- Variations in health care in the NHS are a persistent and ubiquitous problem. But which variations are acceptable or warranted - for example, variations driven by clinical need and informed patient choice - and which are not? The important question is how to promote 'good' variation and minimise 'bad' variation. Variations in health care: The good, the bad and the inexplicable explores the possible causes of variation, shows the different ways in which variations can be measured, and analyses variations by PCT in rates of elective hospital admissions for selected procedures. The causes of variation are complex and inter-related - they may be affected by, for example, differences in geographical patterns of illness, differences in clinicians' behaviour, the effects of incentives in the financing of health care. These causes are mapped and discussed. The data itself is not always easy to collect and analyse. The report outlines the different ways in which variations can be measured and how these measures can be adjusted for need - age and gender, economic and social characteristics. To illustrate geographical variation, the report analyses rates of elective hospital admissions for common procedures across PCTs - eg, hip replacement, cataract removal, tonsillectomy. It also breaks the information down into : variations in common, clinically effective procedures, trends in variation, variation in day case rates, variation in low effectiveness procedures, variation in pre-operative health, preference-sensitive variations. Even after allowing for legitimate factors that affect rates of surgery, significant variations were found between PCTs - even among common interventions known to be effective. Policy-makers, commissioners and providers will benefit from the recommendations made for tackling this problem. The first step is to ensure that data is collected and analysed systematically. The Department of Health's Atlas of Variations may help to focus future reporting of variations; it would be good to extend this to cover the new population groups of GP commissioning consortia. Thereafter the recommendation is that a programme is established to identify causes of variation at local level and to assess which variations and causes have the most impact on equity, effectiveness, efficiency and patient health outcomes. The impact of the current reforms is considered: will GP consortia be able to implement shared decision-making as a way of establishing more appropriate level of warranted variation? Could the new role of the National Institute for Health and Clinical Excellence (NICE) help in generating the right kind of clinical evidence and guidance necessary for clinicians and patients on how to weigh up the trade-offs such evidence inevitably reveals?