Abstract

Background. Culture and climate represent shared beliefs and values that may influence quality of care in health care teams, and which could be manipulated for quality improvement. However, there is a lack of agreement on the theoretical and empirical relationships between climate and culture, and their relative power as predictors of quality of care. This study sought to examine the association between self-report measures of climate and culture in primary care teams and comprehensive measures of quality of care.

Methods. The data were derived from a cross-sectional survey of 492 professionals in 42 general practices in England. Self-report measures of culture (the Competing Values Framework) and climate (the Team Climate Inventory) were used, together with validated measures of quality of care from medical records and self-report.

Results. The majority of practices could be characterized as ‘clan’ culture type. Practices with a dominant clan culture scored higher on climate for participation and teamwork. There were no associations between culture and quality of care, and only limited evidence of associations between climate and quality.

Conclusions. The current analysis would not support the hypothesis that culture and climate are important predictors of quality of care in primary care. Although larger studies are required to provide a definitive test, the results may suggest the need for a more complex model of the associations between culture, climate and outcomes, and further research may be required into the interaction between culture and climate with other determinants of behaviour such as internal and external incentives.

Introduction

Primary health care is provided by teams, and teamwork is characterized by interaction between professionals who are working towards a common goal and show ‘task interdependence’ i.e. the need to develop shared understandings.1 This has led to interest in the social–psychological processes that operate at the level of primary care teams. Two terms relating to important processes at this level are organizational ‘climate’ and ‘culture’.

The meaning, measurement and utility of the concepts of culture and climate is an area of debate,2,3 but the interests of health service researchers in these concepts relate to the suggestion that both may be important determinants of health care performance that might in turn be manipulated for quality improvement.4,5

For the purposes of the present paper, organizational climate can be defined as a team's shared perceptions of organizational policies, practices and procedures.1 Organizational climate is generally seen to refer to a particular issue i.e. a climate for information sharing.1 A widely used definition of culture is that it represents ‘a pattern of shared basic assumptions—invented, discovered or developed by a given group as it learns to cope with its problems of external adaptation and internal integration.2

It is unclear whether culture and climate represent distinct concepts. Measures of climate (which are dependent on perceptions of specific behaviours) may reflect ‘intermediate’ level manifestations of a ‘deeper’ organizational culture,4 although it is not clear whether this distinction has empirical support. The relationship between the two constructs has more than theoretical significance. Measures of both constructs have been found to be associated with outcomes such as team functioning and morale, aspects of quality of care and patient satisfaction.5–11 If the concepts are distinct, it is important to determine which construct is the best predictor of outcomes, or whether each construct is associated with different outcomes, to inform relevant quality improvement interventions. The current study sought to examine this issue by using data from a large quality of care study in primary care, to explore:

  1. The association between two common self-report measures of climate and culture.

  2. The association between measures of climate/culture and quality of care.

Methods

The current analyses were part of a larger longitudinal time series study testing the effects of the new GP contract.12 The new contract (the Quality and Outcomes Framework) uses financial incentives attached to 136 quality indicators to reward the quality of primary care.13 The clinical indicators combine both process measures (e.g. record of cholesterol) and intermediate outcomes (e.g. level of cholesterol achieved). For the time series, data will be collected over multiple time points. The current paper focuses on analysis of data collected during 2003, although some reference is made to previous analyses on the 1998 data set.

Measures

Practice characteristics

The practice sample was formed in 1998 and originally involved a stratified random sample of 60 English general practices (80% response rate).7,14 Fifty-seven practices still existed in 2003, and forty-two (74%) agreed to participate in the longitudinal study. The characteristics of the practices in 1998 and 2003 are shown in Table 1, and compared with characteristics of all English practices. Practices in the study sample in 2003 were more likely to be involved in the training of GPs, and were more likely to be larger practices.

