Geographical variation in glaucoma prescribing trends in England 2008–2012: an observational ecological study

OBJECTIVES To explore (1) the national trend in population-adjusted prescription rates for glaucoma and ocular hypertension (OHT) in England and (2) any geographical variation in glaucoma/OHT prescribing trends and its association with established risk factors for primary open-angle glaucoma (POAG) at the population level. DESIGN Observational ecological study. SETTING Primary care in England 2008-2012. PARTICIPANTS All patients who received 1 or more of the 37 778 660 glaucoma/OHT prescription items between 2008 and 2012. PRIMARY AND SECONDARY OUTCOME MEASURE METHODS Glaucoma/OHT prescription statistics for England and its constituent primary care trusts (PCTs) between 2008 and 2012 were divided by annual population estimates to give prescription rates per 100 000 population aged ≥40 years. To examine regional differences, prescription rates and the change in prescription rates between 2008 and 2012 for PCTs were separately entered into multivariable linear regression models with the population proportion aged ≥60 years; the proportion of males; the proportion of West African Diaspora (WAD) ethnicity; PCT funding per capita; Index of Multiple Deprivation 2010 score and its domains. RESULTS Between 2008 and 2012, glaucoma/OHT prescriptions increased from 28 029 to 31 309 items per 100 000 population aged ≥40 years. Between PCTs, nearly a quarter of the variation in prescription rates in 2008 and 2012 could be attributed to age, WAD ethnicity and male gender. The change in prescription rates between 2008 and 2012 was only modestly correlated with age (p=0.003, β=0.234), and income deprivation (p=0.035, β=-0.168). CONCLUSIONS Increased population-adjusted glaucoma/OHT prescription rates in the study period were likely due to increased detection of POAG and OHT cases at risk of POAG. Between PCTs, regional variation in overall prescription rates was partly attributable to demographic risk factors for POAG, although the change in prescription rates was only modestly correlated with the same risk factors, suggesting potential variation in practice.