Racial and Ethnic Disparities in Prior Authorizations for Patients With Cancer

ABSTRACT

Objective: Prior authorization is used to ensure providers treat patients with medically accepted treatments. Our objective was to evaluate prior authorization decisions in cancer care by race/ethnicity for commercially insured patients.

Study Design: Retrospective study of 18,041 patients diagnosed with cancer between January 1, 2017, and April 1, 2020.

Methods: Using commercial longitudinal data from a large national insurer, we described the racial and ethnic composition in terms of prior authorization process outcomes for individuals diagnosed with cancer. We then used linear regression models to evaluate whether disparities by race or ethnicity emerged in prior authorization process outcomes.

Results: The self-identified composition of the sample was 85% White, 3% Asian, 10% Black, and 1% Hispanic; 64% were female, and the mean age was 53 years. The average prior authorization denial rate was 10%, and the denial rate specifically due to no medical necessity was 5%. Hispanic patients had the highest prior authorization denial rate (12%), and Black patients had the lowest prior authorization denial rate (8%). Regressions results did not identify racial or ethnic disparities in prior authorization outcomes for Black and Hispanic patients compared with White patients. We observed that Asian patients had lower rates of prior authorization denials compared with White patients.

Conclusions: We observed no differences in the prior authorization process for Black and Hispanic patients with cancer and higher rates of prior authorization approvals for Asian patients compared with White patients.

Am J Manag Care. 2024;30(10):In Press

Takeaway Points

Given the importance of the prior authorization process, this study explores whether racial and ethnic disparities emerge in the prior authorization process for patients with cancer.

• Prior authorizations are first submitted with a mean number of 45 days from the diagnosis of cancer, and 10% of prior authorizations are denied.
• The prior authorization process demonstrates little difference for Black and Hispanic patients relative to White patients.
• Asian patients tend to have faster prior authorization approvals and lower denial rates compared with White patients.

Prior authorization is a process to discourage low-value services without hindering health care quality.^1,2 Generally, access to cancer treatment relies on the prior authorization process, a common tool used by insurers to ensure providers are limiting specialty services to appropriate, evidence-based treatments.^3-7 The success of this process depends on physicians and their administrative staff to provide the necessary information to the insurer in a timely manner.^8-11 Because cancer treatments are often time sensitive, the ability of a physician’s office to efficiently navigate the prior authorization process can positively affect the health and care satisfaction of patients.^12-14

A number of studies describe the existence of racial disparities in cancer treatment; however, it is not clear whether the prior authorization process is a factor exacerbating or reducing disparities. Ideally, the submission of cancer treatment prior authorization requests and the adjudication of those requests are equally efficient for all individuals and lead to timely approval.^15-20 One study describing racial disparities in prior authorization in gynecologic oncology found that patients of Asian descent were more likely to have a prior authorization submitted compared with White patients.²¹

In this study, we describe the experience of commercially insured patients with cancer within the prior authorization process by self-reported race and ethnicity using data from a large national insurer. What is known about the provider-insurer prior authorization process is extremely limited because accessing and combining such data are difficult. Our results can be used to inform policy makers, other health plans, providers, patients, and employers on equity in the prior authorization process for specialty care.

DATA AND METHODS

Data Sources and Population

We utilized medical claims and data from the prior authorization process from Elevance Health from January 1, 2017, to April 1, 2020 (henceforth referred to as the study period). We identified commercially insured (self-insured and fully insured) adult individuals (aged 18-64 years) who had at least 2 evaluation and management office visit claims with a cancer diagnosis or 1 cancer diagnosis during an emergency department or inpatient stay during the study period (we excluded cancer diagnoses for basal cell carcinomas, which generally do not require a prior authorization). We chose to focus on the commercially insured because Elevance Health’s commercial utilization management processes are uniform across commercial plans. For Medicare Advantage plans, the utilization management approach differs and can vary by health plan because it is adjusted based on the clinical profile of covered individuals. Varying utilization management approaches in Medicare plans have also been reported elsewhere.²²

We limited the claims to individuals who had a cancer diagnosis and at least 6 months of medical eligibility after the cancer diagnosis date. The first date of a cancer diagnosis was selected to be the index date from which prior authorization outcomes were measured in the following 6 months. Six months was chosen because most patients in our data begin cancer treatment within 6 months. Data were limited to individuals who had a prior authorization, and cancers were identified using International Classification of Diseases, Ninth Revision (140-199) and International Statistical Classification of Diseases, Tenth Revision (C00-C90) codes.

