Behavioral interventions to reduce the pica of persons with developmental disabilities.

The authors read 26 small studies on pica in people with developmental disabilities. They grouped the treatments into packages like DRO, NCR, and overcorrection. They wrote a story-style review, not a meta-analysis.

Most studies had one to four kids or adults. Settings were homes, schools, or clinics. The team asked, "Do these packages stop eating non-food items?"

Every package cut pica by at least half, often to zero. DRO worked when the reinforcer matched the person. NCR worked when items were always available. Overcorrection worked but needed staff who could stay calm.

No single package won. The winner was the one that fit the learner's reason for pica.

Boudreau et al. (2015) extends this story. They ran the same packages in a day-treatment clinic with the kids and saw big drops in pica. They added numbers: effect sizes were large and maintained at home.

Allan et al. (1994) is a grandparent here. They tested one boy and two girls and showed that picking reinforcers and punishers from a short assessment beats using a stock list. Ganz et al. (2004) cites this study as proof that custom picks matter.

van der Miesen et al. (2024) zooms out. Their 2021 meta-analysis folds pica studies into a giant SIB pool. They find caregiver-delivered treatments at home work as well as clinic care. This updates the 2004 view that most packages ran in clinics.

You do not need a magic protocol. You need a quick functional guess, then pick the package that removes the payoff for pica. Start with NCR if it looks automatic; add DRO if attention is the fuel. Track one week of data and adjust. Caregivers can run it at home once you show them the plan.