May 23rd, 2013
It is increasingly well-known that improper payments cost taxpayers as much as $50 billion each year. These include reimbursements for billing for non-existent patients, falsified diagnoses, non-covered procedures, services not rendered or simply upcoded, as well as billing errors in favor of providers. Steps are being taken to address these issues through increased acceptance of approaches, tools and techniques from private industry and from industries outside of healthcare. More than just technology, some of the most powerful ideas to come along are that incentives matter, decentralization may achieve results faster and better, and stretch goals are crucial.
Scale of the problem
Safeguarding taxpayer resources and maintaining access to healthcare are clear public policy priorities. The Government Accountability Office (GAO) has long designated Medicare as a high-risk federal program due to its vulnerability to waste, fraud and abuse. Conservative estimates by the National Health Care Anti-Fraud Association are that improper payments represent 3 percent of national health care spending. The GAO and others estimate nearly 10 percent of the more than $500 billion in current annual Medicare payments are improper. At the same time, Medicare provides necessary — and often much needed — access to health care for 48 million Americans.
Some good early signs
Positive signs are emerging of fraud reduction at multiple levels. The Centers for Medicare and Medicaid Services (CMS) have attempted to improve the program’s integrity using informed beneficiaries as one line of defense against fraud along with external contractors. Medicare Administrative Contractors (MACs) process, adjudicate, and pay claims on behalf of CMS while other contractors such as Program Safeguard Contractors/Zone Program Integrity Contractors (ZPIC) investigate payments and recoup over-payments. Overall, a proactive approach championed by CMS’ new Director of the Center for Program Integrity seeks to shift the agency from trying to recoup improper payments and prosecuting criminals to not making such payments in the first place.
Technology solutions are also delivering real results. Spending on such fraud control has a high return of investment, returning about $16 in reduced spending for every dollar invested. Unfortunately, this high rate of return implies that much low-hanging fruit has not yet been harvested. Representing around .02 percent of the Health and Human Services department’s budget, such solutions clearly have room to expand and improve. Economics being what it is, only if the discounted return of the stream of annual reduced spending is less than the present value of fraud control investment should investment stop. CMS appears to be very far from reaching this point where marginal benefits are less than marginal costs.
Learning from financial services
CMS has increasingly turned to algorithmic-based approaches pioneered and proven in consumer financial services. Consider fraud vulnerability when commercial banks facilitate consumer credit and merchant payments through credit cards. Best-practice solutions include Fair Isaac (FICO) risk-scoring to preemptively identify those merchants, customers, and transactions that present a statistically higher risk of fraud, and preventing a detected potential fraud before or soon after a payment is made. Such fraud-control programs in consumer banking have reduced losses by more than two-thirds over twenty years despite more sophisticated and more global fraud activities.
While the financial services industry has had prolonged success combating fraudulent transactions with technology-based solutions, it is important to realize that the rate of improper payments was perhaps at most 0.1-0.2 percent. The estimated level of improper payments in Medicare is 50-100 times higher. This level of endemic fraud, waste, and abuse is best addressed through incentives for all stakeholders, implementation of multiple, differentiated and decentralized solutions, and establishment of stretch goals for combating improper payments.
Incentives and decentralization matters
Properly incented, firms in industries outside healthcare invest in fraud detection software and the associated contracted services because such systems typically pay for themselves in less than a year in avoided fraud, and they continue to generate on-going savings. Vendors are incented to offer risk-based pricing, to have ‘skin in the game’, and to otherwise compete with one another for the business. Different firms choose different approaches best suited to their types of business and their assessment of the cost-benefit tradeoff of investments in particular anti-fraud products. Rarely will one vendor own the entire market or seek to solve every problem using centralized, heavily architected, anti-fraud algorithms.
Quicker, decentralized initiatives would more easily handle the local nature of health care fraud in which a local pattern or sequence of claims by one provider triggers local attention. For example, identifying potential fraud in a series of claims from several providers in Miami-Dade would benefit far more from comparison data and benchmarks from across the county line in neighboring Broward than from Rochester, MN.
Incentives and decentralized solutions are already visibly playing a role in the private health care sector. Consider one of the largest private health insurers in the country, Highmark of Pennsylvania. It took just 7 months to implement FICO commercial software predictive analytics to reduce fraud in its Blue Cross Blue Shield franchise. Highmark also has a large Medicare Advantage book of business where it is directly exposed to losses from fraud and abuse. In consequence, it is known that they use the same technology to reduce their financial risk in that area as well.
Conversely, MACs are overwhelmingly incented to make payments efficiently and quickly to providers on behalf of Medicare’s fee-for-service program. Accordingly, Highmark cannot be expected to — and doesn’t — invest in implementing similar FICO commercial software applications in its MAC business in several eastern states.
Tougher policies and stretch goals
The federal government should consider the option of revisiting MAC and ZPIC contracts to mandate fraud-prevention technology investments. This could ensure that the current adjudication of claims (i.e. is the claim technically correct) is coupled with local fraud detection (i.e. is the emerging pattern of claims associated with this provider or this area or this patient appropriate). In parallel, allowing MACs and/or ZPICs a direct and meaningfully large share in the savings from reduced improper payments would be a high powered incentive to drive the sort of fraud-control seen in the private sector.
In this era of public sector deficits and potential cuts to entitlement programs, the improved integrity of the Medicare program is paramount. Given the acknowledged scale of the problem, an aspirational stretch goal would be to reduce improper payments from 10 percent to 1 percent over 10 years by aligning the incentives of the many and diverse private companies who serve as the government’s claims handlers with public program objectives. Conservatively, such reductions in fraud would save Medicare $45 billion each year. By demonstrating that CMS is a careful steward of taxpayers’ resources, Medicare can continue to meet its mission of providing high-quality care to those most vulnerable in our society.Email This Post Print This Post
Don't miss the insightful policy recommendations and thought-provoking research findings published in Health Affairs.
to the #1 source of health policy research.