Predictors of phone call attempts needed to complete follow-up assessments in longitudinal mental health research

Background: In longitudinal research paradigms, selective attrition threatens the validity of research findings. Previous research has found that certain socio-demographic variables predict participant retention. However, there has been little published on the number of contacts it takes to reach participants in mental health research and whether this can be predicted by specific baseline characteristics.

Methods: We analyzed longitudinal data from 170 mothers who were enrolled in an RCT studying the effectiveness of a problem-solving intervention to improve maternal well-being. The sample was largely low-income (74% receiving WIC or public insurance) and non-white (56%). Depressive symptoms were assessed every 6 weeks for 9 months. Regression analysis was used to determine whether there was an association between baseline characteristics and the number of phone calls required to complete follow-ups.

Results: Low-income mothers required more phone calls to complete assessments, as compared to higher-income mothers (β=1.36, p-value 0.03). The study completion rate was 81% among low-income mothers; 97% among higher-income mothers. The average frequency of phone calls to complete a follow-up was 3.65 (sd=3.4) and the upper quartile was 4.33 calls. The subsample of low-income mothers in the upper quartile was largely non-white, unmarried, not working, and had a history of depression.

Conclusion: In this sample, mothers who were lower income tended to be more difficult to reach. Given that 25% of our sample required more than 4 phone calls per completed assessment, these data highlight the importance of research planning and allocation of resources to ensure the validity of study results.