The problems and barriers identified in the included studies were categorized into 10 major groups. The included studies usually addressed several different problems and barriers. Multimedia Appendix 2 provides an overview of the categories of problems and barriers that were found in each article. The 10 groups included “validity,” “usability,” “technology,” “use and adherence,” “data privacy and data security,” “patient-physician relationship,” “knowledge and skills,” “individuality,” “implementation,” and “costs” (Figure 2).

Problems with validity were addressed in 16 articles [16,18,21,23-30,33,35,39,40,44]. Of these, 3 described quantitative studies [24,27,35], 4 described qualitative studies [16,18,21,39], 6 were mixed-design studies [25,26,30,33,40,44], and 3 were reviews [23,28,29]. Problems addressed in these studies were mainly in the area of contents, outcomes, and user input.

Problems in validity concerning the contents and outcomes of some mHealth apps were due to a lack of accordance with clinical standards. A fundamental problem was found in missing empirical evidence [24,27,29].

Some content was declared as inappropriate [18], wrong [24], or ambiguous [24,33,40]. Patients and health care staff stated that the depth and quality of information was often not suitable [21,39]. One quantitative study found that a few apps provided details about the underlying formulas used for calculations [24]. In 2 studies, some users criticized that functions did not meet their needs [35,40].

The lack of added value was perceived or assumed in qualitative and quantitative studies [21,25-27,44]. Some studies described mHealth solutions as inferior to usual care [21,23,29,30] or mentioned adverse effects or even harm [16,21,27,28,39]. One mixed-design study found that physiotherapists were skeptical about whether a hybrid setting including an mHealth app could be conducive to building and maintaining a robust working alliance between patients and physiotherapists [25].

In addition to problems with app content, an app assessment study found problems with user inputs and their validation [24]. In another study, patients described problems with changing the entered values [39]. Only a few apps provided guidance based on user-entered data. This was especially important in mental health crises or the risk of suicide [27]. Some apps caused inappropriate alerts after incorrect data entry in the settings component [44]. Patients manipulated the generated results by deliberately entering incorrect values to receive better feedback from the app [39]. Therefore, the medical benefits of the app could be reduced. Apps including physical exercises also faced the problem of validation. Lack of feedback on the correctness of exercise executions led to a feeling of insecurity and incorrect execution of exercises [25].

Health care professionals described app measurements, calculations, and the resulting data provided to the user as imprecise and inaccurate [16]. App assessment and interviews with patients revealed that incorrect results, despite correct input values, were a problem [24,39]. In 1 study, health care providers criticized that mHealth apps were not able to consider each aspect related to the subjects separately [33].

In total, 19 articles described the problems and barriers related to the usability of mHealth apps [16,18-20,22,25,26,29,30,33,35,37-44]. Two quantitative studies [22,35], 5 qualitative studies [16,18,37,39,42], 11 mixed-design studies [19,20,25,26,30,33,38,40,41,43,44], and 1 semistructured review [29] addressed usability as a problem. On the one hand, usability issues related to appearance or contents were seen in the software (the apps themselves), and on the other hand, problems were found in the hardware (executing devices). Some articles cited usability as an unspecified problem, among others [16,20,44].

Both patients and health care professionals described the design of mHealth apps as problematic [16,18,25,26,33,37,39,43,44]. Some apps included nonintuitive navigation or interfaces [26,29,37,38,44] or were generally difficult to use [18,19,22,30,35,41,42]. Three studies found that an insufficient or no user manual was provided [30,39,43]. Further issues with usability were observed in bulkiness or size of devices [19,37,40], visual unattractiveness, and discomfort [20].

Technical problems were mentioned in 18 articles [16,18-20,22,24-27,29,33,37-40,42-44] and were found in both devices and software. Three quantitative studies [22,24,27], 5 qualitative studies [16,18,37,39,42], 9 mixed-design studies [19,20,25,26,33,38,40,43,44], and 1 review [29] described problems related to technology.

A case study with an older user found a dependence on technological support [37]. Patients faced issues related to hardware, such as lack of free storage space on the smartphone, short battery life [20], and use of small devices with small screens, such as smartphones [44]. Some patients still use feature phones with limited functions compared with smartphones and accordingly cannot use apps [22].

