ABSTRACT
Introduction
Medicare Advantage (MA) enrollment has grown rapidly, yet its implications for resource use in complex head and neck cancer (HNC) surgery remain unclear.
Methods
Using 2019–2023 Medicare Provider Analysis and Review data, we identified adults admitted with an HNC diagnosis who underwent a qualifying head and neck surgical procedure. We used 1:1 propensity score matching to balance MA and traditional fee‐for‐service (FFS) beneficiaries on demographics, comorbidity, region, and admitting diagnosis. Outcomes included index length of stay (LOS) and discharge destination (home, home with home health, skilled nursing facility/rehabilitation, or hospice). Negative binomial and multinomial logistic regression models were used to estimate associations between insurance type and LOS or discharge disposition.
Results
The matched cohort included 4618 admissions (2309 MA; 2309 FFS). Mean LOS was similar for MA and FFS patients (7.1 vs. 7.3 days), and discharge destination distributions were not significantly different. MA admissions had lower observed frequencies of selected complex procedure categories, including reconstructive surgery, neck dissection, laryngectomy, and mandibulectomy. In procedure‐adjusted sensitivity models, MA enrollment was not associated with LOS (rate ratio 0.99, 95% CI 0.94–1.03) or discharge destination.
Conclusions
Among Medicare beneficiaries hospitalized for HNC surgery, short‐term inpatient utilization during the index admission was similar for MA and FFS enrollees. MA admissions had lower observed frequencies of selected complex procedure categories; however, this finding should be interpreted as a difference in observed operative mix rather than direct evidence of reduced access. Further studies linking claims to tumor registry and hospital‐level data are needed to determine whether these patterns reflect case mix, treatment selection, hospital networks, or access barriers.
Level of Evidence
3.
Keywords: fee for service, head and neck cancer, length of stay, Medicare Advantage, postacute care
1. Introduction
Head and neck cancers (HNC) often require complex surgery with substantial perioperative resource use. Often, surgical management of HNC requires free flap reconstruction, and many patients postoperatively require enteral feeding, tracheostomy tube, and rehabilitation assistance (speech and language pathology, occupational therapy, physical therapy) prior to discharge. Postoperative length of stay (LOS) and discharge destination are clinically meaningful measures of utilization and recovery. Prior work demonstrates that discharge to SNF after head and neck reconstruction is associated with delayed initiation of adjuvant therapy and higher readmission rates, and that longer LOS strongly predicts non‐home discharge [1, 2, 3, 4].
Given the older age of patients with HNC, many are eligible for Medicare for insurance coverage. More than half of Medicare beneficiaries are enrolled in Medicare Advantage (MA), which now accounts for approximately 54% of all eligible enrollees [5]. Previous studies indicate that Medicare payments to MA plans in 2025 are approximately 20% higher per enrollee than projected fee‐for‐service (FFS) spending for similar beneficiaries [6]. Further, MA plans employ prior authorization requirements and selective provider networks that may influence access and utilization for cancer care [7, 8, 9, 10]. As such, MA beneficiaries may experience differential care patterns and may be less likely to receive complex cancer surgery at top‐ranked hospitals [11, 12, 13]. Other literature reports lower use and shorter duration of post‐acute care among MA enrollees compared with FFS beneficiaries, although results for patient‐reported functional outcomes are mixed [14, 15]. Additionally, insurance status has been associated with differential access to high‐quality HNC care, including treatment at high‐volume or top‐ranked hospitals. These findings raise the possibility that plan type could shape where and how complex HNC surgery is delivered, but claims data alone cannot distinguish network‐related access barriers from differences in tumor stage, recurrence, prior treatment, patient preferences, or hospital selection [16].
Despite the specialty's intensive use of inpatient and post‐acute services, the influence of Medicare insurance type on resource utilization in HNC surgery remains underexplored. While a protective “Medicare effect” on HNC diagnosis and survival has been demonstrated relative to uninsured patients [17], disparities within Medicare by plan type have not been examined. We hypothesized that MA beneficiaries would have shorter LOS and higher rates of non‐home discharge than FFS beneficiaries. We also evaluated operative mix as a secondary descriptive outcome, motivated by the a priori expectation that rates of selected complex procedure categories would be similar after matching.
