The decline in sample size of the Annual Population Survey has impaired our ability to continue using the existing method to estimate age-specific proportions of good health by local authority. Instead of using the survey’s observed percentage prevalence, we used logistic regression to estimate the age-specific probability of reporting good health by sex and local authority. The odds of reporting good general health are modelled as a function of sex, age, and local authority of residence; age is included as a natural cubic spline and interacted with sex and region (England) or country of residence (Wales). Using the model, we estimated probabilities for all combinations of sex, single year of age and local authority, calculating the probabilities for the age groupings required for the Sullivan life table and further aggregation up to larger geographical groupings.