Software-related issues were seen in functionality [27,43], challenges with software updates, and technical issues with operating systems [44]. Patients reported bugs, glitches, or intermittent screen freezes in 2 apps [16,44]. Technological failures may lead to physician-induced errors [33]. In addition to technological problems, it was stated that detecting these potential issues before app distribution is a challenge [24].

A further issue was observed in terms of compatibility. Problems for patients could result from incompatibility or difficulties between apps and running devices [18,20,26,38], between running and external devices [19,20], and connection to a server [40] or electronic health records [42].

Some articles mentioned technical problems that were not described further [19,25,29,37,39,44]. One mixed-design study referred to technical difficulties with access but did not provide further details [38].

Problems in the category of use and adherence were found in 4 quantitative studies [22,24,31,35], 6 qualitative studies [16,18,21,37,39,42], 9 mixed-design studies [19,20,25,26,30,38,40,43,44], and 3 reviews [17,28,29]. Qualitative research with patients as well as health care professionals and 1 review found a common problem of lack of adherence [16,29,39,42]. Many studies have reported high dropout rates [18-20,22,26,28-31,35,38,40]. In 2 studies, patients expressed minimal or no interest in the use of mHealth apps [42,44]. Problems cited regarding use and adherence included lack of motivation, additional burden on patients, social or disease context, lack of time and integration of the app into daily life, and lack of personalized elements.

Patients and health care professionals perceived lack of motivation [18,21] and lack of engagement in users [16,19,25,40,43] as reasons for low adherence and a high number of dropouts. In 1 study, users expressed that they had forgotten to use the app [20].

Some patients perceived the app as an additional burden or found it overwhelming [19,42,44]. Some participants did not update their goals to conform to expert recommendations and maintain achievable goals [19]. Otherwise, patients expressed concerns about being judged if they did not complete or missed lessons [37]. Two qualitative studies revealed that some patients did not use the app based on the given advice [39,42].

Some social situations [19,20,24,25,37] or disease-specific contexts [37] were reported as being problematic. Environmental influences [19,26] and special use cases [19] were further issues. One review stated that distraction by other web-based activities could be a problem [29].

Patients pointed out that the lack of a pause option [26] and difficulties with integration in everyday life can be a major barrier to adoption [44]. A further risk was a possible interference of technology use on relationships [40].

Lack of time was a major factor that was decreasing use [20,31,42,44]. Repetitive, long, complicated, and boring content reinforced the problem and might lead to even less time spent on mHealth apps [19,21,25,26,29,38].

Qualitative studies and reviews have shown that the lack of human factor also affected use and adherence. In other words, mHealth apps lacked personal touch, empathy, and complex aspects of human interaction [17,21]. Communication was sometimes seen as ineffective [29]. Some people will not use mHealth apps and reject them because they see their recovery as a process that depends only on the health care professionals caring for them [16].

Data security and data privacy were addressed in 12 publications [16,18,21,27-29,33-35,40-42]. Thereof, 2 studies had a quantitative design [27,35], 5 studies had a qualitative design [16,18,21,34,42], 3 studies were mixed-design studies [33,40,41], and 2 publications included reviews [28,29]. There are mainly 2 groups of problems related to generated or collected data.

On the one hand, there are data security and data privacy problems [16,18,21,27,29,33-35,41,42]. On the other hand, there are problems with nontransparent communication of data privacy policies [21,27,28,40].

First, concerns regarding weak security arise in the context of data security and privacy [34,35]. Thus, a study of patients who are chronically ill found that 37.2% of them reported being concerned about the disclosure of personal information [35]. Further problems regarding data security and privacy were access without permission [16] and possible breaches of and concerns about confidentiality [16,27,33,39,42]. One app assessment study emphasized in its discussion that when data are stored on provider servers, there is an increased risk that the data will be used or sold for undesirable purposes [27]. Finally, health care practitioners mentioned problems regarding the identification of individuals by unauthorized data access [21] and patients worried about the unpredictable consequences of data leaks [34].

The other problem was the transparency of data handling [40]. Although some apps did not provide privacy policies [27,28], others were not clear or difficult to understand [21,27].