2. Methods
We performed a retrospective cohort study from the Medicare Provider Analysis and Review (MEDPAR) limited dataset files from 2019–2023. MedPAR is a Centers for Medicare & Medicaid Services (CMS) dataset that contains inpatient hospital facility claims for Medicare beneficiaries. Each record summarizes a stay, including admission and discharge dates, diagnoses, procedures, provider identifiers, LOS, and discharge destination. Because MedPAR is derived from Medicare administrative billing data, it contains data for both Medicare beneficiaries enrolled in MA and FFS plans. This study was exempt by the Ohio State University Institutional Review Board given the use of deidentified secondary data.
We identified patients with an admitting diagnosis of head and neck cancer (HNC) using International Classification of Diseases, Tenth Revision (ICD‐10) codes. Inclusion required documentation of a concurrent head and neck surgical procedure, defined by the ICD‐10 Procedure Coding System (ICD‐10‐PCS), as listed in the Table S1. Each MedPAR record represents a single inpatient stay; the unit of analysis was the admission. Qualifying surgical procedures could occur on any day of the admission and were not restricted to the day of admission.
The primary outcomes were utilization variables including (1) LOS and (2) need for post‐acute care. LOS was the number of inpatient days during the index hospital admission when the procedure occurred. Post‐acute care use was determined by discharge location and categorized as (1) home, (2) home with home health, (3) skilled nursing facility (SNF) or rehabilitation, or (4) hospice. Operative mix, defined by selected major procedure categories, was evaluated as a secondary descriptive outcome. Our covariates included age, sex, race, Charlson comorbidity index (CCI), admitting diagnosis, and geographic region.
Patients undergoing a HNC procedure enrolled in MA versus FFS Medicare were compared. Propensity score matching was performed to balance baseline characteristics between MA and FFS patients. A nearest‐neighbor matching approach (1:1 ratio) was used with a logistic regression model estimating the probability of MA enrollment based on age, sex, race, Charlson comorbidity index (CCI) score, geographic region, and primary admitting diagnosis subsite. Matching was conducted on the logit of the propensity score with a caliper of 0.01 to ensure close matches and minimize residual confounding. These variables were selected because they are the patient‐level characteristics available in MedPAR that plausibly influence both MA enrollment and perioperative outcomes. Hospital‐level characteristics (teaching status, volume, bed size, etc.) were not available in the limited dataset file.
All analyses were performed after propensity score matching. Baseline characteristics used in the matching were summarized as means (standard deviations) or frequencies (percentages) and assessed for balance between MA and FFS patients using standardized mean differences (SMDs), with SMDs < 0.1 indicating good balance. Variables not included in the matching model, such as outcomes or surgical procedures, were summarized similarly, and differences between groups were assessed using two‐sample t‐tests for continuous variables or Fisher's exact tests for categorical variables. Associations between insurance type and discharge destination were evaluated using multinomial logistic regression models, and odds ratios (ORs) with 95% confidence intervals (CIs) were reported. LOS, treated as a count variable, was analyzed with negative binomial regression to account for overdispersion, with rate ratios (RRs) and 95% CIs reported.
The matched, unadjusted regression models were interpreted as the primary overall estimates of the association between insurance type and LOS or discharge destination. Because procedure type may lie on the causal pathway between insurance type and downstream utilization, procedure‐adjusted models were considered conditional sensitivity analyses rather than estimates of the total payer association. In these sensitivity models, major procedure categories were included to assess whether associations persisted after conditioning on broad operative type. For discharge destination models, LOS was also included in the fully adjusted specification; for LOS models, discharge destination was included in the fully adjusted specification. These reciprocal adjustments may introduce collider bias and were therefore interpreted cautiously. Because surgical procedures were identified from inpatient ICD‐10‐PCS codes rather than more granular CPT codes or operative reports [18], procedures were kept broad and restricted to a set of operations likely to influence length of stay, including reconstructive procedures, neck dissection, laryngectomy, mandibulectomy, and tracheostomy. These categories were not mutually exclusive; a single admission could include more than one procedure category. All statistical analyses were conducted in SAS 9.4 (SAS Institute Inc., Cary, NC). All tests were two‐sided, and a p‐value < 0.05 was considered statistically significant.
3. Results
From 2019–2023, we identified 4618 matched inpatient admissions for adults undergoing HNC surgery, comprising 2309 MA and 2309 FFS beneficiaries. Propensity score matching yielded strong covariate balance: after matching, absolute standardized mean differences were ≤ 0.04 across age range, sex, race, Charlson Comorbidity Index (CCI), geographic region, and admitting diagnosis subsite. The cohort was predominantly male (69.7%) and White (83.7%); mean CCI was 7.0 (SD 2.9). By subsite, oral cavity was most common (41.8%), followed by larynx (24.1%) and oropharynx (17.6%). These distributions and balance diagnostics are summarized in Table 1.
