Oncology Patients in Clinical Trials: A Health Care Cost Analysis

ABSTRACT

Objectives: To assess health care resource utilization and costs among commercially insured patients with colorectal cancer (CRC) or ovarian cancer (OC) participating in clinical trials (PCTs).

Study Design: A retrospective claims analysis from 2017 to 2022 was conducted using the Milliman Consolidated Health Cost Guidelines Sources Database, a proprietary database that aggregates deidentified claims data from multiple large commercial insurers representing more than 75 million covered lives.

Methods: Participants in CTs were identified with diagnosis, modifier, or procedure codes for a CT and a diagnosis code for either CRC or OC. The first observed CT claim was assigned as the index date (index). For patients with CRC or OC not participating in a CT, proxy episodes were created for each qualified 18-month period following the first observed cancer claim. We compared average allowed costs (per-patient-per-month) incurred in the 9 months following index after adjusting for confounding using inverse probability of treatment weighting.

Results: We identified 69 CT participants with CRC, 40,716 nonparticipants with CRC, 49 CT participants with OC, and 17,041 nonparticipants with OC for the study. Adjusted mean costs for both medical and drugs, in the 9 months following index, were lower for CT participants than nonparticipants. Among CRC patients, mean medical costs were $2806 for the CT cohort vs $3996 for the non-CT cohort, a difference of $1190 (P = .015), and drug costs were $447 and $1635, respectively, a difference of $1189 (P = .0002). Among OC patients, mean medical costs were $2211 and $2508, respectively, a difference of $296 (P = .524), and mean drug costs were $371 and $2294, a difference of $1924 (P < .0001).

Conclusions: Commercially insured patients with OC and CRC enrolled in CTs incurred lower costs following CT initiation than those not enrolled, driven by payers not incurring costs for oncology therapies.

Am J Manag Care. 2026;32(9):In Press

doi:10.37765/ajmc.2026.90012

Takeaway Points

This study analyzes health care resource utilization and costs among patients with colorectal cancer (CRC) and ovarian cancer (OC) enrolled in clinical trials vs those not enrolled.

• Patients with CRC and OC in clinical trials experienced lower per-patient-per-month medical and drug costs compared with nonparticipants.
• Our findings highlight potential cost savings for health care payers when these patient populations participate in clinical trials.
• This evidence demonstrates to payers that covering members in cancer clinical trials does not increase costs compared with matched cohorts who are not enrolled.

Clinical trials (CTs) play an important role in oncology care, providing patients with access to new cancer drugs, surgeries, and therapies when standard regimens are either limited or unavailable.¹ A review of data from the Commission on Cancer, representing approximately 70% of US cancer cases, found an overall CT participation rate of 7.1%, which varied by tumor type.² A study specifically of patients with ovarian cancer (OC) found that between 2011 and 2021, approximately 5% participated in a CT.³ A study that evaluated the participation of patients with colorectal cancer (CRC) and lung cancer found a similar rate of enrollment at 5.3%.⁴ Both studies included patients with and without metastases.^3,4

Section 2709 of the federal Public Health Service Act requires coverage of routine patient costs for patients participating in CTs for select life-threatening conditions, including cancer.⁵ Guidance from CMS indicates that this law is self-regulating and that plans offering group or individual coverage are expected to operate “using a good faith, reasonable interpretation of the law.”⁶ Since the law’s implementation in 2014, insurance denials have persisted through prior authorization requirements and other obstacles that deter patients from enrolling in CTs.^7-11

Although older data suggest CT enrollment may reduce or maintain cancer treatment costs, newer studies done since the approval of expensive targeted therapies show different results.^12-15 A 2021 study evaluated patients enrolled in a second-line trial for metastatic non–small cell lung cancer and found that compared with nonparticipants, participants had a $6663 reduction in total monthly costs for the duration of the trial; this resulted in a total savings of $45,308 shared between the payer and patient, despite participants being more likely to receive a targeted therapy during treatment.¹⁶