In this context, the patient-physician relationship refers to all types of relationships between patients and health care providers. Thus, it includes therapeutic relationships as well. Problems in this category were described in 13 articles [16,21,23,25,27,29-32,34,38,40,44] and included “the attempt to replace the clinician,” “lack of a therapeutic alliance,” “negative impact on the relationship,” “information inequalities,” and “the question of responsibility.” This category was addressed in 2 quantitative studies [27,31], 4 qualitative studies [16,21,32,34], 5 mixed-design studies [25,30,38,40,44], and 2 reviews [23,29].

Both patients and health care professionals mentioned problems with the lack of face-to-face contact. Both assumed preferences for face-to-face communication for some patients and providers [21,34]. Health care professionals have particularly emphasized the lack of nonverbal communication and para-communication associated with face-to-face conversations [21]. Nevertheless, the spectrum of physician replacement ranges from taking over individual decisions [40] to complete replacement [21,23,44]. One review described substantially lower treatment effects due to the substitution of face-to-face intervention [29].

The absence of a treating person resulted in a lack of therapeutic space [38] and therapeutic alliance [21,25] considered vital for successful therapeutic care [21]. Without human support, 1 study found difficulties with user engagement in active components [31].

Problems with the “patient-physician relationship” also occurred when mHealth apps were integrated into the treatment process. Even if technology can assist in health care, concerns regarding interference with relationships are pronounced by patients [34] and health care providers [21]. Physiotherapists saw that mutual trust could suffer from continuously monitoring a patient [25].

The limited capacity to export or download data reports reduces the ability to communicate directly from the app with others [27]. Information asymmetries can arise, and specialists could end up in situations in which patients receive treatment results before they do [32].

Finally, the responsibilities are altered and might lead to new problems. Physicians expressed concern that they have to handle additional data or alerts and that the use of mHealth apps could lead to the detraction of the patient’s self-management [16]. Physicians who did not engage as leaders in digital interventions were also seen as problematic in 1 study [30].

Problems and barriers related to knowledge and skills were identified in 4 quantitative studies [22,24,27,35], 7 qualitative studies [16,18,21,34,37,39,42], 4 mixed-design studies [19,30,33,41], and 3 reviews [17,23,29]. In some studies, it was found that patients had limited digital literacy [41] or abilities and experience regarding the use of mHealth apps [16,18,19,22,27,33,37]. In other cases, wrong perception [21,22,34,35] and a lack of knowledge [19,21,24,33,41] were seen as problematic.

Little, bad, or no experience with apps is seen as a major problem [18,19,42] and fosters the issue of low abilities and confidence with technology use [16,19]. Although young individuals showed few difficulties in app use, older patients, especially those with conditions such as dementia [27] or declining cognitive functions [37], face difficulties in app use [19]. A special problem is that older adult users have more problems because they use their mobile phones for known functions and discourage themselves from learning new technology through trial and error [22].

A major barrier for app use is found in its perception. Irrespective of individual apps, some patients believe that mobile phones [34] and apps [22,35] are complicated and difficult to use. Patients might also feel dismissed because they see inferior care in digital products compared with face-to-face contact [21].

In addition to the abovementioned problems in digital literacy [41], literacy in general and numeracy were found to be a barrier for app use [21,24]. In 2 studies, participants did not understand specific app functions [19,33].

Clinicians experienced similar problems as patients. Low experience and skills were frequently observed [16,19,21,23,29,30,39]. Some clinicians even had a more negative attitude toward this type of intervention than patients [17]. Others expressed a lack of confidence in the integration of technology in health care [21].

A further problem mentioned in 15 articles [16,17,19,21,25,26,29,33,35-37,39,40,43,44] is the intention or capability to customize mHealth apps to the individual needs of patients. One quantitative study [35], 4 qualitative studies [16,21,37,39], 8 mixed-design studies [19,25,26,33,36,40,43,44], and 2 reviews [17,29] included problems and barriers corresponding to individuality. This is, for example, expressed in the fact that mHealth apps are usually not adapted to each individual [17,19,33,35-37,40,44]. Thus, the authors discussed the difficulty in designing attractive and useful programs for all patients, which are at least as effective as standard therapy [17,36], and pronounced the difficulties due to the diversity of the target users, especially in terms of age [33,35,37] and diseases [35]. Furthermore, the authors described individualization of functions due to perceptual impairments [37] and motor or physical issues [44] to be problematic. In 1 study, patients indicated that the goals set by the app were too simple and that the app could not be customized to their needs as much as necessary [40].

