Major depressive disorder (MDD) is a mood disorder characterized by low or depressed mood present for two weeks or more, along with other symptoms such as loss of interest or pleasure, loss of energy, sleep and appetite disorders, diminished ability to concentrate and recurrent negative thoughts as main features (1). The lifetime prevalence of MDD varies from 2 to 21% (2), presenting approximately twice as often in women than in men (3). MDD is one of the medical conditions that generates the largest chronic disease burden adjusted by “years lived with disability” in populations of any socioeconomic status. Additionally, MDD increases the risk of developing other medical conditions [i.e., cardiovascular and neoplastic diseases, (3)]. Social and health care costs associated with MDD are high (4), and patients still have barriers to access to quality treatment. Since the emergence of the COVID-19 pandemic, the prevalence of depression and anxiety has continued to increase globally (5). This points to the need for urgent and cost-efficient solutions for people experiencing this symptomatology.

The progress of new technologies is having an impact on clinical practice in mental health. In regard to developing improvements in clinical attention for people with mental health disorders, smartphone apps are in the spotlight. App interventions delivered via smartphones are low cost and of special relevance for public health (6). Apps are flexible interventions that can be therapist-guided, unguided [i.e., self-help apps; (7)], or applied as a stand-alone or combined treatment.

Regarding psychotic disorders, the literature suggests that app interventions are feasible and well accepted by people with such disorders, especially during early stages of the disease (8), and therefore may improve treatment adherence, symptomatology, and duration of hospital admissions (9). In addition, evidence suggests that people presenting with more severe psychotic and depressive symptoms seem to have a better engagement with digital interventions compared to people experiencing mild symptoms (10). In this vein, apps seem to improve manic and depressive symptomatology in people with bipolar disorder, especially when interventions include psychoeducation techniques (11) without any apparent limitation in their use by these patients (12). On the other hand, to date, possible negative effects associated with the use of digital interventions in patients with depressive disorders have been detected, such as the appearance of new symptomatology, worsening of the existing disorder, trouble when implementing the designated tasks or a subjective feeling of pressure while trying to complete them on time (13). Similarly, other barriers, such as worry about the efficacy of the intervention, difficulties in app use, privacy issues and lack of direct professional feedback, were detected (14).

Regarding anxiety and depressive disorders in particular, the literature suggests that apps may have a role in reducing anxious-depressive symptomatology. One systematic meta-review of meta-analyses revealed that the use of mobile-based interventions is a promising approach to achieving a modest reduction in depressive and anxious symptoms, although there are no clear differences when such interventions are compared to active therapeutic interventions (6). In terms of assessing the extent of the clinical improvement, it is paramount to consider the level of intervention adherence, which is greater when the intervention involves a higher degree of user engagement (15). However, maintaining user engagement is a challenge in relation to app interventions because, often, there is a rapid drop in their use (6). In a previous systematic review, contradictory results regarding the efficacy of apps based on Cognitive Behavioral Therapy (CBT) in patients suffering from mild and moderate depression were found (16). According to Hrynyschyn and Dockweiler (16), the clinical and methodological heterogeneity in the clinical trials interferes with the outcome analyses and prevents us from drawing solid conclusions. However, the review conducted by Hrynyschyn and Dockweiler (16) provides limited evidence about the efficacy of these interventions, since it focuses exclusively on apps based on CBT, includes patients with mild or moderate depression selected by using cutoff scores in questionnaires about depression (without the need of being diagnosed by a health professional) and only analyzes studies carried out from 2015 onward. In their meta-analysis, Serrano-Ripoll et al. (17) found smartphone app interventions had a moderate effect on depressive symptomatology. Nonetheless, these authors include studies conducted with participants with depressive symptoms, and not in all the studies were the patients diagnosed by a health professional. Additionally, in this systematic review, only studies that compared app interventions to non-active control groups (i.e., minimal intervention, treatment as usual, waiting-list control) were analyzed. This could overestimate the effect of app interventions.