Payer type and the prevalence of cancer by age group are important because both factors have an impact on the cost of care.^17,18 For OC and CRC, the median onset ages are 63 and 66 years, respectively, so approximately half of US patients in these groups have commercial insurance, often employer sponsored.^19,20 A 2014-2020 analysis found a cancer incidence of 0.61% and a prevalence of 1.97% among commercially insured patients across all cancer types, not specific to OC or CRC across all cancer types.²¹ And these rates are rising. The World Health Organization projects 29.9 million new cases globally by 2040 vs 20 million in 2022, and the US exceeded 2 million new cases in 2024.^3,22,23 CRC incidence increased in individuals aged 20 to 49 years from 1998 to 2019 but decreased in older individuals.²⁴ Although OC deaths have declined, OC remains the fifth leading cause of cancer death in women.^20,25 Early-onset cancer may pose additional risks, as younger patients often delay care until symptomatic, risking advanced-stage diagnosis.⁴

Despite the clear financial and clinical implications, there has been no evaluation to date of the impact of CT participation on health care resource utilization (HCRU) among commercially insured patients with OC or CRC. To address this gap, this study compared CT participants with nonparticipants to understand the impact of CT participation on patient HCRU and cost to payers and patients in a commercially insured population. CRC and OC were selected due to high unmet needs, rising incidence among younger patients, a lack of viable treatment options at later disease stages, and reasonable study sample sizes.²⁶

METHODS

We conducted a retrospective analysis of administrative claims data from 2017 to 2022 for commercially insured patients with CRC and OC using the Milliman Consolidated Health Cost Guidelines Sources Database. Notably, the methodology employed in this analysis is novel, so literature to guide methods and cancer selection was unavailable.

Cancer types were selected through sensitivity testing across all cancer types with biomarkers. Among cancers with recent (2017-2019) biomarker CTs, CRC, OC, and pancreatic cancer had sufficient trial enrollment. However, patients with pancreatic cancer who met the study criteria had an insufficient sample size (N = 17), precluding analysis. Patients with evidence of CT enrollment were identified as those having either a relevant encounter diagnosis, modifier, or procedure code (eAppendix Table 1 [eAppendix available at ajmc.com]). The index date (index) was defined as the earliest date a CT code was observed. Patients were required to have evidence of active treatment, defined as receipt of at least 1 type of cancer treatment (radiation, antineoplastics, oncologic surgery) between 12 months before and 6 months after the index, with a CT start date between 2017 and 2019. Refer to eAppendix Table 2 for exclusion criteria. For patients without evidence of a CT, proxy CT episodes were created.

Following episode identification, we evaluated baseline characteristics for both cohorts in the 12 months prior to index (Table 2). As the primary end point, all-cause allowed costs (the total paid by all payers and the patient)—evaluated as 2 separate outcomes, medical services and drugs—were evaluated from index through 6, 9, and 12 months. Allowed costs represent the maximum reimbursable amount for medical treatments and services, including outpatient pharmaceuticals, as determined by negotiated provider reimbursement rates, calculated prior to the application of member cost-sharing provisions or coordination-of-benefits adjustments. Evaluation of costs focuses solely on allowed costs, which were censored at the earliest of departure from data, hospice encounter, or 24 months post index.

Propensity scores were calculated for the CRC and OC cohorts via logistic regression to predict trial participation using demographics, baseline HCRU, and Hierarchical Condition Category variables. The final model used stepwise selection (α = .20; eAppendix Table 3). Stabilized inverse probability of treatment weighting (IPTW) reweighted trial and control observations to balance covariates. Weighted linear regressions then estimated the effect of trial participation on 9-month per-patient-per-month (PPPM) medical and drug costs. PPPM differences and P values reported throughout are derived directly from the unrounded regression coefficients; because displayed PPPM values are rounded to the nearest dollar for readability, subtracting the rounded values shown in text or tables may not always exactly equal the reported difference. Sensitivity analyses evaluating metastatic disease gave similar results.