Different functions are affected by a lack of individualization. Patients and health care professionals expressed that exercise programs often consist of a fixed number of different standard exercises [16,25,26], that data input is limited to imprecise standardized possibilities [39], and that communication provided by the mHealth app is unadjusted [21,43,44]. A special problem is the “cold start problem.” It describes the need for time at the beginning of the intervention to personalize the app content to the user profile through artificial intelligence [26].

The implementation of mHealth apps in health care systems faces different problems. Problems related to implementation were found in 4 quantitative studies [22,24,31,35], 4 qualitative studies [16,21,39,42], 5 mixed-design studies [30,38,40,41,44], and 4 reviews [17,23,28,29].

Barriers to access were seen as a problem for implementation. These occurred because of a lack of infrastructure, socioeconomic conditions, or social reasons. Lack of access (eg, lack of smartphones or broadband and computers) is a fundamental barrier for the use of mHealth apps [22,35,38,42]. Disparities in access subsequently foster concerns that only a fraction of users benefit from apps in health care [21,22,29]. Issues in the context of equity may stem from income or disability and result in nonequally distributed devices and connectivity [28]. Further barriers to access concerned stigma and culture [31] as well as language [28,31]. However, no further information was given on these barriers.

Further problems concerned transferability of study effects to real-world care and organizational barriers, such as lack of capacity or preparedness of health care systems and reimbursement structures. Successful transfer into clinical practice was seen as a problem [17,23,41]. Many questions, for example, regarding modes of action or for which target groups app-based therapy is most suitable, are still unanswered. Thus, mHealth programs showing effectiveness in experimental settings do not necessarily show good results in real health care settings [17]. Staff members reported low expectations and low confidence in the ability of national health care systems to implement digital tools [21]. Barriers for implementation were lack of health system readiness, organizational resistance to change, and policy uncertainties [44]. Approval of apps, for example, by the US Food and Drug Administration, focuses on safety and minimal effectiveness thresholds and does not provide sufficient information for decision makers [28]. Reimbursement options are not uniform [28], and a lack of collaboration among stakeholders, such as developers, health care professionals, and patients, in the design and development process affects acceptance and adoption [16].

Low acceptance is a 2-fold problem. On the one hand, some professionals have less interest in information in apps than in paper-based information. This was highlighted by patients commenting that care providers always asked for paper forms despite information being provided in app format [39]. On the other hand, professionals need to be open to the use of mHealth apps because their strong leadership engagement and promotion are fundamental for mHealth use [30,35].

Health care professionals see the use of digital solutions as an additional burden placed on them [21,40,42] and express fear regarding the complexity of, and the responsibility for, identifying and managing risk [21]. Interacting with mHealth apps was frequently seen as obstructive for workflows [21,28,42].

Three further problems concerning the implementation were as follows: first, some app manufacturers were not available and did not respond to requests [24,42]; second, some users expressed the amount of choice being overwhelming [16]; and finally, frequent app updates, requiring evaluation of new and confirmation of old functions, were potential problems [24].

Costs were mentioned in 10 articles [18,21,27-29,31,34,35,42,44]. Of these, 3 had a quantitative design [27,31,35], 4 were qualitative studies [18,21,34,42], 1 had a mixed-design [44], and 2 were reviews [28,29]. On the patient’s side, the use of mHealth apps might be problematic because it always requires a running device [21] and often requires data transmission [34,35]. One study mentioned potential costs for apps as a concern [18]. Furthermore, a problem for patients is the lack of opportunity to test and evaluate apps before they are purchased [27]. These costs usually have to be borne by patients and might lead to socioeconomic inequalities [21]. Another problem for patients is the lack of opportunity to test and evaluate apps before purchasing them [27].

As with traditional health services, health care practitioners need time to integrate mHealth apps into their treatment. However, this effort is often not reimbursed [28,29,42,44]. Therefore, providers demanded that the time used for mHealth interventions be compensated in the same way as face-to-face treatments [28]. Some clinicians questioned the value of investing in mHealth apps and preferred investing in staff training and staff employment rather than digital tools [21]. One article did not specify the problem of costs [31].