- S M Campbell, research fellow ()a,
- M Hann, research associatea,
- J Hacker, researchera,
- C Burns, researchera,
- D Oliver, researchera,
- A Thapar, general practitionerb,
- N Mead, research associatea,
- D Gelb Safran, directorc,
- M O Roland, directora
- a National Primary Care Research and Development Centre, University of Manchester, Manchester M13 9PL
- b Department of General Practice, University of Manchester, Rusholme Health Centre, Manchester M14 5NP
- c The Health Institute, New England Medical Center, Boston, MA 02111, USA
- Correspondence to: S M Campbell
- Accepted 3 August 2001
Objectives: To assess variation in the quality of care in general practice and identify factors associated with high quality care.
Design: Observational study.
Setting: Stratified random sample of 60 general practices in six areas of England.
Outcome measures: Quality of management of chronic disease (angina, asthma in adults, and type 2 diabetes) and preventive care (rates of uptake for immunisation and cervical smear), access to care, continuity of care, and interpersonal care (general practice assessment survey). Multiple logistic regression with multilevel modelling was used to relate each of the outcome variables to practice size, routine booking interval for consultations, socioeconomic deprivation, and team climate.
Results: Quality of clinical care varied substantially, and access to care, continuity of care, and interpersonal care varied moderately. Scores for asthma, diabetes, and angina were 67%, 21%, and 17% higher in practices with 10 minute booking intervals for consultations compared with practices with five minute booking intervals. Diabetes care was better in larger practices and in practices where staff reported better team climate. Access to care was better in small practices. Preventive care was worse in practices located in socioeconomically deprived areas. Scores for satisfaction, continuity of care, and access to care were higher in practices where staff reported better team climate.
Conclusions: Longer consultation times are essential for providing high quality clinical care. Good teamworking is a key part of providing high quality care across a range of areas and may need specific support if quality of care is to be improved. Additional support is needed to provide preventive care to deprived populations. No single type of practice has a monopoly on high quality care: different types of practice may have different strengths.
What is already known on this topic
What is already known on this topic Quality of care varies in virtually all aspects of medicine that have been studied
Most studies look at quality of care from a single perspective or for a single condition
What this study adds
What this study adds Quality of care varies for both clinical care and assessments by patients of access and interpersonal care
Practices with longer booking intervals provide better management of chronic disease; preventive care is less good in practices in deprived areas
No single type of practice has a monopoly on high quality care—small practices provide better access but poorer diabetes care
Good team climate reported by staff is associated with a range of aspects of high quality care
Quality of care varies in most settings in which it has been studied, including in the United States,1–3 the United Kingdom,4–8 New Zealand, 9 10 Australia, 11 12 and Holland,13 and medical errors are a cause of increasing concern.14 In the United Kingdom the government has proposed a range of strategies for improving quality in the NHS.15–17 To respond appropriately to such initiatives it is necessary to understand both the extent of variation in quality of care and its causes, and several authors have examined these relations.18–24 However, data on quality of care are not widely available in the United Kingdom, especially in primary care. Researchers rely largely on information collected from volunteer practices or on the small amount of routinely available data. In a systematic review of quality of care in general practice,4 we found that many studies focus on only one clinical area, precluding comparison of factors affecting different aspects of quality of care.
Quality of care is a multidimensional concept,25 and different aspects of quality need different methods of measurement.26 In this study we used a range of methods to carry out detailed assessments of quality of care in a stratified random sample of practices. The study represents the most comprehensive evaluation of quality of care in general practice in the United Kingdom to date. We have previously defined the components of quality of care as a combination of access (whether patients can get to health care) and the effectiveness of clinical care and interpersonal care (whether care is any good when they get there). 25 27 Our results are presented within this framework. The aims of the study were to assess the extent of variation in quality of care in English general practice and to identify factors associated with high quality care.
Selection of practices
We used a three stage process to select practices. We selected three out of the eight English NHS regions—North Thames, North West, and South West—as being nationally representative in terms of rurality, socioeconomic deprivation, and geographical dispersion of population. From each of these three regions we selected two health authorities as being representative of their region in terms of rurality and socioeconomic deprivation. The six health authorities selected were Bury and Rochdale, West Pennine, Enfield and Haringay, South Essex, Avon, and Somerset. Finally, within each of these six authorities we selected a random sample of 10 practices stratified in terms of practice size, training status, and socioeconomic deprivation. These 60 practices were invited to take part in a detailed assessment of quality. When a practice refused to participate, another with similar characteristics was chosen at random and invited to participate; 60 out of 75 (80%) practices that we approached agreed to take part.
