Turning Pathways on Their Head: Deimplementation of Low-Value Treatment Regimens

J Clin Pathways. 2025;11(2):6-9. doi:10.25270/jcp.2025.02.01

Even before the creation of its founding framework in 1966, quality improvement in health care has traditionally focused on enhancing services that are delivered.¹ Less emphasis has been placed on avoiding low-value care (LVC), although that has begun to change.^2,3 Definitions of LVC vary widely but generally encompass services or treatments that provide little or no benefit to patients and often have the potential to cause harm, including financial toxicity.^2,4 It has been estimated that more than 30% of US health care expenditures go to LVC, with reported costs ranging from $100 billion to $270 billion annually.^2,5 Notably, a significant portion of LVC cost burden falls to patients. One study noted that of the $900 million spent on LVC in four US states, patients directly paid for approximately $90 million.⁶

Given the high costs associated with oncology, value-based reimbursement models and educational campaigns, such as the Oncology Care Model (OCM), Choosing Wisely, and Common Sense Oncology, have sought to reduce LVC and oncology spending while maintaining high-quality care.6-8 Unfortunately, these efforts have not been as impactful as hoped. OCM had no overall impact on che­motherapy spending, and uptake of recommendations by educational campaigns have been limited, so payers have begun shifting emphasis within oncology to de­implementation of LVC.^8,9 However, deimplementation programs have been lim­ited in scope and success. A review by Tabriz and colleagues of 12 unique oncol­ogy studies reported a variety of deimplementation strategies, such as educating clinicians by developing and disseminating new guidelines, auditing and providing feedback, having a clinical champion, and developing a decision support tool, of­ten integrated within the electronic health record system.10 None of these strate­gies included offering an alternative high-value replacement, and only half of these studies found that the deimplementation strategies successfully reduced LVC.¹⁰

In addition, oncology LVC recommendations have typically focused on a lim­ited set of services: imaging, staging or surveillance scans, surgery, and radiother­apy. When initiatives to reduce LVC have sought to improve selection of medica­tions, they have typically not gone beyond generic/biosimilar drugs or supportive care agents.^8,10 This narrow focus leaves out many regimens that would qualify as LVC. Medical oncologists often have multiple treatment options that are supported by guidelines and compendia. However, not all approvable regimens are equal; they offer a range of clinical outcomes, strength of evidence, and potential for toxicity. With the me­dian annual cost of new oncology medications at $260 000, rapid advancements, and the increasing use of surrogate end points in clinical trials, there is a need to identify and focus on treatment regimens that are low value.^11,12

For this study, the Evolent medical team evaluated the im­pact of our active deimplementation strategies to achieve mean­ingful reductions in low-value oncology regimens. Evolent partners with health plans and providers to improve outcomes for people with complex health conditions, including cancer. The cornerstone of our oncology solution is our high-value clinical pathways, which are vetted by an independent scientific advisory board of leading academic and community oncolo­gists. These pathways guide providers to select the most effi­cacious and least toxic regimens, known as Level 1 Pathways. Evolent also categorizes some regimens as Level 2, denoting those that do not meet Level 1 criteria but are still acceptable and approvable per compendia. Yet there is a third group of regimens that, while still approvable, offer substantially poorer outcomes or cost much more than other options that deliver similar clinical benefits. By creating a separate category in our pathways for these inferior regimens, Evolent aims to help pro­viders select preferred regimens while avoiding low value ones.

Clinical Methods

Evolent developed the low-value regimen (LVR) program as a deimplementation strategy to promote improved patient out­comes in higher quality cancer care. Evolent determined LVRs by assessing the following criteria: clinical outcomes, level of clinical toxicity, strength of evidence, and risk of financial toxic­ity relative to other alternative therapy. See Table 1 for more de­tailed criteria. The Evolent medical team performed a review of National Comprehensive Cancer Network (NCCN) treatment guidelines for 10 cancer groups to identify regimens of interest that stood out as potential candidates for more thorough review of determination criteria. Seeking a meaningful impact to im­prove the cost of oncology care, the medical team selected 10 cancers comprising most treatment requests, including breast, colorectal, diffuse large B-cell lymphoma, follicular lympho­ma, mantle cell lymphoma, multiple myeloma, marginal zone lymphoma, non–small cell lung, small cell lung, and prostate cancer. The regimens of interest were then compared with Evolent preferred alternative (Evolent Level 1 Pathway) and/ or Evolent nonpreferred alternative (Evolent Level 2 Pathway). Regimens were determined to be an LVR if clear evidence of inferiority existed across one or more of the criteria compared with a higher value alternative.

