Measuring the Quality of Primary Health Care

Measuring the quality of primary medical care

Introduction
The last decade of the twentieth century has seen an enormous increase in interest in measuring and monitoring the quality of medical care. There are two main reasons for this; first the importance of measurement for improving quality and secondly the need for greater explicit accountability of health professionals and health systems. For many branches of medicine, primary care included, this focus on measurement was new. In this chapter, we describe some of the key issues around measurement in family practice. We start by highlighting some of the political, societal, and professional reasons for focusing on measurement. We then provide a framework for measuring quality in family practice and explore the development and role of quality indicators in this process. We describe some of the ways in which indicators can be used and ways in which we believe they should not be used. Finally, we emphasize the need to move beyond what is easily measured in order to truly reflect the nature and role of general practice.

Why should quality of primary care be measured?
The first reason for measuring quality of care was the recognition of how much variation existed in the quality of care provided. Initially, researchers focused on technical issues, for example, demonstrating large geographic variation in rates for surgical procedures. More sophisticated methods were then developed to assess the appropriateness of these procedures.(1) These showed that a substantial proportion of procedures carried out on patients did not need to be done, while other patients who would benefit from the procedures did not receive them.(2) Inappropriate over-use, under-use, and misuse of medical procedures appeared to co-exist, and this could be found wherever the issue was studied.(3,4 and 5) Although this research was initially done in relation to specialist practice, considerable gaps are also found between the care that primary care physicians aim to give, and the care which they actually deliver.(6)
Recognition of the phenomenon of variation is not new. Over 2000 years ago, Hippocrates commented that:
In acute diseases, practitioners differ so much among themselves that those things which one administers thinking the best that can be given, another holds to be bad… and similar differences are to be found in the examination of entrails.
Hippocrates, 460–377 BC
Variation in medical practice is not of itself a bad thing. However, it may be a bad thing where variation is associated with demonstrable deficiencies in quality of care. The demonstration of variation in medical practice led in the late 1990s to the widespread realization that some aspects of medical care were frankly unacceptable. In several countries, there have been high-profile health scandals and medical error is now recognized as an important cause of harm to patients.(7,
In the public's mind there has been some confusion between the research showing widespread variation in medical practice and the presence of the very small number of doctors whose practice is totally unacceptable. Nevertheless, both of these issues have combined to produce demands for greater accountability from the medical profession. Doctors needed to be able to assure the public that the quality of medical care they provided was of a high standard. This was in line with other societal changes where citizens expected demonstrable high standards from both public services and private industry. The medical profession has not been immune from these expectations.
Some of the drivers for change have been doctors themselves. The rise of evidence-based practice has given doctors the means to examine their own care. Their own professional ethic of doing as well as they can for their patients has been to a large extent been behind the widespread move in the last decade of the twentieth century for doctors to participate in medical audit and quality circles, most often within their own practices or medical groups.
The relative influence of these various factors has played out differently in different countries. In the Netherlands, there has been a strong professional movement to develop national clinical guidelines, and to encourage professional development activities around these guidelines. This has been a ‘bottom up†approach, led by the doctors themselves. In the United Kingdom, on the other hand, government has taken a much more central role in attempting to introduce a national system of mandatory quality reporting and assessment. In the United States, the link between quality improvement and cost containment has become blurred, with the public regarding attempts to reduce the inappropriate over-provision of care (e.g. by strengthening the gate-keeper role of the primary care physician) as depriving them of choice and a reduction in quality.
The impact of these international societal changes on the medical profession has been profound, and doctors struggle with changes in the way society views them. The old English adage ‘Trust me, I'm a doctor now rings hollow, and doctors wonder what their new place in society will be. It is clear that a new type of professionalism is emerging, which Irvine defines as:

• Clear professional values compatible with public expectations.
  
• The use of explicit standards.
  
• The adoption of collective as well as personal responsibility for observing standards of practice.
  
• Systematic process for showing that doctors are up to date and perform well.
  
• Effective local medical regulation, including swift and effective machinery for dealing with dysfunctional doctors.
  

