Healthcare spending projections are often inaccurate. This problem is amplified when (possibly wrong) estimates inform decisions related to coverage of drugs such as weight loss GLP-1s.
getty
The term “dismal science” was coined by Scottish historian Thomas Carlyle in 1849 to describe economics, highlighting its predictive shortcomings. Despite 175 years of development, economists’ ability to forecast remains limited. Nevertheless, accurate predictions are crucial for evaluating healthcare policies, assessing legislative budgets, and implementing regulatory changes.
Two major federal entities, the Centers for Medicare and Medicaid Services Office of the Actuary and the Congressional Budget Office, often produce flawed predictions due to a heavy reliance on assumptions. Economic models are grounded in presuppositions that may not always hold true.
Certain CMS models, especially those concerning prescription drugs, rely on historical baseline averages. This approach can fail to account for the budgetary effects of new medications entering the market. For example, CMS did not foresee the significant increase in prescription drug spending following the introduction of hepatitis C treatments, starting with Sovaldi in 2014. Similarly, early projections underestimated the surge in popularity of GLP-1 medications like Ozempic and Mounjaro within Medicare and Medicaid.
Spending on GLP-1 drugs for covered indications under Medicare and Medicaid has risen significantly in a short time frame and could increase further if coverage for GLP-1s for obesity is expanded, even at reduced net prices.
When the Biden administration proposed expanding Medicare’s Part D to include obesity drugs in 2024, it projected costs between $25 billion and $35 billion over ten years.
The large financial estimates have contributed to the reluctance or unwillingness of plan sponsors to engage in the Trump administration’s BALANCE model as initially conceived. Earlier this year, CMS introduced the Better Approaches to Lifestyles and Nutrition for Comprehensive Health (BALANCE) model, seeking to improve Medicare beneficiary access to GLP-1 medications and promote lifestyle changes to prevent chronic conditions and combat obesity.
Due to low participation from payers, the BALANCE model is currently on hold. Nevertheless, CMS plans to proceed with a Bridge program starting in July, which will provide GLP-1 medications to beneficiaries for $50 monthly, making weight loss drugs like Zepbound and Wegovy accessible at this price.
Currently, about 16 million people qualify for the program through conditions like diabetes or cardiovascular disease risk factors. An additional 13 million overweight and obese individuals without these conditions will also be eligible through 2027.
Interestingly, the CMS documentation omits potential financial impacts on the government from either the BALANCE model or the Bridge program. The lack of publicly available cost estimates raises questions about whether the agency is hesitant to disclose inaccuracies or is waiting for the CBO’s analysis.
The CBO estimates that if all eligible individuals participated in the Bridge program, taxpayer costs could exceed $30 billion annually. However, if only 20% enroll, the yearly cost would be approximately $6 billion, potentially limiting funding for other healthcare initiatives.
This uncertainty prompts the question of prediction accuracy. For over a decade, the CBO has estimated the costs of lifting Medicare’s ban on obesity drugs under the Treat and Reduce Obesity Act, reintroduced over a dozen times since 2012. Their calculations have been criticized as both overly optimistic and pessimistic.
Historically, the accuracy of CBO projections has been questioned concerning major legislation, such as the Affordable Care Act (ACA). In 2012, the CBO predicted 25 million enrollees under the ACA by 2017, but only 10.3 million signed up. Was this due to the model’s flaws or the ACA’s implementation? Or both?
Moreover, despite fewer enrollees than expected, the CBO underestimated the cost of Medicaid expansion in 2015 by $26 billion.
While often considered the best nonpartisan analysis available, CBO projections face challenges in accuracy due to the complexity of the U.S. healthcare system. Behavioral responses, like reactions to new insurance mandates or subsidies, can be misjudged.
Predicting human behavior is notoriously difficult. Despite some improvements in short-term precision over four decades, errors arise during high inflation, economic volatility, or unexpected legislative changes. The Covid-19 pandemic, for instance, upended many predictions.
Long-term forecasts spanning 10 years or more are even less reliable due to the unpredictability of future economic and political landscapes, both domestically and globally.
Unlike natural sciences like meteorology, the social science of economics has made less progress in prediction accuracy. An article in Mother Jones states that weather forecasting has significantly improved, with today’s five-day forecasts as accurate as a one-day forecast in 1980. In contrast, economic forecasting remains challenged, whether concerning GDP, employment, or healthcare budget implications from changes in GLP-1 insurance coverage.
