Cognitive-Based Interventions for Improving Psychological Health and Well-Being for Parents of Children with Developmental Disabilities: A Systematic Review and Meta-analysis.

No included study was judged to be at low risk of bias. All outcomes were parents' own questionnaire answers, and parents knew which group they were in; 13 of the 25 trials gave no information about how group allocation was concealed, and 18 of 25 raised concerns about departures from the intended intervention. Eight of the 25 trials (38.10%) were pilot studies with small samples, and in six of them each group had 20 or fewer participants. The authors say these were probably too small to detect real differences. The participants were mostly women (85.15%) and eight trials included mothers only, so these results mainly describe mothers and may not apply to fathers. The pooled estimates for stress, depression and anxiety came with a great deal of variation between trials (I² = 83%, 90% and 90%), so the average hides trials that did better or worse. Egger's test suggested possible publication bias for depressive symptoms (p = 0.0467). Effects were pooled mainly from measurements taken within four weeks of the programme ending, so the review says little about whether any benefit lasts. Acceptance and commitment therapy rested on four studies and showed no significant effect on parental stress or depressive symptoms, while dialectical behaviour therapy and compassion-focused therapy had one study each and could not be pooled. The reviewers could not compare parents of children with cerebral palsy, intellectual disability or mixed diagnoses because too few studies existed, and format was tangled up with the type of intervention, so the subgroup differences should be read cautiously. The authors note that autism diagnostic criteria changed in 2017, so parents recruited before and after that point may not be comparable.

No funding statement or conflict of interest declaration appears in the text supplied for this review.

Do not read the very large effects in the tables, such as g = 8.64, as the size of benefit you should expect. Those figures come from single studies the reviewers removed from the pooled analyses as outliers, and the pooled effects were much smaller, between 0.29 and 0.95. Every included study had at least some risk-of-bias concern.