TABLE 1.
Propensity score matching variables summary by insurance type.
| Variable | Level | FSS (n = 2309) | MA (n = 2309) | Total (n = 4618) | Std. mean diff. (After matching) a |
|---|---|---|---|---|---|
| Age | 45–64 | 283 (12.3%) | 286 (12.4%) | 569 (12.3%) | 0.0071 |
| 65–69 | 641 (27.8%) | 643 (27.8%) | 1284 (27.8%) | ||
| 70–74 | 564 (24.4%) | 557 (24.1%) | 1121 (24.3%) | ||
| 75–79 | 410 (17.8%) | 408 (17.7%) | 818 (17.7%) | ||
| 80–84 | 255 (11%) | 258 (11.2%) | 513 (11.1%) | ||
| 85–89 | 111 (4.8%) | 114 (4.9%) | 225 (4.9%) | ||
| 90 and over | 45 (1.9%) | 43 (1.9%) | 88 (1.9%) | ||
| Sex | Female | 696 (30.1%) | 704 (30.5%) | 1400 (30.3%) | 0.0075 |
| Male | 1613 (69.9%) | 1605 (69.5%) | 3218 (69.7%) | ||
| Race | Asian | 34 (1.5%) | 47 (2%) | 81 (1.8%) | 0.0402 |
| Black | 206 (8.9%) | 201 (8.7%) | 407 (8.8%) | ||
| Hispanic | 56 (2.4%) | 58 (2.5%) | 114 (2.5%) | ||
| Other | 72 (3.1%) | 80 (3.5%) | 152 (3.3%) | ||
| White | 1941 (84.1%) | 1923 (83.3%) | 3864 (83.7%) | ||
| Charlson index (CCI) |
Mean (SD) (min, max) |
7 (2.9) (3, 18) |
7 (2.9) (3, 20) |
7 (2.9) (3, 20) |
0.0061 |
| Geographic Region | Mid/South Atlantic | 772 (33.4%) | 764 (33.1%) | 1536 (33.3%) | 0.0178 |
| Midwest | 311 (13.5%) | 307 (13.3%) | 618 (13.4%) | ||
| New England | 50 (2.2%) | 47 (2%) | 97 (2.1%) | ||
| Pacific | 302 (13.1%) | 316 (13.7%) | 618 (13.4%) | ||
| Puerto Rico | 72 (3.1%) | 72 (3.1%) | 144 (3.1%) | ||
| South | 530 (23%) | 529 (22.9%) | 1059 (22.9%) | ||
| West | 272 (11.8%) | 274 (11.9%) | 546 (11.8%) | ||
| Admitting diagnosis (Subsite) | Hypopharynx | 41 (1.8%) | 47 (2%) | 88 (1.9%) | 0.0186 |
| Larynx | 562 (24.3%) | 550 (23.8%) | 1112 (24.1%) | ||
| Nasopharynx | 39 (1.7%) | 40 (1.7%) | 79 (1.7%) | ||
| Oral Cavity | 966 (41.8%) | 966 (41.8%) | 1932 (41.8%) | ||
| Oropharynx | 406 (17.6%) | 408 (17.7%) | 814 (17.6%) | ||
| Salivary gland cancer | 245 (10.6%) | 242 (10.5%) | 487 (10.5%) | ||
| Sinonasal | 50 (2.2%) | 56 (2.4%) | 106 (2.3%) |
For multi‐level categorical variables, the reported standardized mean difference (SMD) corresponds to the maximum absolute SMD across all levels for that variable.
Unadjusted utilization patterns, which were treated as the primary overall post‐matching comparisons, were similar between insurance groups. Mean LOS was 7.3 days (SD 8.6) for FFS and 7.1 days (SD 8.1) for MA (p = 0.26). There were no significant differences in discharge destination between groups: home (52.9% FFS vs. 52.4% MA), home with home health services (31.3% FFS vs. 32.8% MA), SNF or rehabilitation (12.6% FFS vs. 11.3% MA), or hospice (3.2% FFS vs. 3.6% MA) (overall comparison, p = 0.41). As a secondary descriptive outcome, operative mix differed by payer: MA admissions had lower observed frequencies of selected major procedure categories, including reconstructive surgery (5.4% of MA vs. 11.0% of FFS admissions), neck dissection (40.8% vs. 52.0%), laryngectomy (7.1% vs. 11.6%), and resection of mandible (5.2% vs. 9.8%; all p < 0.001) (Table 2).
