Emergency readmissions: is 30-day the optimal time interval to capture hospital quality of care?

Meeting people

Authors: Anastasia Arabadzhyan & Adriana Castelli, Centre for Health Economics, University of York

Quality of hospital care is key for assessing hospital performance and designing incentive schemes, with emergency readmissions widely used as a quality metric.

In England, the 2011/12 National Tariff Payment System introduced the 30-day readmission rule to incentivise Trusts to reduce avoidable readmissions. The 30-day cut-off is also used in other healthcare systems such as the US, Germany, Denmark. However, some evidence suggests that it may not be the optimal threshold to capture hospital care quality. Indeed it may penalise hospitals for low quality of care that is, in fact, due to factors outside of hospitals’ control. In this study, we aim to answer two research questions:

1. Which is the optimal time to readmission threshold to capture quality of care in the English NHS?

2. Does this threshold change if we consider sites within Trusts rather than Trusts?

We focus on patients diagnosed with a disease of the circulatory system and construct the pool of index admissions, using the 2018/19 Hospital Episodes Statistics Admitted Patient Care dataset. We estimate basic and extended specifications of three multilevel logistic regression models (admissions within sites, admissions within Trusts, and admission within sites nested within Trusts) for each of the binary outcome variables, indicating whether an index admission was followed by a readmission at day 0 to 90. For each model, we calculate the proportion of total variance explained by the Trust and/or site-level random effects, by means of the intracluster correlation coefficient (ICC). A higher ICC means that hospital factors have more weight in explaining the variation, so the readmission indicator is more

reliable as a quality signal.

Preliminary results show that the ICCs exhibit a decreasing trend as the readmission window widens, with a slight variation in their dynamics depending on the model specification. Between site variation explains a larger proportion of the total variation in the outcome than between Trust variation, suggesting that policy interventions may be more effective at site level. Our approach to defining readmission window cut-off suggests using a 0-3 days readmission timeframe to ensure high reliability of emergency readmissions as a quality signal.

Key words: Emergency readmissions, hospital quality, circulatory system, National Health Service, England.


Adriana Castelli is a Senior Research Fellow at the Centre for Health Economics (CHE), University of York, which she joined in 2004 to work on the pioneering research project (joint CHE and the National Institute of Economic and Social Research, London) on the development and measurement of outputs and their quality, inputs and productivity for the English NHS.

Adriana leads the Efficiency & Productivity policy research area in CHE and has an extensive experience in the measurement of health care system productivity, in which she has gained both national and international recognition as a leading expert. She has collaborated with both UK and international partners. She has published widely in health economics, health policy and social science journals. Her work on the National and Trusts level Productivity of the NHS has had a major policy impact as follows: (1) influence on the Atkinson review of government output; (2) results used by DH in reporting to the House of Commons, Health Select Committees, the Public Accounts Committee and Public Expenditure Inquiries; (3) our 2013 report and interactive spreadsheet informed the 2013 Spending Review.

Adriana’s research interests include health policy reforms evaluation, health system performance with a particular focus on the development of productivity measures of

healthcare goods and services, quality of health care. She has experience on health systems of both high, and low- and middle-income countries.

Link zoom: https://unibocconi-it.zoom.us/j/92893642816?pwd=NWE1b0ozUVRwUnZMNDdXc2V1anBvZz09

Meeting ID: 928 9364 2816

Passcode: 142525

Lunch at the end of the Seminar. For those willing to participate in person, please write to erica.dugnani@sdabocconi.it before Thursday, April 20th.