Abstract
Background:
A prior study reported that over half of patients with HNSCC initiated PORT after 6 weeks from surgery during 2006–2014. In 2022, the CoC released a quality metric for patients to initiate PORT within 6 weeks. This study provides an update on time to PORT in recent years.
Methods:
The NCDB and TriNetX Research Network were queried to identify patients with HNSCC who received PORT during 2015–2019 and 2015–2021, respectively. Treatment delay was defined as initiating PORT beyond 6 weeks after surgery.
Results:
In NCDB, PORT was delayed for 62% of patients. Predictors of delay included age >50, female sex, black race, nonprivate insurance/uninsured status, lower education, oral cavity site, negative surgical margins, increased postoperative length of stay, unplanned hospital readmissions, IMRT radiation modality, treatment at an academic hospital or in the Northeast, and surgery and radiation at different facilities. In TriNetX, 64% experienced treatment delay. Additional associations with prolonged time to treatment included never married/divorced/widowed marital status, major surgery (neck dissection/free flaps/laryngectomy), and gastrostomy/tracheostomy dependence.
Conclusions:
There continue to be challenges to timely initiation of PORT.
Keywords: head and neck cancer, head and neck squamous cell carcinoma, postoperative radiation, quality care, treatment delay
1 |. INTRODUCTION
Timely, guideline-adherent treatment is essential to delivering quality cancer care.1 According to the National Comprehensive Cancer Network (NCCN), patients with locoregionally advanced head and neck squamous cell carcinoma (HNSCC) requiring postoperative radiation (PORT) should initiate therapy within 6 weeks of surgery.2 A National Cancer Database (NCDB) analysis spanning 2006 through 2014 reported that greater than half of patients with HNSCC requiring PORT initiated treatment after 6 weeks from surgery.3 Several subsequent publications attempt to further elucidate reasons for treatment delay and measures to implement targeted quality improvement interventions.4–6
In January of 2022, the American College of Surgeons (ACS) Commission on Cancer (CoC) released a national quality metric for patients with HNSCC to initiate PORT within 6 weeks.7 Given that 8 years passed since the last large-scale analysis of timeliness of care, this study provides an update on adherence to guidelines for time to initiation of PORT for patients with HNSCC. Additionally, we sought to define the current state of the problem, propose additional factors associated with delay, and need for further interventions.
2 |. MATERIALS AND METHODS
The study data was obtained from the NCDB. The NCDB includes patient information from more than 1500 CoC accredited institutions across the United States.8 The study cohort, outcome measures, and study variables were based on methods previously described by Graboyes et al.,3 except for the time period of database review. Graboyes et al. analyzed 2006 through 2014; therefore, we reviewed 2015 through 2019. The cohort was composed of patients with HNSCC who initiated PORT, with treatment delay defined as beginning radiation beyond 6 weeks after the date of surgical intervention. Demographic characteristics collected included age, sex, race, ethnicity, insurance status, area of residence, educational attainment quartile, median household income, and distance to hospital. Educational attainment was determined by comparing patient zip codes at the time of diagnosis to American Community Survey data and is based on the percentage of adults in a given zip code who did not graduate high school. Clinical and surgical covariates included cancer site, American Joint Committee on Cancer (AJCC) clinical and pathological stages, surgical margin status, postoperative length of stay, 30-day hospital readmissions, radiation modality, concurrent chemoradiation, and Charlson–Deyo Comorbidity score. Patients diagnosed before 2018 were staged according to the AJCC 7th edition guidelines, and after 2018 using the 8th edition. Hospital characteristics analyzed included treatment facility type, number of facilities involved in treatment, whether surgery and radiation were conducted at the same facility, and region of care. Treatment facility type is based on the annual volume of newly diagnosed cancer cases and structure of facilities: community cancer program (100–500 annual diagnoses), comprehensive community cancer program (>500 annual diagnoses), academic/research program (>500 annual diagnoses plus resident trainees), and integrated network cancer program (organization owning a group of facilities designed to offer comprehensive and integrative cancer care services).
