We revisited the three interrelated epidemiological concepts of effect modification, interaction and mediation for clinical investigators and examined their applicability when using research databases. The standard methods that are available to assess interaction, effect modification and mediation are explained and exemplified. For each concept, we first give a simple “best-case” example from a randomized controlled trial, followed by a structurally similar example from an observational study using research databases. Our explanation of the examples is based on recent theoretical developments and insights in the context of large health care databases. Terminology is sometimes ambiguous for what constitutes effect modification and interaction. The strong assumptions underlying the assessment of interaction, and particularly mediation, require clinicians and epidemiologists to take extra care when conducting observational studies in the context of health care databases. These strong assumptions may limit the applicability of interaction and mediation assessments, at least until the biases and limitations of these assessments when using large research databases are clarified.

Patients with ASD had a higher risk of first-time AF after closure than the comparison cohort. There was no effect of closure on the use of AF-related medicine in patients with prevalent AF.

Prior psychiatric diagnoses are more common in critically ill patients than in hospital and general population cohorts. Among survivors of critical illness, new psychiatric diagnoses and psychoactive medication use is increased in the months after discharge. Our data suggest both a possible role of psychiatric disease in predisposing patients to critical illness and an increased but transient risk of new psychiatric diagnoses and treatment after critical illness.

Cardiac surgery associated-acute kidney injury (CS-AKI) occurs in 30-50% of patients undergoing surgery for congenital heart disease. Here we determine if CS-AKI is associated with chronic kidney disease (CKD) in patients with congenital heart disease. Using Danish regional population-based registries, our cohort study included patients with congenital heart disease born between 1990-2010 with first cardiac surgery between 2005 and 2010 (under 15 years of age). Utilizing in- and out-patient laboratory serum creatinine data, we identified individuals fulfilling KDIGO stages of AKI within 5 days of cardiac surgery. A unique personal identifier enabled unambiguous data linkage and virtually complete follow-up. The cumulative incidences of CKD stages 2-5 according to presence of CS-AKI were computed utilizing serum creatinine values and Pottel's formula. Using Cox regression, the corresponding hazard ratios were computed, adjusting for sex, age at first cardiac surgery, calendar period of surgery, and congenital heart disease severity. Of 382 patients with congenital heart disease undergoing cardiac surgery, 127 experienced CS-AKI within 5 days of surgery. Median follow-up was 4.9 years. The five-year cumulative incidence of CKD for patients with CS-AKI was 12% (95% confidence interval 7%-20%), significantly higher than the 3% (1%-5%) for those without CS-AKI with a significant adjusted hazard ratio of 3.8 (1.4-10.4). Thus, CS-AKI in patients with congenital heart disease is common and is associated with an increased risk for CKD.

ObjectiveTo investigate whether in utero exposure to antidepressants is associated with increased risk of attention deficit hyperactivity disorder (ADHD).DesignCohort study.SettingDenmark.ParticipantsAll Danish singletons born alive from 1996 to 2009 were included. Using national medical registries, we defined in utero exposure to antidepressants as redemption of an antidepressant prescription by the mother 30 days prior to or during pregnancy. We defined maternal former users of antidepressants as women, who had redeemed a prescription up to 30 days prior to pregnancy, and never users as women who had never redeemed a prescription.Main outcome measuresADHD was defined as redemption of a prescription for ADHD medication or an ADHD hospital diagnosis. Children were followed through 2010, and we used proportional-hazards regression to compute adjusted HRs comparing children exposed in utero and children born to former antidepressant users with children born to never users. To adjust for confounding from family-related factors, we conducted a within-mother between-pregnancy analysis comparing exposed children with unexposed siblings using conditional logistic regression.ResultsWe identified a cohort of 877 778 children, of whom 1.7% were exposed in utero. The overall median follow-up time was 8 years; selective serotonin reuptake inhibitors were the most commonly used class of antidepressant during pregnancy (78% of users). The adjusted HR comparing children exposed to any antidepressant in utero with children born to never users was 1.2 (95% CI 1.1 to 1.4), and 1.6 (95% CI 1.5 to 1.8) comparing children born to former users to children born to never users of antidepressants. In the within-mother between-pregnancy analysis (n=867), the adjusted OR was 0.7 (95% CI 0.4 to 1.4).ConclusionsThis study provides no evidence to support a causal association between in utero exposure to antidepressants and risk of ADHD.