Measuring and improving quality of care in family practice

All British Columbians should have family physicians to provide them with optimal quality of care. Those who currently have family physicians are likely receiving that care now, but as the auditor general of British Columbia pointed out in February 2014, “the current physician compensation models are not linked to quality (so cost-effectiveness cannot be determined).”[1] The resulting recommendation was to rebuild physician compensation models so they align with the delivery of high-quality, cost-effective physician services. In essence, this is a move to a pay-for-performance system. This is justifiable in terms of transparency of the use of taxpayer funds, but the critical question is how we define performance. Discussions so far have focused largely on quantitative measures such as the number of procedures undertaken and other measures put into a core dataset called the Physician Quality Assurance Steering Committee Core Data Set.[2] Without a much more comprehensive measurement system accompanied by an extensive practice improvement process, a counting approach is unlikely to improve quality or cost-effectiveness.

The assessment of performance quality within family practice is not new. In 1985 the UK set up a voluntary self-assessment process that looked at four principal areas of performance: professional values, accessibility, clinical competence, and ability to communicate. This approach was designed to help practising physicians identify improvements they could make to their practice as a whole. Introduced before computers were commonplace in family practice, the approach sought to look at the practice overall. In that sense, it was ahead of its time—emerging before evidence on access and communication was available—and is in contrast to approaches that rely on counting procedures.[3]

The approach recognized that health care delivery is limited in its ability to prevent and cure disease. Diseases often materialize in response to exposures such as cigarette smoke, poor nutrition, lack of activity, and crowded living conditions.[4-6] These are often outside the control of the health care system, and it means that health system performance assessment “should relate to those aspects of care which can be altered by [those] whose performance is being measured.”[7]

Despite the cause of disease often being outside the control of family physicians, they are effective at reducing avoidable hospital admissions for both acute and chronic conditions—reducing the death rate—and first contact with a family physician for a health problem also reduces secondary care expenditure and interventions.[7-9] Family physicians make 2500 diagnoses per year, covering 450 diseases.[10] The average family doctor writes prescriptions for 233 different drugs within 1 year.[11] This is equivalent to 30% of the available drugs on the market. People go to their family doctor with an average of four symptoms, and people with multiple diseases are now the norm.[12] Given this volume of activity, it is not surprising to find variation in the tests and treatments ordered by family physicians. What is acceptable variation in this clinical activity needs to be determined, but this is only one element of the quality framework. 

The UK has used the Quality Outcome Framework for 10 years, which financially rewards practices that meet benchmarks that include some of these aspects, among others.[13] It measured four domains: clinical, organizational, additional services, and the patient experience. These domains have been modified with experience.[14] This approach improves some measures that are related to health outcomes. The percentage of patients with coronary heart disease with their cholesterol in the target range increased from 17.6% in 1998 to 61.4% in 2003, following the introduction of the Quality Outcome Framework. The rate of patients whose coronary heart disease indicators were met rose from 60.5% in 1998 to 78.2% in 2003. For patients with asthma and diabetes, the rates of those with targets achieved rose to 70% and 77% respectively.[15] These are impressive changes, but the challenge in using this information to measure quality for individual doctors or practices is that the context of care is different in many of these practices and using these measures alone is not enough to determine quality, hence the inclusion of the other domains. 

In order for care to be effective, it must be accessible. Despite research and many studies, there is no simple solution to the problem of access.[16] Whether it be online appointment systems for patients or late-night openings, evidence shows that local solutions are needed, and that means practices continuing to monitor and make changes to their appointment systems, which is a major focus of the Australian process of assessing quality of family practice.[17]

Another important part of the background to measuring and improving patient care is the patient medical home model. This is an innovative model for care being introduced across Canada that includes 10 broad concepts, including patient-centred care, access, comprehensive service, and electronic medical records.[18] Each domain has benchmarks within it, but to use all these is probably too much at the initial stage. However, any plan that is used to help improve quality should have these in mind.

Based on the history of quality assurance in primary care, what should a BC family practice quality-of-care measurement process look like? The measures of quality have to be easily and cheaply assessed and need to be relevant to the community. It has to start with discussions with the communities that are part of any quality assessment process. An organization to champion this might be the Institute for Healthcare Improvement, with its teams of UBC and SFU students already involved in promoting quality and safety in BC. With the experiences from the UK and Australia/New Zealand and the patient medical home informing those discussions, it should be possible to develop a BC-specific set of core quality benchmarks that can lead to improved quality.[19] Although many characteristics may be similar, there will be variation from town to town reflecting population needs. That means no single provincial quality assessment process in terms of one set of data, but a distributed model of quality assessment reflecting the distributed model of health care in BC.

Creating a distributed quality assessment model is only one step. If variation in a practice is found to be significant, there has to be a process in place for positively addressing that variation, and the process needs to be in place before the quality measurement process starts. Variation in ordering tests and prescribing has been found to be reduced with face-to-face group feedback as well as the use of computerized prescribing support systems.[20,21]

Family doctors do not wake up every day with the intention of performing less well than their peers or failing to do the best for their patients. Any system introduced has to help physicians and their practices to be better and has to be seen to do that. For that reason, most quality-improvement programs have incentives and use comparators with local and regional norms. We need to establish communities of practices where sharing information is both acceptable and useful. Over the last few years the government has funded the formation of divisions of family practice where local communities of doctors work together to identify health care needs and use the funding to address these needs for the benefit of their patients. Divisions of family practice may be ideal groups to undertake quality assessment, but in many cases they may be too large. The informality and local circumstances vary significantly from one part of a city to another. For any system to be helpful it needs to be undertaken with smaller groups of practices that are confident in sharing information.