In terms of the prior authorization data, we observed the decision (approval, denial), whether the denial was due to medical necessity reasons, and the date of the prior authorization. We limited prior authorization data to those related to a cancer diagnosis, but we were not able to identify the type of service requested in the prior authorization. As such, we could only review whether a cancer-related prior authorization occurred (and the outcome) and not the specific cancer treatment request.

Sample Outcome Measures

We evaluated 3 outcomes related to the prior authorization process. First, we examined the number of days between the cancer diagnosis and the submission date of the first prior authorization. Second, we assessed the prior authorization denial rate, measured by the share of denied prior authorizations relative to the total number of submitted prior authorizations in the 6 months after the cancer diagnosis. And third, we looked at the prior authorization denial rate for the prior authorizations denied due to no medical necessity relative to the total number of submitted prior authorizations.

Independent Variables of Interest

The main variables of interest were self-reported race or ethnicity drawn from the sociodemographic data available from the insurer for covered individuals. Self-reported race and ethnicity data were provided from employers and electronic health records or clinical data. Specifically, we created binary indicators for non-Hispanic White, non-Hispanic Asian, non-Hispanic Black, and Hispanic. Hispanic ethnicity was recorded in our data as either Hispanic-White or Hispanic-Black; we included both groups in the Hispanic category.

Covariates

We included a large set of sociodemographic control variables identified from the medical claims and the American Community Survey. Health plan sociodemographic information included age, sex, plan type (health maintenance organization, preferred provider organization, consumer-directed health plan), insurance coverage type (employer sponsored commercial coverage, or Medicare Advantage coverage), and length of health plan enrollment prior to the cancer diagnosis. From the medical claims, we identified the type of cancer as either 1 of the 13 most common types of cancer, “other cancer,” or multiple types. We also measured the extent of comorbidities for each member with the Deyo-Charlson Comorbidity Index (DCCI) in the 6 months before the index cancer diagnosis.²³ Finally, based on the member’s zip code of residence, we merged in the block group characteristics on household income and education level from the 2013-2017 American Community Survey 5-year estimates.

Statistical Analysis

We used a linear regression model to assess the association of race and ethnicity with the 3 prior authorization outcomes. We also separately estimated this association for each of the 3 most common cancers (breast, prostate, and thyroid). To account for idiosyncratic differences at the state level, we included state fixed effects as well as the sociodemographic control variables described above to account for differences in characteristics by race and ethnicity. For example, we controlled for health plan type because we wanted to compare disparities for individuals enrolled in the same health plan rather than make comparisons across health plans. See the eAppendix (available at ajmc.com) for more information on the regression model and eAppendix Table 1 for the full list of control variables included in the regressions. Standard errors were clustered at the state level. In a set of sensitivity analyses, we utilized hospital referral region fixed effects instead of state fixed effects (but did not report the results in the manuscript because the coefficients remained qualitatively similar).

Due to known differences in cancer care by sex and region,^17,24,25 we stratified the sample in subgroup analyses by sex and geographic region (census region) to assess the impact of race and ethnicity on the days between cancer diagnosis and first prior authorization (eAppendix Figures 1 and 2 provide results for the other outcomes). We did not conduct these analyses for Hispanic patients, as the sample size was not large enough for all subgroup analyses.

All analyses were performed using R 4.0.2 (R Foundation for Statistical Computing). This study was conducted in full compliance with relevant provisions of the Health Insurance Portability and Accountability Act (HIPAA). Because researchers only used analytic files derived from a limited data set to perform the analyses as defined by the HIPAA Privacy Rule 45 CFR 164.514(e), no waiver of informed consent or exemption was needed from an institutional review board.

RESULTS

Descriptive statistics by race and ethnicity for the prior authorization outcome variables are presented in Table 1. The mean number of days from diagnosis to submission of prior authorization request was 45 days; non-White patients’ mean time until submission was 43 days vs 46 days for White patients. The mean prior authorization denial rate was 10% (relative to a mean of 8.5 submitted prior authorizations), and the denial rate specifically due to no medical necessity was 5%. Some variation by race and ethnicity emerged. Hispanic patients had a prior authorization denial rate of 12%, with a 7% denial rate due to no medical necessity, whereas Black patients had a prior authorization denial rate of 8% and a denial rate of 4% for medical necessity.

