Skip to 0 minutes and 6 seconds Step 2.8 concluded with a recommendation to look beyond AMR to evidence about effective behaviour change interventions for other health care outcomes. Audit and feedback is widely used as a strategy to improve professional practice. It is believed that health care professionals are prompted to modify their clinical practice when given performance feedback showing that it is inconsistent with a desirable target.
Skip to 0 minutes and 33 seconds A Cochrane review of 140 randomised trials from a wide range of health care settings, found that feedback was more likely to be effective if it was delivered more than once, both written and verbal, given by a respected peer or colleague rather than an external source regulator, targeted at behaviours with room for improvement, and accompanied by explicit recommendation and action plans for change. In the Cochrane Review Of Interventions To Improve Antibiotic Prescribing to Hospital Inpatients, we found that feedback enhanced the effectiveness of enabling interventions such as reminders or review of patients by stewardship teams. However, feedback was only used in 36% of these 66 interventions, and only half of the feedback interventions included action plans for change.
Skip to 1 minute and 28 seconds These findings influenced the design of an Antimicrobial Stewardship intervention in 47 hospitals in South Africa. The participating hospitals were part of a private health care organisation called Netcare so we will refer to this as the Netcare intervention.
Skip to 1 minute and 50 seconds Results of the Netcare intervention have been published in Lancet Infectious Diseases. A video by Adrian Brink provides more information about participating hospitals and intervention design. This video can be viewed in full on step 6.2 of the antimicrobial stewardship MOOC, which will be linked in the text below this video. We’re going to use the AACTT framework to look at how behaviour is specified. We’re then going to look at seven behaviour change techniques that can be identified in the intervention. Goal setting, feedback on behaviour, action planning, social comparison, reward and threat, salience of consequences, and demonstration of behaviour.
Skip to 2 minutes and 38 seconds The action was to stop antibiotics, and the actors were pharmacists working in continuing care wards or ICUs and HDUs. There were four groups of target patients for stopping antibiotics– those receiving more than 7 or more than 14 days of antibiotics, those receiving 4 or more antibiotics, and those with double or redundant antibiotic coverage. The timing for the action was clearly specified for each of the four groups of target patients.
Skip to 3 minutes and 13 seconds Stopping antibiotics was the primary action for the Netcare intervention. However, there were other key actions that enabled pharmacists to stop antibiotics. The AACTT framework includes both focal and ancillary actions. This example from the AACTT framework paper is about hand hygiene. The focal action is staff cleaning their hands with alcohol gel. The ancillary actions are that cleaning staff refill alcohol gel dispensers, and that hospital administrators order dispensers and hand gel.
Skip to 3 minutes and 52 seconds Five ancillary actions in the Netcare intervention can be identified in the paper and in step 6.3 from the Antimicrobial Stewardship MOOC. The first three are giving pharmacists specific responsibility for driving change, allocating time for the intervention, and prioritising their time for the highest risk patients. Hospital managers were the actors for these three ancillary actions. The fourth ancillary action is giving pharmacists the capability to stop antibiotics for the target patients. This required teaching them whatever they needed to know about stewardship. These were not ID pharmacists like in the USA, or even clinical pharmacists, which they don’t have in South Africa yet. Infection specialists were presumably the actors for training of the pharmacists.
Skip to 4 minutes and 44 seconds The fifth ancillary action was identification of the target patients for the pharmacist to review and stop antibiotics. It’s not clear how this was done or who did it. Most hospitals don’t have access to individual patient prescribing data, so this is a very important issue for planning a similar intervention in other hospitals.
Skip to 5 minutes and 13 seconds The Netcare intervention goal is to achieve a 10% reduction in antibiotic consumption across the participating hospitals within two years. The goal was presented to multidisciplinary teams and endorsed by doctors at each participating hospital. Feedback applied evidence from the Cochrane review on effectiveness of audit and feedback. It was delivered monthly, in both written and verbal formats, by a colleague who was the improvement project manager.
Skip to 5 minutes and 48 seconds Participating hospitals adapted their action plans according to how many of the targeted interventions have been implemented, the improvements that are being made, and the effect on antimicrobial consumption in their hospitals. They receive monthly feedback about comparison of their hospital with other participating hospitals.
