Context Bureaucratic organisational culture is less favourable to quality improvement, whereas organisations with group (teamwork) culture are better aligned for quality improvement.
Objective To determine if an organisational group culture shows better alignment with patient safety climate.
Design Cross-sectional administration of questionnaires.
Setting 40 Hospital Corporation of America hospitals.
Participants 1406 nurses, ancillary staff, allied staff and physicians.
Main outcome measures Competing Values Measure of Organisational Culture, Safety Attitudes Questionnaire (SAQ), Safety Climate Survey (SCSc) and Information and Analysis (IA).
Results The Cronbach alpha was 0.81 for the group culture scale and 0.72 for the hierarchical culture scale. Group culture was positively correlated with SAQ and its subscales (from correlation coefficient r=0.44 to 0.55, except situational recognition), ScSc (r=0.47) and IA (r=0.33). Hierarchical culture was negatively correlated with the SAQ scales, SCSc and IA. Among the 40 hospitals, 37.5% had a hierarchical dominant culture, 37.5% a dominant group culture and 25% a balanced culture. Group culture hospitals had significantly higher safety climate scores than hierarchical culture hospitals. The magnitude of these relationships was not affected after adjusting for provider job type and hospital characteristics.
Conclusions Hospitals vary in organisational culture, and the type of culture relates to the safety climate within the hospital. In combination with prior studies, these results suggest that a healthcare organisation's culture is a critical factor in the development of its patient safety climate and in the successful implementation of quality improvement initiatives.
- Patient safety climate
- quality improvement
- organisational culture
- healthcare quality improvement
- safety culture
Statistics from Altmetric.com
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.
- Patient safety climate
- quality improvement
- organisational culture
- healthcare quality improvement
- safety culture
The introduction of quality improvement into healthcare has brought attention to the relevance of organisational culture. Indeed, models for total quality management emphasise the fundamental importance of gaining leadership support for a psychology of change in order to bring about change.1 A growing body of literature shows that the ability to make improvements depends on organisational context.2–4 Organisational cultures that emphasise teamwork and innovation have been found in alignment with quality improvement, whereas bureaucratic, hierarchical cultures, which inherently promote stability and resist change, are less suited for quality improvement.5–9
Systems engineering models for highly reliable organisations as well as crew resource management principles also emphasise collaborative communication, leadership, organisational learning and teamwork.10 11 Based on the prior findings of a fundamental relationship between an organisation's culture and quality improvement, we hypothesised that organisational culture emphasising teamwork and/or innovation will be aligned with favourable attitudes towards patient safety. This study surveyed personnel in intensive care units on organisational culture and safety climate to characterise whether organisations vary in culture and to assess the association between organisational culture and safety climate. A secondary aim was to produce a greater understanding of the psychometrics of these scales and their use by level of analysis, and to provide a reference for their use.
The study was a cross-sectional analysis of surveys distributed across 61 hospitals managed by HCA (formerly, Hospital Corporation of America). HCA distributed the survey to providers and managers. Instructions stated that participation was voluntary and anonymous. Surveys were sent to 110 adult and paediatric Intensive Care Units (ICUs) between 20 December 2005 and 28 February 2006. HCA provided data on process, outcome and covariate measures aggregated within the medical centre. The project was approved by the Vanderbilt University IRB.
The surveys included four existing survey instruments. The Competing Values Organisational Culture Assessment Instrument (see online appendix) assesses staff perception across five characteristics: organisational character, management, cohesion, emphasis and distribution of rewards.12–14 The particular mix of values shapes the attributes of the organisation and characterises its framework for structure, order, orientation to strategy and development, leadership style and mode of operation that form the organisation's behaviour and common way of thinking.13–17 Measures of organisational culture included the continuous scaled score for each of the four culture types and a category score for dominance. An organisation is classified as having a dominant culture if one of the four culture scales exceeds all others by five or more points; otherwise, the organisation has a balanced culture.8 9 The group culture scale measures values associated with affiliation, teamwork and participation. The developmental culture scale is characterised by risk-taking, entrepreneurial innovation and change. The hierarchical culture scale reflects bureaucracy and chain of command. The rational culture scale emphasises efficiency, production and achievement.
