The American Journal of Managed Care November 2008
Plan Design and Active Involvement of Consumers in Their Own Health and Healthcare
Objective: Underlying consumer-driven health plans (CDHPs) is the belief that the financial incentives, enhanced choices, and increased information will stimulate consumers to become active, informed managers of their own health and healthcare (ie, activated consumers). To examine this assumption, we assessed whether enrollees became more activated after enrolling in a CDHP and the degree to which those who were more activated adopted productive health behaviors.
Methods: This was a longitudinal study of employees of a large manufacturing company where a CDHP was offered along with a preferred provider organization in 2004. Two waves of survey data were collected with a final sample size of 1616 employees.
Results: The hypothesis that enrollees in a CDHP become more activated over time was not supported. However, the data suggest that those who were more activated were more likely to engage in the behaviors that CDHPs seek to encourage and to newly adopt these behaviors over time. This appeared to be true regardless of plan type.
Conclusion: Even though CDHPs do not appear to foster activation, they may provide a supportive environment for those who are more activated to manage their health. Encouraging enrollment based on enrollee readiness to take advantage of the CDHP environment may be more productive than relying on plan designs alone to activate enrollees once they are enrolled.
(Am J Manag Care. 2008;14(11):729-736)
Consumers who are more actively involved in managing their own health and healthcare (ie, activated consumers) are more likely to enroll in a consumer-driven health plan (CDHP), but these plans do not appear to foster activation among their enrollees.
- Those who are more activated are more likely to engage in the behaviors that CDHPs seek to encourage (eg, information-seeking and healthy behaviors), and they are more likely to newly adopt these behaviors. However, this is true regardless of plan design.
- Encouraging enrollment based on enrollee readiness to take advantage of the CDHP environment may be more productive than relying on plan designs alone to encourage enrollees to become more active once they have enrolled.
Evidence is emerging that consumers do change their behavior when in a CDHP. Consumers in CDHPs appear to be more cost-sensitive and reduce utilization and expenditures compared with those who stay in plans with traditional designs.3-5 Studies indicate that at least some of the reductions are for necessary care, including the discontinuation of prescription drugs for chronic diseases. Some studies indicate that CDHP enrollees are more likely to seek out information than those in a preferred provider organization (PPO).1-3 Finally, there is some evidence that more activated consumers are more likely to enroll in a CDHP in the first place.6 Thus, the evidence is mixed as to whether CDHP enrollment stimulates enrollees to become more active, informed managers of their health and healthcare.
In this analysis we examine the degree to which CDHP enrollees become more activated (take a greater role in managing their health and healthcare) after enrolling in a CDHP, and the degree to which those who are more activated adopt productive behaviors (eg, information-seeking, healthy). One hypothesis is that CDHP enrollment, with its incentives and information supports, encourages consumers to be actively in charge of their health. Alternatively, it may be that those who are already more activated are able to better manage within a CDHP, engaging in healthproducing behaviors. Finally, both hypotheses may be true.
Specific research questions were the following:
• Do consumers who enroll in a CDHP become more activated over time compared with those who remain in a PPO?
• Are consumers who are more activated more likely to engage in information-seeking to inform their choices? More likely to engage in healthy behaviors? Do these choices occur more often in a CDHP?
• Are consumers who are more activated more likely to adopt new information-seeking or healthy behaviors over time? Do these choices occur more often in a CDHP?
This is a longitudinal study of salaried and hourly employees of a large manufacturing company where 2 CDHPs were introduced at the beginning of 2004 alongside a PPO. The employer funded the personal care account in both CDHPs at the same level. The plans differed in the size of the deductible: one had a high deductible typical of such plans offered by employers,7 and the other had a more moderate deductible. For more details about the plan differences, see the article by Greene et al.5 Two waves of survey data were collected from a sample of employees, the first wave in the summer of 2004 (first year of enrollment) and the second wave in the summer of 2005.
The survey, which was administered using a mixed-mode approach (both Web and phone), asked respondents about their use of information and healthcare utilization decisions during the calendar year, as well as about their demographic characteristics.
The response rate in the 2004 survey was 79%, and the final sample size was 2104 employees. A follow-up survey in 2005 resurveyed the same respondents. Nine percent were no longer employed at the company at the time of the followup survey. The response rate in 2005 was 80% and the final sample size was 1616 employees. Data on plan enrollment for each employee was obtained from the company’s administrative database and verified by the employee.
The sociodemographic characteristics of the entire study sample are shown in Table 1, as well as the characteristics of enrollees in each plan type. Generally, the CDHP enrollees had more education and higher incomes, and were younger and in better health. The PPO enrollees were more likely to be hourly workers (compared with salaried workers). Finally, the CDHP enrollees had higher activation scores than those enrolled in the PPO.
The dependent measures used in this study relate to health information-seeking and healthy behaviors:
• Health information-seeking: Respondents were asked whether they had done each of the following in 2004 (or 2005 for the follow-up survey): (1) used any Web site for health information, (2) were persistent in asking a doctor to explain something until it was understood, and (3) used a telephone advice nurse or health coach.
• Healthy behaviors: Respondents were asked how often in a typical week they (1) limited fat in diet, (2) exercised regularly, and (3) ate 5 or more servings of fruits or vegetables in a day.