Derived from Irvine(9)
Given this background, there are a number of reasons for measuring quality of care. The main ones relate either to improving quality, or to providing external accountability. Another important reason is to provide consumers or payers (e.g. employers) with information to give them choice about the health care they seek.
Where quality of care is demonstrably deficient, the response of most professionals will be to try to improve it. Thus, for example, clinical audit is an activity undertaken by professionals in order to assess the quality of care they are providing, and, if necessary, to improve it. This is an activity that takes place entirely within the practice or workplace of the individual doctor or group of doctors. Improving quality of care is the first reason for measuring quality of care, and it is the one that is likely to be uppermost in the minds of doctors. Measuring quality is an essential prerequisite to improving care:(10) improvements are unlikely to take place unless data are available on what is already being achieved.
The second reason for assessing quality of care is to demonstrate to the public that acceptable standards of care are being provided—public accountability. This is a driver that is most likely to be given prominence by governments and payers for health care who are looking for value for money and want to be able to demonstrate to the population that their health service is delivering a high quality of care.

Continued below......

Dr. Nayyar R. Kazmi
Public Health Specialist and Site Admin

What aspects of care should be measured?
The first section of this chapter outlined the political, societal, and professional forces behind the drive to measure the quality of medical care. What then should be measured?
This question is a particularly taxing one for primary care. For some branches of medicine, there are fairly clear outcomes. In cardiac surgery, for example, communication with the patient and quality of the hospital food might both be measured, but the assessment of case-mix adjusted success rate, complication rate, and mortality are of prime importance. Most effort will, appropriately, go into assessing these latter aspects of care.
Primary care is different. The work of primary care is diffuse, and addresses a wide range of patients' needs. Quality in primary care is multifaceted and cannot be assessed in terms of a single or simple outcome. Assessment of quality, therefore, needs to be multifaceted in order to avoid an inappropriate focus on narrow areas and to prevent perverse incentives for improvement.
In thinking about what aspects of primary care should be measured and monitored, it is appropriate to start with what patients want most from their primary care physicians. This is summarized in Box 1.

Box 1 What patients want from primary care(11,12)

• Availability and accessibility, including availability of appointments, waiting times, physical access, and telephone access.
  
• Technical competence, including the doctor's knowledge and skills, and the effectiveness of his or her treatments.
  
• Communication skills, including providing time, exploring patients' needs, listening, explaining, giving information, and sharing decisions.
  
• Interpersonal attributes, including humaneness, caring, supporting, and trust.
  
• Organization of care, including continuity of care, coordination of care, and availability of onsite services.
  

There are a number of ways in which the domains of quality of care can be classified. The central domains are access and effectiveness—can patients get to health care and is it any good when they get there? From a general practice perspective, it is important to recognize that effectiveness includes both effective clinical care and effective interpersonal care in order to reflect the prominence of interpersonal care in what patients want. Both clinical care and interpersonal care have technical elements that can be taught and learned, and neither is sufficient on its own. Providing good clinical care is not sufficient to be a good general practitioner,(13) and good interpersonal skills cannot be a substitute for poor clinical skills. Providing good access and providing effective care cannot be delivered without there being good organization of care, and assessing practice organization and management may be an important part of the equation.
For populations, there are two other elements that are important—equity and efficiency. Efficiency is an important marker of quality of care for populations as inefficient care (e.g. prescribing expensive but ineffective drugs) may have opportunity costs for the care that can be provided to other patients. These domains of quality are summarized in Box 2.

Box 2 Domains of quality in primary care(14,15)

• Quality of care for individuals is determined by
  
• Access
  
• Effectiveness of care
  
• Technical care
  
• Interpersonal care
  
• Organization of care
  
• Quality of care for populations is additionally determined by
  
• Equity
  
• Efficiency
  

Any approach to assessing the quality of primary medical care needs to be aware of these domains of care. To focus on one risks devaluing other aspects of care. For example, doctors taking part in a major quality improvement scheme focusing on the quality of chronic disease management noted that access and interpersonal aspects of care sometimes suffered as a result.(16)

How can quality of care be assessed?
Following a description of the domains of quality that are important to primary medical care, the next question is how these should be measured. A wide range of methods are available.(17) To date, much quality assessment has been carried out entirely within practices—doctors have set out to find out about aspects of their care that could be better, and used this as a starting point for improving quality.(18) Medical audit(19) requires that the doctors of the practice set standards and review their own practice against those standards.