TABLE 2.
Descriptive summary by insurance type after PSM.
| Variable | Level | FSS (n = 2309) | MA (n = 2309) | Total (n = 4618) | p |
|---|---|---|---|---|---|
| Discharge destination | Home | 1221 (52.9%) | 1209 (52.4%) | 2430 (52.6%) | 0.410 |
| Home with home health | 723 (31.3%) | 757 (32.8%) | 1480 (32%) | ||
| Hospice | 75 (3.2%) | 83 (3.6%) | 158 (3.4%) | ||
| SNF/rehab | 290 (12.6%) | 260 (11.3%) | 550 (11.9%) | ||
| Length of stay |
Mean (SD) (min, max) |
7.3 (8.6) (1, 147) |
7.1 (8.1) (1, 99) |
7.2 (8.3) (1, 147) |
0.256 |
| Procedure count |
Mean (SD) (min, max) |
6 (4.8) (1, 25) |
4.5 (3.8) (1, 25) |
5.2 (4.4) (1, 25) |
< 0.001 |
| Urbanicity | Rural | 36 (1.6%) | 44 (1.9%) | 80 (1.7%) | 0.367 |
| Urban | 2273 (98.4%) | 2265 (98.1%) | 4538 (98.3%) | ||
| Reconstructive | No | 2055 (89%) | 2185 (94.6%) | 4240 (91.8%) | < 0.001 |
| Yes | 254 (11%) | 124 (5.4%) | 378 (8.2%) | ||
| Neck dissection | No | 1109 (48%) | 1367 (59.2%) | 2476 (53.6%) | < 0.001 |
| Yes | 1200 (52%) | 942 (40.8%) | 2142 (46.4%) | ||
| Laryngectomy | No | 2042 (88.4%) | 2144 (92.9%) | 4186 (90.6%) | < 0.001 |
| Yes | 267 (11.6%) | 165 (7.1%) | 432 (9.4%) | ||
| Mandible | No | 2083 (90.2%) | 2188 (94.8%) | 4271 (92.5%) | < 0.001 |
| Yes | 226 (9.8%) | 121 (5.2%) | 347 (7.5%) | ||
| Tracheostomy | No | 1688 (73.1%) | 1770 (76.7%) | 3458 (74.9%) | 0.005 |
| Yes | 621 (26.9%) | 539 (23.3%) | 1160 (25.1%) |
Note: Matched characteristics of Fee‐for‐Service (FSS) and Medicare Advantage (MA) patients after 1:1 propensity score matching (n = 4618). Discharge patterns and length of stay were similar between groups (mean 7.3 vs. 7.1 days, p = 0.26). However, MA patients underwent fewer complex surgeries, including reconstructive, neck dissection, laryngectomy, and maxilla/mandible procedures (all p < 0.0001).
Conditional sensitivity models that adjusted for procedure groups and LOS also found no significant association between insurance type and discharge destination. Although none of the differences reached statistical significance, the point estimates suggested slightly higher odds of discharge to home with home health for MA beneficiaries (adjusted OR 1.13, 95% CI 0.98–1.31; p = 0.105) and lower odds of discharge to a skilled nursing facility or rehabilitation (0.89, 0.72–1.10; p = 0.266) and hospice (0.78, 0.55–1.10; p = 0.162) compared with FFS beneficiaries. Unadjusted estimates showed similar patterns (Table 3).
TABLE 3.
Association between Medicare Advantage versus Fee‐for‐Service and discharge destination after Head and Neck Cancer surgery.