Skip to 6 minutes and 13 seconds Dr. Dena van den Bergh is the Director of Quality Leadership at Netcare. She has over 17 years’ experience in leading large scale change and improvements in health care in South Africa and internationally. You can view videos in which Dr. Van den Bergh discusses the intervention design on steps 6.3 and 6.6 of the Anti-Microbial Stewardship MOOC, which will be linked in the text below this video. You can identify three BCTs that we encountered at step 2.7 of this Social Science and Behaviour Change MOOC. The first BCT is social comparison. Comparison between hospitals was an important stimulus to change, because it enabled them to see and learn from what others were doing. The second BCT is reward and threat.
Skip to 7 minutes and 5 seconds Hospitals were rewarded by seeing the success of the project and their contribution to that success. But comparison could also be a threat if it identified hospitals that were not contributing, “you didn’t want to be one of the hospitals that was in red”. However, this threat could be used as a powerful stimulus to change by the project leads from these hospitals. The third BCT is salience of consequences. Reduction in antibiotic use was clearly linked to reduction in resistance throughout the intervention. Feedback about progress toward goals is another opportunity to reinforce the importance of the outcomes that will be achieved.
Skip to 7 minutes and 52 seconds Dr. Angelika Messina is a Quality Leadership Manager at Netcare, and was the project manager for their antimicrobial stewardship programme. You can view a video in which Dr. Messina discusses how comparison facilitated improvement on step 6.7 of the Antimicrobial Stewardship MOOC, which will be linked in the text below this video. You can identify another BCT, which is demonstration of the behaviour. This could happen between hospitals when a hospital shared successful changes with other hospitals. It could also happen within a single hospital by demonstrating successful changes between wards or units.
Skip to 8 minutes and 37 seconds There was little change in outcome or process measures in the first year of the Netcare intervention. These graphs show data for the outcome measure, which was total antibiotic use over more than five years, because the graph includes 16 months of pre-intervention and 20 months of post-intervention data. In contrast, the process measures only show data for 24 months of the intervention phase. Interventions take time to change complex systems. There was very little improvement in process or outcome measures until the second year of the Netcare intervention, which was implemented across 47 hospitals. However, in step 1.10, we reviewed an intervention that took over 10 months to achieve 95% reliability for discontinuing surgical antibiotic prophylaxis in any two hospitals.
Skip to 9 minutes and 36 seconds The Netcare intervention applied a model for improvement which was introduced by Walter Shewhart in the 1920s. It’s now widely used for improvement in health care, and there’s an open access guide to applying the model from NHS Improvement. The model asks three questions. What are we trying to accomplish? How would we know that a change is an improvement? And what changes can we make that will lead to improvement? The model uses Plan, Do, Study, Act, or PDSA cycles, to test changes that may lead to improvement. Using PDSA cycles enables you to test out changes on a small scale, building on the learning from these test cycles in a structured way before wholesale implementation.
Skip to 10 minutes and 23 seconds This gives stakeholders the opportunity to see if the proposed change will succeed and is a powerful tool for learning from ideas that do and don’t work. This way, the process of change is safer and less disruptive for patients and staff. The four stages of this PDSA cycle are Plan, the change to be tested or implemented, Do, carry out and test for change, Study, based on the measurable outcomes agreed before starting out, collect data before and after the change, and reflect on the impact of the change and what was learned. Act, plan the next change cycle or full implementation.
Skip to 11 minutes and 8 seconds Note that the detail about the plan stage of the PDSA cycle uses the same language as the AACTT framework– who, what, where, and when.
Skip to 11 minutes and 24 seconds Although widely recommended as an effective approach to improvement, PDSA cycles can be challenging to use. This systematic review of PDSA in health care used the origins and theory of PDSA cycles to identify five key features. Documentation is crucial to support local learning and transferability of learning to other settings. Second feature is iterative cycles. Multiple PDSA cycles must occur. Depending on the knowledge gained from a PDSA cycle, the following cycle may seek to modify, expand, adopt, or abandon a change that was tested. Third feature is prediction-based tests of change. Prediction of the outcome of the changes developed in the plan stage of a cycle. This change is then tested and examined by comparison of results with the prediction.