RESULTS

Colorectal Cancer

Sixty-nine patients (mean age 52 years; 45% female) with CRC initiated enrollment in a CT between 2017 and 2019, and 40,716 non-CT proxy episodes were included in this study (Table 1). The majority of CT episodes identified for study involved patients also diagnosed with a secondary malignancy (metastatic) at the time of index—94% of CT episodes and 82% of non-CT episodes (Table 2). Potential confounders (as evaluated at baseline) for cost comparison are summarized in Table 2, including specific oncology therapy modalities.

Across both CT and non-CT cohorts, high baseline rates of systemic therapies were observed, with 44% and 49.1% receiving bevacizumab (Avastin), 58% and 44% receiving irinotecan (Camptosar), and 68% and 76% receiving other antineoplastics. Raw average baseline allowed costs PPPM were similar and ranged from $7120 to $7807. Average monthly costs for both cohorts mostly stayed between $10,000 and $15,000. Average quarterly costs for CT index episodes declined dramatically for the first 3 quarters (9 months) after the index and then returned to pre-index levels beginning in month 10. Payers not incurring costs for trial drugs drove observed cost differences. For a complete description of quarterly allowed costs, see eAppendix Figure 1.

The modeled mean allowed costs of medical services (excluding physician- and pharmacy-administered drugs) incurred in the 9 months following the index, after controlling for confounders among non-CT episodes, were $3996 PPPM for the non-CT cohort. Comparatively, the modeled mean cost among episodes with CT participation was $2806—a difference of $1190 (P = .015) (Table 3). A similar relationship was observed when evaluating mean costs for physician- and pharmacy-administered drugs (non-CT modeled mean $1635 vs CT modeled mean $447, a difference of $1189 [P = .0002]).

When evaluating allowed PPPM incurred in the 6 months following the index, the modeled mean medical cost across CT episodes was $3348 PPPM, significantly lower than the non-CT mean of $3871 PPPM, a difference of $523 (P = .0046). The modeled mean cost for physician- and pharmacy-administered drugs incurred within CT episodes ($427 PPPM) was also significantly lower than that of non-CT episodes ($1602 PPPM), a difference of $1175 (P < .0001).

Similarly, when evaluating costs incurred in the 12 months following the index, the modeled mean costs across CT episodes ($3274 PPPM) remained significantly lower than those across non-CT episodes ($4298 PPPM), a difference of $1024 (P = .046). The modeled average cost for physician- and pharmacy-administered drugs for CT episodes of $572 was also significantly lower than the $1672 for non-CT episodes, a difference of $1100 (P = .001). For more details on the model outcomes evaluating 6- and 12-month costs, see eAppendix Tables 1a and 1b.

Ovarian Cancer

For OC, 49 patients (mean age 55 years) initiating enrollment in a CT and 17,041 non-CT proxy episodes were included in this study (Table 1). Most episodes identified for study involved patients also diagnosed as metastatic at index—59% of the CT episodes and 49% of the non-CT episodes (Table 4). CT patients had higher baseline oncology surgery rates (37% vs 5%) but lower bevacizumab use (25% vs 40%) than non-CT patients.

Average baseline allowed costs PPPM within the CT cohort ($7189) were lower vs the non-CT cohort ($8583). Average quarterly allowed costs among CT episodes remained below $10,000 following the index and were lower for the first 6 months following index. Comparatively, quarterly costs PPPM among non-CT episodes ranged from $10,000 to $15,000, with the largest contributor being costs for physician-administered drugs. For a complete description of quarterly allowed costs, see eAppendix Figure 2.

The modeled mean allowed costs of medical services (excluding physician- and pharmacy-administered drugs) incurred in the 9 months following the index, after controlling for confounders among the non-CT episodes, were $2508 PPPM for the non-CT cohort. Comparatively, the mean cost among episodes with CT participation was $2211, a nonsignificant difference of $296 (P = .524) (Table 5). A significant difference was observed in modeled mean allowed costs for physician- and pharmacy-administered drugs: CT episodes averaged $371 and non-CT episodes averaged $2294, a difference of $1924 (P < .0001).

