(Explaining differences in English hospital death rates using routinely collected data)
In this week's BMJ Professor Brian Jarman and colleagues from Imperial College School of Medicine and Harvard Medical School report that one of the key determinants of hospital death rates is the number of hospital doctors per bed and the number of general practitioners per head of population in the localities from which hospital admissions were drawn.
Based on a study of over 7.5 million discharges from 183 English NHS hospitals, over the four years 1991-2 to 1994-5, Professor Jarman and his team found that death rates varied across hospitals from 3.4 per cent of admissions to 13.6 per cent. They say that the most powerful predictor of this variation is the percentage of emergency admissions (in this study 60 per cent of admissions were considered emergencies). However, after adjustments for the percentage of emergency cases and for age, sex and primary diagnoses they found that the best predictors of hospital death rates was the ratio of hospital doctors per bed and the number of GPs per head of population served.
The authors report that England has one of the lowest number of physicians per head of population of all the member countries of the Organisation for Economic Cooperation and Development (OECD) - the ratio of doctors to patients in England in 1994 was only 59 per cent of the OECD average (1.6/1000 as opposed to 2.7/1000).
Jarman et al warn that considerable care should be exercised when interpreting hospital mortality data, but based on their findings they calculate that if the number of hospital doctors and general practitioners were to be increased in England, this is likely to be associated with a reduction in hospital mortality rates. They therefore conclude that the higher the doctor to patient ratio the lower the number of deaths but caution that their study is the first of its kind and their findings need to be validated
'"/>
Contact: Jill Shepherd
jshepherd@bma.org.uk
44-171-383-6529
BMJ-British Medical Journal
4-Jun-1999