Measuring NHS activityBMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.7002.454a (Published 12 August 1995) Cite this as: BMJ 1995;311:454
- Allyson M Polloc,
- Neil Vickers
- Senior lecturer, public health sciences Research assistant Department of Public Health Sciences, St George's Hospital Medical School, London SW17 0RE
Admission rates are misleading
EDITOR,--We recently analysed five years of statistics on hospital episodes for all residents in the South Thames region admitted with a diagnosis of colorectal cancer between 1989 and 1993. Using date of birth, age, sex, and postcode as patient identifiers, we were able to link finished consultant episodes to admissions to calculate both the total number of admissions and the total number of patients treated. Data on 35230 finished consultant episodes were obtained; these related to 20434 patients, who had 32364 admissions for colorectal cancer. We then examined trends in hospital activity over time, using three different measures of activity: finished consultant episodes, admissions, and patients. Finally, we derived the ratio of finished consultant episodes to admissions and patients treated (table).
The number of admissions (day cases and ordinary admissions) increased 98% between 1989 and 1993 (table). The number of patients admitted, however, rose by only 14%. Most of the increase in finished consultant episodes and admissions was therefore due to repeat admissions and readmissions of the same patients. In 1989 the ratio of all admissions and finished consultant episodes to individual patients was 1.23 and 1.33 respectively, but by 1993 these ratios had risen to 2.13 and 2.24 respectively. Thus in 1993 there were 2.24 finished consultant episodes and 2.13 admissions per patient treated.
The implications of this for contracting are threefold. Firstly, admission rates are often used as a proxy measure for the population's access to treatment and care, but the paradox is that an area with high admission rates may admit few patients (with numerous readmissions) while, conversely, areas with low admission rates may admit many patients but with fewer readmissions and repeat admissions. Secondly, because the admission or finished consultant episode includes day cases it may give a spurious picture of the use of resources and patterns of care. Thus an area with low admission rates may have higher costs because more patients are admitted as inpatients; conversely, an area with high admission rates where many day case procedures are undertaken may have low costs. We are undertaking further analysis to ascertain how admissions relate to actual treatment of colorectal cancer. Thirdly, our findings support the Radical Statistics Health Group's conclusion that the increase in the number of finished consultant episodes in recent years cannot be used to support ministers' claims that more patients are being treated.1