Some studies have highlighted the advantages of digital interventions, such as the motivational aspects of the apps and their accessibility (18). This is especially relevant to those individuals for whom it is more difficult to access face-to-face clinical appointments (19). Despite these advantages, implementing digital interventions still presents numerous challenges before such interventions can be incorporated into daily clinical practice, such as the lack of integration of apps into computer systems in public health care, or issues surrounding the protection of client privacy (20). This contrasts with the increasing worldwide popularity of apps among smartphone users, which is leading to an overload of the app market (21), along with the proliferation of apps that employ non-evidence-based techniques (22). In line with the aforementioned and as a result of the mental health problems associated with the COVID-19 pandemic, the rise of apps continues to increase (23). Thus, a systematic review to critically analyze and reflect on the efficacy of such apps is needed.

Given the high availability of these apps in app stores (21) and their growth in both the app market and clinical practice (20), it is necessary to critically assess the efficacy of these interventions in patients with depressive disorders diagnosed by a health professional. Therefore, the aim of the present systematic review is to determine the efficacy of smartphone app interventions for patients with depressive disorders.

To achieve the aims of the present systematic review, the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement (24) were followed.

Seven studies that met the inclusion criteria were selected. Figure 1 shows the selection process of the articles included in this systematic review. The main characteristics of the selected studies are shown in Table 1.

The seven studies included a total of 651 patients. All of the participants were adults, and their weighted average age was 35.49 years. In six of the selected articles, the patients had a diagnosis of MDD (according to the ICD-10, DSM-IV or DSM-5), whereas in one of the articles, the diagnosis was other specified depressive disorder [according to the DSM-5; (25)]. The average percentage of females was 69.52%.

The aim of this systematic review was to determine the efficacy of smartphone apps in patients with depressive disorders. After reviewing the seven articles included in this article, it can be concluded that the use of apps does not reduce depressive symptomatology nor improve the quality of life in patients diagnosed with depressive disorders when compared to an active control group. However, the effect of these interventions on perceived stress and anxious symptomatology remains unclear. The studies are characterized by a considerable degree of clinical and methodological heterogeneity, making it difficult to draw firm conclusions.

Regarding the efficacy of apps on depressive symptomatology, the analyzed studies show that app interventions do not reduce depressive symptoms in depressed patients when compared to an active control group. Only the study by Mantani et al. (28) found an improvement in depressive symptoms in a smartphone app intervention group, compared to a control group. However, in six of the seven studies of this review, no significant changes in depressive symptoms were observed in the app intervention group when compared to an active control group. When interpreting this result, the considerable level of heterogeneity among the selected studies should be taken into account. Different methodologies were used in the selected studies (including the application of different interventions, the combination of their components without verifying which ones were responsible for their efficacy, and the different measurements used at different times of the intervention and at different follow-up periods), and they include participants with different clinical characteristics (e.g., clinical severity). Furthermore, as noted by Kerst et al. (32), most studies include some sort of clinical support—although the vast majority is minimal, which makes it difficult to verify the unique effect of app interventions.

The findings of this review contrast with those of Hrynyschyn and Dockweiler (16), who found contradictory results on the efficacy of CBT-based apps in reducing depressive symptomatology in patients with mild or moderate depression, and highlight the considerable degree of heterogeneity of the articles included in their work. The review of these authors only included studies from 2015 onward, which only analyzed the efficacy of CBT-based apps in patients with mild to moderate depression (who in the majority of studies were recruited from a cutoff score on a depressive symptom questionnaire). Unlike the work of Hrynyschyn and Dockweiler (16), in this review, we consider studies published over a longer period of time, in which patients with a depressive disorder (diagnosed by a health professional) took part and received different types of treatment (including pharmacological treatment).

The results of this systematic review are not consistent with those of Serrano-Ripoll et al. (17), who found that app-based interventions resulted in a moderate reduction in depressive symptoms. These authors included in their meta-analysis participants with depressive symptoms, which in some studies had not necessarily been diagnosed by a health professional. In terms of interpreting the findings of Serrano-Ripoll (17), we should be taking into account that these authors analyzed studies comparing an app intervention group to a non-active control group, which could account for the effect found. App interventions rarely outperform active interventions [i.e., therapeutic interventions acting as an active control group; (6, 33)]. In our systematic review, we only included studies comparing an app intervention group to an active control group, which could make it difficult to identify any significant differences between the interventions. In fact, there are studies included in our systematic review that incorporate apps in both the intervention and control group, and although they do not find differences between the conditions, within-group improvements in both the intervention group and the control group in depressive symptoms (25, 26) are observed. Watts et al. (31) also found significant within-group improvement in the app intervention in relation to depressive symptoms (though they found no differences between the app and the same intervention in a computer-based version). The fact that within-group improvements in app intervention groups have been observed offers promise for future treatments. Another difference between our systematic review and that of Serrano-Ripoll et al. (17) is that our work only involves patients with depressive disorders diagnosed by a health professional. This allows us to ensure that the sample of patients between the different studies is as homogeneous as possible.

Regarding the effectiveness of the apps on quality of life, this variable did not improve significantly in any of the five studies that measured it. Therefore, the results of our review suggest that the app intervention does not appear to improve the quality of life in depressed patients. However, these results are tentative and should be regarded with caution. Hrynyschyn and Dockweiler (16) outline that the efficacy of app interventions in improving quality of life is inconclusive. We, however, state that it is possible that longer follow-up periods are required to detect improvements in the quality of life of depressive patients who are undergoing an app intervention. In this way, patients could apply the skills learned and achieve changes in their daily life.

With regard to the efficacy of apps for reducing perceived stress and anxiety, the results are unclear. A reduction in perceived stress was only significant in the study by Raevuori et al. (29), whereas Tønning et al. (30) found no significant improvement in perceived stress. Thus, no conclusions can be established. Additionally, a significant improvement in the reduction of anxious symptoms was found in only one (25) of the four studies that analyzed this variable. This is not consistent with the efficacy of apps to reduce anxiety in meta-analytic studies (15). As is the case with depressive symptomatology, it may be difficult to detect significant improvements in anxiety when an app intervention group is compared with an active control group. Finally, our results are in line with the findings of the review by Hrynyschyn and Dockweiler (16), who found contradictory results regarding the efficacy of apps to reduce anxiety in depressive patients. Therefore, the results regarding the efficacy of apps in reducing anxious symptoms in depressive patients are inconsistent.

In relation to experiential avoidance, in two of the seven studies that measured this variable, no significant differences between the app intervention group and control group were found. To the best of our knowledge, this variable has not been extensively explored in studies that employ apps to treat depression. Taking into account the rise of third-generation therapies (34), it is possible that in the coming years more studies will consider this variable.

With regard to other outcome variables, the studies analyze very different variables, which makes it difficult to draw firm conclusions. However, it is worth highlighting that in the study by Mantani et al. (28), in their per-protocol sample, there was a lower burden of side effects in the intervention group than in the control group at nine weeks. Although this was only analyzed in one study, it opens the possibility of combining apps with pharmacological treatment. The combination of treatments (i.e., app + medication) could be suitable for patients who do not respond to antidepressant treatment or who experience significant side effects while taking medication (28).

This systematic review has the following limitations. First, the selected articles showed a considerable degree of clinical and methodological heterogeneity, which explains the disparity in the results and makes it difficult to draw any firm conclusions. For example, the selected studies use different treatments; include various components (e.g., mindfulness, BA) without knowing which ones are responsible for the efficacy; use different measurement instruments (at different time points), different app interventions, and different control conditions; and include depressed patients with varying levels of clinical severity. This heterogeneity has also been found in studies that analyze the efficacy of app interventions (16, 17). In this review, we attempted to reduce heterogeneity by selecting studies conducted in clinical patients with depression diagnosed by a health professional. Second, although most studies use the PHQ-9 and the BDI-II to assess depressive symptomatology, other authors use different instruments, which makes it difficult to analyze the results regarding the main variable of interest. Third, only English-language studies were considered; therefore, language bias could not be ruled out.