| Discharge destination (MA vs. FSS) | OR | 95% CI | p | |
|---|---|---|---|---|
| Unadjusted | Home with Home Health versus Home | 1.06 | 0.93–1.20 | 0.397 |
| Hospice versus Home | 1.12 | 0.81–1.54 | 0.499 | |
| SNF/rehab versus Home | 0.91 | 0.75–1.09 | 0.294 | |
| Adjusted a | Home with Home Health versus Home | 1.13 | 0.98–1.31 | 0.105 |
| Hospice versus Home | 0.78 | 0.55–1.10 | 0.162 | |
| SNF/rehab versus Home | 0.89 | 0.72–1.10 | 0.266 | |
| Length of stay (MA vs. FSS) | RR | 95% CI | p |
|---|---|---|---|
| Unadjusted | 0.96 | 0.91–1.01 | 0.152 |
| Adjusted a | 0.99 | 0.95–1.04 | 0.739 |
For outcome Discharge Destination, model adjusted for Procedure (Dichotomous variables: Reconstructive, Neck dissection, Laryngectomy, Maxilla/Mandible, Trach) and LOS; For outcome LOS, model adjusted for Procedure (Dichotomous variables: Reconstructive, Neck dissection, Laryngectomy, Maxilla/Mandible, Trach) and Discharge Destination.
MA enrollment was not associated with LOS. The matched, unadjusted LOS ratio was 0.96 (95% CI, 0.91–1.01; p = 0.152). In the corresponding procedure‐adjusted sensitivity model, the negative binomial regression estimated a LOS ratio of 0.99 (95% CI, 0.94–1.03; p = 0.739) for MA vs. FFS (Table 3).
4. Discussion
In this matched national cohort of HNC surgical admissions, we found no statistically significant differences in the primary short‐term inpatient utilization outcomes of LOS or discharge destination between MA and FFS beneficiaries. Unadjusted post‐matching models and conditional models that adjusted for broad procedure categories yielded similar conclusions. MA admissions had lower observed frequencies of selected major procedure categories, including laryngectomy, neck dissection, and mandible resection. These findings are best interpreted as differences in observed operative mix among hospitalized surgical admissions. They do not establish that MA beneficiaries had reduced access to complex surgery, because operative selection may reflect tumor stage, recurrent disease, prior treatment, frailty, patient preferences, hospital selection, or coding differences not captured in MedPAR.
Several mechanisms could explain the observed operative‐mix difference. Complex HNC procedures such as reconstructive surgery, total laryngectomy, and mandibulectomy are typically reserved for patients with more extensive disease, prior treatment, or reconstructive needs. In our matched cohort of surgical admissions, MA beneficiaries had lower observed frequencies of these categories, but the reason for this difference cannot be determined from administrative claims. The finding could reflect true differences in access to high‐volume hospitals or surgeons, selective referral networks, or prior authorization; it could also reflect differences in disease severity, prior treatment, nonoperative treatment pathways, or patient goals. Prior national studies using MedPAR and similar data have reported that MA beneficiaries are less likely than FFS beneficiaries to receive major cancer surgery at high‐volume or high‐quality centers, with higher mortality observed for some complex operations [19]. Our results are consistent with the possibility that payer‐related factors influence operative pathways in HNC, but they remain hypothesis‐generating and should not be interpreted as evidence of underuse without tumor registry, prior‐treatment, frailty, and hospital‐level linkage. Systematic coding differences between MA and FFS claims may also contribute; MA encounter data have historically been subject to different documentation incentives than FFS claims, which could affect ICD‐10‐PCS procedure capture [20, 21].
MA plans have incentives to reduce post‐acute intensity, and prior work demonstrates that MA enrollees receive less and shorter SNF and home‐health care after hospitalization than their FFS counterparts [14, 22, 23]. In our cohort, however, discharge destinations during the index admission were similar between MA and FFS despite differences in observed operative mix. This does not exclude payer‐related differences in post‐acute care after discharge. Our discharge categories capture where patients went, not the duration or intensity of SNF, rehabilitation, home health, or supportive services. Payer‐related differences may also emerge in outcomes outside the index stay, including readmissions, delayed complications, emergency department use, delays in adjuvant therapy, or long‐term functional recovery.
This study has several limitations. Administrative data lack tumor‐specific variables (stage, HPV status, margin status, recurrence), prior treatment, frailty or performance status, functional measures, patient preferences, and granular reconstructive details, leaving residual confounding possible despite propensity score matching and multivariable adjustment. Procedure classification relied on broad ICD‐10‐PCS categories that were overlapping and not mutually exclusive, limiting assessment of surgical nuance and procedure‐specific comparisons. Our analysis focused on short‐term index‐admission outcomes (LOS and discharge destination). We could not evaluate postoperative complications, readmissions, duration or intensity of post‐acute services, delays in adjuvant therapy, feeding tube or tracheostomy dependence, functional recovery, recurrence, or survival. Therefore, similar LOS and discharge destination should not be interpreted as evidence of equivalent longitudinal cancer care or functional recovery. We analyzed only inpatient operations, so differential adoption of outpatient HNC surgery across payers may confound complexity comparisons. Our study period (2019–2023) spans the COVID‐19 pandemic, which may have influenced outcomes. Finally, Maryland's all‐payer hospital rate‐setting system, in which all payers reimburse at the same rate, may attenuate payer‐driven utilization differences for admissions in that state.
