Using propensity score matching with doses in observational studies: An example from a child physical abuse and sleep quality study.
Xiaopeng Ji, Naixue Cui, Jianghong Liu
PMID 31674676WHAT IT FOUND
Chinese adolescents who reported parental physical abuse had worse sleep quality than matched non-abused peers.
The link could disappear if unmeasured factors were considered, so it is not proof of cause.
Key findings
01Among 707 adolescents, 30% (n=209) reported physical abuse by parents in the previous year.
02After pair matching, abused adolescents had worse sleep quality than matched non-abused peers (d=0.679, p=0.005).
03The abuse-sleep association was sensitive to hidden bias: a 1.1-fold increase in odds of abuse from unmeasured factors made it non-significant.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study was observational, not randomised, so it cannot prove that physical abuse causes poor sleep. The sensitivity analysis showed that unmeasured factors producing a 1.1-fold increase in the odds of abuse could make the abuse-sleep association non-significant. Sleep quality was self-reported and may be affected by recall bias. Some PSQI subdomains have low internal consistency in adolescents. The physical abuse scale had internal consistency below 0.70 in this study, so measurement error is possible. Analyses were limited to adolescents with complete data, and several relevant covariates were missing or not collected, including behavior problems, parental stress, medication use, clinical sleep disorders, and depression. The comparison of higher versus lower abuse doses among abused adolescents did not show a significant sleep difference.
Declared interests
The authors declared no conflicts of interest.
The easy way to misread this
Do not read this as proof that physical abuse causes poor sleep. It is an observational association, and the sensitivity analysis showed that unmeasured factors producing a 1.1-fold increase in the odds of abuse could make the finding non-significant.