Appropriate use of cardiac imaging in Australia

Health systems around the world are concerned by the same topic: How their national health budgets can be spent efficiently and economically. Some related factors, such as the ageing population, an increase in sophistication and costs of treatments and an increase in the demand for imaging services, have all contributed to the rise in healthcare spending, placing further burdens on limited resources. The healthcare system of the United States of America has found this particularly challenging, where the use of cardiovascular imaging had grown persistently for the previous two decades. This increase, associated with marked geographical variation of imaging utilisation, the spread of imaging equipment, repetition of use due to poor quality tests, and defensive medicine, have led to the concern about the value of cardiac imaging. Part of the response in the United States of America to the disconnection between the use in imaging and its value has been the development of the Appropriate Use Criteria (AUC) for Cardiac imaging. The first criteria were introduced in 2005, with subsequent versions being published to include different cardiac imaging techniques, aiming to improve utilisation of imaging technologies in an efficient way, contributing to improvement in patient care, doctors' decision making, standardisation of medical practice and control of health expenditure. The concept of appropriate use of cardiovascular imaging has remarkably influenced the relationship between clinicians, health policymakers and insurance companies to the point that from January 2017, all doctors and hospitals must certify their use and adherence to the criteria in order to receive reimbursement under the United States Medicare schedule. Despite the utilisation of the AUC for the past 12 years, several problems are apparent. Issues related to the scientific evidence of the criteria; the uncertainty of the impact of the AUC on clinicians' requesting behaviour, and on health outcomes, question the widespread use of them to achieve appropriate use of cardiovascular imaging in the United States and other countries like Australia. Due to an increased interest in the utilisation of these criteria in the Australian practice, this thesis which comprises quantitative and qualitative research methods, investigates the use of cardiac imaging techniques in Australia and if the American Appropriate Use Criteria are suitable to use in the Australian health care system. The work in this thesis is divided into three parts. The first part (Chapter 3), aims to determine if the use of cardiac imaging in Australia is appropriate based on the analyses of growth and geographical variation of imaging used in different regions of the country. Data from Medicare Australia Statistics, Australian Health Survey, and Health Workforce were analysed to define the growth and regional variation in the use of imaging. The increase of imaging was determined using rate of tests per 100,000 people in each geographical area (Medicare locals) and Spearman correlations, negative binomial regression and zero-inflated negative binomial regressions, were used to investigate univariable and multivariable associations between age-weighted testing and characteristics of each Medicare local4. The results identified the ongoing growth of cardiac imaging use per 100,000 people in Australia, more pronounced in echocardiography modalities (transthoracic, trans-oesophageal, and stress echocardiography). Stress echocardiography was the technique with the highest increase since 2002 (423%), and there was a substantial regional variation in the use of all echocardiography modalities4. Greater regional echocardiography use was associated with females, proximity to major cities, higher socioeconomic status, and local concentration of physicians, but not with disease burden or deaths due to cardiovascular causes. The medical workforce appears to be the strongest independent driver of echocardiography use. The next section (Chapters 4 to 7) aims to determine whether the Appropriate Use Criteria are suitable for use in Australia to achieve appropriate use of imaging. Several studies were undertaken to understand this. First, we related similarities and relationships between cardiac guidelines and Appropriate Use Criteria; second, we assessed the impact of the criteria on clinicians' requesting behaviour. Third, we evaluated the impact on health outcomes, and finally, factors that lead doctors to order inappropriate‚ÄövÑvp echocardiograms. For examining the similarities of Appropriate Use Criteria with guidelines (Chapter 4), the criteria and the published Cardiology guidelines were matched to identify concordance between both. Concordance was determined when appropriate‚ÄövÑvp items in the criteria had recommendation class I or IIa in the guidelines; inappropriate‚ÄövÑvp scenarios had recommendation class III, and uncertain‚ÄövÑvp items had recommendation class