Table 2 presents all sociodemographic characteristics, displaying that the self-identified composition of the sample was 85% White, 3% Asian, 10% Black, and 1% Hispanic; 64% were women; and the mean age was 53 years. In terms of comorbidities, the sample had a mean DCCI score of 2, and 32% had metastatic cancer at the time of diagnosis. Most were enrolled in a preferred provider organization plan (53%), and the vast majority were covered through employer-sponsored health plans (96%), although some were already covered by a Medicare Advantage plan despite being younger than 65 years. eAppendix Table 2 displays all summary statistics by race and ethnicity.

Table 3 shows the regression results for all outcomes. We found no racial or ethnic disparities in the number of days between the cancer diagnosis and the prior authorization submission. However, some differences emerged in the success rate of getting a prior authorization approved. Compared with White patients, Asian patients had a lower rate of denied prior authorizations (–2.6 percentage points; a 25% reduction based on the mean Asian denial rate) and a lower rate of denials due to no medical necessity (–1.7 percentage points; a 32% reduction). We did not observe any differences in the denial rates and medical necessity denial rates for Black patients compared with White patients.

Stratifying the regression sample by cancer type to identify differences in outcomes, Asian patients with breast cancer had a higher number of days until prior authorization submission (7.4 days; 14% greater) than White patients. However, the medical necessity denial rate was lower for Asian (–2.2 percentage points; a 56% reduction) compared with White patients. For prostate cancer, the number of days until prior authorization submission was lower for Asian (–11.1 days; 24% reduction) and Black (–5.8 days; 11% reduction) patients compared with White patients. For thyroid cancer, the rate of prior authorization denials were lower for all groups and the denial rate for medical necessity was lower for Black patients compared with White patients; however, these results should be taken with caution given the small sample sizes.

The Figure shows the regression results for days between cancer diagnosis and first prior authorization stratified by sex and region for Asian and Black patients. Relative to White patients, Asian patients had a longer wait period in the Midwest and West. See eAppendix Table 3 for estimates with CIs. eAppendix Figures 1 and 2 display results for the other 2 outcomes, with lower prior authorization denial rates and medical denial rates for Asian patients across both sexes and most regions compared with White patients.

DISCUSSION

This study examines racial and ethnic disparities in cancer treatment prior authorization outcomes. Previous research has shown that self-identified Black patients experience delays in initiation of cancer treatment.^15-18,20,26 We did not observe racial or ethnic differences in the number of days between the cancer diagnosis and the submission of the first cancer-related prior authorization. We did observe a lower prior authorization denial rate for Asian patients compared with White patients. Subgroup analyses showed some prior authorization outcome heterogeneity by cancer type and region.

Disparities reflected in the prior authorization process may have origins at the insurer, patient, or provider level. However, there is little empirical evidence that provides helpful answers to the cause of disparities. For example, a lower prior authorization denial rate could reflect a lower rate of provider advocacy for some patients. That is, having a physician requesting authorization for emerging or speculative treatments that are not yet considered medically necessary or are considered low-value care could put a patient at higher risk of receiving a denial, which may itself reflect a certain kind of privilege that is experienced less often by Asian patients. It is also possible that the differential experience may have origins in more conservative treatment preferences conveyed by the patient or family based on cultural differences.

There are several implications for future research. First, we observed no evidence of racial disparities in prior authorization submission time for cancer care; however, this work can only speak to existing disparities in cancer care based on data from only a single insurer. Experiences may be different for other payers, although prior authorization guidelines for cancer treatment should be similar across the industry. As such, more research is needed to describe the prior authorization process and potential implications for disparity in cancer care. Second, many factors besides prior authorization can impact the timing of cancer treatment initiation. These can include environmental, geographic, and individual characteristics. Geographic supply-side factors, such as specialist density, may affect appointment times and could influence the time between cancer diagnosis and treatment initiation. This work cannot speak to these factors, so future research could explore how these factors may affect the time to prior authorization. Lastly, cancer care is an important cohort to study, but future research should expand the scope of this analysis to focus on other disease groups requiring prior authorization as well as service types with common prior authorization requirements.

Limitations

This study has several limitations. First, race and ethnicity were self-reported, so the sample was limited to individuals who reported these data, which may not be representative of all commercially insured patients with cancer. Second, identified baseline differences, such as in health plan enrollment by race/ethnicity, may imply the potential for other unobserved differences that could be related to race/ethnicity and prior authorization take-up. Third, the analyses included data from only 1 insurer and are therefore not representative of all commercially insured patients with cancer. However, the profile of commercially insured individuals from this insurer is likely to be similar to those covered by other private insurers, although we cannot rule out an overrepresentation in the 14 states in which Elevance Health sells insurance directly. Finally, our analysis reviewed only patients with at least 6 months of coverage after the cancer diagnosis. Thus, patients who dropped out of the sample immediately after the cancer diagnosis were not included. However, the number of individuals who dropped out of the sample (49%) was similar in magnitude to retention rates in commercial health plans (55%).²⁷

CONCLUSIONS

Our findings describe the association between the cancer treatment prior authorization process and self-reported race and ethnicity. We did not observe that patients identifying as Black and Hispanic were more likely to have a longer wait time from cancer diagnosis to prior authorization submission. However, we found that Asian patients see higher rates of treatment regimen approval and lower rates of prior authorization denials. More research is needed to elucidate how disparities may appear in review of prior authorization patterns, as it is possible that lower or higher rates of denials may both indicate the presence of disparity.

Acknowledgments

The authors would like to thank Vera Lordan and Fiona Van Middlesworth for helping with the data collection and Rebecca Cobb for her support in the data analysis of this study. They also thank Manish Oza, Derrel Grey, and Gosia Sylwestrzak for helpful comments.

Author Affiliations: Elevance Health (BU, DC, ME), Indianapolis, IN; Texas A&M University (BU), College Station, TX; CVS Health (SS), Woonsocket, RI; Carelon (DD, MJF), Morristown, NJ.

Source of Funding: None.

Author Disclosures: Dr Debono and Dr Fisch are both employees of Carelon, a subsidiary of Elevance Health, and both own stock in Elevance Health; Carelon participates in prior authorization of medical procedures and drugs. Dr Eleff owns stock in Elevance Health. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (BU, SS, DC, DD, ME, MJF); acquisition of data (BU, SS, DC); analysis and interpretation of data (BU, SS, DC, DD, ME, MJF); drafting of the manuscript (BU, DC, DD, ME, MJF); critical revision of the manuscript for important intellectual content (BU, DC, DD, MJF); and statistical analysis (BU, SS, DC).

Address Correspondence to: Benjamin Ukert, PhD, Texas A&M University, 212 Adriance Lab Rd, College Station, TX 77843. Email: bukert@tamu.edu.

REFERENCES

1. Brook RH. Assessing the appropriateness of care—its time has come. JAMA. 2009;302(9):997-998. doi:10.1001/jama.2009.1279

2. Schwartz AL, Brennan TA, Verbrugge DJ, Newhouse JP. Measuring the scope of prior authorization policies: applying private insurer rules to Medicare Part B. JAMA Health Forum. 2021;2(5):e210859. doi:10.1001/jamahealthforum.2021.0859

3. Agarwal A, Freedman RA, Goicuria F, et al. Prior authorization for medications in a breast oncology practice: navigation of a complex process. J Oncol Pract. 2017;13(4):e273-e282. doi:10.1200/JOP.2016.017756

4. Dickens DS, Pollock BH. Medication prior authorization in pediatric hematology and oncology. Pediatr Blood Cancer. 2017;64(6):e26339. doi:10.1002/pbc.26339

5. Smith AJ, Mulugeta-Gordon L, Pena D, et al. Disparities in prior authorization for gynecologic cancer care (137). Gynecol Oncol. 2022;166(suppl 1):S85. doi:10.1016/S0090-8258(22)01362-2

6. Anderson KE, Darden M, Jain A. Improving prior authorization in Medicare Advantage. JAMA. 2022;328(15):1497-1498. doi:10.1001/jama.2022.17732

7. Trapani D, Kraemer L, Rugo HS, Lin NU. Impact of prior authorization on patient access to cancer care. Am Soc Clin Oncol Educ Book. 2023;43:e100036. doi:10.1200/EDBK_100036

8. Bhardwaj AS, Liu M, Pintova S, et al. Implementation of a streamlined prior authorization process to improve cancer care delivery. J Clin Oncol. 2023;41(suppl 16):1531. doi:10.1200/JCO.2023.41.16_suppl.1531