Quality of clinical care: chronic disease management—We used computerised disease registers or prescribing records to select 20 patients in each practice receiving maintenance treatment for each of three conditions: asthma in adults, angina, and type 2 diabetes mellitus. Some small practices had fewer than 20 patients with diabetes and angina. After confirming the relevant diagnosis from the medical records, we extracted data from medical records to identify aspects of care previously defined by expert panels as being both necessary to undertake and necessary to record for these conditions.28 We measured the inter-rater reliability for all items and rejected those for which the κ value was <0.6 or which applied to <1% of the relevant sample. The box lists criteria used in the analyses.
Items used in the clinical scores
These criteria were devised by panels consisting largely of general practitioners with a special interest in the three areas, who used a systematic process to combine evidence with expert opinion.28 Italics indicate conditional variables that do not apply to all patients.
Past 14 months, record of:
Frequency or pattern of angina attacks
Prescription or advice to take aspirin unless record of contraindication or intolerance
Prescription of β blocker as maintenance treatment if sole therapy
Action taken on blood pressure if systolic pressure >160 mm Hg, or systolic pressure >140 mm Hg and cholesterol >5.5 mmol/l
Past five years, record of:
Action taken if cholesterol >5.5 mmol/l
Weight advice if overweight
Smoking advice to smokers
Referral for exercise electrocardiography
Referral for specialist assessment
Past 14 months, record of:
Daily, nocturnal, or activity limiting symptoms
Past five years, record of:
Normal or predicted peak flow or record of difficulty using a peak flow meter
Self management plan for patients taking high dose steroids or who have had inpatient treatment for asthma
For patients with recorded exercise induced bronchospasm, prescription of short acting bronchodilators for use before exercise
Smoking advice to smokers
Peak flow during a consultation for an exacerbation of asthma
Speech rate, pulse rate, or respiratory rate during a consultation for an exacerbation of asthma if bronchodilator was used immediately
Prescription of oral steroids if peak flow <60% of normal or predicted
Action taken if patient experienced nocturnal symptoms
Action taken if patient experienced symptoms limiting activity
Referral to a respiratory physician if oral steroids used in maintenance treatment
Type 2 diabetes mellitus
(Criteria developed before publication of United Kingdom prospective diabetes study29)
Past 14 months, record of:
Glycated haemoglobin (HbA1c)
Recording of peripheral pulses or record of visual examination of the feet
Serum creatinine concentration
Examination of fundi or visual acuity
Record of hypoglycaemia symptoms if patient taking sulphonylurea
Past five years, record of:
Serum cholesterol concentration
Documentation of education about diabetes
Advice given to smokers
Under 80 years—offered treatment if average of last three readings shows diastolic pressure >100 mm Hg, or systolic pressure >150 mm Hg and diastolic pressure >90 mm Hg
Over 80 years—offered treatment if average of last three readings shows diastolic pressure >110 mm Hg, or systolic pressure >160 mm Hg and diastolic pressure >100 mm Hg
If patient was prescribed angiotensin converting enzyme inhibitor, creatinine and potassium were measured within one month of starting treatment
If patient is being treated for hypertension and has proteinuria (macroalbuminuria but not microalbuminuria), the patient is taking an angiotensin converting enzyme inhibitor
For patients aged under 70, if the last HbA1c was >9, patient offered a therapeutic intervention aimed at improving glycaemic control
For patients aged over 70, if the last HbA1c was >10, patient offered a therapeutic intervention aimed at improving glycaemic control
Referral to a specialist if serum creatinine is >200 mmol/l
Quality of clinical care: preventive care—For each practice we sent a questionnaire to the appropriate health authority to collect information on rates of uptake for cervical cytology screening; primary childhood immunisation; measles, mumps, and rubella immunisation; and preschool vaccination.
Patient evaluation: access and interpersonal care—We randomly selected 200 adults from each practice list and sent each patient a copy of the general practice assessment survey. 30 31 Patients in five out of the six health authority areas received two postal reminders. We used data from these questionnaires to assess the quality of access, continuity of care, and interpersonal aspects of care.
