The healthcare industry is going through mergers and acquisitions and facing new economic struggles. As hospitals realize the need to consolidate resources regionally, a big challenge takes shape: handling the information necessary to manage cash flow efficiently. Cash flow management is vital for hospital systems to sustain their daily operations. Our team used DMAIC to develop strategies for the California Region of Providence Health and Services to manage cash flow at the hospital level while focusing on regional performance.
In January 2008, Providence Health and Services, California Region, was facing financial hurdles not uncommon to other healthcare systems. Cash-on-hand was limited due to the execution of capital-intensive projects and large, unrealized revenue in discharged-not-final-billed (DNFB) and accounts receivable (AR). The amount of cash on hand was dangerously low, approaching a level that would hinder the ability of hospitals in the region to cover operational expenses. Providence’s leadership launched a regional Six Sigma project to address the problem.
The baseline DNFB was close to 18 AR days. The Six Sigma team identified three major processes that contributed to charges on hold, and therefore to the DNFB: health information management (HIM) coding; treatment authorization requests (TAR) review; and other miscellaneous revenue management processes, such as admitting and back-end hold resolution. A core team was established, in which the revenue business office (RBO) and relevant departments from each facility were represented (HIM, admitting and case management). The team held voice of the customer (VOC) focus groups with RBO managers and performed process, material and information flow analysis along with failure mode and effects analyses (FMEA) in the operating units.
A critical activity within Define was the operational definition of the financial operation’s metrics at the regional and local facility levels. Assigning buckets to each of the DNFB components to avoid duplication was key to the accurate monitoring of the total DNFB.
The team identified DNFB as a key metric. The DNFB is calculated by dividing the unbilled dollar amount for charges to discharged patients by the average daily revenue. Thus, the DNFB is reported in AR days; it represents a normalized metric that can be used to compare performance of multiple hospitals in the region. The team identified three major contributors to the DNFB that could be tracked at the hospital level: the DNFB uncoded (accounts not yet processed by the coders); the DNFB from TAR approvals; and the DNFB from miscellaneous revenue management operations, including admitting errors, non-HIM holds and other identifiable factors. The DNFB uncoded was close to 6 days, the DNFB from TAR approvals was close to 4 days and the DNFB from revenue management was close to 8 days.
During the Analyze phase, the team expanded on the Measure work by developing an infrastructure to support the implementation phase. The following were key success factors of the DNFB reduction program:
The following are seven examples of solutions implemented during the DNFB reduction program and the problems that they resolved:
1. Clear backlogs in admitting, TAR review and HIM – In some cases we encountered backlogs equivalent to more than $5 million in accounts held whose charges could not be realized by the region. This enabled us to recover the initial revenue push (effectively cash that immediately became available).
2. Create one unique and concise procedure to track charts across the facility electronically – Initially, the location of a chart was not visible and in many instances an all-hands-on-deck approach was needed to locate charts. The new procedure helped expedite the TAR review process (chart retrieval for case management), the physician reviews and the quality reviews.
3. Redesign the HIM department flow to reduce waste – We redesigned the flow within the department to minimize the initial chart assembly time and expedite the delivery of charts to the coders. Figure 2 shows the department flow before and after the improvement, when the travel path was reduced by 80 percent.
4. Create tools to monitor the prompt resolution of holds –Many accounts were not being billed because of errors that were not visible to the department managers. These errors occurred along several steps of the billing process. We generated daily reports that enabled the managers to monitor progress toward hold resolution and to route resolution to the appropriate manager.
5. Create tools to actively manage the uncoded accounts – All accounts were originally treated equally. We developed tools that enabled the HIM managers to organize the information and look dynamically at accounts by age, charges, insurance type, patient type (i.e., mom or baby account), or some or all of these factors at the same time.
6. Design a training program and train admitting staff to avoid face sheet errors – A large portion of holds in admitting were traced to inaccurate admitting information entered in the face sheet forms, which are quick summaries of patient histories. An analysis of the main errors and their sources was conducted and was used to develop a training program for the staff.
7. Design a training program and train admitting staff to collect co-pays – The staff was culturally biased against the collection of charges up front. Retraining the staff included monitoring their performance and giving active feedback. Figure 4 shows the improvement specifically in the point-of-service (POS) collections for ER services at Providence Saint Joseph Medical Center. The staff improved by 908 percent.
We were successful in reducing the DNFB from 17.65 AR days to less than 8 AR days for all the hospitals in the region (Figure 5). This represented $11 million of initial revenue push and an excess of $1 million in annual cost of cash savings.
Figure 6 shows that the improvement in the total DNFB amounted to a 50 percent reduction. Table 1 shows the results of the t-test, indicating a statistically significant difference between the data collected before improvement and the data collected after improvement.
In the Control phase, the team developed accountability matrices tailored to each facility’s need to monitor performance in a long-term fashion. Table 2 shows an example control plan for one of the facilities. This plan contains information about control activities, frequency, documentation and accountability.
The improvement in cash flow management in the Providence Health and Services, California Region was the result of the application of data-driven methodologies that enabled tight monitoring and active management of the local and regional financial metrics through the collaborative effort by the hospital and regional teams.
The following individuals also contributed to the DNFB program: Karl Carrier, Dave Mast, Teresa McSpadden, Janice Grankowski, Christina Fuentes, Pam Hodge, Barbara Serrano, Peggy Lewis, Joanna Kuzmak, Diane Taminosian, Suzanne Call, Linda Love, Aracely Villalobos and Bonnie Barnett.