Difficulties in conducting a randomized controlled trial of health service interventions in intellectual disability: implications for evidence-based practice.

Oliver et al. (2002) tried to run a gold-standard RCT of assertive community treatment for adults with intellectual disability. They stopped early because ethics boards, consent rules, and capacity questions made the trial impossible.

The paper is a warning letter to future researchers, not a data report.

The team learned that the biggest barrier was not the treatment itself. It was getting permission to try it. Consent forms, capacity tests, and guardian approvals took months and still blocked random assignment.

They concluded that attitudes, not science, were the main roadblock.

Three years later DeLeon et al. (2005) and Hatton et al. (2005) ran almost the same RCT that Oliver et al. (2002) called impossible. Both teams found no benefit of assertive outreach over standard care. The later studies prove the 2002 fears were real, yet surmountable.

Stott et al. (2017) adds another layer. Their review shows that even when you get into the trial room, most readiness tools for therapy are weak. Poor consent tools can exclude clients who might actually do well.

Griffith et al. (2012) meta-analysis looks brighter: single-case designs in ID already show large effects for challenging behavior. The clash is clear—small-n studies work, but big RCTs stall. The difference is consent burden, not intervention complexity.

If you plan to evaluate any program for adults with ID, budget more time for ethics than for treatment. Build easy-read consent forms, train guardians, and consider single-case or stepped-wedge designs that need fewer participants. The later RCTs show you can finish the trial, but only if you front-load the consent work.