Fumikazu Hiyoshi, Takumi Kanata, Mari Oba, Takeshi Fujii, Ryo Iwata, Akira Iino, Keitaro Murayama, Toshiaki Kikuchi, Asuka Yoshimi, Shinsuke Kito, Hironori Kuga, Hideki Oi, Koichiro Watanabe, Norio Ozaki, Tomohiro Nakao, Kazuyoshi Takeda, Kazuyuki Nakagome
Online support systems may help address unmet mental health needs after the COVID-19 pandemic. In this uncontrolled, real-world longitudinal pre–post study, we described one-month changes in mental health outcomes among help-seeking users of KOKOROBO, a Japanese severity-guided web-based mental health support platform, which recommends a wait-and-see approach, AI chatbot use, or online-counseling depending on the user’s mental health severity; we explored baseline factors associated with change ( n = 3,325). Participants completed measures of depression (PHQ-9), anxiety (GAD-7), insomnia (ISI), and health-related quality of life (EQ VAS) at baseline and one month. Service-use subgroup analyses were descriptive because service use was self-selected and subgroups overlapped. Missing data were addressed using multiple imputation, and multivariable linear regression examined associations between changes and baseline characteristics. At one month, PHQ‑9, GAD‑7, and ISI scores decreased, and EQ VAS increased (PHQ‑9 mean change = − 1.61; 95% CI, − 1.85 to − 1.37). Improvements were generally observed across service-use subgroups. Frequent versus almost no communication with others was associated with more favorable changes across outcomes, whereas communication with family was not associated with change. These findings characterize short-term real-world outcomes; controlled studies are needed to evaluate the effect of the platform and its severity-guided recommendation strategy.