BACKGROUND: Serious gaming could support patients in learning to cope with chronic pain or functional somatic syndromes and reduce symptom burdens. OBJECTIVE: To realize this potential, insight is needed into how, why, for whom, and when it works in actual treatment circumstances. METHODS: Following a realist approach, process evaluations were performed before, during, and after a two-armed, natural quasi-experiment (n=275). A group of patients with interfering chronic pain or fatigue symptoms received a short additional blended mindfulness-based serious gaming intervention during a multidisciplinary rehabilitation program. A control group only received the regular rehabilitation program. During two sessions before and one session after the experiment, expectations about serious gaming processes were discussed in focus groups with local care providers, implementers, and experts. Patients participated in a survey (n=114) and in semistructured interviews (n=10). The qualitative data were used to develop tentative expectations about aspects of serious gaming that, in certain patients and circumstances, trigger mechanisms of learning and health outcome change. Hypotheses about indicative quantitative data patterns for tentative expectations were formulated before inspecting, describing, and analyzing-with regression models-routinely collected clinical outcome data. An updated program theory was formulated after mixing the qualitative and quantitative results. RESULTS: Qualitative data showed that a subset of patients perceived improvement of their self-awareness in moments of daily social interactions. These results were explained by patients, who played the serious game LAKA, as a "confrontation with yourself," which reflected self-discrepancies. Important characteristics of serious gaming in the study's context included innovation factors of relative advantage with experiential learning opportunity, compatibility with the treatment approach, and the limited flexibility in regard to patient preferences. Perceived patient factors included age and style of coping with stress or pain. Learning perceptions could also depend on care provider role-taking and the planning and facilitating (ie, local organization) of serious gaming introduction and feedback sessions in small groups of patients. Quantitative data showed very small average differences between the study groups in self-reported depression, pain, and fatigue changes (-.07