SIIM 2017 Scientific Session Quality & Performance Improvement Part 2 Friday, June 2 7:00 am 8:00 am HCC Screening Compliance Module: A Tool for Automated Compliance Monitoring, Clinician Notification, and Efficacy Research Joseph C. Wildenberg, MD, PhD, University of Pennsylvania; Gregory J. Nadolski, MD, MTR; Michael C. Soulen, MD; Tessa Cook, MD, PhD Background Individuals with cirrhosis and some with chronic Hepatitis B infection are considered high risk for the development of hepatocellular carcinoma (HCC). Despite strong endorsement for biannual screening with ultrasound by the American Association for the Study of Liver Diseases (AASLD) compliance with these recommendations remains low [1,2]. As with many screening recommendations the underlying causes of this poor compliance are many, including confusion about need for screening or the proper screening interval, the cost, time and anxiety associated with regular imaging, and an inability to easily monitor compliance at the individual patient, provider or system level [3]. The most publicized screening recommendations, such as lipid profiles, annual mammography for women and colonoscopies, are commonly incorporated into a "Health Maintenance" list within the patient summary screen of many electronic medical record (EMR) systems. Unfortunately, the ability to add or modify the entries in this list is often limited in scope or restricted to the EMR vendors. Furthermore, the recommendations within this list utilize only minimal logic such as a patient's age or gender to determine eligibility. Current versions of most EMRs do not have a method for specifying particular diagnostic codes or allowing more complex logic to evaluate if screening is necessary or compliance has been satisfied. Even when properly implemented there is increasing scrutiny from clinicians, patients and payers regarding the usefulness of screening at the population level. Prostate specific antigen (PSA) and annual mammography are only two examples of recommendations that have courted controversy within the past few years [4,5]. The ability to both monitor compliance as well as perform robust analyses of efficacy will be necessary to properly justify continuation of impactful recommendations while altering or discontinuing others. Case Presentation We developed a modular application using Python, JQuery, and HTML5, powered by a MS-SQL database to automate the identification of patients eligible for HCC screening, monitor their compliance and initiate communication with referring clinicians (Figure 1).
Figure 1 The data import module of the application interfaces with the EMR to identify eligible patients by International Statistical Classification of Diseases and Related Health Problems (ICD)-9/10 encounters and extract pertinent clinical data. This data is then assembled into the relational database, located behind the health system's firewall, for use in calculating compliance. The data for patients already present in the database are updated as necessary (e.g. a change in clinical providers, a new diagnosis of HCC, or death). The data import module also interfaces with an existing imaging database to identify any exams the patient received that would qualify as a screening exam for HCC. A separate logic module is run daily to calculate the current compliance of all eligible patients within the database. This logic is specific to the screening recommendation, in this case 6 month intervals for all eligible patients, and allows a customized grace period such that communication with the responsible clinical provider is not immediately initiated when a patient falls out of compliance. Additionally, prior to deeming a patient out of compliance the logic module checks for a scheduled imaging study that would satisfy compliance but which has not yet been performed. After the logic module has run, any patients deemed out of compliance are passed to the communication module. This algorithm sends a message to the responsible clinical provider, though the EMR, reminding them that the patient is due for their screening exam. A link within the message opens an internal webform and allows the provider to report that an appropriate imaging study has been ordered or indicate reasons the patient may be out of compliance, such
Outcome as an eligible imaging study performed at an outside institution (with a field to enter the date). If two weeks have passed without a response from the clinician a repeat message is sent. Within the target webform there is also the ability for the provider to indicate that the patient should not be considered eligible for screening by the application (e.g. incorrect diagnosis, expected lifespan less than 1 year, prior liver transplant, or the patient declines screening). This information is then used to update the database. A browser-based dashboard, adapted from an existing initiative for tracking incidental abdominal lesions [6], interfaces with the MS-SQL database to display pertinent information and statistics. This dashboard can be used to filter by provider or hospital to compile provider/location-specific statistics on patient compliance. There is also the ability to manually review provider responses to previously-sent messages in order to understand why patients are removed from screening or to update inconsistent data within the database. The initial data import identified 6783 patients eligible for screening collected over 5 years. The database at the time of initiation allowed for an immediate retrospective review of compliance by duration from qualifying diagnosis. At the time of the first data import compliance with screening was 23.9%. Figure 2 shows that many patients were in compliance for the first 6 months after eligibility, likely from an exam that diagnosed cirrhosis, but with a sharp decline 6 months later consistent with poor adherence to the screening recommendations. Note that the gradual attrition of patient numbers at longer durations after eligible diagnosis is secondary to limited follow-up for patients diagnosed close to the date of the data import. Figure 2 Discussion The application is initially in a pilot to select hepatologists to determine the accuracy of the import and logic modules and ensure the messaging system fits within the clinical workflow. In addition to increasing the number of hepatologists, we expect to expand this pilot to include primary care providers to capture patients that do not follow with a hepatologist. A more robust approach to screening is necessary to increase compliance, justify the costs and ensure that the screening tests are impactful on patient outcomes. We have designed this application as a tool to help clinicians identify patients eligible for HCC screening and remind
Conclusion References them if their patient is no longer compliant with current recommendations. It is also designed to aid in retrospective research on overall disease prognosis and the impact of screening on morbidity/mortality. The modular design of this application will allow for easy extension to other disease processes for which there are screening recommendations and that require more complex logic than is currently implemented in common EMRs. These could include recommendations for lung-cancer screening in chronic smokers, abdominal aortic aneurysm screening or colon cancer screening. The data import and logic modules, as well as the back-end database, may also be adapted to address recommended follow-up that is specific to particular patients. Incidental findings, which continue to increase in frequency as imaging improves, may necessitate a single follow-up exam scheduled many months in the future or require multiple follow-up exams at variable intervals such as those recommended by the Fleishner society for lung nodules [7]. These types of recommendations can be challenging for patients and providers to follow without a tracking tool. At this time we do not have sufficient longitudinal data to determine if compliance rates increase through implementation of this application. We anticipate approximately 6 months of follow-up will be necessary to directly compare the pre- and post-implementation screening rates. Our initial experience has suggested that a HCC compliance monitoring tool, designed with a modular architecture, can integrate with both the EMR and imaging databases while remaining flexible for potential uses in other disease processes. Furthermore, the back-end database is immediately available for research into screening behavior and the natural history of the disease process. 1. Bruix J & Sherman M. Management of hepatocellular carcinoma: an update. Hepatology. 2011; 53(3):1020-1022. 2. B-H Zhang, B-H Yang, Z-Y Tang. Randomized controlled trial of screening for hepatocellular carcinoma. J Cancer Res Clin Oncol, 130 (2004), pp. 417-422 3. Hoffman RM, Lewis CL, Pignone MP, et al. Decision-making processes for breast, colorectal, and prostate cancer screening: the DECISIONS survey. Med Decis Making. 2010. 30(5 Suppl):53S-64S. 4. Pharoah PDP, Sewell B, Fitzsimmons D, Bennett HS, Pashayan. Cost effectiveness of the NHS breast screening programme: life table model. BMJ 2013. 5. Andriole GL, et al. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst. 2012;104(2):125-32. 6. Zafar HM, Chadalavada SC, Kahn CE Jr, et al. Code Abdomen: An Assessment Coding Scheme for Abdominal Imaging Findings Possibly Representing Cancer. J Am Coll Radiol. 2015. 12(9):947-50. 7. MacMahon H, Austin JHM, Gamsu G, et al. Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society. Radiology 2005; 237: 395-400. Keywords
hepatocellular carcinoma, screening, automated tracking, provider communication