Automated audit and feedback as a tool to measure and improve the quality of primary care

Summary This thesis investigated automated audit and feedback towards patients and professionals as a tool to measure and improve the quality of primary care for diabetes type 2 and chronic kidney disease. Introduction Although audit and feedback is a well-studied intervention to assess and improve the quality of healthcare, there is no gold standard available for its design and implementation. Involving patients in an audit and feedback intervention can be a challenging task and little is known about the needs and preferences of patients when interacting with their health data that is stored in the electronic health record of the general physician. Moreover, it remains unclear if patients even want to receive feedback about their health data. In addition, recent research is not benefiting the field of audit and feedback research in general and of automated audit and feedback research in particular. New trials are not building on existing knowledge to explore factors that are responsible for improving the effectiveness of (automated) audit and feedback intervention. The question is not anymore whether audit and feedback works (it does), but rather how it works, especially in primary care. Furthermore, quality indicators that can be used in an automated audit and feedback intervention for chronic diseases in primary care are also lacking. Finally, the data-completeness of the electronic health record of the general physician and more particularly the correct registration of diagnoses needs to be improved. Automated audit and feedback can also be used in this regard. Methods To investigate the health information needs and expectations of patients when interacting with their digital health data a cross-sectional study was performed. In addition, patient related characteristics were investigated to examine if they had an influence on 2 different variables, namely the importance people attributed to obtaining health information through a patient portal and their expectations concerning their personal healthcare when consulting a patient portal (chapter 2). The effectiveness of automated audit and feedback in primary care and facilitating features important in the design of this intervention, were explored with the help of a systematic review (chapter 3). Furthermore, 2 sets of quality indicators, extractable out of the electronic health record of the general physician, were developed with a Rand-modified Delphi method for diabetes type 2 and chronic kidney disease (chapter 4). Finally, to improve the data-completeness of the electronic health record of the general physician, and in particular the complete and correct registration of diagnoses, a protocol for a cluster randomized controlled trial was designed that described a pilot implementation strategy for an automated audit and feedback intervention in primary care based on our previous findings (chapter 5). Results Patients were interested in the use of a patient portal to access their digital health data and they wanted to receive feedback based on this health data stored in the electronic health record of their general physician. More specifically, they liked to receive alerts or some form of communication through a patient portal if they needed to take action to manage their health. Moreover, people who were middle-aged and who considered patient empowerment important, attached a greater importance to obtaining health information through a patient portal. In addition, the expectations people had concerning their personal healthcare when using a patient portal were influenced by their level of education, the interest in shared-decision making and by the difficulty people experienced in finding and understanding relevant healthcare information. Overall, automated audit and feedback interventions in primary care were effective in improving the quality of care in three quarters of the included studies. Moreover, automated audit and feedback was effective in all studies targeting preventive medicine and medication safety. The use of benchmarks as comparators in the delivery of feedback and the evidence based aspect of feedback were important and well-used features of automated audit and feedback in primary care. However, other important features of automated audit and feedback in primary care, such as the frequency of feedback provision and the cognitive load of feedback were described more irregularly, indicating the need for a framework or methodology to design future audit and feedback interventions. In addition, some benefits of automated audit and feedback in primary care were its potential to be cost-effective, the scalability of the interventions and the fact that automated audit and feedback could target many different conditions and procedures. Two sets of quality indicators, extractable out of the electronic health record of the general physician, were developed for diabetes type 2 and chronic kidney disease, respectively. The first set of quality indicators for diabetes type 2 included 39 quality indicators and included categories such as screening and prevention (3 QIs), diagnosis (1 QI), management (24 QIs), complications of diabetes type 2 (8 QIs) and finally 3 quality indicators that were added by the panel members. The second set of quality indicators for chronic kidney disease included 36 quality indicators on different aspects of primary care for chronic kidney disease and included categories such as definition and classification (1 QI), diagnosis and screening (1 QI), management (8 QIs), treatment (8 QIs), medication and patient safety (6 QIs) and referral to a specialist (12 QIs). A protocol for a cluster randomized controlled trial was developed based on our previous findings. This protocol described a pilot implementation strategy of an automated audit and feedback interventions to improve firstly the complete and correct registration of diabetes type 2 and secondly the registration of chronic kidney disease in the electronic health record of the general physician. The quality indicators concerning diagnoses we designed in chapter 4 were used in this protocol. In addition, 4 other quality indicators were chosen from the 2 sets we developed so that a baseline evaluation of these 4 quality indicators could be performed. The automated audit and feedback intervention that was designed also incorporated the 4 features we examined in our systematic review namely the use of benchmarks as comparators, evidence-based feedback, feedback provided more than once and feedback with a low cognitive load. Furthermore, some exploratory secondary outcomes were also included in this protocol. These exploratory outcomes were the correct and complete registration of heart failure and of a few lifestyle and biometric parameters in the electronic health record, namely smoking habits, physical exercise, alcohol use, weight and height. Conclusion This thesis presented new research concerning the impact and optimal conditions of automated audit and feedback towards patients and professionals in primary care. Patients were identified as critical healthcare users who wanted feedback based on their digital health data. Automated audit and feedback was in general effective in primary care, especially for preventive medicine and medication safety. Two sets of quality indicators that could be extracted out of the electronic health record of the general physician were developed for diabetes type 2 and chronic kidney disease. Furthermore, a pilot implementation strategy to improve the data-completeness of the electronic health record of the general physician was designed. These findings contributed to the design of a Covid surveillance system in primary care and could assist in identifying priority risk groups for Covid vaccination. In the future this research could also contribute to monitor the Covid vaccination status of a population.

Jaar van publicatie:2021

Toegankelijkheid:Open