Purpose: To explore and describe the enactment of user involvement and combined care in a Danish clinic that aimed at providing integrated diabetes and mental health care. Design: An ethnographic study. Data Sources and Methods: Data consisted of field notes from 96 hours of participant observations and field notes from 32 informal conversations with healthcare providers, users and relatives as well as 12 semistructured interviews with users. Data were analysed using a thematic analysis. This study reports to the SRQR guidelines. Results: Treatment was not combined as intended if only one healthcare provider handled the consultations. Here, the healthcare providers' focus was often on their own area of expertise—either mental health or diabetes. If more than one healthcare provider handled consultations, the consultations were often divided between them, focussing on one condition at the time. Healthcare providers noted, that learning from peer colleagues was a way to increase the possibility for combined care. Furthermore, combined care was highly dependent on the healthcare providers' ability to involve users' illness experiences in their own care planning. Here, a high level of user involvement increased the levels of combined care during consultations. Conclusion: This study set out to explore and describe user involvement and combined care in a specialised diabetes and mental health outpatient clinic. Combined care is complexed and requires that healthcare providers are well-equipped to manage the complexity of delivering care for people with both conditions. The degree of combined care was linked with the healthcare providers' ability to involve users and their knowledge on the condition outside there are of expertise. Relevance to Clinical Practice: A peer-learning environment in combination with clinical guidelines and joint display could support healthcare providers in involving users in own care and when delivering care outside their area of expertise. Public contribution: No patient or public contribution. Due to the COVID-19 pandemic, the original user council withdraw their consent to participate due to health-related worries and anxiety concerning the pandemic. The user council consisted of three members diagnosed with diabetes and severe mental illness. They were invited to participate in physical meetings, phone or online meetings. Presenting findings from the study to the study participants were also hindered by the second lockdown. This influenced the possibility for data triangulation.