Table 1

Characteristics of the QuIP practices (1998/2003) and comparison practices in England

 1998a (n = 60) 1998b (England) 2003c (n = 42) 2003d (England) 
Training 17/60 (24.3%) 2204/9090 (31.0%) 13/42 (25.7%) 2272/8832 (28.3%) 
Whole-time equivalent (≤2 GPs) 29/60 (48.3%) 4360/9090 (48.0%) 15/42 (35.7%) 4221/8832 (47.8%) 
Single-handed practices 18/60 (30.0%) 2779/9090 (30.6%) 10/42 (23.8%) 2577/8832 (29.2%) 
Receiving deprivation payments 34/60 (56.7%) 5390/9090 (59.3%) 21/42 (50.0%) 4956/8832 (56.1%) 
 1998a (n = 60) 1998b (England) 2003c (n = 42) 2003d (England) 
Training 17/60 (24.3%) 2204/9090 (31.0%) 13/42 (25.7%) 2272/8832 (28.3%) 
Whole-time equivalent (≤2 GPs) 29/60 (48.3%) 4360/9090 (48.0%) 15/42 (35.7%) 4221/8832 (47.8%) 
Single-handed practices 18/60 (30.0%) 2779/9090 (30.6%) 10/42 (23.8%) 2577/8832 (29.2%) 
Receiving deprivation payments 34/60 (56.7%) 5390/9090 (59.3%) 21/42 (50.0%) 4956/8832 (56.1%) 
a

All practices in England in 2003.

b

Original 1998 study sample.

c

All practices in England in 1998.

d

Current 2003 QuIP sample.

Climate and culture

Climate was measured using the Team Climate Inventory (TCI), which has demonstrated construct, predictive and discriminant validity,1 and predicts aspects of quality of care.7 The TCI is a 65-item measure with six subscales rated on 5-point scales (from ‘strongly agree’ to ‘strongly disagree’). The subscales include participation (the ‘safety’ of the decision-making environment), support for innovation, reflexivity (discussion and review of procedures), task orientation (emphasis on monitoring quality), clarity of objectives and teamworking. An overall measure of climate based on the summed subscales was also calculated.

Culture was assessed using the Competing Values Framework (CVF).15,16 This instrument was introduced into the 2003 data set, and was chosen on the basis of a previous review of culture measurement instruments17 which demonstrated that the measure had an explicit theoretical framework,16 was reasonably reliable18–20 and demonstrated associations with relevant outcomes in health care.5,18,20,21 This measure assesses two dimensions. The first describes the organization in terms of how processes are carried out, either ‘organic’ (characterized by flexibility and spontaneity) or ‘mechanistic’ (characterized by control and stability). The second describes the relationship of the organization to the outside world, either ‘internal’ (characterized by activities designed to ensure smooth functioning of the organization) or ‘external’ (characterized by competition and differentiation from other organizations). These two dimensions then define four basic cultural ‘types’: clan (internal, organic), hierarchy (internal, mechanistic), developmental (external, organic) and market (external, mechanistic).

Respondents were asked questions about five aspects of culture: characteristics, leadership, cohesion, emphases and rewards. Respondents apportioned 100 points across four sets of statements, according to which best described their organization. Each statement related to one of the four culture types. Team culture was deduced by summing, separately, the points apportioned to each of the four culture types on the five aspects of culture. The ‘dominant’ culture was that with the highest points total. In addition, culture ‘balance’ was calculated using the Blau index.11 Higher values on this index indicated a more even distribution of points between the four culture types, whereas lower values indicated a predominance of one culture type.

Quality of care

Objective data were collected from medical records, using previously developed review criteria for coronary heart disease, asthma and diabetes.22 In addition, data were collected relating to recent guidelines concerning coronary heart disease and diabetes on levels of blood pressure, cholesterol and HBA1c. Patients were selected at random from lists of patients receiving relevant drugs for each condition on repeat prescription within the previous 6 months. Data were collected for up to 20 patients per condition per practice in 1998 and for 12 patients per condition per practice in 2003.