However, there is an increasing focus on using methods that more readily allow information to be shared outside the practice. They may focus on assessing competence (what a doctor is capable of doing) or performance (what he or she actually does). Consultations may be observed directly or on videotape (a requirement for gaining a licence as a general practitioner in the United Kingdom) or assessed indirectly from patient questionnaires (now a routine part of much US practice). The doctor's work can be examined by reviewing medical records (a regular part of peer assessment visits in the Netherlands) or by data provided for payers (a method used routinely in the United States). Sometimes, specific payments are tied to information about quality of care, for example, the attainment of specified levels of cervical cytology or immunization coverage in the United Kingdom.
When quality assessment is required for external review, it is most frequently measured by applying quality indicators. A quality indicator relates to a specific measurable aspect of care that is sensitive to change and can be applied retrospectively to assess quality of care.(20) Quality indicators also need to represent aspects of care that are under the control of the practitioner. Quality indicators highlight areas of performance that may require further investigation: they do not make definitive judgements about quality. Indicators are different from guidelines, targets, and standards (Box 3). In many cases, it is not appropriate to set targets for quality indicators—minimum standards may be set locally if appropriate but the onus for the majority of practices should be to show sustained improvement over time. This is more important than absolute levels of achievement, since some practices will have greater difficulty achieving specific targets, often for reasons outside their control (e.g. patient characteristics or lack of resources).

Box 3 Examples of guidelines, indicators, and targets or standards

• Guideline: Eligible women are offered routine cervical screening.
  
• Indicator: The proportion of eligible women who have had cervical screening carried out within the recommended period.
  
• Target or standard: The proportion of eligible women who have had cervical screening carried out within the recommended period should exceed 80 per cent.
  

Quality indicators can assess structure, process, or outcomes of care. In practice, outcomes are rarely suitable as quality indicators for primary care—they often occur long after the care given by the primary care physician, and may be confounded by factors outside his or her control. So, for example, ensuring that blood pressure is monitored and controlled is a more appropriate quality indicator for primary care of diabetes than the incidence of stroke or renal failure.
In practice, few indicators satisfy all the criteria for a good quality indicator, and many look more like guidelines–that is, they are more suitable for prospective guidance than retrospective assessment. In the appendix to this chapter, some examples are provided of quality indicators that could be applied to primary medical practice. These are mapped to the domains of care outlined in Box 2. In addition to these indicators, others are available relating to clinical care for some of the most common conditions seen in general practice.(21)
Apart from quality indicators, there are other established ways of approaching quality assessment in primary care. Colleges and Associations of family physicians have been among the leaders of promoting quality assurance systems for primary care. Available measures of quality in primary care include:
United Kingdom: The Royal College of General Practitioners. This has a number of schemes, including Membership by Assessment of Performance, Fellowship by Assessment, the Quality Practice Award, and Quality Team Development. These are described at www.rcgp.org.uk. The UK College is the only English-speaking college that published full details of all quality criteria and methods of assessment on its website. A range of quality indicators for the problems seen most commonly in British general practice are also available.(21) In future, general practitioners' remuneration in the United Kingdom is likely to depend on their performance against a broad set of quality indicators.
Australia: The Royal Australian College of General Practitioners publishes standards for general practice that can be ordered from its website. The website also contains downloadable guidelines for an ‘enhanced primary care program†and details of its practice accreditation scheme. These are available at www.racgp.org.au.
United States: The American Association of Family Practice website contains a limited number of clinical policy statements, relating especially to childhood disorders and immunization, www.aafp.org. The HEDIS measures, which are widely used in the United States are described in outline at www.ncqa.org and can be ordered from the National Committee for Quality Assurance website at www.ncqa.org. The US government also maintains a national guideline clearing house at http://www.guideline.gov/index.asp. Quality indicators relevant to primary care have also been developed by the RAND Corporation.(22,23 and 24)
Netherlands: The Dutch College of General Practitioners has been among the most active in taking a systematic approach to developing clinical guidelines for general practitioners, and disseminating them among its membership. They are published in two books,(25,26) and are currently available in English on the website of the college, www.artsennet.nl/nhg. From 60 of these guidelines about 140 indicators have been produced and are currently being validated in Dutch general practices.
Both the College of Family Physicians of Canada, www.cfpc.ca, and the Royal New Zealand College of General Practitioners, www.rnzcgp.org.nz, currently publish a small selection of web based guidelines.
In other cases, academic groups have led efforts to develop sets of quality indicators that could be used on a widespread basis, either based around interpersonal aspects of the consultation,(27) or by applying the principles of evidence-based practice.(28)
Patient evaluation is an important part of quality assessment, especially for interpersonal aspects of care and assessment of patients' access to primary care. A number of instruments are available for assessing patients' views. Those available in the public domain include:
The Primary Care Assessment Survey(29,30) (PCAS) has been used widely in the United States, and has evidence of both predictive validity and sensitivity to change.(31,32)
The General Practice Assessment Survey(33,34) (GPAS) was developed from the American PCAS for use in the United Kingdom. Its website (www.gpas.co.uk) contains the instrument, manual, national benchmarks, and software for analysis using both Excel and SPSS.
EUROPEP (www.equip.ch/groups/pep/europep.pdf) takes a similar approach to GPAS, and is available in 15 European languages.

Continued Below.....
Dr. Nayyar R. Kazmi
Public Health Specialist and Site Admin

Methods of developing quality indicators
In the past, quality indicators have most frequently been developed by experts meeting and forming a consensus about what care should be given and how it should be assessed. Sometimes, this can lead to strikingly inappropriate results, for example, when a group dominated by specialists produces measures for use in primary care. This process can be improved by deriving indicators directly from clinical guidelines. This method has been used extensively in the Netherlands where detailed guidelines have been produced from research evidence and professional consensus. The emergence of evidence-based practice offered an opportunity to introduce new rigour into the indicator development process, and some groups have explicitly concentrated on evidence in producing quality indicators.(25) However, the evidence base, certainly in terms of randomized controlled trials, is weak for much of the medicine which primary care physicians practice. So, methods are needed to combine both what is known from the literature (including but by no means exclusively trials) and professional opinion from appropriate sources—for example, primary care practitioners with particular expertise in an area. These approaches can be combined systematically to develop quality indicators for primary care. The importance of making use of both evidence and professional opinion is well demonstrated by the finding that of a set of quality indicators produced for asthma, angina, and diabetes management in primary care, only 26 per cent were judged by the panelists to have a strong evidence base.(35) Furthermore, some important aspects of care may simply be unmeasurable with methods currently available—it is, therefore, important to recognize the extent to which any set of quality indicators can only offer a partial window on quality of care.
When quality indicators are developed, it is important that the details of the method used should be explicit. There should be a description of how evidence was collected from the literature, when the literature was searched (quality indicators may change as new evidence accumulates), who assessed the evidence, and how they developed indicators from the information they had available. There is no perfect method, and the way in which the process is carried out can have a substantial impact on the end result.
The need for caution when using quality indicators While there may be some benefits to measuring quality of care, there are also substantial potential drawbacks. Before thinking about using any quality indicators, there are some very important issues to bear in mind:
Any set of quality indicators for primary care that sets out to be comprehensive needs to reflect the many and varied functions fulfilled by primary care. This will prevent an inappropriate focus on narrow areas of activity. However, if indicators attempt to be comprehensive, they will almost certainly be too cumbersome for practical use.
There may be substantial data collection costs associated with measuring the quality of primary care without any clear benefit in terms of quality improvement. Measuring quality on its own does not improve care. Therefore, there is little point in setting out to assess quality unless there is some sort of infrastructure to ensure that the data are used.
Some indicators will inevitably be more ‘evidence basedâ€. Even for disorders like angina, diabetes, and asthma, the proportion of quality indicators that can truly be evidence based is quite small. However, all indicators should have face validity, for example, reflecting professional consensus in medical practice.
The data available for assessing quality, for example, clinical data in medical records may vary greatly between physicians and practices. This can make it very difficult to compare data between sites.
Practices vary considerably in the ways in which they are organized and how they provide services for their patients. A number of the indicators may not be appropriate for every practice, or for practice in every country.
Within localities, it can still be difficult to interpret apparent differences between the performance of physicians or practices or to know what, if any, action needs to be taken.
It is important to remember that these are indicators of quality, not definitive judgements. As such, caution is required when making comparisons between practices or over time. The great strength of indicators is in promoting discussion about the possible reasons for any differences found and ways of improving the quality of care provided.
Quality indicators will always be subject to gaming or mis-representation by some individuals or organizations. This is a particular risk when indicators are used inappropriately to make judgements, rather than to encourage engagement with a quality improvement process. The risk of gaming should be recognized and minimized where possible but it should be clear that there is more to be gained than to be lost by trusting primary care professionals to use quality indicators honestly.