The main predictor variable was the Patient Activation Measure (PAM). The PAM, which assesses patient knowledge, skill, and confidence with respect to managing one’s health and healthcare, was developed using qualitative methods, Rasch analysis, and classical test theory psychometric methods. The resulting measure is a unidimensional, intervallevel, Guttman-like scale. The research to date has found the PAM to have strong psychometric properties, including content, construct, and criterion validity. Findings indicate the PAM predicts a range of behaviors, including healthy behaviors (eg, diet and exercise); disease-specific self-management behaviors (eg, adherence to drug regimens, monitoring, managing symptoms); behaviors in the medical encounter; and consumer-related behaviors (eg, using quality information, reading about side effects associated with a new drug).8-10 The PAM is scored on a theoretical 0-100 scale. Most scores fall within the range of 39-85. Activation has been shown to be changeable, with changes of 4 points on average after a 6-week intervention.11 A 4-point change also is significantly linked with changes in behaviors. For example, Fowles found in an employed sample that individuals who ate breakfast, exercised regularly, or followed a healthy diet scored 4-5 points higher on the activation than did those who did not engage in each of the behaviors.12
The tables show the CDHP plan with higher and lower deductible options collapsed. Additional analyses were performed with the 2 CDHP plans separated out, and the differences were minimal. Although the analysis shows the collapsed version, where differences occur they are mentioned in the text. Thus, throughout the analysis, CDHP enrollees, shown as a single group, are compared with the PPO enrollees.
The analysis begins with bivariate assessments and moves to multivariate approaches. The control variables for multivariate analysis fall into 3 categories: health status, sociodemographic characteristics, and mode of survey administration. We used 2 measures of health status: a measure of self-rated health and number of chronic conditions. Sociodemographic measures include age, education (high school graduate or less, some college, college graduate, or more), race/ethnicity, sex, work type (salaried or hourly), and household income (<$35,000, $35,000-$74,999, $75,000+).
Other research has found that the mode of survey administration can have an independent effect on responses, particularly questions sensitive to the influence of social desirability.13 To control for any influence on responses that may be caused by mode of administration (Web or telephone,) we included mode as a control variable in multivariate analyses.
We began with an examination of degree to which these behaviors occurred in the baseline year (2004) in a CDHP-enrolled population and in the PPO-enrolled population. This analysis establishes the base rate of the behaviors in the 2 plan designs. Then we assessed the degree to which activation predicted each of the behaviors within each plan design. Next we examined the degree to which activation predicted the adoption of a new behavior in the second year of observation within each of the plan designs. That is, if the behavior was not performed in 2004, was it newly adopted in 2005? Also, to what degree does baseline activation predict the adoption of a new behavior? A sizable number of employees switched plan enrollment between 2004 and 2005. Of the 623 employees who were in the PPO in 2004, 269 (43%) switched to the CDHP in 2005. Of the 960 employees who were in the CDHP in 2004, 12 (1%) switched to the PPO in 2005. Because we wanted to observe what happens to people over time, we excluded the switchers from the analysis and only included those whose plan enrollment was stable from 2004 through 2005. Thus, the analysis followed a cohort of enrollees over the study period.
The characteristics of those who switched from the PPO to the CDHP in 2005 (n = 269) also were examined. Plan switchers tended to have more education and income than those who stayed in the PPO. The characteristics of switchers were more similar to those of the original CDHP enrollees than to those of employees who continued PPO enrollment (data not shown).
CDHP Enrollment and Activation Over Time Table 1 indicates that those who chose the CDHP had higher activation scores than those who chose to stay in the PPO, suggesting that a CDHP may be more attractive to those who are more adept at managing their health. However, the findings indicate that enrollment in a CDHP did not result in significant gains in activation after a year of enrollment. The activation score of 62.6 to 63.0 during the same time period. None of the differences—either over time, within plan design, or across plan design—were statistically significant. Even after controlling for age, education, income, and self-rated health, there were no significant changes in activation scores from 2004 to 2005 for either CDHP enrollees or PPO enrollees.
Table 2a shows the percentage of CDHP and PPO enrollees who engaged in information-seeking and healthy behaviors in 2004 (base rate of behaviors). The CDHP enrollees were significantly more likely to have engaged in 1 of the information- seeking behaviors and 2 of the healthy behaviors. Table 2b shows the percentage of CDHP and PPO enrollees who adopted new behaviors in their second year of enrollment. Only those enrollees who did not perform the behavior in 2004 are included in the analysis. The CDHP enrollees were more likely to adopt only 1 of the information-seeking behaviors; adoption of a new healthy behavior was no more likely in either plan design.
Activation and Behaviors
In the next step in the analysis, we examined activation as a predictor of the performance of any of the examined behaviors and in the adoption of any of them as new behaviors. CDHP enrollees went from an average 2004 activation score of 64.1 to a score of 64.5. PPO enrollees went from an average Table 3a shows the percentage of respondents scoring in the upper half of the PAM who either engaged in or did not engage in each of the examined behaviors in 2004. The data were broken out by plan enrollment type. Those who were more activated were significantly more likely to engage in 2 of 3 of the information-seeking behaviors whether they were enrolled in either the CDHP or the PPO. Those who were more activated also were more likely to engage in all 3 of the healthy behaviors than were those who were less activated. However, this was true only for the CDHP enrollees.