Patient-level quality scores were computed based on a simple ratio of the number of indicators for which care was actually provided divided by the number of indicators for which care should have been provided. The score represents the percentage of ‘necessary care’ that was actually provided to each patient. Practice-level quality scores were computed as the simple average of the scores for the individual patients within each practice.

Patient report data were collected using a questionnaire based on the General Practice Assessment Survey,23,24 but which included items from the latest version of that scale, the General Practice Assessment Questionnaire which has some amended item content. The combination was used to maximize comparability with previous data sets in the longitudinal study, but to take advantage of developments in the instrument. The questionnaire included multiple scales, although only those relating to continuity (one item concerning how often patients see their usual doctor, answered on a 6-point Likert scale), communication (nine items answered on a 6-point Likert scale) and overall satisfaction (one item, answered on a 7-point Likert scale) were used in the present study in line with theoretical predictions (see methods). A random sample of 200 patients were selected by each practice, and questionnaires were sent out in May 2003 with a postal reminder 3 weeks later. The overall response rate was 47%. Practice response rates varied: nine practices (21%) had response rates greater than 60%, 24 (58%) had rates between 40% and 60% and nine (21%) rates less than 40%.

Analysis

Research question 1

The first research question concerned the association between climate and culture. Because there are multiple dimensions of climate and multiple culture types, it was decided to test the following specific hypotheses. These were based on previous empirical studies11 and the theoretical model underlying the CVF.15 The hypotheses and their rationales are given below.

Hypothesis 1: Teams with a dominant clan CVF type will score higher on the TCI participation and teamworking scales.

It was hypothesized that a cultural focus on smooth internal functioning would also be characterized by climates which supported teamworking.

Hypothesis 2: Teams with a good balance of cultures (i.e. high Blau index) are more likely to have a high overall TCI score.

This hypothesis was based on previous research which suggested that balance was related to perceived team effectiveness.11

Research question 2

The second research question concerned the associations between climate, culture and quality of care. As with research question 1, specific hypotheses were tested, based on examination of published theory15 and previous empirical findings.7,19

Hypothesis 3: Higher TCI scores will demonstrate an association with quality of diabetes care and patient satisfaction.

Previous analyses on the 1998 data set indicated that practices with higher TCI scores demonstrated higher quality of diabetes care (controlling for number of practice staff and routine booking interval) and overall satisfaction on the patient questionnaire.7 These analyses were repeated on the 2003 data set.

Hypothesis 4: Practices with higher scores on clarity of objectives and task orientation on the TCI will demonstrate better chronic disease management.

It was hypothesized that an understanding of the key aims of service delivery, combined with an emphasis on ongoing monitoring of quality of care would be conducive to the delivery of structured methods of chronic disease management.25

Hypothesis 5: Practices with a clan culture, and those with higher scores on TCI participation and teamworking scales will demonstrate more positive patient perceptions of general satisfaction, communication and continuity.

It was hypothesized that a focus on the smooth functioning within practice teams would be reflected in more positive interpersonal relationships outside that team.26,27

It should be noted that other hypotheses about associations with CVF types were proposed, but could not be tested because the sample contained so few practices of that culture type. In the interests of transparency, these are listed in Appendix.

Statistical analysis

Although the CVF and TCI instruments were completed by individual practice staff members, the conceptual basis of the measures is concerned with ‘shared’ perceptions of culture/climate within the practice team. Therefore, CVF and TCI data were aggregated across respondents within practices in line with the wider culture and climate literature18,19,28 and used as practice-level variables in the analysis. Practices with TCI response rates of less than 30% were removed on advice from the instrument developer (Michael West, personal communication).

