摘要
Background: Mental disorders contribute substantially to the global burden of disease, while mental health care systems frequently fail to meet demand due to limited capacity, delayed access, and insufficient continuity of care. In line with Andersen’s Behavioral Model of Health Services Use, these challenges reflect interactions between contextual and individual determinants, health behaviors, and outcomes. Smartphone-based interventions (SBIs) may help reduce existing gaps by providing scalable, low-threshold support as standalone tools, by augmenting ongoing treatment as adjuncts, and by extending care into high-risk transition periods. Beyond scalability, SBIs can leverage smartphone-specific affordances, including continuous monitoring and biofeedback, to support context-sensitive self-regulation and to provide sustained support during critical transitions, such as continuing care after inpatient treatment. This dissertation evaluated both the evidence base and the innovation potential of SBIs within the context of the care gap. Methods: The efficacy of SBIs was examined using two complementary approaches. First, systematic reviews and meta-analyses synthesized randomized controlled trials (RCTs) on (1) standalone SBIs for mental health outcomes and (2) adjunctive SBIs delivered alongside established treatments, focusing on additive effects compared with equivalent treatment alone. Risk of bias and heterogeneity were assessed to support interpretation. Second, two RCTs tested innovative SBIs: (1) a three-arm trial (n = 166) evaluating a brief stress management SBI that combined cognitive-behavioral content with heart rate biofeedback derived from built-in smartphone sensors versus the same SBI without biofeedback, and a waitlist control (WLC) condition, and (2) a multi-center randomized trial (n = 356) that evaluated a smartphone-based continuing care intervention for alcohol use disorder (AUD) following inpatient detoxification versus a WLC condition. Results: The updated meta-analysis of standalone SBIs (72 RCTs, n = 20,784) found small-tomoderate effects for depression, anxiety, and sleep problems, whereas evidence for substance use disorders, suicidality, and self-injury was limited or inconsistent. Considerable heterogeneity and variable study quality further constrained confidence in the pooled estimates. The meta-analysis of adjunctive SBIs (46 RCTs, n = 4,869) identified small additive effects for depression, mania, smoking cessation, and AUD, with more uncertain evidence for anxiety due to limited data and heterogeneity. Across adjunctive trials, symptom or behavior monitoring was common; however, therapist-facing data integration was infrequent. In the stress management RCT, the biofeedback-enhanced SBI reduced perceived stress compared with the WLC condition (d = 0.41) but did not show a significant advantage over the identical SBI without biofeedback. In the AUD continuing care RCT, the intervention reduced risk of relapse over six months (hazard ratio = 0.72), while effects on secondary outcomes were not significant. Conclusion: Across evidence syntheses and two innovation-focused trials, SBIs showed promise but also clear constraints. Standalone SBIs provide modest symptom reductions for high-prevalence disorders such as depression, anxiety, and sleep disturbances; however, evidence remains insufficient for several other clinically significant conditions. Adjunctive SBIs can yield small additional benefits, but their impact likely depends on deeper therapeutic integration and stronger feedback loops between patient-generated data and clinical decision making. The present RCTs underscore both the feasibility of leveraging smartphone-specific technologies and the need for replication and optimization to disentangle specific effects from non-specific influences such as expectancy bias and use intensity. Looking ahead, advances in artificial intelligence may further enhance personalization, adaptivity, and scalability, but only if embedded within rigorous evaluation, robust privacy protection, and human oversight that augments rather than replaces clinical expertise.