- The challenges of current cost allocation techniques are concern with data accuracy and the use of retrospective reporting.
- Staffing and scheduling processes are the low-hanging fruit in cost-based accounting initiatives.
- Real-time clinical and financial data flow applied to predictive analyses and machine learning allows for improved prediction of cost anomalies and actionable events.
The increased sophistication of recent cost accounting techniques has not eliminated frustrations with the resulting cost data
in the healthcare industry because the accuracy of the information is suspect and often outdated.
The challenges that persist with respect to historical cost allocation techniques are threefold:
1. Many procedure level costs are ultimately calculated as historical averages, which can bring suspicion and concern of the accuracy of these calculated costs when compared to the actual cost for any individual patient.
2. The analysis is inherently retrospective in nature because the data represents activity from the prior fiscal period. This only allows the user to analyze the historical data for problems to predict if they could occur again in the future — an effort to engage preventive action. But it is not actionable for the activity that is being reported because that activity has already occurred in the past.
Getting the same data faster, while the current activity is being contemplated or occurring, allows healthcare leaders to act on the very activity that is generating the alerts.
3. The ability to balance the decision-support discharge-based cost results with the historical general ledger costs is another common problem. The levels of accruals on the general ledger, coupled with the amount of distortion caused by long length of stay (LOS) patients present a constant frustration in balancing the results of financial data from these reporting sources.
Without regular adjustments to account for these differences, the confidence in the accuracy of the financial decision support data will be under constant suspicion and its use hindered by a lack of confidence in its accuracy.
Real-time decision support allows for quick action
Leading-edge decision support shops nationwide are looking to other industries for a clue on the next steps for the evolution of decision support in healthcare.
The investment and application of technology in healthcare has historically been focused on the clinical aspects of patient care delivery that touches the patient from a treatment or care delivery perspective. Given this focus, investment in traditional financial decision support technologies has lagged behind other industries where leading-edge techniques provide decision-makers with timely data needed to optimize performance.
Highly capitalized service industries (i.e., airlines) and manufacturing firms have evolved from cost allocation to a more accurate and direct job order costing approach when delivering data to decision-makers. The result is a decision support data flow that provides real-time, or near real-time delivery of actual cost data to decision-makers, who can cause change in a nimble and effective fashion. Efforts to mimic this approach in healthcare have historically been frustrated due to the lack of systems and infrastructure to deliver the data needed for real-time analysis.
Recent advancements in technology and systems evolution have provided the infrastructure that sets the stage for the next evolution in decision support data in healthcare, resulting in a foundation that will now support delivery of real-time, or near real-time information to decision-makers.
The following are the major advancements that will help spur this evolution:
- Government funding for electronic health record systems has prompted most acute care institutions to build the components to accumulate the relevant data needed by decision-makers.
- Supply chain and enterprise resource planning (ERP) systems have increased in sophistication and now are tracking supplies and other expenses in a near real-time fashion, regardless of whether they are chargeable items.
- Hardware and internet infrastructures have also contributed to a ubiquitous, cloud-based access to the relevant data needed to manage the assets and expenses for items used in the quality care of the patient.
Labor and staffing lead the way
Making the move to real-time decision support data will occur in multiple stages. The financial return from additional investment in applications will drive prioritization, and the low-hanging fruit in many areas will be the labor premium associated with traditional staffing and scheduling processes. By linking the actual and scheduled case volume in near real-time 15-minute increments to the staffing schedule, premium pay (e.g., overtime, agency, call-back, etc.) can be predicted and avoided and staffing bottlenecks smoothed.
Even when constrained by regulatory and union requirements, aligning staffing to the predicted patient activity provides a strong financial return. Pilot projects for this phase of real-time cost savings are underway in several locations, and initial proof of concept results have demonstrated that foundational systems are in place to drive the real-time consumption of productivity data and provide a solid ROI.
After the operational and department view of real-time costs are developed, the linkage between clinical activities and actual costs will follow.
- Actual labor costs by day and by unit will link the nursing costs to the specific patient receiving services.
- Actual supplies ordered and used for the patient will be accessed and compared to the standard or historical protocols under clinician review.
By using the actual costs for labor and supplies, the assignment of direct costs will be accomplished in near-real time and delivered to the caregiver to enrich treatment analyses. The assignment of indirect or overhead costs can be applied as a factor or “burden rate” that is based on the most recent fiscal period calculation and reconciliation of costs to the general ledger, providing total or fully burdened costs.
The actual patient costs are also fed into the historical cost allocation engine each month as the direct patient costs and used to improve the accuracy in historical cost allocation, replacing relative value units (RVUs). This has the effect of continuous refinement in the accuracy of the cost allocation process and links the two decision support sources in one continuous source of truth.
Machine learning drives actionable data
Finally, the application of predictive analyses and machine learning to this real-time clinical and financial data flow allows for improved prediction of cost anomalies and actionable events. Although artificial intelligence is in its infancy in many industries, it is proving extremely valuable in balancing costs with quality of production output, a critical need in healthcare. The data becomes available in a timelier fashion — closer to the actual provision of care.
That will be the point where predictive analysis itself becomes actionable, subject to required clinical review and judgement. Delivering decision support data in a real-time or near real-time fashion will truly be the hallmark of the next generation of financial decision support systems. If history is a guide, the evolution will occur in healthcare in the next few years.