When evaluating PPPM costs through 6 months following index, the $1957 modeled mean medical cost for CT episodes was not significantly different from the non-CT episodes’ modeled mean medical cost of $2490, a difference of $533 (P = .1782). However, the CT episode PPPM of $251 for modeled mean cost for physician- and pharmacy-administered drugs was significantly different from the non-CT episode PPPM of $2219, for a difference of $1968 (P < .0001).

Similarly, when evaluating 12-month PPPM costs, the modeled mean cost among CT episodes of $3084 was no longer significantly different from the non-CT episodes’ modeled mean cost of $2788, a difference of $296 (P = .5610); however, the modeled mean cost for physician- and pharmacy-administered drugs for CT episodes of $328 remained significantly different from the non-CT episodes modeled mean cost of $2310, a difference of $1982 (P < .0001). Model results evaluating the 6-month and 12-month PPPM costs are provided in eAppendix Tables 2a and 2b.

DISCUSSION

Overall, CT participation observed in this analysis—6.4% for patients with CRC and 7.9% for patients with OC—was similar to national averages. Although these patients represent a subset of the overall oncology population, the rate of metastatic disease observed in this study was high for both types of cancer: 94.2% for CRC and 59.2% for OC.

Among 69 CT patients with CRC and 40,716 non-CT proxy episodes, patients with evidence of CT enrollment had lower PPPM all-cause health care costs by $2379 (the sum of the medical and drug cost differences modeled separately in Table 3: $1190 + $1189) during the 9-month follow-up period vs the weighted sample of patients without evidence of CT enrollment. Similarly, among 49 CT patients with OC and 17,041 non-CT proxy episodes, PPPM costs were $2220 (the sum of the medical and drug cost differences modeled separately in Table 5: $296 + $1924) lower for those enrolled in a CT over the 9-month follow-up period vs those not enrolled in a trial.

The observed cost differences reflect that trial sponsors are providing investigational drugs at no cost to patients or payers. These drugs and their administration costs do not appear in PCT claims. The absence of drug costs did not lead to compensatory increases in other medical service categories. Medical service costs (including office visits, laboratory tests, imaging, hospitalizations, and supportive care) were not higher for PCTs. In CRC, total allowed costs were significantly lower; in OC, they were similar. This demonstrates that CT participation does not impose offsetting utilization burdens that would erode the drug cost savings. Prescription drug costs remained lower over the 12 months following the index (eAppendix Figure 1 and Figure 3), supporting our algorithm’s identification of drug-related trials. These results demonstrate that, from a payer perspective, CT participation does not increase overall costs.

The length of CT study periods can vary depending on factors such as the treatment being evaluated. This study required a definitive time period to observe the primary end point. Because definitive information on the exact start and end dates of CTs is unavailable in claims data, several factors were considered when deciding the time frame. One study of advanced CTs indicated a median duration of 6 months.²⁷ Further, many trials experience attrition as patients voluntarily drop out.²⁸ This study evaluated 6, 9, and 12 months following the index to account for the variation in CT duration. Notably, the OC patient population in this study had a lower percentage of patients with evidence of a secondary malignancy vs the non-CT comparator group, indicating they may have been included at a less severe stage, which may impact CTduration.

The rolling-proxy episode methodology employed for the non-CT cohort represents a pragmatic approach to addressing inherent data limitations. Because the timing of initial cancer diagnosis for the control patients often preceded the study observation window, we were unable to match groups on time from diagnosis. Instead, non-CT patients contributed observation windows at varying points in their cancer trajectories, weighted using propensity scores to balance observable characteristics. This differs from traditional matched cohort designs and introduces potential temporal heterogeneity between groups. Although IPTW methods help control for measured confounders, unmeasured factors related to the timing of disease progression may remain. The observed cost differences should therefore be interpreted as reflecting the impact of trial participation among patients who met study eligibility criteria, rather than as a comparison at identical disease time points. Future research with more comprehensive longitudinal data would be valuable for validating these findings.