Overall, our findings suggest that among older adults admitted for inpatient HNC surgery, MA enrollment was not associated with shorter LOS or differences in discharge destination in matched analyses. MA admissions had lower observed frequencies of selected complex procedure categories, but this pattern should be viewed as an operative‐mix difference rather than direct evidence of access barriers. Clarifying whether these patterns reflect case mix, treatment selection, hospital networks, or inequitable access to complex oncologic surgery will be important as MA enrollment continues to grow.
Funding
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: ICD‐10‐CM diagnosis codes and ICD‐10‐PCS procedure codes used for cohort identification. Diagnosis codes are organized by anatomic subsite (hypopharynx, larynx, nasopharynx, oral cavity, oropharynx, salivary gland, and sinonasal). Procedure codes are organized by surgical category (reconstructive, neck dissection, laryngectomy, mandibulectomy, tracheostomy, and other ablative). Other ablative procedures were used for cohort inclusion but were not modeled as a distinct procedure category in adjusted analyses.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Dang S., Patel T., Lao I., et al., “Discharge Disposition After Head and Neck Reconstruction: Effect on Adjuvant Therapy and Outcomes,” Laryngoscope 133, no. 11 (2023): 2977–2983, 10.1002/lary.30648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Duckett K. A., Kassir M. F., Nguyen S. A., et al., “Factors Associated With Head and Neck Cancer Postoperative Radiotherapy Delays: A Systematic Review and Meta‐Analysis,” Otolaryngology ‐ Head and Neck Surgery (United States) 171, no. 5 (2024): 1265–1282, 10.1002/ohn.835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Tucker J., Hollenbeak C. S., and Goyal N., “Discharge Destination and Readmissions Among Patients With Head and Neck Cancer,” Laryngoscope Investigative Otolaryngology 7, no. 5 (2022): 1407–1429, 10.1002/lio2.890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Cramer J. D., Patel U. A., Samant S., and Shintani Smith S., “Discharge Destination After Head and Neck Surgery: Predictors of Discharge to Postacute Care,” Otolaryngology ‐ Head and Neck Surgery (United States) 155, no. 6 (2016): 997–1004, 10.1177/0194599816661514. [DOI] [PubMed] [Google Scholar]
- 5. Ochieng N., Freed M., Biniek J. F., Damico A., and Neuman T., “Medicare Advantage in 2025: Enrollment Update and Key Trends,” 2025, accessed November 19, 2025, https://www.kff.org/medicare/medicare‐advantage‐enrollment‐update‐and‐key‐trends/.
- 6. Medicare Payment Advisory Commission , “Report to the Congress: Medicare Payment Policy,” MedPAC, 2025.
- 7. Neprash H. T., Mulcahy J. F., and Golberstein E., “The Extent and Growth of Prior Authorization in Medicare Advantage,” American Journal of Managed Care 30, no. 3 (2024): E85–E92, 10.37765/ajmc.2024.89519. [DOI] [PubMed] [Google Scholar]
- 8. Chino F., Baez A., Elkins I. B., Aviki E. M., Ghazal L. V., and Thom B., “The Patient Experience of Prior Authorization for Cancer Care,” JAMA Network Open 6, no. 10 (2023): E2338182, 10.1001/jamanetworkopen.2023.38182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gracie J., Jimenez R., and Winkfield K. M., “The Burden of Insurance Prior Authorization on Cancer Care: A Review of Evidence From Radiation Oncology,” Advances in Radiation Oncology 10, no. 1 (2024): 101654, 10.1016/j.adro.2024.101654. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Schleicher S. M., Mullangi S., and Feeley T. W., “Effects of Narrow Networks on Access to High‐Quality Cancer Care,” JAMA Oncology 2, no. 4 (2016): 427–428, 10.1001/jamaoncol.2015.6125. [DOI] [PubMed] [Google Scholar]