All analyses were conducted within Stata version 9.29 Hypothesis 1 was tested using analysis of variance to compare between practices with different culture types. Hypothesis 2 was tested by calculating the Pearson correlation. Hypotheses 3–5 were tested using linear regression. The dependent variables in these analyses (condition-specific quality scores and patient questionnaire scores) were measured at the level of individual patients, while all predictors were practice-level variables. To account for this structure, the analysis used the practice as the unit of analysis and took the mean across patients in a practice as the dependent variable, then applied analytical weights based on the number of patients per practice to allow for the dependent variable being a mean.29 This approach is essentially equivalent to an analysis at the patient level with patients clustered by practice.

An alpha of 5% (two tailed) was used throughout. In view of the sample size and non-normality of some outcome measures (e.g. the patient questionnaire scales), where statistical significance testing resulted in a P value of less than 0.1, bootstrap resampling (with 10 000 replications) was used to confirm the result. This approach (which estimates the sampling distribution of an estimator by sampling with replacement from the original data) provides a more robust method of estimation in situations where the assumptions underlying conventional parametric tests may not be met.

To aid interpretation, the scores for all scales (with the exception of CVF, which is categorical), have been transformed to a range of 0 (minimum possible score) to 100 (maximum possible score). Effects of TCI are expressed as the differences in quality score (e.g. diabetes care score, out of 100) expected for a 10-point difference in climate scores (out of 100). Practice overall TCI scores had an interquartile range of 9.8 points, therefore the effect coefficient approximates to the expected difference in quality scores between practices at the 25th and 75th percentiles of TCI.

Results

The sample consisted of 42 practices and 736 professionals. Useable responses were received from 492 professionals (66.9%). However, respondents from four practices were excluded because of low response rates (<30%) from those practices. Of the remaining 38 practices, 29 (76%) were found to have a dominant clan culture. Of the remainder, six had a dominant hierarchical culture, three a dominant market culture and none had a dominant developmental culture. Culture type was therefore categorized as ‘clan’ or ‘non-clan’. Practice TCI scores were distributed over a fairly narrow range, from a minimum of 49 points to a maximum of 79 (out of 100), with the middle 50% of practices scoring between 60 and 70.

Research question 1—association between culture and climate

In support of hypothesis 1, practices with a dominant clan culture scored significantly higher than practices with non-clan cultures on both the participation subscale [clan mean 71.6, SD 7.2; non-clan mean 56.3, SD 9.2; mean difference = 15.3, 95% confidence interval (CI) 9.2–21.2] and the teamworking subscale (clan mean 69.5, SD 4.7; non-clan mean 59.8, SD 4.0; mean difference = 9.7, 95% CI 6.2–13.2).

Hypothesis 2 was not supported, as culture balance and overall TCI score were significantly negatively correlated (r = −0.42, P = 0.01). However, further analysis indicated that the high proportion of clan cultures among the sample meant that the negative correlation between TCI and balance simply reflected a positive association between TCI and the percentage of points assigned to the clan dimension (r = 0.6, P < 0.001). When the effect of this association was partialled out, the correlation changed from −0.42 to +0.23 (P = 0.18).

Research question 2—association between culture/climate and quality of care

There was no evidence of an association between overall TCI score and diabetes care or overall satisfaction on the patient questionnaire, and thus hypothesis 3 was not supported, failing to replicate previous work (Table 2).7,14

