Automating revenue reconciliation through development of an open-source tool

Abstract

Study objectives: The complexity of medical billing leads to errors and potential lost revenues. The errors constitute a breakdown in the billing cycle. When a breakdown in the billing cycle has occurred, the potential governmental and commercial payers do not receive accurate invoices, and, predictably, reimbursement is not actualized. The billing process for emergency department (ED) visits often relies on multiple complicated computer systems from individual vendors, as well as cooperation between many organizations, departments, and other third parties. At each step in the process, there exists a potential for omission of charges, and this is represented by discrepancies in the data between systems. Reconciling of data to other sources reveals process errors within the billing system, which range from omission of patients to incomplete or incorrect billing. Although there exist manual methods for completing this process, they are time-consuming and repetitive and require close attention to detail. We present EDReconciler, an open-source, freely distributed application designed to reconcile data from the individual data sources that represent the billing cycle pathway. Methods: The application was designed to import and process billing data from the author's institution, and it can be extended to function in other institutions and other reconciliation settings. Each step in the billing cycle is subject to reconciliation. We found 6 useful data formats from 5 computer systems, and each was imported for reconciliation: (1) ED admission data from hospital admission system; (2) patient tracking data from clinical system; (3) HL7 data received from medical record coding service; (4) demographic information from patient research information system; (5) physician billing records; and (6) hospital facility billing records. The tool imports patient records and calculates daily census according to data from each data source. When discrepancies arise, the user may use the tool to discover names and information about patient encounters that may not have completed the billing cycle. Dollar estimates of the unbilled encounters can be computed (according to average charges per patient), which can be useful to demonstrate the need for administrative financial support of further inquiries. Global queries can be made to search for problem items such as specific billing codes or unbilled accounts. Results: The errors discovered in this analysis resulted from omitted patient records, incomplete clinical documentation, and lost charges from automatic data transfers. Early estimates show a value of $195,000 during the 4-month period studied at our institution, which is based on 650 noninvoiced patients, 60% specificity, and $500 per patient. We expect the final number to be higher with improvement of imported data reports, as well as indications that lost charges tend to come from complicated, higher-billed patients. Current work includes finalizing an estimate of losses to justify continued auditing and submission of invoices for services that remain billable. Also, implementation of long-term control procedures that include the developed application are under way. Conclusion: Reconciliation of the billing cycle using reliable data from other sources is essential to ensure appropriate revenue capture. Automated reconciliation can estimate errors, discover lost revenue, and most important, recognize process errors that cause the issues.