The UK NHS has a number of important strengths. Its costs are relatively low compared to the health care systems of other developed countries due in part to cash limited central budgeting. It is extremely popular with the electorate and surveys show overall satisfaction with the NHS despite some dissatisfaction with waiting lists and a public perception of underfunding. The NHS model of general medical care provided by independent contractors has been acclaimed as "a British success" (General Medical Services Council, 1983). The role of the UK GP combines providing primary care and acting as a gatekeeper to secondary care. This increases equitable access to care for the population and assists in cost containment. As a model, it is currently being emulated in other countries including Sweden and US Health Maintenance Organizations but, as in these countries, the UK primary care model has been evaluated poorly. There are of course continuing weaknesses in the UK health care system. There is insufficient knowledge upon which to base health care services and increase efficiency. In the future, if a knowledge-based health care service is to be created, a considerable amount of research and evaluation is required to identify "what works" in health care (i.e., what is effective) and also the cost effective ways of altering provider behaviour to maximise the amount of health gain which can be achieved using a limited budget. The NHS reforms created a lot of enthusiasm and energy but its effects are difficult to disentangle from the simultaneous increases in funding. There is little evidence from the UK or elsewhere that competition in health care produces efficiency or improvements in resource allocation. Evaluation is required to identify which of the reforms are increasing efficiency. Competition needs to be used with caution and recognised as a mean and not an end in itself. It is remarkable how both clinical practice and health policy reform, in the UK and elsewhere, is poorly evaluated. Medical practice varies substantially locally, regionally, and internationally, e.g., patients with similar age and stage of cancer receive very different levels of radiotherapy across Europe. For most interventions, the appropriate level of treatment may be asserted but is not based on cost effectiveness knowledge. Health policy analysts, like clinicians, make assertions about competition and other health care reforms which are value- rather than knowledge-based. Both groups of decision-makers should be more cautious, informing their choices with research rather than relying on unsubstantiated optimism!