Table 2

Associations between culture/climate scales and quality of care

 Hypothesis number Regression coefficienta 95% CI P value Bootstrap P value 
Angina versus      
    TCI clarity of objectives 4.15 −0.70, 8.99 0.09 0.13 
    TCI task orientation 2.40 −1.58, 6.39 0.23  
Asthma versus      
    TCI clarity of objectives 5.22 −2.42, 12.9 0.17  
    TCI task orientation 5.15 −0.95, 11.3 0.10 0.14 
Diabetes versus      
    TCI 3.98 −1.49, 9.45 0.15  
    TCI clarity of objectives 1.60 −4.23, 7.44 0.58  
    TCI task orientation 3.78 −0.74, 8.29 0.10 0.10 
Patient questionnaire overall satisfaction versus      
    TCI 2.65 −0.63, 5.93 0.11  
    TCI participation 2.03 −0.35, 4.41 0.09 0.02 
    TCI teamworking 2.38 −1.95, 6.71 0.27  
    Cultureb 1.51 −4.00, 7.03 0.58  
Patient questionnaire communication versus      
    TCI participation 2.14 −0.17, 4.44 0.07 0.01 
    TCI teamworking 2.83 −1.35, 7.01 0.18  
    Cultureb 1.03 −4.40, 6.46 0.70  
Patient questionnaire continuity versus      
    TCI participation 3.72 0.56, 6.87 0.02 0.01 
    TCI teamworking 4.18 −1.72, 10.1 0.16  
    Cultureb 3.74 −3.82, 11.3 0.32  
 Hypothesis number Regression coefficienta 95% CI P value Bootstrap P value 
Angina versus      
    TCI clarity of objectives 4.15 −0.70, 8.99 0.09 0.13 
    TCI task orientation 2.40 −1.58, 6.39 0.23  
Asthma versus      
    TCI clarity of objectives 5.22 −2.42, 12.9 0.17  
    TCI task orientation 5.15 −0.95, 11.3 0.10 0.14 
Diabetes versus      
    TCI 3.98 −1.49, 9.45 0.15  
    TCI clarity of objectives 1.60 −4.23, 7.44 0.58  
    TCI task orientation 3.78 −0.74, 8.29 0.10 0.10 
Patient questionnaire overall satisfaction versus      
    TCI 2.65 −0.63, 5.93 0.11  
    TCI participation 2.03 −0.35, 4.41 0.09 0.02 
    TCI teamworking 2.38 −1.95, 6.71 0.27  
    Cultureb 1.51 −4.00, 7.03 0.58  
Patient questionnaire communication versus      
    TCI participation 2.14 −0.17, 4.44 0.07 0.01 
    TCI teamworking 2.83 −1.35, 7.01 0.18  
    Cultureb 1.03 −4.40, 6.46 0.70  
Patient questionnaire continuity versus      
    TCI participation 3.72 0.56, 6.87 0.02 0.01 
    TCI teamworking 4.18 −1.72, 10.1 0.16  
    Cultureb 3.74 −3.82, 11.3 0.32  

All analyses use ‘number of practice staff’ and ‘routine booking interval’ as additional covariates in keeping with previous analyses.

a

For all TCI analyses, the regression co-efficient represents the estimated change in each outcome for a 10-point difference in TCI score (see methods section).

b

Clan versus non-clan—regression coefficient represents the estimated difference in average patient questionnaire scores between clan practices and non-clan practices.

No significant associations were found between clarity of objectives or task orientation and quality of care for angina, asthma or diabetes, and thus hypothesis 4 was also not supported.

There were no significant differences between clan cultures and non-clan cultures in terms of patient reports of general satisfaction, communication or continuity, and thus this aspect of hypothesis 5 was also not supported. However, TCI participation was significantly related to perceptions of continuity, and also to overall satisfaction and communication on the evidence of the bootstrap. No associations were found between other TCI scales and patient questionnaire scales.

Conclusions

Overall, the study found some significant associations between aspects of culture and climate, but neither of these measures was predictive of quality of care. A published review found 10 studies of the link between culture and performance, and reported that only four of these 10 studies reported evidence for such a link.5 The results suggest that current policy which emphasizes the importance of culture and climate in quality improvement may be misplaced. However, there are major complexities associated with the measurement of both culture/climate and performance, and thus statements about the overall importance of climate and culture in quality improvement requires several caveats.

The association between climate and culture

As predicted, there was a significant association between clan culture and climate for participation and teamworking. Previous work on the balance index in relation to care for people with chronic illness showed it to be associated with perceived team effectiveness,11 which is itself related to climate,6,7 but the association between balance and climate in the current data was opposite to that predicted, although is most likely an artefact of the high proportion of clan cultures.