Team climate and team effectiveness—Because of the importance now ascribed to teamwork in general practice, we sent the team climate inventory to all staff employed by the practices32; 48 (80%) practices took part in this assessment. In line with previous applications of this method and on the recommendation of the questionnaire's main developer (M West, personal communication, 2000), we excluded from the analyses any practices where less than 30% of the staff completed questionnaires. The analyses included data from 42 (70%) practices, representing 387 (60%) members of staff. The team climate inventory assesses perceptions of staff members of how people work together, how frequently they interact, whether teams have identified aims and objectives, and how much practical support and assistance are given towards new and improved ways of doing things. For the analyses reported in this paper we combined the team climate subscales into a single score.
For each criterion for angina, asthma, and diabetes we recorded whether the necessary aspect of care was recorded. We analysed these binary variables with an item response model within a multilevel framework (items within patients) by using GLLAMM-6 within Stata version 6.33 For each condition, we calculated a score for each practice by using a random intercept constant only multilevel model (patients within practices). This is equivalent to calculating a mean score for each practice but adjusting for different pools of patients in different practices and the fact that many items were conditional variables that did not apply to all patients (for example, action to be taken if cholesterol exceeded a certain value). Only items that were applicable for individual patients were included in the score for the practice. Higher clinical scores (maximum=100) therefore reflected better clinical care measured with evidence based process measures.
We then used the scores for angina, asthma, diabetes, preventive care, access, continuity, and interpersonal care as dependent variables in a series of backwards stepwise regression models to identify predictors of high quality care. Clinical scores at the level of the patient were analysed with a multilevel model to account for the potential of clustering within practices. We analysed the scores from the survey of patients within a survey framework to allow for clustering, by using an ordered logistic regression model. We analysed the indicators for preventive care by looking at the achievement of higher target rates (90% for immunisations, 80% for cervical cytology) with logistic regression. All analyses were undertaken with Stata.
We regressed a common set of independent variables on each dependent variable. These independent variables were practice size (based on whole time equivalent general practitioners), routine booking interval for consultations (5 minutes, 7.5 minutes, 10 minutes), overall team climate, and deprivation score. We derived the deprivation score for each practice by using NHS deprivation bands, calculating the weighted sum of patients in each band (with census based deprivation payments as weights) divided by total list size. We included the training status of the practice in early analyses, but we subsequently excluded this as it was not a significant predictor of any of the outcomes.
Quality of clinical care: chronic disease management
Variation in quality of chronic disease management—Data were collected in all 60 practices. Table 1 summarises practice scores for these and other variables. The practice scores for asthma, angina, and diabetes were significantly, but only moderately, correlated (angina v asthma r=0.43, P<0.001; angina v diabetes r=0.32, P<0.001; asthma v diabetes r=0.55, P<0.001).
Predictors of quality of chronic disease management—Compared with practices with five minute consultation booking intervals, practices with 10 minute booking intervals had higher scores for all three chronic diseases (table 2). Adjusted mean scores in practices with routine 10 minute booking intervals were 10.0 points higher for diabetes (95% confidence interval 1.06 to 18.95, P=0.028), 10.2 points higher for angina (3.83 to 16.58, P=0.002), and 21.6 points higher for asthma (12.30 to 30.91, P<0.001) than in practices with five minute intervals. For diabetes, two other variables were significantly associated with differences in quality of care. Larger practices had higher scores for diabetes than did smaller practices (adjusted difference 2.16 (0.22 to 4.10), P=0.029), as did practices where staff reported better team climate (2.37 (0.36 to 4.38), P=0.021).
Quality of clinical care: preventive care
Complete data for all five indicators were available for 42 (70%) practices. Table 3 shows summary statistics for the preventive care indicators. Practices in deprived areas had lower uptake rates for cervical cytology—odds ratio 0.65 (0.48 to 0.89, P=0.008). Preventive care and other practice variables showed no significant independent associations.
Access and interpersonal aspects of care
Copies of the general practice assessment survey were sent to 11 831 patients, and 4493 (38%) were returned after, for most practices, two reminders. We compared the results with those of studies with response rates of between 60% and 90% (other published data,34 and data held by the National Primary Care Research and Development Centre) and found that the mean and median survey scores and relations between scale scores and sociodemographic factors were similar to ours. We therefore decided to include the survey data in our analyses despite the low response rate, although these results, which are summarised in table 1, should be treated with considerable caution because of the low response rate.
Smaller practices had higher scores for access (adjusted odds ratio 0.87 (0.76 to 0.99), P=0.038), as did practices where the staff reported better team climate (1.23 (1.09 to 1.38), P=0.001). Practices with higher scores for team climate also had higher scores for continuity of care (1.33 (1.18 to 1.50), P<0.001).
