Parental recall bias in observational studies: Child dental service use.

Background: Parents are frequently used to recall events of exposure and disease as a proxy for their children in observational health research.

Aim: To assess the validity of parental recall of children's utilisation of dental services.

Design: Parents in the Longitudinal Study of Australian Children (LSAC) were asked to recall their children's overall dental service use in the previous year and whether the study child had received fillings or extractions due to decay in the previous 2 years. True positives were identified through parents who correctly recalled child dental service utilisation with actual dental treatment received in a linked dataset.

Results: Of the 10 090 participants from the LSAC, 1290 study children had linked dental treatment data eligible for inclusion in this study. The absolute true-positive rate for parental recall of dental service use in the previous year was 82% (n = 1263). Overall true-positive percentages were lower for recall of fillings (40%) and extractions (7%) in the previous 2 years. Increasing number of recall days was associated with the rate of true-positive recall adjusted for other factors in all three models.

Conclusion: Accuracy of parental recall for tooth fillings and extractions was low. The use of parental recall for specific dental treatment over a time span of 2 years in further research is not recommended.