Using Ordered Probit Analysis to Assess Treatment Satisfaction: An Economic Interpretation

An important aspect of evaluating patent satisfaction with health services is using global satisfaction measures. Global satisfaction questions are usually worded “All things considered, how satisfied are you with the care you received…?” The responses are typically given on an ordinal scale with very dissatisfied at the lowest end and very satisfied the highest end. The economic interpretation of global satisfaction is higher satisfaction corresponds to higher utility. Although measuring utility is controversial in economics, when thought of as utility, patient satisfaction is tied to the assumptions of utility-maximizing imbedded in consumer theory and most specifically to the concept of demand. It is clear from the literature that attempts to measure global satisfaction lacks theoretical guidance and suffers from inconsistency in measuring and operationalizing variables. Accordingly, the purpose of this research is not to delve into the institutional, patient, or psychological aspects of health care delivery that lead to global satisfaction; it is rather to introduce a methodological approach that makes it possible to view global satisfaction from the economic point of view. This research shows how a single treatment satisfaction variable and two other treatment-relevant and economically meaningful covariates can be identified and used in an ordered probit model, and how the results of the analysis can elucidate valuable information for treatment providers. The ordered probit model, described in this research is specified as follows: ln(y^*_u)=ln(X^'_u1b₁)+ln(X^'_u2b₂)+e_u This model uses the log-linear functional form. Since the model is under-specified the error term is set to zero. The present analysis and results are based on data from an outcomes study of substance abuse treatment services (n=509), however this model can be used for any treatment population. The dependent variable is a latent variable representing global satisfaction. The covariates are the outcome variable weeks of abstinence from drug use (Est. 0.205; S.E. 0.041; X²<0.0001), and an adjustor variable, GAF, measuring overall life functioning (Est. 0.722; S.E. 0.431; X² N.S.). This analysis yields a set of indifference curves showing the relationship between global satisfaction and treatment outcome. These findings show global satisfaction levels are relative to treatment outcomes and that the effort (and cost) necessary to obtain higher levels of satisfaction increase on the margin. This technique can be incorporated into the on-going quality assurance activities of any health care organization including monitoring global treatment satisfaction levels intertemporally.