Validating a Health Consumer Segmentation Model: Behavioral and Attitudinal Differences in Disease Prevention-Related Practices

Abstract

While researchers typically have segmented audiences by demographic or behavioral characteristics, psychobehavioral segmentation schemes may be more useful for developing targeted health information and programs. Previous research described a four segment psychobehavioral segmentation scheme—and a 10-item screening instrument used to identify the segments—based predominantly on people's orientation to their health (active vs. passive) and their degree of independence in health care decision making (independent vs. dependent). This study builds on this prior research by assessing the screening instrument's validity with an independent dataset and exploring whether people with distinct psychobehavioral orientations have different disease prevention attitudes and preferences for receiving information in the primary care setting. Data come from 1,650 respondents to a national mail panel survey. Using the screening instrument, respondents were segmented into four groups—independent actives, doctor-dependent actives, independent passives,and doctor-dependent passives. Consistent with the earlier research, there were clear differences in health-related attitudes and behaviors among the four segments. Members of three segments appear quite receptive to receiving disease prevention information and assistance from professionals in the primary care setting. Our findings provide further indication that the screening instrument and corresponding segmentation strategy may offer a simple, effective tool for targeting and tailoring information and other health programming to the unique characteristics of distinct audience segments.

This study was funded by the Robert Wood Johnson Foundation (RWJF) and conducted by RWJF, Porter Novelli, and experts in the area of health behavior. In addition to RWJF, we would like to thank Dr. Erica Frank and Dr. Larry Green for their expert advice and guidance on this project.

Notes

∗Bolded figures are significant at p < 0.05.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.

∗Bolded figures are significant at p < 0.05.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.

∗Bolded figures are significant at p < 0.05.

In some instances, upper or lower confidence intervals are rounded to 1.00. In cases where bolded, exact figures did not cover 1.00.

∗Numbers are means (and standard deviations) from 5-pt. scale, with “1” equaling strongly disagree and “5” equaling strongly agree. Bolded figures are significant at p < 0.05.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.

∗Bolded figures are significant at p < 0.05.

ⁱ Percent reporting scores of 8–10 on a 10-point scale, with “1” equaling not at all confident and “10” equaling extremely confident.

ⁱⁱ Percent reporting scores of 6–7 on a 7-point scale, with “1” equaling not at all likely and “7” equaling very likely.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.

∗Bolded figures are significant at p < 0.05.

ⁱ When considering “family,” respondents were asked to think only of blood relatives such as parents, grandparents, or brothers/sisters.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.

∗Bolded figures are significant at p < 0.05.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.

∗Bolded figures are significant at p < 0.05.

^a Differed significantly from IA.

^b Differed significantly from DDA.

^c Differed significantly from DDP.

^d Differed significantly from IP.