The Safety Attitudes Questionnaire (SAQ), a validated 30-item survey, was used to measure ICU safety climate across six scales: (1) Teamwork Climate, (2) Safety Climate, (3) Job Satisfaction, (4) Perceptions of Management, (5) Stress Recognition and (6) Working Conditions.18 The Safety Climate Scale (SCSc) is a 10-item survey adapted from aviation safety and used to assess patient safety climate.10 19 20 Information and Analysis (IA) is a four-item scale used in the RAND ICICE study that assesses how the organisation uses data and information related to quality improvement.8 9 21
Data and analysis
Analysis at the respondent level included Cronbach alpha as a measure of internal consistency, comparisons of means and medians for respondent and hospital characteristics, and Pearson correlation coefficients across survey scales and subscales. Scoring rules for subscale and scale scores imputed means for missing values, provided the majority of subscale items are present; otherwise, the scale was scored as a missing value. A random-effects mixed model with covariates was used to model the effect of organisational culture on safety climate after adjusting for provider job type, ICU type and hospital characteristics (location, hospital admissions per year, percentage of Medicare/Medicaid patients and percentage of emergency department admissions). Because all multivariable models were nearly identical to and replicated our results, we choose to present the simpler analysis for ease of presentation and readability. Data were also aggregated at the ICU and the hospital level for analyses on organisational culture and climate. One-way analysis of variance (ANOVA) was used to analyse for differences between job types and hospitals classified by dominant type of culture.
Sixty-seven ICUs (61%) from 41 hospitals (67%) returned 1502 (43.4%) completed surveys. Three of these ICUs had fewer than six respondents accounting for deletion of 9 surveys; 86 surveys had missing data for organisational culture, and one survey had missing data for the safety climate scales. Therefore, the data used in analysis included 1406 respondents representing 64 total ICUs (58%) from 40 hospitals (66%). The hospitals were primarily suburban but included rural (18%) and inner city (28%) hospitals. ICUs were 17% surgical, 22% medical and remainder combined medical–surgical. Survey respondents were nurses (66.2%), ancillary nursing staff (8.3%), allied clinicians (14.2%) and physicians (5.3%). The HCA hospitals participating in the survey did not differ from the non-participating hospitals in administrative variables (data not shown).
Psychometrics of organisational culture instrument
The descriptive statistics for the four scales of organisational culture are shown in table 1. The group and hierarchical scales had Cronbach alpha values of 0.81 and 0.72, respectively, whereas the developmental and rational scales obtained unreliable values, excluding further analyses on these subscales.22 The correlation matrix (table 2) shows a significant negative correlation of −0.73 between the group and hierarchical culture scales.
Variation of organisational culture across hospitals
Pooling across respondents and using conventional scoring rules,8 9 15 of the 40 hospitals (37.5%) were classified as hierarchical culture organisations, 15 (37.5%) as group culture organisations and 10 (25%) as balanced-culture organisations.
Psychometrics of the climate scales
The descriptive statistics and Cronbach alpha for the SAQ, SCSc and IA scales are shown in table 3. The correlations between the overall SAQ climate score and the subscales are displayed in table 4. The Cronbach alpha of the SAQ and subscales is greater than 0.70, and each item displayed substantial item-to-scale correlations. Except for Stress Recognition, the SAQ subscales show positive correlations ranging between 0.67 and 0.81. Stress Recognition is negatively correlated with all of the other SAQ subscales and is not correlated with the SAQ overall score. The correlations at the respondent level of analysis were 0.86 between the SAQ and SCSc, 0.68 between the SAQ and the IA, and 0.70 between the SCSc and IA.
Ratings by job type
Physicians had higher scores on group culture and lower scores on hierarchical culture (table 5). Nurses and ancillary providers scored group culture lower and hierarchical culture higher. Nurses also had lower ratings on the SAQ and SCSc compared with physicians but did not differ on the IA scale.
Association between organisational culture and patient safety
The univariate correlations between the group and hierarchical culture scales and the SAQ, SCSc and IA scales are shown in table 6. Except for the stress recognition subscale of the SAQ, there was consistently a high positive correlation between the patient safety climate and group organisational culture scales and a high negative correlation between safety climate and hierarchical organisational culture scales. The partial correlations between organisational culture and safety climate adjusted for administrative hospital characteristics, provider job type and hospital location resulted in correlations and statistical significance equivalent to the unadjusted respondent level results shown in table 6.