Skip to 12 minutes and 22 seconds The fourth feature is small scale testing. PDSA starts small in scale and builds as confidence grows. This allows the change to be adapted according to feedback, which facilitates rapid change and learning. The final feature is use of data over time. This is necessary to understand the impact of a change on the process or outcome of interest. The framework was then used to explore the consistency with which the method has been applied in peer reviewed publications from health care. The review included 73 papers. Only 47 of the 73 publications reported in enough detail to assess the application of the framework.
Skip to 13 minutes and 8 seconds Note there is not enough detail to assess application of the framework to PDSA cycles in the Netcare intervention, either in the Lancet ID paper or in the videos. Of the 47 papers in the review, only 30% fully documented the application of the sequence of iterative cycles. Furthermore, a lack of adherence to the nation of small scale change was apparent. Only 8% of studies reported small scale testing at less than monthly intervals, and only 15% use quantitative data at monthly or more frequent data intervals to inform progression of cycles.
Skip to 13 minutes and 53 seconds Reporting of Antimicrobial Stewardship interventions could be improved by using the TIDieR checklist. This was published in 2014 and continues to be accessed over 2000 times a month in 2020. The original publication has very helpful additional online resources, including detailed examples of different formats, which can be used to describe and provide study intervention material through websites and smartphone apps, in addition to peer reviewed publications. Note that this checklist emphasises the need to report on adaptation and modification of an intervention. Use of PDSA cycles enables documentation of these adaptations and modifications.
Skip to 14 minutes and 45 seconds In conclusion, Netcare’s collaborative model for improvement enabled pharmacists to achieve sustained reduction in antibiotic use across 47 hospitals. Netcare attributed their success to applying skills beyond those of infectious diseases and microbiology, particularly in quality improvement. The Netcare study and others that we’ve reviewed showed that change takes time. Improvement in process may take a year, and improvement in outcome may take even longer. Reporting of interventions is critical to successful replication. Remaining questions for the Netcare intervention are, what were the ancillary actions, and who were the other actors that enabled pharmacists to stop antibiotics? How were changes tailored and modified, especially in the first year?
Skip to 15 minutes and 39 seconds Structure reporting of antimicrobial stewardship interventions could be improved by using the TIDieR checklist and guide to report both the intervention design and its implementation. You can use the Behaviour Change Wheel to document three key steps in the intervention design. First use the AACCT framework to specify the Behaviour that you’re trying to change. Second, use the Com-B model and the Theoretical Domains Framework to ask why is current Behaviour as it is, and what would it take to bring about change? And third, identify Behaviour Change Techniques to match the intervention components to the behavioural diagnosis. Then finally use PDSA cycles to document small scale testing, modification, and adaptation of changes over time.
The Netcare intervention across 47 hospitals in South Africa
In this video, Professor Peter Davey introduces Behaviour Change Techniques and their use in AMS interventions. He then takes you through the Brink intervention and analyses factors such as Goal specification, Participant involvement, Feedback and Action plans for change.
The Cochrane Review of Interventions to Improve Antibiotic Prescribing to Hospital Inpatients revealed that both the content and reporting of interventions for antimicrobial stewardship fell short of scientific principles and practices. Additionally, there is a strong evidence base regarding BCTs in other contexts that should be applied to antimicrobial stewardship now.
Peter discusses a clip on the Brink Intervention by Dr Adrian Brink, which is featured in the Antimicrobial Stewardship: Managing Antibiotic Resistance FutureLearn course, on step 6.2. The clip features a discussion on the importance of implementing AMS in daily practice using a quality improvement model and behaviour change techniques.
You will then be introduced to PDSA Cycles for testing change for improvement and theoretical frameworks made based on them.
Please find below links to the papers referenced in the video.
Please find a downloadable PDF of the PowerPoint slides below.
In the comments below, please share your thoughts on the Brink Intervention, particularly relating to the following issues:
- Goal specification
- Participant involvement
- Action plans for change
© BSAC & UoD