Internationally, patient safety is a growing concern. Patient harm occurs in 10% of hospital admissions. A million adverse events occur in general practice each year in Australia. Overseas data reports that patients receive recommended care only 50% of the time. We will significantly advance this work by investigating how and why this occurs, with a focus on the roles of teamwork, safe medication use and the application of information technology to support improved decision-making.
Professor Braithwaite and his co-investigators have been awarded an National Health and Medical Research Council (NHMRC) research grant totalling $8.4 million. The NHMRC Program Grant [Chief Investigators, Professor Jeffrey Braithwaite, Director Centre for Clinical Governance Research in Health (CCGR); Professor Johanna Westbrook, Director Centre for Health Systems and Safety Research; Professor Enrico Coiera, Director Centre for Health Informatics (CHI), Professor William Runciman, Director Australian Patient Safety Foundation (APSF) and Professor Ric Day, Director of Clinical Pharmacology & Toxicology, St Vincent's Hospital, Sydney] is examining, across Australia, various facts of patient safety. The grant is valued at $8.4 million. The program administered by the Australian Institute of Health Innovation, the University of New South Wales, commenced in January 2009 for five years providing exciting career and higher degree opportunities.
Despite widespread recognition of the need for reliable measures of the state of health care, we still have a poor understanding of what goes wrong in health care, and, particularly, how and why things go wrong. There is an urgent need to develop genuinely safe health services based on sound theoretical foundations, grounded in accurate, relevant measurements of the health system, and an understanding of behavioural and practice change, using technology shaped by an understanding of clinical work.:
The Chief Investigators are internationally recognized for their leadership in the field of patient safety. They bring together various interrelated areas of expertise in health informatics, medicine, medication errors and the cultural determinants of health care problems. The scope of the project includes collaboration with national and international partners. Some 15 research staff and ten higher degree students are engaged in the grant, with multiple other colleagues and partners involved. A brief description of the aims of each of the four cross-linked sub-program is given below:
Which plans are being used to treat patients, and why are they chosen?
Aim 1: To extend our capabilities to measure the adoption of clinical best practice and to determine for the first time the extent to which Australians receive care consistent with the standards suggested by evidence- and consensus-based best clinical practice.
Aim 2: To determine the reasons that underlie provision of care that deviates from best practice.
Stage 1: A random population survey of around 6600 to ask about use of health services and consent to access medical records. Stage 2: Review of consenting patient records. Stage 3: Telephone interviews with around 2000 practitioners involved in the treatment of people identified in stage 1. Stage 4: A subset around 3000 participants from stage 1 being re-interviewed.
What system problems perpetuate flawed plans and failures of their execution?
Aims: To identify and measure the impact of clinicians’ work and communication patterns, social networks, and team and organisational factors, on safe plan execution.
Which information technology interventions are most likely to enhance the selection of the right plan, and its effective execution?
Aim 1: To study the determinants of safe and effective clinical decision-making mediated by decision support technologies.
Aim 2: To determine the design parameters for safe and effective decision support system use in real world clinical settings.
Can a theoretical synthesis of safety research build a safety model that predicts the dynamic and complex interactions of health service performance?
Aim 1: To use information from the literature and our research programs to develop and iteratively refine a computational model of each of the layers of health care and of the interactions between them.
Aim 2: To use modelling and simulation to help identify practices and organisational components that fail tests of safety and quality, and to use the model as a predicative tool to guide research and policy about the safety and quality remedies most likely to succeed in given contexts.
The lessons learnt are in the diverse patient safety domains of medication safety, appropriateness, e-health, incident classification, uptake of evidence, and systems change. The twelve studies outlined below are a small sample of over 150 peer-reviewed papers that have been published.
We demonstrated empirically for the first time that interruptions to nurses during medication administration are associated with a significant increase in both the incidence and severity of medication errors.
We published the first large-scale Australian study demonstrating the effectiveness of electronic prescribing systems in two hospitals to reduce prescribing errors. Overall, there was a significant 60% reduction in total prescribing error rate and a significant decrease in the proportion of serious prescribing errors. We examined two commercial e-prescribing systems and found both introduced new ‘system-related’ errors which constituted 35% of all post-intervention prescribing errors. The paper has received wide national and international media attention.
