Published: 19/08/2004, Volume II4, No. 5919 Page 23

Readmission rates are used in many healthcare systems as a measure of quality of care. Two measures of readmissions are included in the balanced scorecard used by the Healthcare Commission in assessing NHS acute trusts for star-ratings. These are a broad measure of readmissions for adults, and a readmission rate for people with broken hips.

A wide range of factors relevant to quality of care can lead to higher readmission rates. The quality of care given both surgically and on the ward in rehabilitating the patient before they are discharged is one factor. Pressure to discharge patients early may be another. The level of support provided to patients at home after discharge may be yet another. Understanding the pattern of readmission rates can be important in interpreting these. For example, large numbers of patients being readmitted within the first few days following discharge would suggest that early discharge may be the problem.

Readmission rates for trusts can be calculated from the national inpatient records held as hospital episode statistics. By linking spells of care in different hospitals it is possible to calculate the number of patients discharged from one hospital who are readmitted as an emergency to any other accident and emergency unit in the country. The measure normally used is the number of patients readmitted as emergencies within 28 days of discharge.

Dr Foster has calculated readmission rates for a range of diagnoses and procedures for all acute trusts. The first chart shows the distribution of readmission rates for people who have had a hip replacement operation, taken from all hospitals performing more than 100 such operations over the three years to March 2003. The figures range from 2 per cent to 12 per cent, with the average at 7.5 per cent.

Standardised readmission ratios are calculated as a ratio of the observed number of readmissions to the expected number (multiplied by 100). Thus a figure of 120 means the number of admissions was 20 per cent higher than expected given the age, sex and deprivation of the patients.

The second chart shows the distribution of standardised readmission ratios by size of unit in terms of the number of procedures performed. The units where the readmission ratio is significantly higher or lower than average are marked in yellow and purple respectively.

Most of those with significantly high or low results are among the smaller units where the spread of outcomes is wider. However, there are a number of larger units that also have significantly high readmission ratios.

Comparing readmission rates between diagnoses and procedures shows little correlation: units with high readmission rates for hip replacement do not show any great tendency to have high readmission rates for stroke patients.

But there is a greater consistency looking at departments, particularly among larger units. The last chart plots readmission rates for knee replacement surgery against hip replacement surgery for units that do more than 300 hip replacements a year. Units with high readmission rates for one procedure tend to have a high rate for the other.

Roger Taylor is research director of health information specialists Dr Foster.