Additionally, although there is a lack of information on the nature of the trial in which each patient in the treatment group enrolled, the selection criteria requiring chemotherapy administration without evidence of billing for an antineoplastic drug increase confidence that CT patients participated in a drug-related trial. The use of these criteria resulted in the identification of a relatively small sample size of treated (CT) patients across both cancer types with a total of 118 patients. Despite this, the models still demonstrated statistical significance, and the consistency of model outcomes across evaluation periods further supports the validity of our findings.

Finally, commercial payers may have coverage policies that stipulate what costs are covered when a patient is enrolled in a CT. For example, UnitedHealthcare defines approved CTs and qualified individuals and provides corresponding rates for routine patient costs that, when incurred, will be covered by the plan.²⁹ Although payers may be concerned that an increase in CT participation may lead to an increase in the plan’s cost burden, these results demonstrate otherwise.

Limitations

This retrospective analysis used administrative claims submitted for reimbursement, which lack clinical characteristics such as disease stage (except secondary malignancies), performance score, and progression-free survival. Patients were retrospectively observed as CT or non-CT, so our analysis may have inherent biases and lack information on the intent or nature of the trial. Further, this study used a rolling-proxy episode approach for non-CT patients, creating multiple comparison periods throughout treatment. Although this addresses unknown trial start dates for those patients, it creates a key limitation: CT patients were observed from enrollment, whereas non-CT patients contributed multiple 18-month windows starting at various treatment points. Also, due to data constraints, cohorts weren’t matched on time from diagnosis, potentially influencing observed cost patterns. To identify trials that were likely drug related, CT episodes were required to indicate chemotherapy administration without evidence of billing. We note that modeling drug and medical costs separately may lead to an overstatement of drug costs and cost savings. Moreover, because comparisons were performed separately for all-cause medical and drug costs to address potential bias when comparing with non-CT episodes that included costs for oncology drug therapies, differences in drug costs may be overstated. Any costs related to impacts on life expectancy are not explicitly evaluated. Additionally, an algorithm approximated the population of patients receiving active treatment, so misclassification is possible. Finally, because this analysis was in a commercially insured population, any conclusions may not be generalizable to other populations.

CONCLUSIONS

Patients incurred significantly lower medical and drug costs during CT participation in the first 9 months, driven by lower medical and pharmacy costs in CRC and lower drug costs in OC. In the 9 months after index, patients with CRC enrolled in a CT had an average total allowed PPPM cost $2379 (the sum of the modeled medical and drug cost differences in Table 3) lower than patients not enrolled in a CT and patients with OC in a CT had an average allowed PPPM cost $2220 (the sum of the modeled medical and drug cost differences in Table 5) lower than non-CT patients. Although these results are based on specific cancers and use administrative claims data, this analysis suggests that patients enrolled in a trial are not more expensive than those not enrolled. Payers should consider these results when evaluating requirements of prior authorizations for trial-related HCRU or denials of participation in CTs.

Acknowledgments

The authors would like to thank the other Access to Comprehensive Genomic Profiling team members, Mary Nesline, Naleen Raj Bhandari, Andrea Stevens, and John Fox, for their contributions to this manuscript, including review of the methodology and initial data outputs.

Author Affiliations: Milliman, Inc (CF, KF, PDB, PE), New York, NY; Illumina, Inc (BB), San Diego, CA; Eli Lilly and Company (AMG), Indianapolis, IN; Invitae (CM), San Francisco, CA; Thermo Fisher Scientific (RHD), Waltham, MA.

Source of Funding: Ms Ferro, Ms Fitch, Dr Bhatt, and Mr Ellenberg are employees of Milliman who received consulting fees from Access to Comprehensive Genomic Profiling (ACGP) to complete this work. Illumina, Eli Lilly and Company, Invitae, and Thermo Fisher Scientific are members of ACGP.

Author Disclosures: Ms Ferro, Ms Fitch, and Mr Ellenberg report being employees of Milliman, Inc and received consulting fees for this work. Ms Bapat reports being a full-time employee at Illumina, Inc and owning stock in the company. Dr Gilligan reports being employed by Eli Lilly and Company and owning stock in the company. Mr Dumanois reports employment at Thermo Fisher Scientific and owning stock in the company.