For consideration of clinical outcomes, NCCN evidence blocks and published literature (eg, meta-analyses) were used as reference material for review. NCCN category designa­tions, availability of randomized controlled trial data, real-world evidence, and accelerated approval status were factors considered in a regimen’s strength of evidence. For clinical toxicity, Evolent used NCCN’s evidence blocks, specifically “safety of regimen/agent” to compare the relative toxicities of each regimen. Additional adjustments were made for sig­nificant toxicity factors necessitating black box warnings or a Risk Evaluation and Mitigation Strategy (REMS) program.

Financial toxicity in oncology refers to the economic bur­den patients and their families face due to the high costs of cancer care. Calculations of total cost of care were performed using average sales price for part B medications and wholesale acquisition cost for part D medications. If applicable, cost of supportive care medications (eg, myeloid growth factor) was also factored in. Financial toxicity risk to the patient was based upon an assumption of a 20% co-insurance. When running these calculations across multiple treatment options, Evolent found instances where the cost of identified LVRs was nearly 70-fold compared with the cost of the high-value alternative. This underscores the need for careful consideration of treat­ment options based on clinical and cost effectiveness. The use of LVRs can lead to ineffective care with significant financial strain for patients and payers alike.

The regimens determined to be low value through this anal­ysis were then vetted internally with the larger medical oncol­ogy team to ensure alignment among Evolent oncologists. The list of LVRs was then presented to Evolent’s external Oncol­ogy Scientific Advisory Board (OSAB), where previous tasks were to evaluate new or existing regimens/agents to determine if they would become a Level 1 (preferred) or Level 2 Pathway. The OSAB comprises board-certified oncologists from across the country who practice in a variety of practice settings and specialties. The final determination of a regimen’s low value status was dependent upon an OSAB strong consensus agree­ment with the Evolent designation.

Educational Methods

In February 2023, the LVR program was launched with an Evolent major national partner in a selected portion of the on­cology membership. Together, both organizations partnered to create multifaceted strategies to improve treatment quality through the LVR program. Key components of the program included targeted provider awareness through the Evolent pro­vider solution and medical teams, system notifications, case-based peer-to-peer discussions, practice scorecards and a pay-for-performance program. Since its inception, Evolent teams have educated clinical and administrative practice leaders on evidence-based reasons for categorizing certain treatments as low value. The medical team reviewed the LVR, the evidence, and the high-value alternatives. This allowed providers to be­come familiar with the LVRs and share feedback on operational or patient population considerations specific to the practice. Campaign materials, including informational videos, were cre­ated for the four most used LVRs across common tumor types to guide ordering providers. Tech enablement through system identification of the LVR displayed a visual designator showing the status as low value. Peer-to-peer conversations also empha­sized patient benefits from alternative treatments and avoided negative impacts on quality of life and financial toxicity. These efforts aimed to prioritize patient outcomes and cost effective­ness by promoting high-value medicine.