An additional analysis was conducted using the TriNetX Research Network. TriNetX is a global federated health research database that provided access to electronic medical records (diagnoses, procedures, medications, laboratory values) from health care organizations (HCOs) in the United States.9 The database was queried to identify patients with head and neck cancer diagnosed during 2015 through 2021 who initiated PORT. Similar exclusion criteria were applied as the NCDB analysis, that is, those who underwent surgery beyond 6 months after initial diagnosis or radiation beyond 6 months postoperatively, and those who underwent induction chemotherapy or received stereotactic radiation or brachytherapy. The diagnosis and procedure codes utilized to execute this search criteria are listed in Data S1, Supporting Information. Those who met inclusion and exclusion criteria were stratified according to whether they received PORT within or beyond 6 weeks of surgery. Demographic (age, sex, race, ethnicity, region, marital status), clinical (tumor site, concurrent chemoradiation, and presence of tracheostomy or gastrostomy tube), and surgical variables (neck dissection, free flaps, laryngectomy) were compared between groups.
3 |. STATISTICAL ANALYSIS
For the NCDB analysis, bivariate associations were first assessed using chi-square tests to compare between those with initiation of PORT less than or equal to 6 weeks versus beyond 6 weeks. The factors with significant bivariate associations were then considered for the multivariable logistic regression modeling process. All factors were included in the full model, which was subsequently reduced using a manual backwards selection procedure to only include those factors which remained significant at the 0.05 level. Analysis was conducted utilizing SAS statistical software version 9.4 (SAS Institute Inc., Cary, NC). For the TriNetX subanalysis, odds ratios and 95% confidence intervals were calculated to determine the odds of receiving substandard care for those who belonged to a particular demographic or had certain clinical or surgical variables coded in their medical record. Statistical analyses were performed within the TriNetX platform, which is based on R, JAVA, and Python.
4 |. ETHICAL APPROVAL
Both TriNetX and the NCDB only use aggregated counts of de-identified information, with no protected health information or personal data made available. Therefore, this study was exempted by the Penn State Institutional Review Board review (STUDY00018629 for TriNetX and STUDY00018678 for NCDB).
5 |. RESULTS
5.1 |. NCDB analysis
In the NCDB, 47 331 patients were identified as potentially eligible for analysis. Excluded were 4770 patients who underwent brachytherapy, stereotactic radiosurgery, or had an unknown radiation modality, 570 patients who received palliative treatment, and 249 patients who underwent neoadjuvant therapy. There were an additional 1244 patients who underwent surgery beyond 6 months of diagnosis and 334 who began radiation beyond 6 months of surgery who were also excluded. The final cohort for analysis included 40 164 patients, with demographics described in Table 1.
TABLE 1.
NCDB analysis of demographic characteristics
| Patient variable | Total patients (n = 40 164) | Initiation of PORT ≤6 weeks (n = 15 263) | Initiation of PORT >6 weeks (n = 24 901) | OR [95% CI] |
|---|---|---|---|---|
| Age | ||||
| <50 | 5258 | 2122 (40.4%) | 3136 (59.6%) | 1.00 [Ref] |
| 50–59 | 11 424 | 4296 (37.6%) | 7128 (62.4%) | 1.12 [1.05–1.20] |
| 60–69 | 13 397 | 5063 (37.8%) | 8334 (62.2%) | 1.11 [1.04–1.18] |
| >70 | 10 085 | 3782 (37.5%) | 6303 (62.5%) | 1.28 [1.05–1.20] |
| Sex | ||||
| Male | 29 739 | 11 858 (39.9%) | 17 881 (60.1%) | 1.00 [Ref] |
| Female | 10 425 | 3405 (32.7%) | 7020 (67.3%) | 1.36 [1.30–1.43] |
| Race | ||||
| White | 35 080 | 13 722 (39.1%) | 21 358 (60.9%) | 1.00 [Ref] |
| Black | 3130 | 929 (29.7%) | 2201 (70.3%) | 1.52 [1.40–1.64] |
| Asian/Pacific Islander | 1193 | 361 (30.3%) | 832 (69.7%) | 1.48 [1.30–1.67] |
| