Article Text

Download PDFPDF

The association between quality of primary care, deprivation and cardiovascular outcomes: a cross-sectional study using data from the UK Quality and Outcomes Framework
  1. T Kiran1,2,
  2. A Hutchings3,
  3. I A Dhalla1,4,
  4. C Furlong5,
  5. B Jacobson6
  1. 1Kennan Research Centre, Li Ka Shing Knowledge Institute of St Michael's Hospital, Toronto, Ontario, Canada
  2. 2Regent Park Community Health Centre, Toronto, Ontario, Canada
  3. 3Health Services Research Unit, London School of Hygiene and Tropical Medicine, London, UK
  4. 4Departments of Medicine and Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada
  5. 5Harrow Primary Care Trust, Harrow, UK
  6. 6London Health Observatory, London, UK
  1. Correspondence to Dr Tara Kiran, 30 Bond Street, Toronto, ON M5B 1W8, Canada; tara.kiran{at}utoronto.ca

Abstract

Background The Quality and Outcomes Framework, a financial incentive scheme for general practitioners in the UK, seems to have improved the quality of primary care and reduced inequalities in primary care delivery. It remains unclear, however, whether higher-quality primary care improves health outcomes or reduces health inequalities.

Methods We conducted a cross-sectional study examining the association between quality of cardiovascular care and coronary heart disease (CHD) outcomes in 1531 general practices in London. We calculated CHD quality achievement scores (ranging from 0 to 100) for each practice using the 2006–2007 data from the Quality and Outcomes Framework. We used weighted linear regression models to assess the practice-level association between the CHD quality score and CHD admissions and deaths.

Findings Overall, practices with higher CHD quality achievement scores had better CHD outcomes. Each one point increase in the CHD quality achievement score was associated with 4.28 (95% CI 1.19 to 7.38; p=0.007) fewer admissions per 100 000 for practices serving highly deprived populations and 2.11 (95% CI 0.68 to 3.55; p=0.004) fewer admissions per 100 000 for practices serving populations of average deprivation. There was no association between the CHD quality achievement score and the CHD admissions for practices serving affluent populations (p=0.906). We observed a similar deprivation-dependent gradient between quality achievement and CHD deaths.

Interpretation High-quality primary care is associated with improved health outcomes. This association is strongest in deprived areas, suggesting that high-quality primary care may play an important role in reducing health inequalities.

  • Primary healthcare
  • health services research
  • quality of healthcare
  • socioeconomic factors
View Full Text

Statistics from Altmetric.com

Improving the quality of primary care in the UK is an important objective, and its significance was reinforced in the recent review of the National Health Service (NHS).1 Historically, there have been large inequalities in both the availability and quality of primary care delivered in the UK. Specifically, deprived areas with greater health needs generally have poorer access to high-quality primary healthcare.2 3 A large body of evidence supports the hypothesis that improving the availability of primary care results in fewer health inequalities.4 However, very little evidence has been produced in support of the hypothesis that improving the quality of primary care reduces health inequalities, and the work that has been done has largely focused on soft indicators of quality of care such as the patient's perception of his or her physician rather than hard indicators such as myocardial infarction or death.5

In 2004, the UK government and the British Medical Association negotiated the Quality and Outcomes Framework (QOF), an incentive scheme that financially rewards general practices depending on their achievement as measured against specific quality indicators.6 7 Approximately 60% of the quality indicators from the incentive scheme relate to clinical care. Most of these reflect recommendations from clinical guidelines that are widely accepted by the medical community and have been confirmed in evidence-based guidance from the UK's National Institute of Clinical Excellence. Some of the indicators measure processes and others relate to treatment or intermediate outcomes. For example, practices are rewarded for keeping a register of their patients with hypertension, for the proportion of these patients who have had a blood pressure recorded in the past 15 months, and for the proportion who have a blood pressure of 150/90 mm Hg or less. The remaining 40% of indicators relate to practice organisation or patient experience factors such as whether patient records contain an up-to-date clinical summary or the minimum length of booked appointments offered by the practice.

Evidence suggests that linking financial incentives to quality achievement has improved the overall quality of primary care in the UK.8 9 However, there is less evidence whether patients served by practices that provide higher quality primary care as measured by performance on the financial incentive scheme have better health outcomes. Although Shohet and colleagues10 showed that a subset of English practices with better performance on an epilepsy-related quality indicator had fewer epilepsy-related hospitalisations, Bottle and colleagues11 found no association between the number of points obtained on the QOF scheme and coronary heart disease (CHD) outcomes at either the practice level or the primary care trust level. Similarly, Downing and colleagues12 also found no consistent associations between achievement on quality indicators and a variety of healthcare outcomes.

Early data from the incentive scheme showed that practices serving deprived populations had slightly lower quality scores.13–17 More recent data has shown that scores in these practices have improved over time and that the incentive scheme has reduced inequalities in primary healthcare provision.18 19 However, there is no evidence to date that improving the quality of primary care in the UK will reduce inequalities in health outcomes.

We, therefore, explored the association between quality of primary care and health outcomes using data from 2006 to 2007 for general practices in London, a city with large health inequalities within and between local authorities including large variations in cardiovascular mortality.20 We focused on cardiovascular disease because it is the leading cause of death in the UK21 and because there is good evidence to support the treatment of cardiovascular risk factors. We tested two related hypotheses: first, that high quality primary care for cardiovascular disease is associated with lower CHD morbidity and mortality and, second, that this association is stronger in more deprived areas.

Methods

Data sources

CHD morbidity and mortality

Practice codes were used to link individual-level data from the Hospital Episodes Statistics and Death Registration databases to general practitioner (GP) practices. We then calculated age-and-sex standardised CHD admission and death rates for each practice in the study using the indirect method that is preferred when outcomes are relatively rare.22 We used admissions data from 1 April 2006 to 31 March 2007 and CHD deaths from 1 April 2004 to 31 March 2007. Deaths were aggregated over a 3-year period because of the small number of deaths in each practice per year.

Quality achievement scores

Publicly available QOF data from 1 April 2006 to 31 March 2007 were used to ascertain performance on a set of quality indicators outlined in table 1 that reflect appropriate prevention and management of CHD. We explicitly included quality indicators that reflected an objective reduction in one of the four major modifiable cardiovascular risk factors: high blood pressure, high blood cholesterol level, diabetes and smoking, and indicators that reflected evidence-based treatment of CHD. Because the incentive scheme does not include a quality indicator that reflects an objective reduction in smoking, we used an indicator that measures whether patients were counselled to quit smoking instead.

Table 1

Summary of the subset of quality indicators included in the study

We summarised the quality of care delivered by the practice for the prevention and management of CHD by calculating a practice's mean achievement on the 12 quality indicators specified in table 1. In other words, we calculated the percentage of patients in the practice who met the target for each of the 12 quality indicators and took the mean, weighting the achievement on each quality indicator equally. This CHD quality achievement score could fall anywhere between 0 and 100. A score of 0 means that none of the indicators were satisfied for any of the patients in a practice, and a score of 100 means that all indicators were satisfied for all the eligible patients. Rationale and details for the calculation of the CHD quality achievement score and the relevant data sources are presented in the Supplementary Appendix.

The financial incentive scheme allows practices to exclude certain patients when reporting their performance on quality indicators for reasons such as extreme frailty, medication intolerance or repeated missed appointments, a practice known as exception reporting.23 While acknowledging that the legitimate purpose of exception reporting is to protect patients from inappropriate treatment and ensure that practices are remunerated fairly, some health policy analysts have noted that exception reporting permits “gaming” by providing practices with the opportunity to increase their income by inappropriately excluding patients who are not receiving recommended care. To account for the potential influence of exception reporting on quality achievement, we calculated two different CHD quality achievement scores. The first measured the proportion of patients in whom the indicator was achieved, with exception-reported patients excluded from the denominator. The second measured the proportion of patients in whom the targets were achieved with exception-reported patients included in the denominator. Except where noted, we focus on the first measure in this paper.

Practice-level deprivation

Practice-level deprivation scores were calculated by the London Health Observatory.24 Individual patient postcodes were obtained from the NHS Strategic Tracing Service for all patients registered to a GP practice in London. The patient postcodes were mapped to the Lower Super Output Area where the patient resided and the corresponding Index of Multiple Deprivation 2007 score. (Each Super Output Area consists of about 1500 people within a defined geographical locality.25) The Index of Multiple Deprivation 2007 score for each patient in the practice was then averaged to produce the practice-level deprivation score. The Index of Multiple Deprivation uses seven domains of social deprivation such as income and employment deprivation to estimate the deprivation level of individuals living in the area, with higher scores indicating greater deprivation.26

Other practice characteristics

We considered several practice characteristics as potential a priori confounding variables. Practice size, the number of GPs per 10 000 patients, GP age and GP country of qualification have all been previously found to be significantly associated with a practice's achievement on the GP incentive scheme.15 In addition to these variables, we also included the percentage of South Asian patients in a practice as a potential confounder because of the high burden of CHD in people of South Asian ancestry.27 Practice size was estimated from the QOF data. Information on GP characteristics was derived from GP payment data extracted on 30 September 2006. The ethnic composition of the practice population was estimated using a model developed by the London Health Observatory.24

Inclusion and exclusion criteria

All the GP practices in the London Strategic Health Authority with reported data for the QOF for 2006–2007 were eligible for inclusion in the study (n=1567). Practices were excluded from the study if they had <1000 patients (n=10), if the practice closed after April 2007 (n=9), if data were missing (n=15) or if the practice had no registered patients in one of the chronic disease categories included in the study (CHD, hypertension, diabetes or stroke; n=2).

Analysis

We developed weighted linear regression models to measure the associations between quality achievement scores and CHD outcomes. Although all variables included in the models were pre-specified,28 we transformed three variables to improve the normality of model residuals: practice size was transformed by taking its logarithm and two variables measuring GP characteristics (per cent qualified in the UK and per cent aged >50 years) were categorised (see table 2). The data for each practice was weighted by the inverse of the variance of the outcome variable to account for uncertainty in the standardised CHD admission and mortality rates.

Table 2

Summary of practice characteristics

To investigate our hypothesis that the association between quality achievement and CHD outcomes would be modified by the deprivation level of the practice population, we included interaction terms between deprivation and quality achievement in the regression models. Both the deprivation and CHD quality achievement scores were kept as continuous variables in the primary analysis. Therefore, to aid interpretation of the interaction effect, we used the regression model output to estimate the association between the CHD quality achievement score and the CHD outcomes at three particular levels of deprivation29:

  1. Low deprivation—deprivation score equal to 10, similar to Richmond upon Thames, the least-deprived local authority in London.

  2. Average deprivation—deprivation score equal to 27, the mean deprivation level of practices included in the study.

  3. High deprivation—deprivation score equal to 46, similar to Hackney, the most-deprived local authority in London.

We also categorised practices into deprivation tertiles and then determined whether there was an association between CHD quality achievement and outcomes within each tertile.

All the analyses were performed using Stata V.10.

Results

One thousand five hundred and thirty-one practices, with 8 345 353 registered patients, were included in the analysis, representing 98% of the GP practices in the London Strategic Health Authority region with reported data from the QOF for 2006–2007. The summary statistics for all the variables are presented in table 2.

The association between the CHD quality achievement scores and the CHD admissions

Overall, practices with higher CHD quality achievement scores had lower CHD admission rates after controlling for potential confounding variables (table 3). Overall, there was relatively weak evidence that this association was more marked as deprivation increased (p=0.116 for the interaction term). The association between the CHD quality score and the CHD admission rates at various levels of deprivation is illustrated in figure 1.

Table 3

Estimated regression coefficients for weighted linear regression showing the association between the CHD quality achievement score and the standardised CHD admissions per 100 000

Figure 1

Association between CHD quality achievement score and standardised CHD admissions per 100 000 (regression coefficient and 95% CI) at different levels of deprivation. The asterisk indicates the effect of a 1-point improvement in the CHD quality achievement score on the standardised CHD admissions rate per 100 000.

For practices serving populations with low levels of deprivation, there was no evidence of an association between the CHD quality achievement score and the CHD admissions (p=0.906). However, for practices serving populations with average and high levels of deprivation, a 1-point increase in the CHD quality achievement score was associated respectively with a 2.11 and 4.28 per 100 000 decrease in the number of CHD admissions (p=0.004 and p=0.007, respectively). The latter figure represents an approximate 0.9% reduction from the mean CHD admissions rate observed in the study (434 per 100 000) for each 1-point increase in the CHD quality achievement score. Details of associations between other variables included in the model and the CHD admissions rate are presented in table 3.

The association between CHD quality achievement scores and CHD mortality

Overall, practices with higher CHD quality achievement scores had lower CHD mortality rates after controlling for potential confounding factors (table 4). This association was modified by the deprivation level of the practice population (p=0.011 for the interaction term). The association between the CHD quality score and the CHD mortality rates at various levels of deprivation is illustrated in figure 2.

Table 4

Estimated regression coefficients for weighted linear regression showing the association between the CHD quality achievement score and the standardised CHD mortality rate*

Figure 2

Association between CHD quality achievement score and standardised CHD mortality rate (regression coefficient and 95% CI) at different levels of deprivation. The asterisk indicates the effect of a 1-point improvement in the CHD quality achievement score on the standardised CHD mortality rate per 100 000 over the 3-year study period.

For practices serving populations with low levels of deprivation, there was no association between the CHD quality achievement score and CHD mortality (p=0.600). In contrast, for practices serving populations with average or high levels of deprivation, a 1-point increase in the CHD quality achievement score was associated with a 0.58 and 1.40 decrease in CHD deaths per 100 000 (p=0.001 and p<0.001, respectively). The latter represents an approximate 1.5% reduction from the mean CHD mortality rate observed in the study (94.4 per 100 000 over the 3-year period) for each 1-point increase in the CHD quality achievement score. Details of associations between other variables included in the model and the CHD mortality rate are presented in table 4.

Effect of exception reporting

With the exception-reported patients included in the analysis, there was still evidence that the association between the CHD quality achievement score and CHD mortality becomes stronger with increasing levels of deprivation (p=0.049 for the interaction term; Supplementary Appendix). However, there was no evidence that deprivation levels modified any association between the quality achievement score and the CHD admissions (interaction term p=0.302) and little evidence of an overall association between the quality achievement score and the CHD admissions (p=0.127 when the interaction term was left out of the model).

Analysis by deprivation tertile

Table 5 summarises the characteristics of practices divided into three groups by deprivation tertile. The most deprived group of practices had the highest mean rate for CHD admissions and was the only group to show a statistically significant association between the CHD quality achievement score and the CHD admission rates (mean decrease of 5.06 admissions per 100 000 for a 1-point increase in achievement score). An interaction test showed that this association was stronger than the (lack of) association in the least deprived group of practices (p=0.02). Similarly, CHD mortality rates were highest in the most deprived group of practices. These practices also showed the strongest association between the CHD quality achievement score and CHD mortality (mean decrease of 1.02 deaths per 100 000 for a 1-point increase in achievement score).

Table 5

Practice characteristics, CHD quality achievement scores and CHD outcomes by deprivation tertile (n=1531)

Discussion

We found that general practices in London with higher achievement on CHD-related quality indicators had lower CHD admission and mortality rates. However, this association was strongest for practices serving populations with high levels of deprivation and disappeared for practices serving populations with lower levels of deprivation. Consistent with previous work, we also found that CHD admissions and deaths were more common in practices serving more deprived populations. Notably, however, this association was attenuated by higher-quality achievement scores: high-quality primary care seemed to mitigate the association between deprivation and CHD.

As expected, the CHD quality achievement score was more strongly and consistently associated with CHD mortality and morbidity when exception-reported patients were excluded from the analysis than when they were included.

Only two previous studies have examined whether practices with higher-quality scores, as measured by performance on the QOF, have better cardiovascular health outcomes.11 12 In contrast to the results presented here, neither of the previously reported studies revealed an association between quality scores and health outcomes. However, our study differed from previous research in two important ways. First, we developed an original CHD quality achievement score that summarised performance only on evidence-based quality indicators that assessed objective reductions in cardiovascular risk factors or appropriate treatment of CHD. Furthermore, we used the proportion of patients in whom the target was achieved rather than the number of QOF points obtained. In contrast, other studies included performance on process-related indicators11 12 or quality indicators for asthma, chronic obstructive pulmonary disease and other diseases not related to CHD outcomes12and assessed quality achievement by the number of points attained in the incentive scheme.11 12 Second, we considered more potential confounders than either study such as ethnic composition and GP practice characteristics.

The only study that did find a relationship between quality achievement and health outcomes in epilepsy was like ours, conducted at the practice level and used a quality indicator that was both disease specific and outcome based.10

Previous analyses have found that CHD admission and mortality rates in the UK are higher in more deprived areas.11 30 31 Although studies from the USA have suggested that better access to primary care is associated with reduced ethnic and socioeconomic health inequalities,4 there has previously been little evidence to support the assertion that better quality primary care can reduce socioeconomic health inequalities.

Limitations

Our study has four important limitations. First, because our study was cross-sectional, we cannot definitively infer causation. However, given that the quality of primary care delivered by a practice is unlikely to change dramatically over a small number of years and given that in many cardiovascular disease prevention trials, a treatment effect is observed within the first year, a cause-and-effect relationship is certainly plausible. Nevertheless, a longitudinal analysis conducted 5 to 10 years after the introduction of the financial incentive scheme would better assess the temporal relationship between quality scores and health outcomes and increase confidence in our findings. Second, the variation in QOF scores was small and may not reflect the true extent of variation in the quality of care. As a result, we were limited in our ability to test our hypothesis. However, this limitation would only serve to increase the likelihood of a negative finding. Third, our analysis was limited to London, and our findings may not be generalisable to other regions in the UK. However, the individual-level deprivation data that we used to generate practice-level deprivation scores are not readily available in many other parts of the country. Finally, we assume that higher CHD quality scores do in fact reflect higher quality care. Although there is some evidence that measured performance on quality indicators relating to stroke care does not correlate with other quality standards,32 the CHD quality targets chosen for our study have excellent face validity and are based on clinical practice guidelines. However, financial incentive schemes may have unintended consequences such as the fragmentation of care and the neglect of conditions for which financial incentives are not provided,33 so assessing the relationship between achievement on QOF quality indicators and actual quality of care is an important area for future research.

Implications for policy

Population-wide financial incentives have the potential to reduce inequalities in healthcare provision if designed appropriately.34 There is already evidence that the QOF incentive scheme has contributed to reduced inequalities in healthcare in the UK.18 19 Results from this study suggest that the financial incentive scheme may have also contributed to reducing inequalities in health outcomes.

Our study showed a positive association between performance on objective, evidence-based quality indicators and health outcomes. There is still little evidence, however, on whether better performance on the process indicators included in the QOF scheme improves health outcomes or patient satisfaction. Our findings, coupled with other research that has shown inconsistent associations between performance in the QOF scheme and clinical outcomes, supports the need for a review of the indicators to be included in the QOF.35

Improving the quality of primary care services is part of the UK government's national strategy to reduce health inequalities36 37 and was reaffirmed as a key priority in the recent Next Stage Review of the NHS.1 Differences in CHD outcomes are responsible for approximately 20% of the gap in life expectancy between the most disadvantaged areas in the UK and the general population.36 Our study provides evidence that higher quality primary care for cardiovascular disease is associated with fewer CHD admissions and deaths in more deprived populations. Primary care may play a more important role in reducing morbidity and mortality in more deprived populations because they have a higher burden of preventable illness and disease. Our findings suggest that if health inequalities are to be reduced, then the focus of attention should be on improving quality in practices serving deprived populations that performed less well on the QOF incentive scheme. Better-performing practices serving deprived populations, where morbidity and mortality rates are similar to those serving less-deprived populations, may offer insights into how this might be achieved.

What is already known on this subject

  • Deprived areas with greater health needs usually have poorer access to high-quality primary healthcare.

  • Introduction of financial quality of care incentives in the UK has improved the quality of primary care and reduced inequalities in primary care delivery.

  • There is conflicting evidence on whether practices with higher-quality scores have better clinical outcomes.

What this study adds

  • GP practices that better meet evidence-based quality targets related to cardiovascular disease and risk factor management have lower hospital admission and death rates because of CHD.

  • The association between quality of primary care and CHD outcomes is strongest for practices serving deprived populations.

  • High-quality primary care appears to reduce inequalities in health outcomes, particularly for cardiovascular disease.

Acknowledgments

The authors thank the following: L Jones, J Van der Meulen and C Sanderson from the London School of Hygiene and Tropical Medicine (London, UK), for their help in developing the study protocol; D Osborne, P Deponte and L Warren from the London Health Observatory (London, UK) and J Goodall from the Health and Social Care Information Centre (London, UK), for their help in preparing the data; J Kwong and L Rosella from the Institute of Clinical and Evaluative Sciences (Toronto, Canada), for their advice on statistical methods and P Aspinall from the University of Kent (Kent, UK) and J Green from the London School of Hygiene and Tropical Medicine (London, UK), for their comments on the manuscript.

References

View Abstract

Supplementary materials

  • Web Only Data jech.2009.098806

    Files in this Data Supplement:

Footnotes

  • Linked articles 109603.

  • Funding This study was funded through the usual operating funds of the London Health Observatory. IAD is supported by a Felllowship Award from the Canadian Institutes of Health Research. The London Health Observatory provided administrative and logistic support including providing access to the data and use of statistical software and office facilities.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the London School of Hygiene and Tropical Medicine.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles