At a Glance
By identifying and applying advanced revenue cycle analytics, healthcare providers can:
- Free up cash
- Find new revenues without harming core services
- Improve productivity, profitability, and patient care
Ineffective revenue cycle management has long had a major impact on the bottom line for many health systems. Today, providers find themselves struggling to keep their heads above water—or, more accurately, above the paper flood created by last-generation approaches that can’t keep up with the growth in billing complexity. Most hospitals simply don’t have the right technology and processes in place to identify and recover missing or erroneous charges.
In the past, approaches that were presented as solutions have actually been a part of the problem. Most health systems use a combination of rules-based approaches and manual audits to find billing anomalies. Traditionally, this approach has been too aggressive—flagging too many invoices for review—or too conservative—
failing to detect missing or erroneous charges. Also, rules-based systems are highly time-consuming to maintain. And they don’t prioritize, take medical and patient context into account, or “learn” on their own; rather, they only “follow the rules.” The bottom line: Traditional rules-based systems miss charges they could bill or correct, and identify many false positives while demanding constantattention and updating through human intervention.
How Advanced Analytics Can Help
Advanced analytic approaches that use machine learning, predictive modeling, pattern detection, anomaly detection, and other sophisticated techniques can
successfully address the weaknesses of rules-based systems. These techniques use the powerful capabilities of machines to find hidden connections and patterns and effectively ferret out records with the highest probability of missing or erroneous charges.
Revenue leakage is a great area to start with the application of advanced analytics, for several reasons. First, advanced analytics address a significant pain point for many hospital systems, and any investments can generate a solid ROI with a short payback period. Second, the machine-learning algorithms that identify billing record outliers form an excellent “signal hub” from which other analytic solutions can be spun off and implemented. Those solutions include, but are not limited to, overcharges and cost containment, collections, readmissions reduction, and staff schedule optimization. All are driven from the same core data asset, supplemented, in some cases, with additional source data. Collectively, these solutions can transform hospital productivity and improve financial results while mitigating risk and financial exposure through use of a sequential, self-funding advanced analytics rollout approach.
That doesn’t mean hospitals should turn everything
over to the machine and hope for the best. The recommended approach is not to rely solely on machine learning, but rather, to pair the machine with humans, following general guidelines for selecting and implementing the new system, as described in the sidebar below.
The big win is when pattern-based machine learning approaches present a “watch list” of opportunities to the auditing professionals, prioritized based on expected value. Such a tool can dramatically improve the auditors’ productivity by helping them to avoid chasing down false positives or missing out on big opportunities. In essence, the machine can do the heavy analytic lifting by cutting millions of records down to human scale, thereby allowing auditors to do their job more effectively and in less time.
Dual Approach: A Case Study
This dual approach is being used successfully.Consider, for example, the appraoch used by a large health system that has applied advanced analytics and a pattern-based approach to address its revenue leakage problem.
Prior to implementing this approach, the health system was using a rules-based system to identify anomalies. But it was omitting charges and losing revenue, despite expensive audits. Understandably, the on-site auditors were unable to manually review every file flagged by the rules or manually selected for audit; there were just too many files for them to assess. By using advanced analytics, the health system significantly increased the targeting and efficacy of the auditor reviews. Every day, the machine learning algorithms scanned all of the files and provided targeted recommendations. A feedback loop fed actual results and auditors’ responses into the models, thus continuously improving the machines’ ability to spot savings opportunities. Over time, the auditors had an increasingly more targeted subset of files to review, with a much higher probability of detecting missed charges and generating higher net-revenue impact.
Overall, the solution identified up to 2 percent of outpatient revenue that had previously gone unbilled. The health system also found an immediate opportunity to reduce audit expenses by 75 percent. The system standardized and centralized the audit process across hospital facilities so auditors were able to better identify common root causes and remedy them at the source.
This approach continues to boost the bottom line at this health system. As you read this, a machine is silently and tirelessly analyzing 100 percent of the system’s daily flow of invoices and identifying missed opportunities and erroneous charges
The ROI of Advanced Analytics
In an industry with increasingly challenging economics, using advanced analytics helps hospitals and health systems significantly increase profitability and operating efficiencies by identifying missing charges and reducing audit expenses. Based on industry averages, hospitals using an advanced analytics solution can increase their operating income by 10 percent or more. By using machine learning-based approaches and combining the power of humans and machines, health systems can not only transform their revenue cycle, but also generate substantial efficiencies in areas such as staffing, procurement, and readmissions reduction.
Pieter Schouten is general manager of healthcare solutions, Opera Solutions, New York (firstname.lastname@example.org).
Selecting and Implementing a Solution
In selecting a revenue cycle analytics solution, hospitals should follow these guidelines.
Don’t discard existing rules and audit practices. Instead, incorporate existing systems and build on them. The existing systems represent years of relevant experience and provide a strong foundation for new solutions.
Limit the scope and impact of change. Adopting the new solution need not—and should not—require major changes in workflow or investments in new IT and infrastructure.
Leverage human-machine interaction. The system should be able to automatically re-prioritize items for follow-up on a daily basis based on probability and expected cash impact, thus enhancing auditors’ productivity and focus. The system also should include an automated loop that captures auditors’ feedback and uses it to continually adjust the models. Ultimately, the auditors make the final call on modifying charges; their feedback is critical in helping models improve.
When implementing the machine learning-based solution, the best approach is to roll it out in defined stages. After the models are trained on historical data, health systems should pilot the solution in one or a few locations before moving to a full execution. This pilot phase allows for fine-tuning the models, tweaking rule integration, and training staff. Design this phase to permit extensive human input, ensuring that auditors’ experience and expertise informs and complements machine learning models. This dual approach hastens adoption and strengthens the entire system’s performance. After the pilot, which should last two to three months, the health system can rapidly expand it to full scale, reaping the benefits systemwide.
Publication Date: Friday, February 01, 2013