- 11. Maganty A., Liu X., Dall C., et al., “Surgery at High‐Quality Hospitals Among Medicare Advantage Beneficiaries Undergoing Cancer Surgery,” JAMA Surgery 160, no. 12 (2025): 1341–1347, 10.1001/jamasurg.2025.4320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Raoof M., Jacobson G., and Fong Y., “Medicare Advantage Networks and Access to High‐Volume Cancer Surgery Hospitals,” Annals of Surgery 274, no. 4 (2021): E315–E319, 10.1097/SLA.0000000000005098. [DOI] [PubMed] [Google Scholar]
- 13. Raoof M., Philip J., Ituarte H. G., et al., “Medicare Advantage: A Disadvantage for Complex Cancer Surgery Patients,” Journal of Clinical Oncology 41 (2022): 1239–1249, 10.1200/JCO.21. [DOI] [PubMed] [Google Scholar]
- 14. Achola E. M., Stevenson D. G., and Keohane L. M., “Postacute Care Services Use and Outcomes Among Traditional Medicare and Medicare Advantage Beneficiaries,” JAMA Health Forum 4, no. 8 (2023): E232517, 10.1001/jamahealthforum.2023.2517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. McGarry B. E., Wilcock A. D., Gandhi A. D., Grabowski D. C., and Barnett M. L., “Extended Hospital Stays in Medicare Advantage and Traditional Medicare,” JAMA Internal Medicine 185, no. 11 (2025): 1362–1369, 10.1001/jamainternmed.2025.4411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Megwalu U. C., Ma Y., Divi V., and Tian L., “Insurance Disparities in Quality of Care Among Patients With Head and Neck Cancer,” JAMA Otolaryngology. Head & Neck Surgery 150, no. 8 (2024): 641–650, 10.1001/jamaoto.2024.1338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Smith J. B., Jayanth P., Hong S. A., Simpson M. C., and Massa S. T., “The “Medicare Effect” on Head and Neck Cancer Diagnosis and Survival,” Head & Neck 45, no. 7 (2023): 1663–1675, 10.1002/hed.27379. [DOI] [PubMed] [Google Scholar]
- 18. Utter G. H., Cox G. L., Owens P. L., and Romano P. S., “Challenges and Opportunities With ICD‐10‐CM/PCS: Implications for Surgical Research Involving Administrative Data,” Journal of the American College of Surgeons 217, no. 3 (2013): 516–526, 10.1016/j.jamcollsurg.2013.04.029. [DOI] [PubMed] [Google Scholar]
- 19. Kim D., Hubbard R. A., Gao Y., et al., “Comparison of the Use of Top‐Ranked Cancer Hospitals Between Medicare Advantage and Traditional Medicare,” JAMA Network Open 5, no. 3 (2022): e221809, 10.1001/jamanetworkopen.2022.1809. [DOI] [Google Scholar]
- 20. Skopec L., Huckfeldt P. J., Wissoker D., et al., “Home Health and Postacute Care Use in Medicare Advantage and Traditional Medicare,” Health Affairs 39, no. 5 (2020): 837–842, 10.1377/hlthaff.2019.00844. [DOI] [PubMed] [Google Scholar]
- 21. Meyers D. J., Trivedi A. N., and Mor V., “Limited Use of Medicare Advantage Encounter Data in Research: Methodological Considerations and Policy Implications,” American Journal of Managed Care 28, no. 3 (2022): e79–e83, 10.37765/ajmc.2022.88830. [DOI] [Google Scholar]
- 22. Prusynski R. A., D'Alonzo A., Johnson M. P., Mroz T. M., and Leland N. E., “Differences in Home Health Services and Outcomes Between Traditional Medicare and Medicare Advantage,” JAMA Health Forum 5, no. 3 (2024): E235454, 10.1001/jamahealthforum.2023.5454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Jung J., Carlin C., Feldman R., Song G., and Mitchell A., “Use of Low‐Value Cancer Treatments in Medicare Advantage Versus Traditional Medicare,” Journal of Clinical Oncology 43 (2025): 2245–2254, 10.1200/JCO-24-01907. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: ICD‐10‐CM diagnosis codes and ICD‐10‐PCS procedure codes used for cohort identification. Diagnosis codes are organized by anatomic subsite (hypopharynx, larynx, nasopharynx, oral cavity, oropharynx, salivary gland, and sinonasal). Procedure codes are organized by surgical category (reconstructive, neck dissection, laryngectomy, mandibulectomy, tracheostomy, and other ablative). Other ablative procedures were used for cohort inclusion but were not modeled as a distinct procedure category in adjusted analyses.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.