Association between culture/climate and quality of care

Previous analyses on the same sample of practices in 1998 had reported a significant association between overall climate, diabetes quality of care and patient satisfaction.7 However, when those analyses were repeated in 2003, the original results were not replicated, and there were no associations between culture and any outcome.

The reasons for the failure to replicate these results are not clear. Although the exact analysis methods differed between the two time points, reanalysis of the 1998 data using the new methods duplicated the earlier results. It is possible that changes in the practices between the two time periods (such as reductions in the proportions of small practices—Table 1) accounted for some of the differences.

Strengths and limitations of the study

Although no formal power analysis was conducted, the relatively small sample size means that the power of the study to detect associations was limited. Although the effects reported in Table 2 were generally of a modest magnitude, the CIs were wide and for some outcomes the analysis was unable to rule out the possibility that important associations exist. Clearly, the results would benefit from replication using a larger sample of practices. The response rate to the patient questionnaire was modest, and may raise issues of representativeness.

Measuring any complex psychological construct raises significant methodological challenges, and there is an ongoing debate in the wider literature concerning whether these constructs are amenable to quantification, or whether qualitative methods are required.17,30 If quantification is possible, there exist particular difficulties in the measurement of climate and culture. Both are shared attributes and are therefore analysed at the aggregate level. This reduces power, and simple aggregation procedures ignore variation within organizations,30 and may miss the importance of certain individuals (such as leaders).31 Future studies may benefit from more use of combined qualitative and quantitative methodology.5

The high proportion of practices classified as clan made some hypotheses untestable. Clearly, the utility of the CVF and the TCI will be low if they show limited variation among organizations.4 As a result it may be useful to seek more nuanced measures of differences in cultural characteristics within those broadly characterized as clan.30 Although typologies have their advantages,5 it is possible that the categorical measure of culture types is crude and is based on arbitrary thresholds.32 There are alternative typologies of organizational culture which might be useful to explore.31,33

Although the outcomes included in the current study are valid, it should be noted that both climate and culture may demonstrate larger associations with alternative outcomes such as team morale, stress and professional satisfaction and longer term team viability.34

The implications of the study for quality improvement in primary care

Policy reform is increasingly discussed in terms of changes in culture, which assumes that culture and climate are important predictors of quality of care and quality improvement.30 The current analysis would not support this model.

Other models of the association between culture/climate and quality of care are possible. For example, both constructs may determine the ‘impact’ of quality improvement initiatives. For example, the new contract may cause changes in climate and culture, which will in turn influence outcomes, so that culture and climate are important mediators of change. Alternatively, quality improvement initiatives may interact with practice climate and culture, so that particular climates and cultures will be associated with particular patterns of change in response to the contract (i.e. they are moderators of change).35 Furthermore, it is possible that climate and culture are important motivating forces in the absence of specific external financial incentives (reflecting the situation in England in 1998), but that the effects are swamped by the introduction of financial incentives and their direct effects on GP behaviour. Further research (including qualitative methods) may be required into the complex ways in which processes like climate and culture interact with other determinants of behaviour such as internal and external incentives.

Declaration

Funding: None.

Ethical approval: None.

Conflicts of interest: None.

Appendix

Other research hypotheses

The hypotheses below were proposed at the beginning of the study but could not be tested as the number of practices displaying developmental, hierarchical and market types were so low.

  1. Teams with a dominant developmental CVF type will score higher on the TCI innovation scale. It was hypothesized that teams with a focus on flexibility and a desire to compete with other organizations would be characterized by climates which supported innovation and entrepreneurship.

  2. Teams with a dominant hierarchical CVF type will score lower on the TCI innovation scale and higher on the TCI clarity of objectives scale. It was hypothesized that teams with a focus on smooth internal functioning and control would also be characterized by climates which provided clear objectives, but also reduced support for innovation.