Authorship Information: Concept and design (CF, KF, PDB, PE, BB, CM, RHD); acquisition of data (PDB, PE); analysis and interpretation of data (CF, KF, PDB, PE, BB, AMG, CM, RHD); drafting of the manuscript (CF, KF, PDB, PE, BB, AMG, CM, RHD); critical revision of the manuscript for important intellectual content (CF, KF, PDB, PE, BB, AMG, CM, RHD); statistical analysis (KF, PE, BB); provision of patients or study materials (CF); obtaining funding (RHD); administrative, technical, or logistic support (CM); and supervision (CF, KF).

Address Correspondence to: Robert H. Dumanois, AB, Thermo Fisher Scientific, 180 Oyster Point Blvd, San Francisco, CA 94080. Email: robert.dumanois@thermofisher.com.

REFERENCES

1. Clinical trials: a significant part of cancer care. Mayo Clinic Comprehensive Cancer Center Blog. 2025. Accessed August 24, 2026. https://cancerblog.mayoclinic.org/2025/05/01/clinical-trials-a-significant-part-of-cancer-care/

2. Unger JM, Shulman LN, Facktor MA, Nelson H, Fleury ME. National estimates of the participation of patients with cancer in clinical research studies based on Commission on Cancer accreditation data. J Clin Oncol. 2024;42(18):2139-2148. doi:10.1200/JCO.23.01030

3. Smith AJB, Alvarez R, Heintz J, Simpkins F, Ko EM. Disparities in clinical trial participation in ovarian cancer: a real-world analysis. Gynecol Oncol. 2023;175:25-31. doi:10.1016/j.ygyno.2023.05.066

4. Fouad MN, Lee JY, Catalano PJ, et al. Enrollment of patients with lung and colorectal cancers onto clinical trials. J Oncol Pract. 2013;9(2):e40-e47. doi:10.1200/JOP.2012.000598

5. Affordable Care Act provision requiring insurance coverage of clinical trials. AmericanSociety ofClinicalOncology; 2014.

6. Affordable Care Act implementation FAQs - set 15. CMS. Updated March 13, 2026. Accessed August 12, 2026. https://www.cms.gov/cciio/resources/fact-sheets-and-faqs/aca_implementation_faqs15

7. Mackay CB, Antonelli KR, Bruinooge SS, Saint Onge JM, Ellis SD. Insurance denials for cancer clinical trial participation after the Affordable Care Act mandate. Cancer. 2017;123(15):2893-2900. doi:10.1002/cncr.30689

8. Mackay CB, Antonelli K, Bruinooge SS, Saint Onge J, Ellis SD. Predictors of insurance denials for cancer clinical trial participation after ACA mandate. J Clin Oncol. 2016;34(suppl 15):e14047. doi:10.1200/JCO.2016.34.15_suppl.e14047

9. Zion TN, Berrios CD, Cohen ASA. Insurance denials and diagnostic rates in a pediatric genomic research cohort. Genet Med. 2023;25(5):100020. doi:10.1016/j.gim.2023.100020

10. Cooper KE, Abdallah KE, Angove RSM, Gallagher KD, Bonham VL. Navigating access to cancer care: identifying barriers to precision cancer medicine. Ethn Dis. 2022;32(1):39-48. doi:10.18865/ed.32.1.39

11. Kumar G, Chaudhary P, Quinn A, Su D. Barriers for cancer clinical trial enrollment: a qualitative study of the perspectives of healthcare providers. Contemp Clin Trials Commun. 2022;28:100939. doi:10.1016/j.conctc.2022.100939

12. Wagner JL, Alberts SR, Sloan JA, et al. Incremental costs of enrolling cancer patients in clinical trials: a population-based study. J Natl Cancer Inst. 1999;91(10):847-853. doi:10.1093/jnci/91.10.847