We used a highly innovative observational method to understand how decision support influenced prescribing decisions during ward rounds. We showed that there was a mis- match between the decision-makers (senior clinicians on the ward rounds) and those doctors using the electronic prescribing systems (junior doctors). As such the decision-makers did not have access to the electronic alerts, and junior doctors ignored decision alerts in this situation. This is one of the first studies to identify the context in which decisions and electronic information systems are used may greatly influence the impact of decision support.
We have undertaken the first population-based study (“CareTrack”) on the appropriateness of health care received by Australians. The study is using a retrospective medical record review on over 1,000 patient records with 22 common conditions and 522 indicators. The indicators have been developed from national and international guidelines.
Major nations are spending tens of billions of dollars on different approaches to the design and implementation of eHealth, with limited success. We showed that top-down centralized (UK) and bottom-up laissez-faire (US) approaches to structuring national IT programs were failing, and proposed a new ‘middle out’ approach to co-ordination. Within six months, the final report of Australia’s National Health and Hospital Reform Commission specifically recommended Australia’s eHealth program be governed using this middle-out model; similar recommendations have since been made in England and Canada.
Although incident reporting is one of the foundations for patient safety improvement, the large volume of data and the highly resource-intensive process for manual review can limit timely responsive action. We demonstrated that automatic categorisation of incident narratives is highly comparable to resource-intensive expert classification with high accuracy rates for incidents related to handover and patient identification.
We have developed the first classification for incidents involving e-health systems. This paper was cited in the 2011 US Institute of Medicine (IOM)’s report on the safety of e-health as providing “new data” about the risks of e-health to patient safety. Our classification is currently being used to categorise incident reports from the UK National Health Service in England and Wales.
In our study on clinical trials involving cholesterol-modifying drugs, we examined the differences between industry and non-industry funding on trial design. The study showed that industry-funded trials are larger, faster, less likely to consider safety outcomes and just as unevenly-distributed across the classes of drugs as non-industry trials.
In a second study we looked how industry-affiliated authors contribute to the evidence base. We found that these authors were more central in co-authorship networks, and tend to receive a greater number of citations. The study raises concerns about the level of influence industry-based authors have over the evidence used to change clinical decision-making.
The development of effective interventions to minimize the risks of poor quality care remains elusive and progress in effecting substantial change in the quality of care has been slow internationally. The publication of our novel proposal for a system-wide model to control adverse events, based upon ‘cap and trade’ has made a significant contribution at the highest level internationally to exploring how we can model system- level interventions, and how they might be engineered.
To assist in determining the efficacy of aviation-style Crew Resource Management (CRM) training in improving teamwork and patient safety, we conducted a randomised controlled trial of a CRM training intervention. Main outcome measures consisted of pre- and post-test quantitative participant teamwork attitudes, and post-test quantitative trainee reactions, knowledge and behaviour. Positive changes were found in knowledge (mean difference 1•50, 95% CI 0•58 to 2•43, P=0•002), self-assessed teamwork behaviour (mean difference 2•69, 95% CI 0•90 to 6•13, P=0•009) and independently observed teamwork behaviour (mean difference 2•30, 95% CI 0•30 to 4•30, P=0•027) when the classroom trained group was compared with the control. Our study provides a substantial contribution to the small body of high quality evidence in relation to the efficacy of classroom CRM training in health care.
Globally health service accreditation is a strategy being used by governments to promote and assess the quality and safety of care. The evidence to support the investment in effort and resources in accreditation is indeterminate. We have undertaken the largest study to examine the relationships between accreditation outcomes and organisational and clinical performance and consumer involvement. Accreditation results predict leadership behaviours and cultural characteristics of healthcare organisations but not organisational climate or consumer participation, and a positive trend between accreditation and clinical performance is noted. The findings from this study, and associated accreditation research work, have been drawn up by accreditation agencies nationally and internationally, the International Society for Quality in Health Care (ISQua) and informed the Australian Commission on Safety and Quality in Health Care (ACSQHC) work in reforming the accreditation programs nationally.
For further information about the program or degrees and employment opportunities contact Professor Jeffrey Braithwaite at firstname.lastname@example.org or call +61(2) 9385 3861.