This retrospective pre-post analysis used prior authorization data to study the impact of Evolent’s LVR program. Evolent compared the volume and rate of LVR prior authorization re­quests and approvals from a preintervention period (February 2023 to May 2023) to a postintervention period 1 year later (February 2024 to May 2024), adjusting for membership and prevalence changes (see Table 2). In total, there were 356 LVR requests in the pre-period and 309 in the post-period. Of these, 234 and 192 were approved, respectively. Average monthly pa­tient count in the pre-period was 8336 compared with 8564 in the post-period. The rate of LVR requests decreased by 15.5% (10.68 vs 9.02 requests per 1000 patients per month), and the rate of LVR approvals decreased by 20.1% (7.02 vs 5.60 approv­als per 1000 patients per month) between the pre- and post-pe­riods. These reductions resulted in the prevention of 5.07 LVRs per 1000 patients per month (10.68 per 1000 requests submit­ted pre vs 5.60 per 1000 requests approved post), or 47.5%, in the post-period. The associated prior authorization interven­tion rates for the pre- and post-periods were 34.3% and 37.9%, respectively. The most used LVRs with respect to cancer type were lurbinectedin for small cell lung cancer, nivolumab plus ipilimumab for non–small cell lung cancer (NSCLC), ramuci­rumab plus docetaxel for NSCLC, and albumin-bound paclitax­el for breast and NSCLC. The LVR program led to a reduction in requests and approvals, indicating a positive shift toward more efficient and evidence-based health care practices. The LVR pro­gram also produced an increase in intervention rates, suggesting better engagement and adherence to the program criteria.

The pursuit of quality in health care has long been a prior­ity, even preceding the establishment of its framework nearly 60 years ago. Although service enhancement has tradition­ally taken center stage, recent efforts have begun focusing on minimizing LVC. Health care systems have implemented value-based reimbursement models and educational initia­tives. These endeavors aim to reduce LVC while maintaining high-quality care. However, the impact remains inconclusive, and uptake of LVC recommendations has been limited. Con­sequently, attention has shifted toward strategies for deimple­menting LVC, including educating clinicians through the de­velopment and dissemination of guidelines, conducting audits with feedback, appointing clinical champions, and integrating decision support tools within electronic health record systems. Despite these efforts, only half of the studies have successfully reduced LVC, and viable high-value alternatives were lacking.

Within oncology, clinical pathways are commonly used to define the treatment options that provide the greatest efficacy, least clinical toxicity, and all else being equal, lowest risk of financial toxicity. To further increase use of high-value care, Evolent created the LVR program to help achieve meaning­ful reductions in low-value oncology treatment. By address­ing LVRs head-on, through a multifaceted approach includ­ing identification, system technology, and provider education offering high-value alternatives, health care systems can help optimize patient outcomes and treatment expectations. These strategies resulted in an overall reduction of 20.1% in approved LVRs compared with baseline to follow-up. If the baseline rate of 10.68 requests per 1000 is held constant, a total of 5.07 per 1000, or 47.5%, of LVR requests, were prevented in the follow-up period through the interventions. A financial analysis estimated program savings by comparing the savings opportunity in the baseline period to the savings opportunity in the follow-up period, specifically the total cost of LVR au­thorization vs total cost of LVR alternatives. Collaboration among medical teams and alignment with guidelines—both in LVRs and therapeutic high-value alternatives—are essen­tial keys for effective deimplementation efforts.

Although the analysis sheds light on strategies for deimplement­ing LVC in oncology, it is essential to acknowledge its limita­tions. These may include generalizability, data sources, tem­poral factors, and behavioral aspects. The study’s findings may be specific to the context in which they were conducted with respect to health care systems, regions, and patient populations. Extrapolating these results to other diseases may not directly apply. The study’s results and conclusions rely on data sources, such as electronic health records and authorization data, which may have limitations if applied to other settings. As for tem­poral factors, the study’s timeframe matters. Health care prac­tices evolve over time, which results from changes in guidelines, technology, and even patient demographics. Behavioral aspects, such as clinician preferences, play a significant role in the out­comes. Overcoming resistance to change, addressing biases, and fostering a culture of evidence-based practices is an investment of resources. Although the Evolent LVR program shows prom­ise, ongoing research and a holistic approach are necessary to address LVRs comprehensively. However, these results may suggest that future deimplementation strategies may help curb LVR use to provide the highest value care to patients.

Affiliation:

¹Evolent, Arlington, VA

Correspondence:

Sang Chau, PharmD, BCOP

1812 N Moore St, Arlington, VA 22209

Email: schau@evolent.com

Acknowledgements:

The authors would like to thank Jordan Silvergleid, Poonam Chadha, Ryan Tieu, and Dr Juhee Sidhu of Evolent for their contributions to the Low-Value Regimen program.

Disclosures:

The authors disclose no financial or other conflicts of interest.