Speed PresentationsDoes academic output correlate with better mortality rates in NHS trusts in England?
Introduction
The importance of academic medicine and research has always been emphasised in medicine. Recently, the National Health Service (NHS) has publicly expressed the importance of research in The Handbook to the NHS Constitution,1 and stated that ‘All NHS organizations must play their full part in supporting health research,’ in the most recent Operating Framework for the NHS in England.2 The NHS has also developed organizations such as the National Institute for Health Research, and encouraged strategic health authorities to support innovation and aid in developing research opportunities within the NHS.
Academic medicine has a responsibility for maintaining and producing high standards of health care. Doctors (both in training and who have completed training) are encouraged to publish academic work in order to further medical expertise and also as part of professional development. It has become mandatory for many training specialities.
However, this does raise the question as to whether doctors and institutes that are involved in academic medicine and research have better outcomes for their patients. The more specific question of whether the academic output of an institute or trust affects patient care has only been studied once in the literature. Pons et al.3 looked at in-hospital risk-adjusted mortality for acute myocardial infarction and congestive heart failure in 50 acute Spanish public hospitals, to compare a weighted citation ratio. There were a number of exclusion criteria including at least five citable papers per year in the field of heart disease. Pons et al. found a low-to-moderate negative correlation between risk-adjusted mortality and the weighted citation ratio: −0.43 (95% confidence interval −0.17 to 0.63) for congestive heart failure and −0.37 (95% confidence interval −0.10 to −0.59) for acute myocardial infarction.
Since the Bristol inquiry, there has been increasing accountability of health care to the public.4 Dr Foster Health is an NHS performance monitor available to the public. It has become a joint venture with the Department of Health and the Dr Foster Unit at Imperial College London. Dr Foster Health annually publishes The Hospital Guide and hospital report cards. These reveal adjusted mortality rates and more specific data regarding stroke, orthopaedic care, urological care and the recording of safety incidents.5
Dr Foster Health uses hospital standardized mortality ratios (HSMRs), which were first described in 1999 in the NHS6 and have since been used internationally.7 The data for calculation of HSMRs are gathered from Hospital Episode Statistics Online8 and are collected quarterly. The HSMR itself is the ratio of the actual number of deaths in a hospital to the expected number of deaths for that hospital for conditions accounting for 80% of deaths. Adjustments are made for the case mix in terms of: age group; sex; emergency or elective admission; primary diagnosis number of emergency admissions in the previous year; post code of patient admission to a palliative care speciality; and comorbidity in terms of a Charlson index.9
Section snippets
Methods and materials
A retrospective observational study was undertaken to compare the number of citations for each individual acute hospital NHS trust in England with HSMRs. Citation numbers were obtained from the Medical Literature Analysis and Retrieval System Online database (Medline), which is maintained and provided by the US National Library of Medicine (http://www.nlm.nih.gov/). It contains bibliographic information from many healthcare-related academic journals.
Medline was queried using the Internet-based
Results
From 2006 to 2010, 37,510 citations were listed on Medline for all 147 NHS acute hospital trusts in England. The total numbers of citations per year are shown in Table 1.
HSMRs were obtained from Dr Foster hospital report cards.5 The inverse of these ratios was calculated so that higher values would correspond to better mortality rates.
Spearman's rank analysis was performed to identify any correlation between citations per admission and the inverse of four types of mortality rate: high-risk
Discussion
This preliminary study shows a significant, albeit weak, correlation between overall mortality rates and academic output in NHS trusts in England.
Why should we think that the academic output of an institution or trust affects, or is correlated with, better patient outcomes or mortality rates? A hypothesis would be that institutions that are publishing more frequently have healthcare staff that are more involved in academic medicine and the latest guidelines and practices. Some work has been
Conclusion
This study found weak correlation between the academic output of a trust and their mortality rates. However, as a method, it could be developed and modified to allow more detailed analysis of specific departmental and speciality academic/research output and their respective morbidity and mortality data using information sources such as Dr Foster Health. The results of this could shape and influence the importance and role of academia in general medical and speciality training.
Ethical approval
None sought.
Funding
None declared.
Competing interests
None declared.
References (17)
- et al.
Do the investigative sites that take part in a positive clinical trial translate that evidence into practice?
Am J Med
(2002) - et al.
Should meta-analysts search embase in addition to medline?
J Clin Epidemiol
(2003) The handbook to the NHS constitution
(2010)The operating framework for 2010/11 for the NHS in England
(2009)- et al.
Is there an association between the quality of hospitals' research and their quality of care?
J Health Serv Res Policy
(2010) Learning from Bristol: the report of the public inquiry into children's heart surgery at the Bristol royal infirmary 1984–1995
(2001)- Dr Foster Hospital Guide. London: Dr Foster Research Limited. Available at: www.drfosterhealth.co.uk; 2011 [last...
- et al.
Explaining differences in English hospital death rates using routinely collected data
BMJ
(1999)