Healthcare expenditures in the United States represent nearly one in every five dollars spent across the American economy and are showing no signs of slowing down. In a landscape defined by an affordability crisis, hospital administrators are constantly searching for sustainable methods to manage expenses without compromising patient care.
A significant portion of this focus naturally lands on pharmaceutical spending, which accounts for approximately 10% of overall healthcare expenditures. Traditional formulary management has long served as the primary mechanism for controlling these costs by curating an inventory of medications designed to be the most cost effective options for care. However, the foundational information used to make these critical formulary decisions falls short of the complexities found in daily clinical practice. Too often, this process is stifled by an add-only trend, where formularies grow without the counterbalance of meaningful therapeutic reviews. Consequently, many committees rely on subjective, manual chart reviews and expert clinical opinion, which are inherently biased and unsustainable in a high-performing health system.
For decades, pharmacy and therapeutics committees have relied almost entirely on data derived from randomized controlled trials (RCTs) required for federal drug registration. While these trials are essential for establishing basic efficacy, they possess structural limitations that hinder their utility across a broader health system population. Clinical trials are designed with strict inclusion criteria that purposefully screen out complex individuals to isolate a clean treatment effect. In reality, only about 14% of daily medical decisions are backed by high-quality evidence. The remaining 86% of clinical scenarios exist in an evidence desert because approximately 70% of clinical trials exclude patients with comorbidities, despite the fact that 70% of real world patients suffer from multiple conditions. Furthermore, standardized clinical trials rarely capture the reality of combination drug therapies, variable dosage escalations or the true utilization costs that health systems often absorb.
This massive gap between published literature and clinical reality represents a multimillion-dollar inefficiency for health systems. Fortunately, the emergence of advanced clinical analytics and artificial intelligence is creating a profound shift from a model of retrieving old published papers to generating new, patient-matched evidence on demand from real-world data. At our company, Atropos Health, we leverage a federated network that contains hundreds of millions of patient timelines. With it, we can analyze vast clinical datasets to produce publication grade observational studies in a matter of seconds. For pharmacy and therapeutics committees, this capability allows them to evaluate 12 times more protocols in roughly six months less time than traditional manual review processes. This allows health systems to build a precision formulary that analyzes competing therapies regarding concrete clinical outcomes and total annual care costs.
The fiscal impact of deploying real-world evidence to guide formulary decisions becomes clear when looking at specific high cost drug classes. Autoimmune biologics, including tumor necrosis factor alpha inhibitors used to treat conditions like rheumatoid arthritis and Crohn’s disease, account for up to 20% of total outpatient pharmacy benefit drug spending. The per patient annual cost for these therapies frequently ranges from $20,000 to $40,000. Real-world evidence generation allows health systems to evaluate whether transitioning patients to lower cost biosimilars maintains health outcomes. Observational data across diverse patient cohorts has demonstrated that transitioning to biosimilars can cut annual medication costs in half without any statistically meaningful difference in markers of inflammation, disease related hospitalizations or subsequent physician office visits. This translates to an achievable annual savings of $10,000 to $20,000 per patient, representing an immediate multi million dollar optimization opportunity for health systems.
Beyond outpatient biologics, real-world evidence can drive significant cost reductions within acute inpatient settings. Consider the prevention of post operative complications such as post operative ileus following major abdominal surgeries. Traditional protocols often utilize highly expensive medications to enhance post surgical recovery. However, a review of real world evidence revealed that utilizing a significantly less expensive alternative, such as naloxegol instead of alvimopan, actually resulted in superior condition specific outcomes, including reduced inpatient length of stay, a lower incidence of ileus and decreased 30-day readmission rates. Because the clinically superior drug is orders of magnitude less expensive, prioritizing it can save a health system more than $2,000 per patient over a standard hospital stay, scaling massive savings across high volume surgical departments.
Similarly, real world evidence provides an analytical anchor for managing the explosive demand and high cost of GLP-1s for weight management and diabetes care. While these innovative drugs offer substantial clinical benefits, their high upfront pricing forces health systems and payers to ask precise questions about which specific patient populations derive the most value. Rather than enacting blunt, sweeping coverage exclusions, AI-based evidence tools allow health systems to conduct highly granular subgroup analyses based on age, race, initial body mass index and existing comorbidities. For example, data analyses have identified specific high risk patient populations where utilizing these advanced therapies reduced the mean hospital length of stay by almost five days. In these carefully targeted groups, the downstream reduction in inpatient and intensive care utilization costs is exponentially larger than the annual cost of the medication, allowing health systems to maximize the clinical return on their pharmaceutical capital and improve patient outcomes.
Evidence is the true currency of value in healthcare, serving as the essential connective tissue between clinical decisions and economic outcomes. When health systems operate with incomplete, outdated or non representative clinical trial data, their formulary decisions are not fully informed. AI-driven real world evidence generation provides the clarity required to eliminate administrative guesswork, replace adversarial gatekeeping and build a precision framework for coverage decisions. By embracing the power of real time data insights, health systems can confidently navigate rising market costs, optimize total enterprise spending and deliver highly personalized, evidence backed care to every patient who walks through their doors.
.jpg "When AI Joins the Drug Discovery Lab")