What information is routinely available for assessing quality care?
Most health care systems have had little or no information available on quality of care until quite recently. The United States has been most active in developing systems for providing information on clinical activity that can be used to assess quality of care. Although the United States is relatively data rich in terms of what goes on in primary care, such data have in the past been almost exclusively collected for the purposes of billing. Information collected for financial reasons is unlikely to be well designed for quality assessment, but in recent years, systems have been developed (e.g. HEDIS, discussed earlier) which allow individual health plans to be compared by those purchasing health care (mainly employers) or those who use them (the patients). Report cards that describe the quality of care given by health plans are now widely available in the United States, though their validity and reliability are uncertain.(36)
The UK government has also taken a lead in disseminating information on quality of care. Unfortunately, the first attempts were based entirely on routinely available data. Since these data were not collected with quality in mind, it was hardly surprising that they were a very poor reflection of quality of care—with one commentary regarding ‘the remedy as worse than the maladyâ€.(37) These indicators included aggregates of hospital admission rates for conditions where good primary care might be expected, in theory at least, to prevent admission (e.g. diabetes, epilepsy, and heart failure), and a number of prescribing indicators. Where routine data for payment do relate to quality of care, then such data may allow widespread comparisons to be made validly. Also in the United Kingdom, target payments have been made to primary care practitioners for some years for meeting cervical cytology and immunization targets—measures that have broad acceptance as being at least partial measures of quality of care in those clinical domains, albeit crude ones.
Where this field moves in future remains to be seen. If payers for medical care (e.g. insurance companies or governments) are serious about wishing to assess quality of care on a routine basis, then information systems could certainly be designed to provide such information.(38) Whether the benefits in terms of quality improvement would be worth the data collection costs remains to be seen.
One further source of information on quality of care relates to the licencing of individual doctors. For reasons outlined earlier in this chapter, many developed countries are developing systems for accrediting or revalidating their doctors.(39) While re-certification has been practiced in the United States for some years, systems are now being introduced in the United Kingdom, the Netherlands, Australia, Canada, and New Zealand. The United States is the only country to rely heavily on external testing. The other countries are developing systems that rely more on the demonstration of planned professional development (e.g. developing and following a personal learning plan), and demonstration of standards achieved (e.g. clinical audit). In the Netherlands and Australia, this is combined with peer assessment carried out through a series of practice visits.(40) As with other aspects of quality assessment, it remains to be seen what mixture will provide the best balance in terms of assuring the quality of medical care at an acceptable cost.

