Patient-centred care is the latest fashion in health policy, yet attempts to make health services more responsive to the people they serve have been on the agenda for some years. The emphasis has shifted though and is now leaning towards using surveys to gain feedback from patients and incorporating the results into performance assessment frameworks. Do these surveys work?
The Picker Institute has some considerable experience in the matter. We use carefully designed surveys to obtain detailed reports of patients' experiences of specific dimensions of patient care, including information and communication, co-ordination of care, respect for patient preferences, emotional support, physical comfort, and involvement of family and friends.
We ask patients whether or not certain processes and events occurred during the course of a specific episode of care. In other words, we want to know whether such or such an incident happened or not - for example, whether patients knew which doctor was in charge of their overall care. These surveys have been used since 1987 in hospitals in the United States and since 1997 in Germany, Sweden, Switzerland and the UK.
Analysis of recent results reveals many common problems and wide discrepancies in the quality of hospital care in these countries. In all countries the most commonly reported problems concerned communication about clinical issues. These include insufficient information in the accident and emergency (A&E) department, insufficient information about tests and treatments, and not enough involvement in treatment decisions. (See graph p 30.)
The differences between countries in the extent of problems reported were striking. As a rough rule of thumb, in our view, a problem score of more than 20% requires attention. The Swiss surveys achieved problem scores of more than 20% in only seven cases out of 44 possible problems. The figure in Germany was 15 out of 44, among Swedish patients the problem scores exceeded 20% in 18 out of 41 cases (three questions were excluded from the Swedish surveys), and in the US it was 20 out of 44. The equivalent figure in the UK was the highest, at 24 out of 44 in answer to specific questions such as whether patients were given sufficient information to help look after themselves after they leave hospital. But when asked to rate the overall level of care they received, only 9% of UK patients surveyed classified it as "poor" or at best "fair" (see graph).
One has to be very cautious when making international comparisons with these figures. Considerable effort was made to ensure that our questionnaires would be interpreted the same way in the different countries, but different expectations on the part of patients in each country may still have been responsible for some of the discrepancies.
Another note of caution is that, with the exception of the US, the number of hospitals surveyed in each country was small and may not be representative.
Nevertheless, it is clear from these results that problems exist in all the hospitals surveyed. The same topics were given high problem scores across the board - for instance lack of information to help patients look after themselves after discharge from hospital was a problem everywhere - but the rates varied between countries, with Switzerland achieving the best results and the UK the worst.
Most surveys tend to ask patients to give satisfaction ratings about treatment. The Picker approach of asking factual questions about events helps to make results easier to interpret. After all, having to wait more than 15 minutes after pressing the call button before someone comes to help is a fact, not an opinion.
These surveys are consequently much more useful when it comes to setting priorities for quality improvement since they can help to pinpoint any problems and help identify precise action.
Interestingly, few patients respond critically when asked to rate overall quality of care, even if responses to specific questions paint a negative picture. Our surveys confirm that patients often want more information than they are given and that a significant proportion would like to be more actively involved in their own care. Surely more should now be done to address these problems. The surveys also tell us more about systemic problems. One reason for the differences observed between, say, Switzerland and the UK, may lie in the fact that British hospitals are more under-staffed. This point is glaring, since UK respondents to our surveys were seven times more likely to report problems with the availability of doctors (30%) than the respondents from Switzerland (4%), and 15 times more likely to report such problems with nurses (30% as against 2% in Switzerland).
Of course, Switzerland spends nearly twice as much per capita on healthcare. But it may also be the case that Swiss hospitals place greater emphasis on improving the quality of patients' experiences. All Swiss hospitals are now required to survey their patients on a regular basis, and hospitals in England will have to carry out annual patient surveys from 2002.
Knowing that a problem applies in your hospital or your ward can motivate change. Comparisons between countries, between hospitals, or even between departments within a hospital, can provide useful benchmarks against which to judge progress too.
People who commission Picker surveys are encouraged to compare their results against the best from our surveys in Europe and the US, giving them a ¡§stretching target¡¨ for improvement. They can then aim to be better than the best. And if the results are made public, they may prove to be a particularly powerful lever for change.
Note on the Methods
The initial development work for the surveys reported here was carried out mainly in the USA.
The questionnaires used in the four European countries were for the most part direct translations of the US questionnaires, which were then tested with patients for cultural and linguistic relevance and comparability of meaning in each of the countries and adapted where necessary.
Responses have been translated into problem scores indicating the percentage of patients reporting a less than satisfactory response to each of the specific questions. Data came from postal surveys carried out in acute care hospitals in each of the countries. Questionnaires were mailed to patients¡¦ home addresses within one month of discharge from hospital.
The analysis was restricted to data collected from in-patients over a 12-month period in each of the countries.
Completed questionnaires were received from 2 249 patients from five hospitals in the UK, 7 163 from nine hospitals in Switzerland, 2 663 from six hospitals in Germany, 3 274 from nine hospitals in Sweden, and 47 576 from 272 hospitals in the USA.
Response rates varied but were greater than 50% in all cases.
• Cleary, P. D., Edgman-Levitan, S., Walker, J. D., Gerteis, M. and Delbanco, T. L., "Using patient reports to improve medical care: a preliminary report from 10 hospitals", Quality Management in Health Care, Vol. 2, 1993.
• Coulter, A. and Cleary, P., "Measuring and improving patients' experiences: How can we make health care systems work for patients?", OECD conference in Ottawa: http://www1.oecd.org/els/pdfs/HEAOTTDOCA010.pdf
• OECD Health Data 98: A comparative analysis of 29 countries, 1998.
©OECD Observer No 229, November 2001