Years ago, I was part of a quality-care process where we asked practices to share measures of diabetic care, ranging from the proportion of patients who had blood pressure checks to the attendance at the clinic for diabetic care. At that first meeting we presented the significant variation in rates of patients who received optimal diabetic care. Although the information was anonymous, each family doctor and practice knew his or her own data. The first comment was from a family doctor who, pointing to the lower end of the scale on the screen said, “That is my practice and I want to know what the practices at the other end of the scale are doing differently.” This led to a discussion about the difference in recall systems and follow-up for patients with diabetes. Being shown that there was scope for improvement, and that local practices were doing things differently—which led to better care—was enough of a trigger to break the anonymity and to change practices for the better.

Rather than creating a top-down benchmarking system, a more positive, distributed approach of working together to help practices improve relevant quality measures is more likely to result in better health outcomes for British Columbians, and at the same time provide the auditor general with evidence of cost-effectiveness.

References

1.    Office of the Auditor General of British Columbia. Oversight of physician services. February 2014. Accessed 6 November 2014. www.bcauditor.com/pubs/2014/report9/oversight-physician-services.
2.    Physician Quality Assurance Steering Committee. Core Data Set. Accessed 4 November 2014. http://qaframework.ca/projects/coredataset/.
3.    Houghton G. General practitioner reaccreditation: Use of performance indicators. Br J Gen Pract 1995;45:677-681.
4.    Jenner W. An address on the etiology of acute specific diseases. Br Med J 1875;1:233-236.
5.    Rose G. Sick individuals and sick populations. Int J Epidemiol 2001;30:427-432.
6.    Fiscella K, Williams DR. Health disparities based on socioeconomic inequities: Implications for urban health care. Acad Med 2004;79:1139-1147.
7.    Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q 2005;83:457-502.
8.    Macinko J, Starfield B, Shi L. The contribution of primary care systems to health outcomes within Organization for Economic Cooperation and Development (OECD) countries, 1970-1998. Health Serv Res 2003;38:831-865.
9.    Forrest CB, Starfield B. The effect of first-contact care with primary care clinicians on ambulatory health care expenditures. J Fam Pract 1996;43:40-48.
10.    Hodgkin K. Diagnostic vocabulary for primary care. J Fam Pract 1979;8:129-144.
11.    de Bakker DH, Coffie D, Heerdink ER, et al. Determinants of the range of drugs prescribed in general practice: A cross-sectional analysis. BMC Health Serv Res 2007;7:132.
12.    Fortin M, Bravo G, Hudon C, et al. Prevalence of multimorbidity among adults seen in family practice. Ann Fam Med 2005;3:223-228.
13.    Walker S, Mason AR, Claxton K, et al. Value for money and the Quality and Outcomes Framework in primary care in the UK NHS. Br J Gen Pract 2010;60:e213-220.
14.    British Medical Association. Quality and Outcomes Framework. Accessed 4 November 2014. http://bma.org.uk/practical-support-at-work/contracts/independent-contra....
15.    Campbell SM, Roland MO, Middleton E, et al. Improvements in quality of clinical care in English general practice 1998-2003: Longitudinal observational study. BMJ 2005;331:1121.
16.    Fallon CW, Hamilton I, Bhopal JS, et al. Introduction of an appointment system in a general practice: Surveys of patients and staff. Health Bull (Edinb) 1990;48:232-237.
17.    Gardner K, Sibthorpe B, Longstaff D. National quality and performance system for Divisions of General Practice: Early reflections on a system under development. Aust New Zealand Health Policy 2008;5:8.
18.    Gutkin C. Adapting the medical home concept to Canada. Can Fam Physician 2010;56:300.
19.    Palmer C, Bycroft J, Healey K, et al. Can formal collaborative methodologies improve quality in primary health care in New Zealand? Insights from the EQUIPPED Auckland Collaborative. J Prim Health Care 2012;4:328-336.
20.    Gillaizeau F, Chan E, Trinquart L, et al. Computerized advice on drug dosage to improve prescribing practice. Cochrane Database Syst Rev 2013;11:CD002894.
21.    Litvin CB, Ornstein SM, Wessell AM, et al. Use of an electronic health record clinical decision support tool to improve antibiotic prescribing for acute respiratory infections: The ABX-TRIP study. J Gen Intern Med 2013;28:810-816.

hidden

Professor Dawes is head of family practice at the University of British Columbia. He started his clinical practice as a family physician in Oxford, England, and looked after 1800 patients for 17 years. He provided full office practice, obstetrics, home care, and in-hospital care. In 1992 following the completion of his PhD he helped develop a master’s program in evidence-based health care. He ran the Centre for Evidence-Based Medicine in Oxford and was professor and head of family practice at McGill University before coming to UBC in 2010. His research includes genomics in primary care and lifestyle interventions to prevent diabetes.