To whom should information on quality of care be available?
Given the increasing moves towards measuring quality of care, there has been considerable debate about who the information should be made available to. The United States has led on making information available to the general public. This is ostensibly to provide external accountability, and to give the public information so that they can take quality into account when choosing where to get their medical care. However, even in this most consumerist of societies, the public has actually taken little notice of information on
quality of care. Moreover, doctors have been highly mistrustful of the quality of data provided. Thus, from either the doctor or the patient end, making information on quality of care publicly available appears to have little impact at present. However, providing this information to management in health care organizations (e.g. health plans) appears to have had greater effect, and it is perhaps at this level that information will have its greatest impact.(41)
The United Kingdom is also moving down the road of making information on quality of care routinely available outside the individual practice. In 2000, 50 per cent of primary care groups and trusts were making or planning to make comparative information on quality of care available in identified form to the practices within their group, and 9 per cent were planning to make such information available to the public.(42) However, patients may differ from professionals in the aspects of care that they consider important for quality and the benefits of making such information available,(43) either in terms of improved quality or in terms of public confidence in medical care remain to be seen. Both public and doctors are at present suspicious of the motives and value of making such information available.(44)

Conclusion
Primary care physicians cannot escape the current trend to want to measure quality. For the reasons outlined in this chapter, it is desirable that general practitioners engage with and influence the measurement process. Nevertheless, the trend towards explicit measurement poses considerable problems for a discipline such as general practice. Much of what general practitioners do is diverse and difficult to define. Good communication, attention to interpersonal care, integration of care, and coordination of care are all important aspects of what general practitioners do. Yet, they are much more difficult to measure than more mechanical tasks such as compliance with chronic disease management protocols. Furthermore, reducing general practice to identifiable component parts risks misunderstanding the patient focus of general practice care(45) and hence missing the benefits of a comprehensive generalist service.
The measurement culture, therefore, creates a dilemma for general practice. Will the things that are measured simply be those that can be measured easily? If so, any accountability based on quality indicators will be a very partial activity. Will the unmeasured parts then be forgotten and devalued? Alternatively, should general practice develop measures of some of its fundamental and defining features? In doing so, there is a risk that this will simply perpetuate the reductionist environment in which health services throughout the world increasingly operate. We believe that crucial aspects of what general practitioners do, such as providing high-quality interpersonal care, must be included in future performance indicators for primary care. This would do much to preserve their importance at a time when many fear that traditional values are being lost from general practice. In rapidly changing times, general practitioners need to keep sight of their core values that will ensure that the contribution of general practice to patient care is appropriately valued and rewarded.

Examples of quality indicators for family practice

Access to care
For urgent matters, patients should be able to see or speak to a doctor or nurse on the day of the request.
Routine appointments with doctors and nurses should be booked at intervals of no less than 10 min.
Effectiveness of care: clinical care
The childhood immunization rate should be over 90 per cent, or if below, efforts should be made to improve the rate year on year.
Patients with coronary artery disease should be advised to take 75–150 mg aspirin per day indefinitely unless contraindicated.
HbA1c levels should be checked at least every 14 months as part of an annual check in all diabetic patients.
Drug therapy should be offered to all patients with sustained blood pressure of more than 160/100, if non-pharmacological measures have been applied.

Effectiveness of care: interpersonal care
Practices should seek information about patient/carer experience at least once every 3 years, using a recognized technique, and the results should be made available to the practice population.
A report of the number and type of complaints made to the practice, and of the response of the practice should be available for scrutiny.

Organization of care
Practices should be organized so that members of the primary care team have an opportunity to meet at regular intervals, appropriate to the team needs.
Patient records should be in chronological order, with a clear summary of key conditions and medication.
Practices should have a written management guideline for conditions identified as priorities, for example, diabetes, enduring mental health problems, coronary artery disease, hypertension and these guidelines should be updated as required.
Practices should have and use a prescribing formulary (developed within or outside the practice).

Efficiency
The generic prescribing rate of individual doctors should be known and monitored.
The rates of admission and referral to specialists should be known and discussed.

Equity
Practices should accept all patients living within their practice area or be able to provide an acceptable explanation for refusal to do so: practices should not systematically refuse to register particular categories of patient, for example, the homeless, the severely mentally ill, or those with problems of substance or alcohol misuse.