What does the evidence show on NHS quality? Leatherman and Sutherland scoured the NHS for data to show what is happening with quality of care. They may have assembled more data than anybody else, but the evidence still falls far short of a complete, validated, interpretable, uncontested picture–even though they entered rooms where people had for years been collecting data that nobody before had ever asked to see. The data that are available show a mixed picture of improvement, stasis, and deterioration. Access to care is the most politically contentious aspect of quality, and the number of patients waiting 12 months or longer for admission to hospital has fallen from 50 000 in 1999 to only 73 in the fourth quarter of 2002-3. But a fifth of people still wait for more than six months. The number of operations cancelled at the last minute, which is frustrating for both patients and staff, increased from around 56 000 in 1998-9 to 68 000 in 2002-3. Waiting times in emergency departments have improved, but a third of patients still wait more than two hours. Access to general practitioners seems to have deteriorated, with 13% of patients waiting more than two days in 1998 and 23% in 2003. Meanwhile, calls to NHS Direct increased from 110 000 in 19989-9 to over 6 million in 2002-3. Effectiveness also shows a mixed picture. The percentage of children being immunised–particularly against measles, mumps, and rubella–has declined, but the proportion of hospitals giving thrombolysis to 75% of patients with heart attacks has increased from 24% to 45%. Other targets of the national service frameworks for heart disease and cancer are being met, and mortality from both circulatory disease and cancer is falling. We don't know, of course, whether these changes–both favourable and unfavourable–would have occurred without billions of pounds of quality initiatives. Nor do we have good data on what is happening with conditions not covered by national service frameworks. There must be anxiety that services that were not prioritised may have deteriorated. The capacity of the system has generally improved. The number of nurses increased from 256 000 in 1997 to 291 000 in 2002, but the numbers of general practitioners and consultants have not increased anywhere near as fast. Facilities for caring for patients who have strokes have increased, but three quarters of patients still spend less than half their time in a stroke unit (down from 83% in 1998). Ultimately of course it must be patients and the public who determine the quality of care. Data purporting to show improvement will be worthless if it doesn't feel that way to patients and the public. The percentage of the public who think that the NHS needs a “complete rebuild” or “fundamental change” fell from 78% in 1988 through 69% in 1998 to 72% in 2001. Dissatisfaction with the system (measured in this way) thus seems to be high and has increased since the quality initiative began. Complaints to the health service ombudsman have increased from 2500 in 1999-2000 to 4000 in 2002-3–although the number of investigations has declined and the overall number is a minute proportion of all patient encounters with the NHS. Few other patient measures allow comparison across time, but a quarter of patients did not feel that “hospital staff did everything they could to control pain,” and 40% said they were not told about the danger signals regarding their illness or treatment to watch for after they went home. Nevertheless, 80% or more of patients felt that they were “always” treated with respect and dignity when in hospital or seeing their general practitioners. Leatherman and Sutherland not only reviewed published reports and gathered data but also spoke to over 50 “experts and leaders” (I was one of them), and their conclusion is cautiously positive. They see a willingness to admit problems, a favourable context for policy, reasonable resources and organisational capacity for improving quality, and a multipronged strategy. They don't, however, see a common understanding of the “state of quality” (hence the NHS consultant who emailed me, insisting it is lamentable), sufficient involvement of public and patients, adequate leadership at all levels, and enough involvement of the clinical professions. Greater involvement of public, patients, and professions will be essential for long term success. Footnotes See News p 1250 Competing interests RS was interviewed by the authors of the report but had no say on its conclusions. The BMJ Publishing Group, of which he is the chief executive, has a contract with the NHS to supply Clinical Evidence and it hopes to finalise a contract to supply evidence based information to NHS Direct. RS is paid a fixed salary and will not benefit or lose financially from the contracts. The full competing interests of RS can be accessed at http://bmj.bmjjournals.com/aboutsite/comp_editorial.shtml>.