  3. Teams with a dominant market CVF type will score higher on the TCI clarity of objectives scale. It was hypothesized that teams with a focus on control and competition with other organizations would be characterized by climates which were achievement oriented and supported clear team objectives.

  4. Practices with a hierarchical culture will demonstrate better chronic disease management. It was hypothesized that a focus on policies and procedures and the smooth functioning of operations would be conducive to the delivery of structured methods of chronic disease management.25

References

1
Anderson
N
West
M
Measuring climate for work group innovation
J Organ Behav
 , 
1998
, vol. 
19
 (pg. 
235
-
258
)
2
Schein
E
Organisational Culture and Leadership
 , 
1985
San Francisco, CA
Jossey-Bass
3
Hofstede
G
Culture's Consequences
 , 
1980
Beverly Hills, CA
Sage
4
Scott
T
Mannion
R
Davies
H
Marshall
M
Healthcare Performance and Organisational Culture
 , 
2003
Oxford
Radcliffe
5
Scott
T
Mannion
R
Marshall
M
Davies
H
Does organisational culture influence health care performance?
J Health Serv Res Policy
 , 
2003
, vol. 
8
 (pg. 
105
-
117
)
6
Poulton
B
West
M
The determinants of effectiveness in primary health care teams
J Interprof Care
 , 
1999
, vol. 
13
 (pg. 
7
-
18
)
7
Bower
P
Campbell
S
Bojke
C
Sibbald
B
Team structure, team climate and the quality of care in primary care: an observational study
Qual Saf Health Care
 , 
2003
, vol. 
12
 (pg. 
273
-
279
)
8
Goni
S
An analysis of the effectiveness of Spanish primary health care teams
Health Policy
 , 
1999
, vol. 
48
 (pg. 
107
-
117
)
9
Peiro
J
Gonzalez-Roma
V
Ramos
J
The influence of work-team climate on role stress, tension, satisfaction and leadership perceptions
Eur Rev Appl Psychol
 , 
1992
, vol. 
42
 (pg. 
49
-
56
)
10
Carter
A
West
M
Firth-Cozens
J
Payne
R
Sharing the burden: teamwork in health care settings
Stress in Health Professionals: Psychological and Organisational Causes and Interventions
 , 
1999
Chichester
John Wiley and Sons
(pg. 
191
-
202
)
11
Shortell
S
Marsteller
J
Lin
M
, et al.  . 
The role of perceived team effectiveness in improving chronic illness care
Med Care
 , 
2004
, vol. 
42
 (pg. 
1040
-
1048
)
12
Shekelle
P
New contract for general practitioners
Br Med J
 , 
2003
, vol. 
326
 (pg. 
457
-
458
)
13
Roland
M
Linking physicians' pay to the quality of care. A major experiment in the United Kingdom
N Engl J Med
 , 
2004
, vol. 
351
 (pg. 
1448
-
1454
)
14
Campbell
S
Hann
M
Hacker
J
, et al.  . 
Identifying predictors of high quality care in English general practice: observational study
Br Med J
 , 
2001
, vol. 
323
 (pg. 
784
-
787
)
15
Cameron
K
Freeman
S
Culture, congruence, strength and type: relationship to effectiveness
Res Organ Change Dev
 , 
1991
, vol. 
5
 pg. 
58
 
