Efficacy of Mental Health Smartphone Apps for Symptoms of Depression and Anxiety: Meta-Analysis of 176 RCTS

Smartphones are ubiquitous and frequently used, allowing mental health apps to provide accessible, on-demand therapeutic content. Apps can deliver personalized treatment through passive data collection and symptom tracking.

Mental health apps are garnering substantial interest due to their potential to scale access to care. However, there are also risks regarding privacy, effectiveness, engagement, and exclusion of the patient-clinician relationship.

Linardon, J., Torous, J., Firth, J., Cuijpers, P., Messer, M., & Fuller‐Tyszkiewicz, M. (2024). Current evidence on the efficacy of mental health smartphone apps for symptoms of depression and anxiety. A meta‐analysis of 176 randomized controlled trials. World Psychiatry, 23(1), 139-149.

Key Points

• Mental health apps have small but significant effects on reducing symptoms of depression (g=0.28) and generalized anxiety (g=0.26).
• Apps with features like cognitive behavioral therapy (CBT), mood monitoring, and chatbots tended to have larger effects than those without.
• Effects were larger when apps targeted specific symptoms like social anxiety or OCD, although the evidence was weaker.
• Apps had similar dropout rates (24%) to face-to-face therapy.
• Apps performed similarly to other validated treatments like online CBT, although more research is needed.

Rationale

• Prior research found mental health apps can reduce depression and anxiety symptoms, but effects varied across trials (Firth et al., 2017; Linardon et al., 2019).
• Over 100 new trials have been conducted since the last meta-analysis in 2019, including larger and higher-quality studies.
• An updated meta-analysis was needed to provide more precise effect estimates and understand what app and trial features influence efficacy.

Method

• The study was conducted as a systematic review and meta-analysis adhering to PRISMA guidelines.
• The authors pre-registered a review protocol.
• Several databases (Medline, PsycINFO, Web of Science, ProQuest Dissertations) were searched, combining terms related to smartphones, RCTs, anxiety, and depression.
• The last search was conducted in June 2023.
• Relevant reviews and reference lists were hand-searched for additional eligible trials.
• 176 RCTs of apps for depressive or anxiety symptoms were included.
• Outcomes were symptoms of depression, generalized anxiety, social anxiety, PTSD, OCD, panic, and acrophobia.

Inclusion criteria

• Randomized controlled trials (RCTs) that tested the effects of a stand-alone, smartphone-based, mental health app against a control condition (e.g., waitlist, placebo, information resources) or active comparison (e.g., face-to-face treatment) for symptoms of depression or anxiety
• Trials that conducted one psychoeducational or information session before the app program
• Published and unpublished trials

Exclusion criteria

• Blended web and app-based programs
• Apps not focused on mental health
• Text-message only interventions
• Adjunctive treatments (e.g., apps incorporated within face-to-face therapy)
• Secondary analyses
• Use of single unvalidated items to measure depression/anxiety
• Studies without sufficient data to calculate effect sizes

Statistical Analysis

• Random effects models calculated effect sizes (Hedges’ g) and heterogeneity (I2).
• Number needed to treat (NNT) was computed.
• Ratio of variance in outcomes between groups was calculated.
• Univariate subgroup analyses used mixed effects models.
• Meta-CART identified combinations of moderators linked to effects.

Results

1. Study Characteristics

• 176 RCTs (174 papers) met inclusion criteria, with over two-thirds conducted recently between 2020-2023.
• 43% had unselected convenience samples. Some trials recruited people meeting diagnostic criteria (6%) or scoring above symptom cut-offs (26%). Fewer trials targeted PTSD, social anxiety, OCD or panic (10%).
• Nearly half of the apps were based on CBT (48%). Other apps used mindfulness (21%) or cognitive training (10%).
• 34% of apps had mood-monitoring features, and 5% had chatbots.
• 60% of trials used an inactive control like waitlist or assessment only. 23% used a placebo app control. 11% used care as usual.
• Most trials employed short-term follow-up of 1-4 weeks (56%) or 5-12 weeks (40%).

2. Depression Symptoms

Apps versus control conditions

1. The pooled effect size of apps (N=16,569) vs control conditions (N=17,007) for depressive symptoms was g=0.28, representing a small but significant effect.
2. The number needed to treat (NNT) was 11.5. This means for every 11.5 participants receiving an app intervention, 1 would have positive symptom change relative to the control group.
3. These effects were robust across sensitivity analyses limiting to lower risk of bias and larger sample trials. Effects were also similar at different follow-up durations.
4. There was less variability in post-test outcome scores for app interventions relative to control conditions (RoV of -0.14). However, high heterogeneity for differences in variance was observed (I2=78%), reflecting variable effects of apps.

Apps versus active interventions

• When apps were compared to active interventions, the pooled effect size was small and non-significant (g=-0.08, p=0.340) based on 8 comparisons.
• Specifically, apps performed similarly to face-to-face treatments (g=-0.12, p=0.257) based on 5 comparisons.
• Apps also performed similarly to web-based interventions (g=-0.01, p=0.962), but this was only based on 3 comparisons.

