Using choice-based conjoint to determine the relative importance of dental benefit plan attributes

Abstract

The purpose of this study was to use conjoint analysis to determine the importance of specific dental benefit plan features for University of Iowa (UI) staff and to build a model to predict enrollment. From a random sample of 2000 UI staff, 40 percent responded (N = 773). The survey instrument was developed using seven attributes (five dental benefit plan features and two facility characteristics) each offered at three levels (e.g., premium = $20, $15, $10/month). Pilot testing was used to find a realistic range of plan options. Twenty-seven hypothetical dental benefit plans were developed using fractional factorial combinations of the three levels for each of the seven attributes. For all of the hypothetical plans, dental care was to be provided in the UI predoctoral dental clinic. Plan profiles were arranged four per page by combining the existing plan with three hypothetical plans, for a total of nine pages. Respondents' task was to select one plan from each set of four. A regression-like statistical model (Multinomial Logit) was used to estimate importance of each attribute and each attribute level. Relative importance (and coefficients) for each of the seven attributes are as follows: maximum annual benefit (.98), orthodontic coverage (.72), routine restorative (.70), major restorative (.67), time to complete treatment (.61), clinic hours of operation (.47), premium (.18). For each attribute, relative importance of each of three levels will also be presented. These coefficients for each level are used to predict enrollment for plans with specific combinations of the dental benefit plan features.