16
Quinn
R
Rohrbaugh
J
A competing values approach to organizational effectiveness
Public Prod Rev
 , 
1981
, vol. 
5
 (pg. 
122
-
140
)
17
Scott
T
Mannion
R
Davies
H
Marshall
M
The quantitative measurement of organisational culture in health care: a review of the available instruments
Health Serv Res
 , 
2003
, vol. 
38
 (pg. 
923
-
945
)
18
Shortell
S
O'Brien
J
Carman
J
, et al.  . 
Assessing the impact of continuous quality improvement/total quality management: concept versus implementation
Health Serv Res
 , 
1995
, vol. 
30
 (pg. 
377
-
401
)
19
Meterko
M
Mohr
D
Young
G
Teamwork culture and patient satisfaction in hospitals
Med Care
 , 
2004
, vol. 
42
 (pg. 
492
-
498
)
20
Goodman
E
Zammuto
R
Gifford
B
The Competing Values Framework: understanding the impact of organisational culture on the quality of work life
Organ Dev J
 , 
2001
, vol. 
19
 (pg. 
58
-
68
)
21
Wakefield
B
Blegen
M
Uden-Holman
T
Vaughn
T
Chrischilles
E
Wakefield
D
Organisational culture, continuous quality improvement and medication administration error reporting
Am J Med Qual
 , 
2001
, vol. 
16
 (pg. 
128
-
134
)
22
Campbell
S
Hann
M
Hacker
J
Durie
A
Thapar
A
Roland
M
Quality assessment for three common conditions in primary care: validity and reliability of review criteria developed by expert panels for angina, asthma and type 2 diabetes
Qual Saf Health Care
 , 
2002
, vol. 
11
 (pg. 
125
-
130
)
23
Bower
P
Mead
N
Roland
M
What dimensions underlie patient responses to the General Practice Assessment Survey? A factor analytic study
Fam Pract
 , 
2002
, vol. 
19
 (pg. 
489
-
495
)
24
Ramsay
J
Campbell
J
Schroter
S
Green
J
Roland
M
The General Practice Assessment Survey (GPAS): tests of data quality and measurement properties
Fam Pract
 , 
2000
, vol. 
17
 (pg. 
372
-
379
)
25
Davis
R
Wagner
E
Groves
T
Advances in managing chronic disease
Br Med J
 , 
2000
, vol. 
320
 (pg. 
525
-
526
)
26
Nievaard
A
Communication climate and patient care: causes and effects of nurses' attitudes to patients
Soc Sci Med
 , 
1987
, vol. 
24
 (pg. 
777
-
784
)
27
Yeatts
D
Seward
R
Reducing turnover and improving health care in nursing homes
Gerontologist
 , 
2006
, vol. 
40
 (pg. 
358
-
363
)
28
Shortell
S
Jones
R
Rademaker
A
, et al.  . 
Assessing the impact of total quality management and organisational culture on multiple outcomes of care for coronary artery bypass graft surgery patients
Med Care
 , 
2000
, vol. 
38
 (pg. 
207
-
217
)
29
Statacorp
Stata Statistical Software: Release 9.0
 , 
2005
College Station, TX
Stata Corporation
30
Davies
H
Nutley
S
Mannion
R
Organisational culture and quality of health care
Qual Health Care
 , 
2000
, vol. 
9
 (pg. 
111
-
119
)
31
Westrum
R
A typology of organisational cultures
Qual Saf Health Care
 , 
2004
, vol. 
13
 
suppl II
(pg. 
ii22
-
ii27
)
32
Gangestad
S
Snyder
M
‘To carve nature at its joints’: on the existence of discrete classes in personality
Psychol Rev
 , 
1985
, vol. 
92
 (pg. 
317
-
349
)
33
Krawelski
J
Dowd
B
Heaton
A
Kaissi
A
The influence of the structure and culture of medical group practices on prescription drug errors
Med Care
 , 
2005
, vol. 
43
 (pg. 
817
-
825
)
34
Firth-Cozens
J
Celebrating teamwork
Qual Health Care
 , 
1998
, vol. 
7
 
suppl
(pg. 
S3
-
S7
)
35
Baron
R
Kenny
D
The moderator-mediator distinction in social psychological research: conceptual, strategic and statistical considerations
J Per Soc Psychol
 , 
1986
, vol. 
51
 (pg. 
1173
-
1182
)

Author notes

Hann M, Bower P, Campbell S, Marshall M and Reeves D. The association between culture, climate and quality of care in primary health care teams. Family Practice 2007; 24: 323–329.