Small practices had higher scores on the receptionist scale (0.82 (0.74 to 0.90), P<0.001), as did practices with fewer deprived patients (0.88 (0.83 to 0.94), P<0.001). More deprived practices had lower scores for interpersonal care (0.92 (0.86 to 0.98), P=0.015) and overall satisfaction (0.90 (0.82 to 0.98), P=0.019). Practices where the staff reported better team climate also had higher scores for satisfaction (1.11 (1.05 to 1.19), P<0.001).
The findings of this study confirm that English general practice varies widely in quality of care, as measured from a range of perspectives. Most studies assess quality of care from a single perspective or for a single condition. Our findings highlight the importance of assessing quality of care with a range of measures, as each approach illuminates different aspects of quality of care.
Predictors of quality of care
Four variables stood out as predictors of quality of care. The largest effect was the relation between the booking interval for routine consultations and the quality of management of chronic disease. Other authors have emphasised the importance of adequate time for consultations.23 The effect was greater for asthma than for diabetes and angina, possibly because the last two conditions are more likely to be treated in separate clinics than in routine surgeries. These data provide strong support for the view that general practice should be structured to allow time for the increasing complexity of the work required of general practitioners.
Secondly, we found significant associations between size of practice and quality of care, as has been seen in other studies, 19 24 although the relation was not simple. Smaller practices scored better than larger ones for access to care, but for diabetes care larger practices had higher scores than smaller ones. This emphasises that no single type of practice has a monopoly on high quality care—different types of practice may have different strengths. This is an important finding at a time when small practices in the United Kingdom are coming under particularly close scrutiny from the government.35 As others have found, there may be a trade-off between high quality clinical care and interpersonal care.36
Thirdly, deprivation predicted poorer uptake of preventive care, highlighting that quality of care in general practice is influenced by environmental factors. 18 37 Preventive care is one area in which patients' actions influence the quality of care that can be provided. In other areas where practices had the main control, no significant associations between deprivation and quality of care were found.
Finally, team climate was associated with quality of care for diabetes care, access to care, continuity of care, and overall satisfaction. This was the only variable that was associated with high quality care across a range of aspects of care. The associations are not necessarily causal: it is possible, for example, that staff felt better in practices where good care was given because they received fewer complaints from patients. However, the measure of team climate is intended to reflect how people actually work together and how much support is given towards maintaining high standards of care. High quality care in general practice needs effective teamwork, and this is emphasised in the awards of the Royal College of General Practitioners, which assess the performance of practice teams rather than individuals.
Limitations of the study
Although this is one of the most comprehensive surveys of quality of care in British general practice, the study looked at only limited aspects of overall quality. For example, the clinical data represented only three chronic conditions, a small part of the clinical work undertaken in general practice. Ongoing work by three of the authors (SC, MR, JH) has developed, and is currently field testing, clinical indicators for 19 common conditions presenting in general practice in the United Kingdom.38
The clinical scores were derived from information available only from medical records. Although the expert panels that derived the review criteria selected only aspects of care that they believed needed to be recorded,28 a considerable gap may still exist between what doctors do and what they record. For example, medical records have been found to underestimate preventive or counselling activities.39 Although there is evidence from the United States that quality of record keeping is positively correlated with quality of care, 40 41 similar analyses have not been carried out in the United Kingdom.
The analytical approach and framework (access and effectiveness) used in this study were experimental. Other workers studying the same phenomena using a different approach may not reproduce the findings from this research. Despite the limitations of the methods used and the low response rate of the survey of patients, which mean that the scores for access and interpersonal care should be treated with some caution, this study confirms that wide variation in the quality of care exists in English general practice. The study also identifies important predictors of high quality care that need to be considered as general practice is restructured to meet the needs of the 21st century.
We thank the staff in all 60 practices and six health authorities who took part in the study, and Emma Ruff and Andrew Pickles for their contribution to the project.
Contributors: The project was devised by SC and MR and managed by SC. Data were collected by SC, JH, and CB, with the assistance of AT, NM, and DO. The general practice assessment survey was developed by MR, DGS, CB, and SC. SC and MH undertook the analyses. SC, MR, and MH wrote the paper, with SC as the principal author. SC is the guarantor of the paper.
Funding National Primary Care Research and Development Centre core funding from the Department of Health.
Competing interests None declared.