Nearly two-thirds of internet users are estimated to access health information on the Web [3]. With such demand for accessible health information, electronic health (eHealth) has become a popular way to provide patients with health information, recommendations to self-manage their health, and access to their health records and data [3,4]. eHealth is defined as “an overarching term used today to describe the application of information and communications technologies in the health sector. It encompasses a whole range of purposes from purely administrative through to health care delivery” [5]. eHealth tools, therefore, are technologies that may include electronic medical records (EMRs), personal health records (PHRs), mobile apps, patient portals, information repositories, and many other internet-based programs or software used to help patients monitor and manage their health. eHealth tools may help decrease fragmentation of care by compiling patient health information from multiple providers into 1 easily accessible location [6], while also streamlining patient-provider communication and promoting shared decision making [3,4].

Well-functioning eHealth tools can help patients better understand their health [7] and may lead to improvements in patient-physician relationships [8]. eHealth tools can encourage patients to play a larger role in shared decision making and might increase focus on self-management and preventative care [8,9]. As technology advances, the use of eHealth tools can provide a level of convenience for both patients and providers [10]. These tools can generally be accessed from any internet-capable device and often provide a method of asynchronous communication such as emails and short message service (SMS) text messaging. These methods allow patients and providers to ask and answer questions at their convenience, creating less of a burden on physician workflow [8].

The ability of patients to use eHealth tools to better manage medication by reporting feedback on symptoms and use of medications directly to health care providers has not been comprehensively explored in the literature. Similarly, there is little evidence showing how eHealth tools might provide prompts to patients and health care providers to make appropriate changes to medication use based on this feedback. A synthesis of this literature will provide greater understanding of what eHealth tool design features may be helpful in patient reporting of medication-related experiences and outcomes.

The objective of this systematic review was to determine the impact of patients’ use of eHealth tools on self-reporting adverse effects and symptoms that promote changes to medication use. The PICO model was used to focus the objective of the review, as seen in Figure 1.

This systematic review was performed following steps outlined by Cochrane’s Effective Practice and Organization of Care group and reported based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement [11]. A total of 3 biomedical and health science databases were searched: MEDLINE/Ovid, EMBASE/Ovid, and CINAHL. References of all included articles were also searched. All 3 databases were searched from January 1, 2000, to April 25, 2018. The search was limited to articles published in English using terms representing eHealth (eg, Web-based applications), symptoms and adverse drug reaction reporting (eg, drug-related adverse effects and adverse reactions), and patient self-monitoring (eg, self-management; see Multimedia Appendix 1 for full search strategy and Multimedia Appendix 2 [12-17] for definitions for terminology used). The search date began from 2000, which generally marks the start of scientific reporting of eHealth interventions that would have relevance to the current use of eHealth tools. As a result of the aforementioned search strategy, studies were included in this review if they determined the effectiveness and impact of changes to medication regimens as a result of using eHealth tools. As such, this review investigated these effects using a comparative quantitative methodological approach.

For the purposes of this review, an eHealth tool was considered to be any internet-based intervention, including mobile health apps, used by patients for clinical purposes that focused on improving patient health and clinical outcomes. The term PHR refers to an eHealth tool wherein a patient has access to and can enter or edit their own health data. The population investigated was community-dwelling individuals of any age in an outpatient setting.

For a study to be included, the eHealth tool must have allowed patients (or caregivers) to enter information directly (as opposed to information being entered by a health care provider); included self-reporting functionalities focusing on medication monitoring, contain a medication monitoring or use component, or specifically incorporating the option for the patient or caregiver to enter symptoms including adverse effects; and needed to focus specifically on medication use, clinical outcomes, or symptom reporting following use of the eHealth tool. Any eHealth tools involving changes in medication reconciliation and recommendations made to changes in drug therapy were also included.

Exclusion criteria were conference abstracts; qualitative studies; articles without a comparator group; articles that did not report on at least one medication-related outcome; articles where self-management strategies focused on lifestyle modification, behavioral interventions, or nondrug interventions; articles focused solely on the validation of an eHealth tool; articles focused on methodological or technical aspects of eHealth interventions; articles containing nonempirical information; articles that synthesized information about multiple eHealth tools in an article (ie, review articles); and eHealth tools used by regulatory agencies to report adverse drug events (ADEs).

All potentially relevant articles were uploaded into DistillerSR software, which was used throughout the selection process. Potentially relevant articles underwent title, abstract, and full-text review. Articles that met inclusion criteria proceeded to data abstraction and risk of bias assessment. Articles not meeting inclusion criteria were excluded at both levels. Figure 2 represents the flow of articles through the selection process.

Title and abstract review were performed independently by 2 reviewers from a pool of 5 reviewers. Of these, 1 reviewer went through the reference lists of all the articles included in this study. Another reviewer went through reference lists of relevant systematic reviews identified during the literature search. Potentially relevant articles were identified. These articles went through abstract review by 2 reviewers. Studies found not to fit inclusion criteria after abstract review were excluded. Full-text review was performed independently by 5 reviewers. The kappa scores were calculated to determine agreement among reviewers who conducted review of titles and abstracts. All kappa scores calculated were greater than .93. Conflicts were resolved by consensus.

Data extraction and risk of bias assessment were performed for each study independently by 2 reviewers. Data extracted included study design and setting, participant demographics, number of participants in each group, intervention components, comparator group components, eHealth tool functionality measured, and results and significance levels for each outcome measure. Conflicts in data extraction were resolved by consensus. Risk of bias assessment used questions recommended by the Agency for Healthcare Research and Quality’s 2014 publication Methods Guide for Effectiveness and Comparative Effectiveness Reviews [18] and was performed for each study independently by 2 reviewers. All conflicts were resolved by consensus. The risk of bias assessment questions are presented in Multimedia Appendix 3.

The primary and secondary outcomes are listed in Figure 1. The primary outcome was the number of changes made to medication regimens following patient self-reporting via an eHealth tool. The included studies varied considerably in populations, eHealth tool functionality, outcomes measured, and study design. Due to high heterogeneity, meta-analysis of outcomes was not feasible. Therefore, results for each outcome were synthesized descriptively and presented as narrative. Available data on barriers to implementation were extracted from the article text and summarized qualitatively so as to heighten awareness of implementation issues.

Subgroup analyses were performed to investigate differences in treatment effect present because of (1) age of participants, (2) patients with specific conditions targeted by intervention, and (3) different features and functionalities of the included eHealth tools.

A total of 3515 articles were generated from database and reference searching, resulting in 2489 potential articles that were screened based on their titles and abstracts, after duplicates were removed. Furthermore, 75 full-text articles were assessed for eligibility, of which 14 were included in this systematic review (see Figure 2 for more details).

Of the included articles, 13 were randomized controlled trials (RCTs) [19-31] and 1 was an open-label intervention [32]. A total of 10 studies were conducted in the United States [20-29], and 1 study was conducted in each of South Korea [19], Canada [30], Finland [31], and Denmark [32]. Dates of publication ranged from 2006 to 2017. The majority were published in 2007 or later (n=13). This distribution mirrors the increase in both internet and eHealth tool usage beginning in the late 2000s [1,33]. Further details on the characteristics of these studies can be seen in Table 1. Details regarding the design and outcomes of included studies are presented in Multimedia Appendix 4.

Of the 13 RCTs included in this review, 2 studies were cluster RCTs [22,26]. The remaining 11 RCTs used participants as the unit of randomization [19-21,23-25,27-31].

A total of 4 RCTs focused on pediatric and adolescent asthma patients [21,23-25]; 1 study focused on adult asthma patients [30]. Moreover, 1 trial included only elderly patients (aged 65 years and older) and focused on medication self-management and safety [20]. The remaining 7 trials all included both adult and elderly participants. From these studies, 1 focused on patients with depression [27]; 3 on patients with type 2 diabetes [19,22,31]; 2 on medication safety and use, including identification of ADEs [26,28] using eHealth tools; and 1 on identifying adverse effects in patients receiving chemotherapy [29].

In 3 studies, use of an eHealth tool in the intervention group was compared with usual care plus links to relevant websites [24,25,28]. Simon et al [27] compared their Web-based messaging eHealth tool for depression and Web-based messaging system to usual care with Web-based messaging between patients and health care providers. Gustafson et al [23] used nearly identical interventions in both groups; the control group was restricted from accessing the eHealth tool but participated in other aspects of the intervention (clinical visits, interviews, patient education, etc). Fiks et al [21] also used a usual care group with no access to the intervention Web portal; however, all health care providers had access to a computerized decision support system. Moreover, 2 studies [22,26] utilized a double-dummy style intervention, where both groups used Web-based PHRs to record information that differed only in content. Cho et al [19] compared an electronic blood glucose (BG) monitoring system with an informal paper-based monitoring system, with both groups receiving diabetes education and regular clinical visits. Chrischilles et al [20] utilized a conventional usual care group without supplementary information or resources. Mooney et al [29] used a self-monitoring tool to manage chemotherapy symptoms. Ahmed et al [30] developed an asthma portal to view patient’s personal health information, monitor patients, and provide feedback on self-management strategies. Karhula et al [31] used a management system for patient self-monitoring of diabetes. Only 1 study was identified as an open-label intervention study, which included a comparator group [32]. Carlsen et al [32] used an eHealth tool to monitor responses of patients with inflammatory bowel disease to determine the need to adjust treatment interval or dose.

Figure 3 displays a summary of the risk of bias assessment. The studies, overall, were of moderate quality, with studies ranging from poor to good. Common issues included small numbers of participants, lack of blinding, poor description of interventions, and contamination of intervention. Many trials relied only on patient self-reported data (as would be expected based on the topic), which can introduce bias if methods to ensure validity and reliability are not demonstrated.

From the 14 included studies, 11 unique eHealth tools were described. The 2 RCTs by Joseph et al [24,25] utilized the same asthma management eHealth tool. The cluster RCT by Schnipper et al [26] used a Web-based PHR to record information that differed only in content, which was nested within the larger RCT by Grant et al [22]. Two studies by Fiks and Ahmed used a Web-based portal for asthma symptom management [21,30]. Each study and eHealth tool is described in Multimedia Appendix 4. Features and functionalities of the eHealth tools are also presented in Table 2. Although Schnipper [26] and Grant [22] use the same eHealth tool, Schnipper’s study [26] investigates a specific medication module. Thus, they have been counted separately here.

All 11 eHealth tools from all 14 studies included a component where patients could self-report medication management information or changes, including symptoms, health data, adverse effects, or ADEs. A total of 12 studies included Web-based patient questionnaires or surveys [20,21,23-32]. Many studies used validated patient questionnaires but several developed their own. A list of patient questionnaires utilized by each study is presented in Table 3. In addition, 10 eHealth tools included patient educational resources [20,21,23-26,28-31]. Details of these resources are also listed in Table 3. Cho et al [19] measured outcomes by having a patient record their BG readings in the eHealth tool. Patients were also interviewed in person by clinicians every 3 months. Schnipper et al [26] had participants complete medication electronic journals, where they would note discrepancies between a Web-based medication list and their actual medications as taken. Karhula et al [31] used a 36-question survey for patients to report their health-related quality of life (HRQoL) score at baseline and post intervention.

The results of each study by outcome can be seen in Multimedia Appendix 5.

A total of 4 RCTs investigated the use of eHealth tools in children and teens with asthma [21,23-25]. There is evidence that eHealth tools may have the potential to reduce symptoms of asthma, frequency of asthma flare-ups, and number of days of school or work missed because of asthma [21,23-25]. They may also promote better asthma control, availability and use of rescue inhalers, and may have the potential to improve asthma symptoms in vulnerable groups (ie, African-American adolescents living in urban centers) [24,25].

Subgroup analysis also found that multifaceted interventions combining use of eHealth tools with clinician support or case management and eHealth tools utilizing direct patient-provider communication might be more effective at improving some aspects of patient self-management and self-efficacy [23,26,30,31]. Both studies utilizing multifaceted interventions and direct patient-provider communication that measured these outcomes found positive significant results [23,26], whereas both studies using only eHealth tools with no patient-provider communication found no significant differences [20,21]. Detailed results from the subgroup analyses can be found in Multimedia Appendix 4.

Evidence from 4 RCTs and 1 open-label intervention [20,22,24,27,32] show that eHealth tools focusing on symptom and adverse effect self-reporting can prompt positive changes in medication prescribing and use. In addition, the majority of eHealth tools studied were able to improve patient symptoms, regardless of functionalities, complexity, and differences in intervention. Those eHealth tools were particularly studied and found to be beneficial for improving signs and symptoms in children and adolescents with asthma [21,23-25]. This review supports the inclusion of patient entry or editing of symptoms into eHealth tools for the purposes of monitoring and reporting outcomes from the use of medications.

Evidence was found that eHealth tools improved the outcome of patient self-management and self-efficacy. Subgroup analysis found that eHealth tools that allow patients and clinicians to communicate directly, and multifaceted interventions combining eHealth tools with clinician support and case management might lead to greater increases in patient self-management and self-efficacy. It is notable that more significant improvements were found for more objective outcome measures, such as number of medication changes and clinical signs, and less were found for more subjective outcome measures such as self-management and self-efficacy. It may also be that sample sizes were too small to detect differences, particularly if this was not the primary objective for these studies. It is likely that the eHealth tools under investigation either did not provide effective content or functionalities to help participants improve self-management and medication management in participants or the tools used to measure these outcomes were not able to detect any differences between groups. Another possibility is the lack of patient understanding of chronic disease and poor perception of health goals.

It was thought that eHealth tools that focus on improvement of patient self-efficacy and self-management might lead to improved medication-use behavior, which in turn may lead to changes in medication use, identification of real or potential ADEs, improvement in signs and symptoms, and overall improvement in HRQoL. However, there is not enough evidence to draw conclusions as to the effectiveness of eHealth tools for identification of adverse effects, improving medication-use behavior, increasing recommendations to medication therapies and improving medication reconciliation, improving health service utilization, and improving overall health status and patient satisfaction. Only a small number of included studies investigated these outcomes; it is likely that with such a small overall sample size, it was not possible to find differences between groups.

As with most systematic reviews on the subject of eHealth tools [7,8,33,54-57], this review found at best moderate evidence that patient reporting via eHealth tools can lead to improved clinical outcomes such as symptom reduction.

A 2012 systematic review by Ammenwerth et al found that use of patient portals linked to a PHR led to significant increases in medication adjustments in diabetic patients [55]. Other reviews and primary articles have also indicated that the use of eHealth tools may be more effective in specific patient populations such as patients with cancer [4,29,34,58]. This review found evidence that use of eHealth tools might increase the number of medication adjustments in diabetic patients [19,22]. In addition, patient-reported symptoms and adverse effects were used to identify toxicities in cancer patients, and in several instances, it lead to medication changes. It was also found that eHealth tools might improve signs and symptoms of asthma in children and teens [21,23-25].

Overall, this review supports findings by Ammenwerth et al [55] that interventions may be more effective at improving health outcomes if they combine eHealth tool features such as patient-provider communication and interactive coaching with eHealth tool use (see, eg, [21,22,26,28] as well as Table 2 and Multimedia Appendix 4 considering subgroup analyses). Evidence from this review indicates that eHealth tools in combination with clinician support or case management, and eHealth tools that encourage provider-patient communication may improve patient self-management and self-efficacy when compared with tools without these features [22,23,26].

