Primary care doctors and population health.

Dr Gulliford is to be congratulated on reporting the first examination of the relationship between the availability of primary care doctors and population health in England. His work follows the calculations in the USA by Starfield that each extra family physician is associated with a reduction of deaths of 34 US citizens per 100 000 population, a finding not replicated for specialists. His findings that, in England, the availability of more local general practitioners is associated with reduced hospital admissions and reduced teenage pregnancy are important. It will be worth while to undertake health economic analyses and calculate the average cost savings to society per practitioner from these two findings. Dr Gulliford is correct, in his summary, to show the standardized ratio for all-cause mortality (15–64) with and without adjusting for limiting long-term illness. It is academically unwise to introduce this particular adjustment, as good general practitioner care will reduce limiting long-term illness. Limiting long-term illness, as a dependent variable, cannot be used to adjust the model. In a worthy attempt to reduce bias, he may inadvertently have introduced some. Once this is agreed, then his work will become even more important. After adjusting for deprivation score, ethnic group and social class, he has shown that an increased number of general practitioners is significantly associated with a reduction in the standardized all-cause mortality (15–64) in the population in England. He has therefore succeeded, for the first time, in replicating Starfield’s work on this side of the Atlantic. Finally, Gulliford does not cite important analysis, which also used data from England, which lends support to his work. Sir Brian Jarman and colleagues showed that death rates in hospitals fell significantly in association with increased numbers of general practitioners in the communities outside the hospitals.