Response to “Are Student Attitudes Toward Family Medicine Research Associated With Family Medicine Career Choice at the Individual Level?”

We agree that evaluating the independent association between a student’s perception of research and their likelihood of choosing family medicine is an additional important construct, as the population-level analysis was not designed to explore individual-level associations. The population-level model, as described in Beinhoff et al, ¹ explored how students’ perception of a specialty at an institution may influence the proportion of its graduates entering family medicine. The individual-level model, as suggested in the letter from Dr Parente, would instead explore an individual’s perception and its influence on their own career choice.

As suggested, using the same data set, we created a new variable, FMAQ except research, by subtracting the research subscale score from the total score for each individual student participant in the sample (n=1,188). We next assessed for multicollinearity and found that FMAQ except research was correlated with the total score 0.617, but the correlation was not so strong as to violate the assumptions of binary logistic regression. No other variables had notable correlations. We then conducted the logistic regression, where the dependent variable was family medicine specialty choice (a binary outcome) and the predictor variables included research subscale score, FMAQ except research, gender, age, race/ethnicity, family income, expected educational debt, marital status, and hometown size. All variables were self-reported and based on survey responses, as described in the original study. In the regression analysis, student perception of family medicine research was not independently predictive of choosing family medicine, beyond the effect of FMAQ except research.

This finding is useful because it suggests that an individual student’s perception of family medicine research does not change their likelihood of choosing family medicine, after controlling for their perception of family medicine more generally. However, our recent study ¹ found that the perceptions of family medicine research among the student body as a whole correlated with the proportion of students choosing family medicine from an institution. This suggests that student perceptions of family medicine research are associated with family medicine careers on an institutional, not individual, level, thus supporting that strong family medicine research is an important part of an institutional culture that is positive toward family medicine.