9. Newcomer LN, Weininger R, Carlson RW. Transforming prior authorization to decision support. J Oncol Pract. 2017;13(1):e57-e61. doi:10.1200/JOP.2016.015198

10. 2021 AMA prior authorization (PA) physician survey. American Medical Association. Accessed August 1, 2022. https://web.archive.org/web/20220802174915/https:/www.ama-assn.org/system/files/prior-authorization-survey.pdf

11. Lin NU, Bichkoff H, Hassett MJ. Increasing burden of prior authorizations in the delivery of oncology care in the United States. J Oncol Pract. 2018;14(9):525-528. doi:10.1200/JOP.18.00428

12. Murphy CT, Galloway TJ, Handorf EA, et al. Survival impact of increasing time to treatment initiation for patients with head and neck cancer in the United States. J Clin Oncol. 2016;34(2):169-178. doi:10.1200/JCO.2015.61.5906

13. Kim H, Srivastava A, Gabani P, Kim E, Lee H, Pedersen KS. Oncology trainee perceptions of the prior authorization process: a national survey. Adv Radiat Oncol. 2021;7(2):100861. doi:10.1016/j.adro.2021.100861

14. Brot-Goldberg ZC, Burn S, Layton T, Vabson B. Rationing medicine through bureaucracy: authorization restrictions in Medicare. National Bureau of Economic Research working paper 30878. January 2023. doi:10.3386/w30878

15. Morris AM, Rhoads KF, Stain SC, Birkmeyer JD. Understanding racial disparities in cancer treatment and outcomes. J Am Coll Surg. 2010;211(1):105-113. doi:10.1016/j.jamcollsurg.2010.02.051

16. Howell EA, Egorova N, Hayes MP, Wisnivesky J, Franco R, Bickell N. Racial disparities in the treatment of advanced epithelial ovarian cancer. Obstet Gynecol. 2013;122(5):1025-1032. doi:10.1097/AOG.0b013e3182a92011

17. Ward E, Jemal A, Cokkinides V, et al. Cancer disparities by race/ethnicity and socioeconomic status. CA Cancer J Clin. 2004;54(2):78-93. doi:10.3322/canjclin.54.2.78

18. Fiscella K, Humiston S, Hendren S, et al. Eliminating disparities in cancer screening and follow-up of abnormal results: what will it take? J Health Care Poor Underserved. 2011;22(1):83-100. doi:10.1353/hpu.2011.0023

19. Chapman EN, Kaatz A, Carnes M. Physicians and implicit bias: how doctors may unwittingly perpetuate health care disparities. J Gen Intern Med. 2013;28(11):1504-1510. doi:10.1007/s11606-013-2441-1

20. Pallok K, De Maio F, Ansell DA. Structural racism—a 60-year-old black woman with breast cancer. N Engl J Med. 2019;380(16):1489-1493. doi:10.1056/NEJMp1811499

21. Smith AJB, Mulugeta-Gordon L, Pena D, et al. Insurance and racial disparities in prior authorization in gynecologic oncology. Gynecol Oncol Rep. 2023:46:101159. doi:10.1016/j.gore.2023.101159

22. Gupta R, Fein J, Newhouse JP, Schwartz AL. Comparison of prior authorization across insurers: cross sectional evidence from Medicare Advantage. BMJ. 2024;384:e077797. doi:10.1136/bmj-2023-077797

23. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. doi:10.1016/0895-4356(92)90133-8

24. Charlton M, Schlichting J, Chioreso C, Ward M, Vikas P. Challenges of rural cancer care in the United States. Oncology (Williston Park). 2015;29(9):633-640.

25. Alimena S, Davis M, Pelletier A, Terry K, King M, Feldman S. Regional variation in access to oncologic care and racial disparities among cervical cancer patients. Am J Clin Oncol. 2022;45(10):415-421. doi:10.1097/COC.0000000000000944

26. Khorana AA, Tullio K, Elson P, et al. Time to initial cancer treatment in the United States and association with survival over time: an observational study. PLoS One. 2019;14(3):e0213209. doi:10.1371/journal.pone.0213209

27. Fang H, Frean M, Sylwestrzak G, Ukert B. Trends in disenrollment and reenrollment within US commercial health insurance plans, 2006-2018. JAMA Netw Open. 2022;5(2):e220320. doi:10.1001/jamanetworkopen.2022.0320