The costs in social and public health policies associated with depression are high in all countries and continue to grow (4), and apps are accessible tools for most citizens. Despite the findings of our systematic review, we consider that the potential beneficial effect of apps merits investigation in future studies since they represent a low-cost and potentially far-reaching intervention of interest in the public health system. If they are shown to be effective, they could represent an opportunity for the prevention and treatment of depressive disorders. To date, however, the increased use of these interventions in clinical practice has not been accompanied by an increase in any associated scientific evidence (35), which generates a certain degree of mistrust on the part of users and professionals when it comes to recommending such interventions. In the future, it will be necessary to continue looking into the efficacy of apps for depressive patients and explore the mechanisms of change and moderators of the efficacy of the intervention (e.g., gender, age, socioeconomic and educational levels, symptom severity, medication, therapist-guided or unguided). Studies that aimed to address this question have not been able to identify moderating variables (6). To determine the mediators of change, it will be necessary to carry out dismantling studies that consider the components of the treatment in isolation (alone and/or in combination). Another step will be to analyze the efficacy of apps based on different effective treatments (e.g., psychotherapy and medication) and the role that apps play in changing the pharmacological treatment received by patients. Taking into account that intervention via apps is often applied to patients with subclinical depressive symptoms, it will be necessary to look into the efficacy of these interventions not only at an intervention level but also at a prevention level. Given the considerable degree of heterogeneity observed in the studies, it will be necessary to refine clinical trial methodology in the coming years.

On the other hand, in regard to applying apps in the clinical setting, some barriers have been found, such as a lack of engagement, which could explain the high dropout rates in some studies (36). Finding strategies that improve adherence to these interventions—such as offering feedback by a professional, daily monitoring of mood, or applying the principles of behavioral economics (15)—is a pending issue (36). In addition, the protection of personal data is a widespread problem when using apps to treat depressive disorders (14, 37) and is an important issue to consider. Multidisciplinary teams (e.g., psychologists, psychiatrists, computer scientists) should work on this issue. Lastly, the efficacy of using apps as a complementary intervention to the usual pharmacological and/or psychological treatment should be analyzed. It is possible that the inclusion of apps in conventional treatments could improve adherence to them (38).

According to the results of this systematic review, app interventions do not appear to reduce depressive symptomatology in depressed patients. However, in recent years, there has been an increase in their use both in the clinical setting and by mobile phone users (21, 23). Despite the potential beneficial effects of apps, we should bear in mind that apps are a double-sided coin, which also carries potential risks such as privacy and security concerns, and adverse effects of psychotherapy, among others (39). These adverse consequences could be minimized if the health professional keeps in mind the risk-benefit ratio of app interventions (39) and only recommends apps incorporating evidence-based interventions (40). When evidence-based apps are used in the clinical setting, health professionals should monitor their use by depressed patients to prevent them from taking risks, especially in the case of severely depressed patients. In depressed patients at risk of suicide, the use of an app could be of interest for suicide prevention (41) as well as to facilitate professional contact in emergency situations. However, the use of apps in patients at risk of suicide should always be complementary to usual treatment, and should support, but not replace, professional judgment (42). Lastly, we must not forget that, to date, app interventions cannot be regarded as a single treatment or be indiscriminately recommended.

Based on the results of this systematic review, it is concluded that app interventions do not appear to reduce depressive symptomatology nor improve the quality of life in depressive patients. However, the effect of these interventions on perceived stress and anxious symptomatology remains unclear. Additionally, the studies are characterized by a considerable degree of clinical and methodological heterogeneity, which makes it difficult to draw solid conclusions and points toward the need to improve the methodology of any future clinical trials.

Any inquiries can be directed to the corresponding author.

AH-G, MJV-F, and NA-G conceived the study, participated in its design, and coordination. AH-G and MJV-F carried out the literature searching and data extraction. All authors were involved in the interpretation of the data, drafting the manuscript, and revising it critically for important intellectual content. All authors contributed to manuscript revision, read, and approved the submitted version.

GL has been a consultant to or has received honoraria or grants from Janssen-Cilag, Otsuka-Lundbeck, Lilly, Astra-Zeneca, Angelini, CIBERSAM, and Instituto de Salud Carlos III. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.