This post was authored by Meenakshi Subbaraman, PhD, an epidemiologist and regular contributor to rhvouchers.org.
Healthcare fraud and error amount to 5-6% of global healthcare expenditure (US$236 billion) every year. US$236 billion could ensure safe water for the entire world, control malaria in Africa, provide vaccines for 23 million children, and quadruple the budgets of the WHO and UNICEF, with US$144 billion to spare (European Healthcare Fraud & Corruption Network). Because health systems in low-income countries generally lack systematic tracking and reporting mechanisms, fraud and error in low-income countries tend to be discussed relative to specific initiatives or projects. A recent example is the debate over the results of Global Fund audits.
Fraud and error are key issues for voucher programs, too, but these output-based aid programs lend themselves to better detection and management of fraud and error. According to “Defining Effective Voucher Management Information Systems” – an upcoming report to be published by the RH Vouchers project based on evaluation of voucher management information systems (MIS) in multiple countries – the potential for fraud and error within voucher programs could be more systematically addressed. In addition to what voucher programs are already doing, we explore what more can be done.
All major voucher programs have anti-fraud and error control measures in place, using voucher design, claims vetting, and spot-checks of reimbursement request patterns. Because these efforts are implemented differently across programs, the ability to make useful cross-country performance comparisons is limited. While all voucher programs – like other large scale initiatives – have had to address accounting errors and respond to potential fraud, reporting can be improved and prevention efforts can be strengthened. Under current practices, the extent of fraud/error in the system can be estimated; however, more systematic methods could improve both the precision in those estimates and response times.
While all voucher programs have reported fraud, reports are largely anecdotal and claims are not always verified. As a result, the extent of the fraud/error problem can only be estimated.
In “Defining Effective Voucher Management Information Systems,” report authors Mahad Ibrahim, Ben Bellows, and Jaspal Sandhu recommend implementation of fraud and error detection mechanisms in relevant software modules and at multiple levels of deterrence (voucher, patient, facility, distributor). They recommend that claims are verified at these various levels as well.
Fraud and error are not unique to health systems in low- and middle-income countries, nor are fee-for-service programs exempt from fraud. In 2010, Medicare, the US national social health insurance program for the elderly and disabled spent US$528 billion – three times the annual budget of the UK’s National Health Service. In serving 48 million people, Medicare rejected 3.82% of claims. That same year, the “improper payment rate” for Medicare’s Fee-for-Service was 8.6%, representing $28.8 billion in spending of a total $336 billion in Medicare Fee-for-Service program.
Prevention & detection
In their 2010 systematic review of reproductive health vouchers in developing countries, Bellows, Bellows, and Warren note that prevention and detection systems are built directly into voucher programs. Prevention measures include the use of logos or barcodes to authenticate individual vouchers. Systematically verifying service delivery (e.g. unannounced facility visits or following up at home with sampled patients to confirm they received services) can also deter fraud since the awareness of enforcement systems is often enough to prevent people from committing fraudulent acts.
Bellows and colleagues also recommend detection systems, such as monitoring voucher sales and services for unusual patterns, such as spikes in distribution. As noted in “Defining Effective Voucher Management Information Systems,” voucher information systems that capably track vouchers, clients, distributors, and providers should additionally implement inexpensive, automated algorithms to identify errors and abnormal trends. Anna Gorter and Corinne Grainger in the Quick Guide to Developing Voucher Programmes (2011) as well as the World Bank’s 2005 Guide to Competitive Voucher Programs outline prevention and detection measures for program design, as well as tips for program managers. For example, Gorter and Grainger present recommendations for: prosecution of offenders, recovery of proceeds from fraudulent activity, staff training, and appropriate contracting measures to hold organizations accountable for fraudulent activity.
In addition to the measures employed by the large voucher programs, there is an opportunity to leverage information systems to identify and manage both fraud and error. The current voucher claims processing software systems in the five countries reviewed by the “Defining Effective Voucher Management Information Systems” report – Bangladesh, Cambodia, Kenya, Tanzania, and Uganda – can build strong detection algorithms to augment existing anti-fraud and error-control measures. Certain limitations of the underlying systems will first have to be addressed, but there is significant potential for such an approach as voucher programs scale.
Voucher programs could also share information across sites, and use this information to both prevent fraud and detect error. US public and private insurance programs, such as Medicare, Medicaid, and the Blue Cross and Blue Shield Association, are now using software for detecting anomalous patterns in shared claims data, and closely examining claims to see if multiple agencies are billed for the same patient services. A “trusted third party” often leads data analyses and pass on questionable claims for further investigation.
Although software systems offer a high-tech, automated solution to fraud and error detection problems, they are not sufficient. Most importantly, managers, government regulators, donors and researchers need to understand the extent and types of fraud and error that exist, especially before investing significant resources on a system that may provide little value in addressing fraud and error. Furthermore, efforts should not outweigh the price of fraud and error, since enterprise-level software systems typically cost significant time, money, and human resources.
Well-built information management systems should be combined with simpler measures, such as voucher design, spot-checks, and making regulation highly visible. Neither prevention strategies nor analytic methods alone are enough. The combination of prevention, detection, and enforcement measures will effectively minimize fraud and error, and improve the efficiency of resources brought to bear on pressing health needs.