Using a hospital-level one-way ANOVA, mean SAQ, SCC and IA scores were compared across hospitals categorised into their dominant category of hierarchical, balanced or group culture (table 7). The hierarchical culture hospitals had significantly lower average safety climate scores than did the group culture hospitals. The balanced-culture hospitals scored higher than the hierarchical-culture hospitals on all scales but lower than the group-culture hospitals on the overall SAQ and IA scales. A multivariate mixed model accounting for nesting of respondents and adjusting for covariates replicated the statistical significance of the one-way analysis of variance results shown in table 7. In addition, differences between ICUs within the same hospital were minor, amounting to only 2.7% of total variation.
The findings of this study support our premise that hospitals vary in organisational culture and that the type of culture within the hospital would relate to perceived safety climate. We found a substantial variation in hospital organisational culture resulting in almost equal percentages of dominance among group, hierarchical and balanced types of culture. There was a significant positive association between group-oriented organisational culture and the attitudes and climate for patient safety, and a significant negative association between hierarchical culture and these same factors.
Our findings replicated the high internal consistency for the group and hierarchical scales, and the low consistency of the rational scale,8 9 21 23 and did not show internal consistency for the developmental scale.24 25 In addition, our negative correlation of −0.73 between the group and hierarchical scales replicates the inverse correlation of −0.65 previously reported.7 Our findings also replicate the psychometric findings for the SAQ, SCC, and IA.18 19 26 The pattern of our SAQ findings with regard to the stress recognition subscale is consistent with prior work,18 indicating that the stress recognition items do not contribute positively towards the construct of safety climate as intended and should be excluded from the SAQ. Physicians and administrators provide higher safety climate ratings than do nurses and ancillary staff. Furthermore, nursing staff viewed culture as more bureaucratic5 27 and rated their organisation lower in safety climate.11 19 20 28 Comparing organisations and interpretation using these scales must consider the differences between professions in perception of culture.
Organisations characterised as having group culture are more effective, and hospitals with hierarchical culture are less effective at implementing quality improvement,8 9 24 29 team functioning and staff morale,5 15 patient satisfaction2 23 24 and overall safety climate.25 30 Collaborative teams from balanced hospital cultures have greater perceived team effectiveness.8 9 Our findings extend the results of these prior studies; we show that group culture is associated positively with patient safety teamwork climate, safety climate, job satisfaction, perceptions of management, working conditions, and information and analysis. These findings provide significant criterion validity to the theoretical framework that group culture is supportive of patient safety climate and quality improvement.
Restrictions to our sampling produced some study limitations. Our participating hospitals may differ substantially from non-participants. Moreover, ICU respondents may not be representative of the hospital in general.16 In addition, study physicians were community practitioners comprising a relatively small portion of the sample. Our response rate was 43%, and our measures were restricted to a single point in time. Finally, HCA facilities represent only a private, for-profit hospital sector of the healthcare industry. Only further research can clarify the effects of these limitations, but reproducing study results similar to those reported in the literature enhances our confidence regarding the generalisability of our findings about organisational culture.
The influence of context on quality improvement and patient safety is widely appreciated.29 31 Quality improvement and patient safety studies cite contextual factors involving the micro and macro systems of healthcare organisations.29 31–37 However, measures of organisational structure are generally crude, focussing on such things as teaching status, hospital size, or percentage capitated care. The Competing Values Model provides a conceptual framework and instrument for measuring organisational culture that has been employed in the study of firms but rarely in the industry of healthcare.13 The current study's findings in combination with prior research suggest that a healthcare organisation's culture is a pivotal force to contend with in the implementation of quality improvement and the development of a patient safety climate. Culture's role in facilitating innovation may be the enabling factor for the success of quality initiatives and key to effective implementation of evidence-based findings derived from translational research. Future healthcare system intervention studies should include evaluation of the culture underlying the structure of the healthcare sector to understand better how we can bring about the positive impacts of healthcare reform the nation hopes to achieve.
We would like to thank the AHRQ collaborative investigators for their work in this study: XL Liu, L Brewer, J Hickok, S Horner, S Littleton, P McFadden, S Mok, J Perlin, J Reischel, SG Chernestky Tejedor and all the HCA medical centres that participated in this project.
Funding This project was supported by Partnerships in Implementing Patient Safety (PIPS) from the Agency for Healthcare Research and Quality (AHRQ), Grant Award Number: U18 HS015934.
Competing interests None.
Ethics approval Ethics approval was provided by the Vanderbilt University School of Medicine.
Provenance and peer review Not commissioned; externally peer reviewed.