The validity of self-reported information about hypertensive disorders of pregnancy in a population-based survey: The Tromsø Study

Objective: To investigate the validity of self-reported information about hypertensive disorders in previous pregnancies among women participating in the fourth survey of the Tromsø Study. Design: Retrospective cohort study with case–control design. Population: Parous women participating in the fourth survey of the Tromsø Study. Methods: Medical records including partograms of 200 randomly selected women who had answered positively to whether they had hypertension and/or proteinuria during one or more of their previous pregnancies (cases) and 200 women who had answered negatively (controls) were studied. The cases and controls were matched for age. The investigators were blinded to the allocation of cases and controls until data collection was finished. Main outcome measures: Self-reported and actual prevalence of hypertensive pregnancy disorders and predictive value of self-reported hypertension and/or proteinuria in previous pregnancies. Results: Clinical data were missing for 23.5% (94/400) of the participants (50 cases and 44 controls). A total of 80% (120/150) of cases and 57.1% (89/156) of controls had answered the question on whether or not they had high blood pressure and/or proteinuria during their pregnancies correctly (positive predictive value 0.800; negative predictive value 0.571). The proportion of false-positive cases declined with increasing age while the proportion of false-negative controls increased until 55 years of age. Conclusions: We found good concordance between self-reported hypertension and/or proteinuria during previous pregnancies and actual clinical findings among the cases. However, there was a tendency towards under-reporting among controls.