Background: National policy recommends intensive support teams for all areas of England for adults with intellectual disabilities who display challenging behaviour. However, to the best of our knowledge, there has not been a systematic evaluation of intensive support teams to date. Objectives: Our objectives were to identify and describe the geographical distribution and characteristics of intensive support teams in England; to create a typology of intensive support teams; to investigate clinical and cost outcomes of intensive support team models and factors associated with those outcomes; and to explore professionals’, service users’ and carers’ experiences and describe the wider system context in which they operate. Design: This was a two-phase mixed-methods study. In phase 1, a national survey examined the intensive support team models in operation in England. In phase 2, an observational study of adults with intellectual disabilities investigated the clinical effectiveness and cost-effectiveness of the two intensive support team models. Semistructured interviews with intensive support team managers and professionals, carers, and adults with intellectual disabilities explored their experiences of intensive support team care. In parallel, we examined service-level outcomes related to the function of intensive support teams. Setting: Phase 1 included 80 intensive support teams serving 242 community intellectual disability services in England. Phase 2 included 21 intensive support teams, half of which were in the enhanced intensive support teams model and half of which were in the independent model. Participants: In phase 1, a total of 73 intensive support team managers provided data. In phase 2, a total of 226 participants with intellectual disabilities from 21 intensive support teams (enhanced: teams, n = 11; participants, n = 115; independent: teams, n = 10; participants, n = 111) were enrolled in the study. A total of 42 stakeholders were interviewed. Main outcome measure: The main outcome measure was the Aberrant Behaviour Checklist-Community, version 2, total score. Additional data sources were the carer and self-reported questionnaires, qualitative interviews and focus groups. Results: Two intensive support team models were identified in England – enhanced and independent. Challenging behaviour at 9 months was reduced in both intensive support team models (β 3.08, 95% confidence interval –7.32 to 13.48; p = 0.561), but the observed Aberrant Behaviour Checklist- Community, version 2, score reduction appeared larger in the independent model than in the enhanced model (21% vs. 13%, respectively). No statistically significant differences were found in the secondary outcomes [Psychiatric Assessment Schedule for Adults with Developmental Disabilities Clinical Interview organic condition (odds ratio 1.09, 95% confidence interval 0.39 to 3.02), affective or neurotic disorder (odds ratio 0.91, 95% confidence interval 0.32 to 2.59), or psychotic disorder score (odds ratio 1.08, 95% confidence interval 0.21 to 5.50); risk score (β 1.12, 95% confidence interval –0.44 to 2.68); or quality of life questionnaire score (β –2.63, 95% confidence interval –5.65 to 0.40)]. Similarly, no differences were observed between models in relation to cost-effectiveness (health and social care costs mean difference £3409.95, 95% confidence interval –£9957.92 to £4039.89; societal costs mean difference –£4712.30, 95% confidence interval –£11,124.85 to £2106.36). The experiences of stakeholders did not differ between the intensive support team models, with carers and adults with intellectual disabilities valuing service accessibility, person-centred care and engagement. All stakeholders reported a range of barriers to intensive support team care. Service-level data and the operational policies from intensive support teams showed variation in organisational function and the roles of intensive support teams. The most commonly delivered intervention was positive behaviour support. Conclusions: The study describes the operation of intensive support teams in England and identified two distinct models.We did not find advantages or disadvantages associated with clinical outcomes between models, nor did we find cost differences. On this basis, we recommend that local services decide which model best suits their circumstances.