13. Bennett CL, Stinson TJ, Vogel V, et al. Evaluating the financial impact of clinical trials in oncology: results from a pilot study from the Association of American Cancer Institutes/Northwestern University clinical trials costs and charges project. J Clin Oncol. 2000;18(15):2805-2810. doi:10.1200/JCO.2000.18.15.2805

14. Fireman BH, Fehrenbacher L, Gruskin EP, Ray GT. Cost of care for patients in cancer clinical trials. J Natl Cancer Inst. 2000;92(2):136-142. doi:10.1093/jnci/92.2.136

15. Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med. 2006;3(11):e442. doi:10.1371/journal.pmed.0030442

16. Merkhofer C, Chennupati S, Sun Q, et al. Effect of clinical trial participation on costs to payers in metastatic non-small-cell lung cancer. JCO Oncol Pract. 2021;17(8):e1225-e1234. doi:10.1200/op.20.01092

17. Tomicki S, Dieguez G, Latimer H, Cockrum P, Kim G. Real-world cost of care for commercially insured versus Medicare patients with metastatic pancreatic cancer who received guideline-recommended therapies. Am Health Drug Benefits. 2021;14(2):70-78.

18. Dieleman JL, Cao J, Chapin A, et al. US health care spending by payer and health condition, 1996-2016. JAMA. 2020;323(9):863-884. doi:10.1001/jama.2020.0734

19. Cancer stat facts: colorectal cancer. Surveillance, Epidemiology, and End Results Program. National Cancer Institute. Accessed January 6, 2025.https://seer.cancer.gov/statfacts/html/colorect.html

20. Cancer stat facts: ovarian cancer. Surveillance, Epidemiology, and End Results Program. National Cancer Institute. Accessed January 6, 2025.https://seer.cancer.gov/statfacts/html/ovary.html

21. Cong Z, Ye X. EPH188 the burden of cancer among enrollees insured through commercial plans in the US (2019-2020). Value Health. 2022;25(suppl 12):S227. doi:10.1016/j.jval.2022.09.1109

22. Cancer statistics. National Cancer Institute. Updated May 7, 2025. Accessed August 12, 2026.
https://www.cancer.gov/about-cancer/understanding/statistics

23. Estimated number of new cases from 2024 to 2050, males and females, age [0-85+]. International Agency for Research on Cancer. World Health Organization. Accessed January 6, 2025. https://gco.iarc.fr/tomorrow/en/dataviz/trends?types=0&sexes=1_2&mode=cancer&group_populations=0&multiple_populations=0&multiple_cancers=1&cancers=40&populations=840&apc=cat_ca20v1.5_ca23v-1.5&group_cancers=0&scale=linear&min_zero=1

24. Colorectal Cancer Facts & Figures 2020-2022.American Cancer Society. 2020. Accessed August 24, 2026. https://www.cancer.org/research/cancer-facts-statistics/colorectal-cancer-facts-figures.html

25. Ovarian cancer statistics. Ovarian Cancer Research Alliance. Accessed January 8, 2025. https://ocrahope.org/for-patients/gynecologic-cancers/ovarian-cancer/ovarian-cancer-statistics/

26. Facing the challenges of late-stage ovarian cancer: navigating limited treatment options and advocating for new therapies. Cure. November 12, 2024. Accessed January 13, 2025. https://www.curetoday.com/view/facing-the-challenges-of-late-stage-ovarian-cancer-navigating-limited-treatment-options-and-advocating-for-new-therapies

27. Haslam A, Olivier T, Thawani R, Prasad V. Duration of treatment in oncology clinical trials: does the duration change when the same drug moves from the experimental arm to the control arm? ESMO Open. 2022;7(3):100480. doi:10.1016/j.esmoop.2022.100480

28. Fogel DB. Factors associated with clinical trials that fail and opportunities for improving the likelihood of success: a review. Contemp Clin Trials Commun. 2018;11:156-164. doi:10.1016/j.conctc.2018.08.001

29. UnitedHealthcare. Clinical trials – commercial and individual exchange medical policy. 2023.