Identifying Hospital Peer Groups According to Case Mix

Given the vast scope of clinical activity, evaluating the work undertaken in hospitals is problematic. . The number of possible diagnoses and interventions runs into thousands, and analysis at such a detailed level is not feasible. In recent years, various case mix classifications have been developed, combining treatments and diagnoses into a more manageable number of homogeneous groups. The aim of such classifications is generally that patients in a particular group should be similar clinically and also in the amount of healthcare resources that they are likely to consume, although groupings of patients with similar expected outcome also exist. In the United States of America, Diagnosis Related Groups (DRGs) were developed by Prof. Robert Fetter at Yale University (Fetter 1991) and have since been used widely in the management of Healthcare both in the USA and world-wide. DRGs were evaluated for use in the National Health Service (NHS) in England but they were found to be unsuitable, mostly because of significant differences in clinical practice. Therefore over the past four years the National Casemix Office has developed Healthcare Resource Groups (HRGs). Patients are allocated to groups on the basis of operative procedures performed, diagnoses (primary and secondary), age ·and in a few circumstances specialty. of treatment and discharge method. Version 2 of HRGs was released in September 1994, consisting of 528 groups which cover all in-patient hospital activity. It is intended that a third version will be produced-0{or release in April 1997. There are many practical applications of HRGs to healthcare management, as they are powerful in summarising activity whilst retaining valuable clinical and resource information. One of the obvious applications is in the contracting process of the NHS internal market. Every healthcare provider is currently required to use HRGs in the contracting process in one of three specialties (Ophthalmology, Trauma and Orthopaedics or Gynaecology). In April 1996 this will be extended to General Surgery, Urology and Ear/Nose/Throat, with the use of HRGs being mandatory in all six. The use of HRGs is discussed here in the context of analysing and comparing the case mix (Le. the mix of patients and treatments) of different hospitals, and developing clusters or 'peer groups' of hospitals with similar activity. Whilst such analysis is interesting in itself, it is also vital if fairer comparisons of efficiency and effectiveness between hospitals are to be made. It would be unfair to compare costs of treatment or outcome of care without first considering the types of patients being treated and the types of care provided. For example patients suffering from cancer will cost more to treat and will be more likely to die than patients in hospital for the removal of cataracts. The development of peer groups is one way of ensuring comparison of 'like with like'. There are of course other methods of allowing for variations in casemix. Standardisation is possible, whereby figures are adjusted to allow for variations. This is similar to the technique employed by epidemiologists in adjusting for age and sex variations across populations. Whilst this is