To the Editor The findings on paternal age and offspring psychiatric morbidity by D’Onofrio and colleagues1 published in JAMA Psychiatry might be distorted by substantial increases in the rate of diagnoses for several disorders during this period. As a result, there may be a lack of time invariance for many key outcome variables used in the authors’ fixed-effects model. This could potentially contribute to some of the reversals between the results of the baseline data and fixed-effects models and to part of the differences between the adjusted model and fixed-effects results.
Holt GB. Fixed-Effects Models and Diagnosing Psychiatric Disorders. JAMA Psychiatry. 2014;71(9):1077-1078. doi:10.1001/jamapsychiatry.2014.564