3. Generalized Anxiety Symptoms

Apps versus control conditions

• The pooled effect size was g=0.26 (NNT = 12.4), indicating a significant small effect of apps in reducing generalized anxiety.
• Effects were robust across sensitivity analyses.
• There was less variability in outcome scores among app groups versus controls (RoV=-0.21). But heterogeneity for differences in variance was high (I2=75%), reflecting variable efficacy.
• In subgroup analyses of all trials, no significant moderators emerged.
• But for trials where anxiety was the primary target, inactive controls, pre-selected samples, CBT apps, and mood monitoring apps had larger effects. Mindfulness and cognitive training apps had smaller effects.
• Multivariate meta-CART analysis found apps with mood monitoring features performed best in pre-selected samples. Apps performed better when anxiety was the primary target.

Apps versus active interventions

• When apps were compared to active interventions, the pooled effect size was small and non-significant (g=0.11, p=0.537) based on 6 comparisons.
• Specifically, apps performed similarly to face-to-face treatments (g=0.16, p=0.441) based on 5 comparisons.
• The one comparison of an app versus a web-based program found a small non-significant effect (g=-0.11, p=0.575).

Insight

• This meta-analysis provides the most comprehensive evidence to date that mental health apps can reduce symptoms of common conditions like depression and anxiety.
• Apps significantly reduced generalized anxiety and depression symptoms versus control conditions but displayed high variability in benefits across trials and users. Certain app features and sample characteristics predicted larger effects.
• Apps using CBT had larger effects than other apps like mindfulness. This aligns with the broader evidence showing CBT is more effective for depression/anxiety. CBT apps may better translate the “active ingredients”.
• There was some evidence that depression apps with chatbots and anxiety apps with mood monitoring had larger effects.
• Possible reasons are that these innovative features increase personalization, engagement, self-awareness, and accountability.
• However, these analyses were post-hoc and based on small samples. More rigorous randomized experiments are needed to confirm the added benefits of chatbots and mood-monitoring features.
• The overall effects are small but likely meaningful at the public health level, given the scalability of apps. Also, apps could be an initial treatment option before escalating care.

Strengths

• Very large pooled sample size and number of trials provided robust estimates and subgroup analyses.
• Extensive sensitivity analyses supported the consistency of main effects.
• Novel analyses of variability and interactions among moderators offered unique insights.

Limitations

• Follow-up was restricted to post-intervention period due to inconsistent reporting. Sustainability of benefits is uncertain.
• Outcomes like remission and recovery could not be analyzed due to rare and inconsistent reporting.
• Considerable unexplained heterogeneity persisted in main analyses.

Clinical Implications

• Results support mental health apps as an evidence-based treatment option that improves access to care. Apps appear comparable to face-to-face therapy in dropout rates and, potentially, efficacy.
• Larger effects emerged when depression was the primary target, suggesting apps could be a suitable initial treatment option before escalating care for those receptive or lacking access.
• Evidence for apps targeting specific anxiety conditions remains weak given considerable risk of bias and small samples in those trials.
• Knowledge of optimal app features and user characteristics can guide tool development and clinical recommendations.
• Apps utilizing evidence-based approaches like CBT, along with newer technological capabilities like chatbots and mood tracking, may produce larger reductions in depression and anxiety symptoms. But further research is required to test their specific, additive contributions.

References

Primary reference

Linardon, J., Torous, J., Firth, J., Cuijpers, P., Messer, M., & Fuller‐Tyszkiewicz, M. (2024). Current evidence on the efficacy of mental health smartphone apps for symptoms of depression and anxiety. A meta‐analysis of 176 randomized controlled trials. World Psychiatry, 23(1), 139-149.

Other references

Firth, J., Torous, J., Nicholas, J., Carney, R., Pratap, A., Rosenbaum, S., & Sarris, J. (2017). The efficacy of smartphone‐based mental health interventions for depressive symptoms: a meta‐analysis of randomized controlled trials. World Psychiatry, 16(3), 287-298.

Linardon, J., Cuijpers, P., Carlbring, P., Messer, M., & Fuller-Tyszkiewicz, M. (2019). The efficacy of app-supported smartphone interventions for mental health problems: A meta-analysis of randomized controlled trials. World Psychiatry, 18(3), 325-336. https://doi.org/10.1002/wps.20673

Keep Learning

• How might mental health apps fit into a stepped care approach to treating conditions like depression and anxiety? What are the advantages and disadvantages of having apps as an initial treatment option?
• This study found apps have small average effects on depression/anxiety but larger effects for more specialized apps targeting particular symptoms. How might this knowledge influence the development and recommendation of future mental health apps?
• What ethical considerations should be weighed regarding the privacy, transparency, and regulation of mental health apps as they continue to grow in popularity? How might policy catch up to guide responsible innovation in this space?