IIb. The results identified that 91% of the criteria had a counterpart on the guidelines for the use of echocardiography. However, only 82% of them were concordant, and the rest were inconsistent. These results suggest that the potential incorporation of an AUC process into Australian practice might face challenges to achieve appropriate use and to improve quality of care. Chapter 5 was a systematic review of published manuscripts related to the AUC use in cardiovascular imaging to study the impact of the appropriateness criteria on clinicians' ordering behaviour. The aim was to determine the proportion of appropriate‚ÄövÑvp and inappropriate‚ÄövÑvp testing over time using meta-regressions. More than 5,200 manuscripts were found in online databases and were analysed. The results showed that there was a positive association with the percentage of appropriate tests for Transthoracic Echocardiography (TTE), as well as a negative association of the ratio of uncertain tests over time, but inappropriate testing remained the same, that is no decrease in the proportion of inappropriate tests over time. These results suggest no impact on clinicians' ordering behaviour. Chapter 6 aimed to determine the effect of the definition of appropriateness on survival time in stable heart failure patients. The analysis was made using survival analysis with time-dependent variables for the combined endpoint of Heart Failure (HF) readmission and death, and a separate analysis for HF readmission, with death as a competing risk. The results showed no differences in the event-free time for combined outcomes; HF readmission was not associated with routine follow-up TTE timing; there were no differences in the cumulative incidence of death between groups. For chapter 7, a qualitative analysis was performed using semi-structured face-to-face interviews with open-ended questions to analyse the factors that lead doctors to order echocardiograms and the relationship with inappropriate testing. Seventeen physicians (cardiologists and non-cardiologists) participated in the study. Personal factors such as lack of expertise, limited experience and inability to manage uncertainty were the most important factors that impacted the decisions of doctors leading to the ordering of inappropriate tests. Other factors included accessibility of tests and adherence to protocols. These results suggest a mismatch between the clinical reasoning of physicians and the AUC for echocardiography. Due to the results of the previous chapters indicating that the AUC are not the ideal tools to achieve appropriate use, to improve health outcomes, and to change doctors' ordering behaviour, a third and final section evaluated a proposed methodology to determine possible inappropriate testing and the way to give answers to doctors without the need for performing a formal test. This third section includes chapters 8 and 9. For Chapter 8, general characteristics of inappropriate TTE were determined and summarised in a questionnaire of four binary questions most commonly associated with inappropriate tests. These questions are related to the absence or change of cardiovascular symptoms or signs, routine surveillance as the purpose of the exam, echocardiograms in the previous twelve months, and if endocarditis with no murmur or positive blood cultures was the reason for the scan. These questions were applied to two different validation groups. Analyses of specificity, sensitivity, and predictive values were performed to determine the accuracy of the prediction of inappropriate requests. Two or more affirmative questions had a high sensitivity and specificity to determine inappropriate requests according to the criteria. The time to answer the questions was limited to up to 2 minutes when it could not be answered directly from the echo request. In only 19% of the requests was there a need to the review the digital record to respond to the questions. Only 18% of the total amount of requests to the echo laboratory, warranted a comparison with the Appropriate Use Criteria. Two approaches to inappropriate requests (standard TTE versus hand-held echocardiography) were compared in a case-control study (Chapter 9). Patients were followed for a period of 6 months and outcomes such as time until the scan, repeat echocardiography, the cost of each strategy, new findings, and impact on management were determined9. The results of this study showed no differences in the finding of new pathologies or change in management between hand-held echocardiography and standard TTE. However, people in the portable echo group had less time to scan if they were inpatients and the average cost of the approach was lower than the standard echocardiography group ($145 AUD vs $241 AUD). Statistical analyses in all chapters were performed using R software; geo-mapping for Chapter 2 was done using QGIS 2.4.0.; hand-held ultrasonography was performed using the GE Vscan V 1.2 portable device and standard TTE was performed GE Vivi...