Using routine data to monitor inequalities in an acute trust: a retrospective study

BACKGROUND: Reducing inequalities is one of the priorities of the National Health Service. However, there is no standard system for monitoring inequalities in the care provided by acute trusts. We explore the feasibility of monitoring inequalities within an acute trust using routine data. METHODS: A retrospective study of hospital episode statistics from one acute trust in London over three years (2007 to 2010). Waiting times, length of stay and readmission rates were described for seven common surgical procedures. Inequalities by age, sex, ethnicity and social deprivation were examined using multiple logistic regression, adjusting for the other socio-demographic variables and comorbidities. Sample size calculations were computed to estimate how many years of data would be ideal for this analysis. RESULTS: This study found that even in a large acute trust, there was not enough power to detect differences between subgroups. There was little evidence of inequalities for the outcome and process measures examined, statistically significant differences by age, sex, ethnicity or deprivation were only found in 11 out of 80 analyses. Bariatric surgery patients who were black African or Caribbean were more likely than white patients to experience a prolonged wait (longer than 64 days, aOR = 2.47, 95% CI: 1.36-4.49). Following a coronary angioplasty, patients from more deprived areas were more likely to have had a prolonged length of stay (aOR = 1.66, 95% CI: 1.25-2.20). CONCLUSIONS: This study found difficulties in using routine data to identify inequalities on a trust level. Little evidence of inequalities in waiting time, length of stay or readmission rates by sex, ethnicity or social deprivation were identified although some differences were identified which warrant further investigation. Even with three years of data from a large trust there was little power to detect inequalities by procedure. Data will therefore need to be pooled from multiple trusts to detect inequalities.