To our knowledge, this is the first systematic review of eHealth interventions focusing on patient self-reporting of symptoms and adverse effects. The review’s search strategy was augmented by reference searching. This review was limited to studies that included medication-related outcomes. The majority of studies included in this review were RCTs, most of which were of moderate quality.

This review also has a number of weaknesses. It was limited to studies published in English, which may have excluded relevant articles. No searching of grey literature was performed, as we focused on empirical work published in academic journals, and so it is possible that some early non–peer-reviewed reports may have been missed. We acknowledge the lack of definitional clarity surrounding the term eHealth and believe future research should focus on establishing better consensus for this term. There was considerable variety among the interventions in the studies, some of which included features such as direct health care provider follow-up, thus making it more complicated to determine which outcomes could be specifically attributable to using an eHealth tool. As this review also examined different populations of varying sample sizes and medical conditions for eHealth tools, it may be difficult to detect differences and generalize findings and conclusions. We did not include qualitative studies in our review because the goal of our study was to better understand the effectiveness and impact of changes to medication regimens based on quantifiable differences in using eHealth tools for self-reporting adverse effects and symptoms that promote changes to medication use versus a comparator. We value the insight of qualitative studies that have been investigated elsewhere [59,60]. Additional exploration of qualitative literature to better understand how use of these types of eHealth tools can generate impacts on medication use and health would be helpful.

Where possible, health care providers should encourage patient use of eHealth tools for symptom and adverse effect self-reporting. eHealth tools may be especially useful for reducing symptoms in certain populations, for example, children and teenagers with asthma. eHealth tools might also encourage patients to improve self-management behaviors and participate in shared decision making with clinicians. Having information from the EMR entered directly into the eHealth tool may reduce the burden on the patient to routinely update their clinical information (something that only highly motivated patients are likely to do regularly) [6]. Clinicians should be encouraged to communicate with patients via eHealth tools where possible, especially where patients are experiencing worsening of symptoms or medication-related adverse effects. Evidence suggests that using technologies such as mobile apps and SMS text messaging may improve patient engagement by allowing quick, convenient communication without a computer or internet connection [23,27]. Clinicians should be supported in their eHealth tool use, and interventions should focus on clinician training and engagement. Ensuring that interventions can be successfully incorporated into physician workflow is important [22,26].

There is a paucity of primary research articles investigating eHealth tools and their impact on medication use. Studies are generally small and of moderate quality. Large-scale RCTs focusing on the use of eHealth tools for medication and symptom management should be undertaken to establish more high-quality evidence. This is especially important given how ubiquitous the use of medication is. Furthermore, the effects of patient self-management and self-efficacy on medication use and symptom experience are not well studied; more research in this area could help drive creation of medication-focused eHealth tools. Low patient engagement and eHealth tool utilization were commonly noted implementation barriers; it could be that patients were not engaged in eHealth tool use enough for them to feel an impact on their satisfaction with health care or overall quality of life. Descriptive evidence shows low proportions of patients felt that eHealth tools improved their care or communication with providers, indicating that development of eHealth tools should focus on functionalities and outcomes that are important to the patient. This may be achieved by utilizing research on patient motivation and behavior change to increase patient engagement [20,24,25].

The results of this review show initial and promising findings that specialized eHealth tools can be used for reporting and monitoring of symptoms and medication-related adverse effects and some evidence that use of eHealth tools have the potential to identify instances where changes in medication use may be appropriate. A modest amount of mixed evidence was found, demonstrating that eHealth tools can improve patient self-management and self-efficacy. Very little or no evidence was found to demonstrate that use of eHealth tools could increase numbers of medication recommendations or improve medication-taking behavior, health services utilization, identification of adverse effects, overall health status, and patient satisfaction. eHealth tools may be more effective at promoting medication changes and improving patient self-management and self-efficacy if they provide mechanisms for direct patient-provider communication and may be more effective in certain populations such as children and teenagers with asthma.