Scaling methods for categorical self-assessed health measures

The lack of a continuous health valuation is a major drawback in health analyses over broad populations. The use of categorical health variables to estimate a continuous health variable is an usual procedure in healthstudies. The most common approaches (ordered probit/logit model and interval regression model) do not admit any skewness in the distribution of health. In the present study a new procedure is suggested, that is attaching a log-normal distribution to health values. Different scaling procedures have been compared, with data obtained from the Catalan Health Survey (2006). The validity of the scaling approaches is assessed by measuring to what extent the health values derived from categorical health variables suit the actual health values. Two different health tariffs have been used for each procedure (VAS tariff and TTO tariff), so that the results are robust to the selection of a metric. In general, models under log normality outperform the other approaches.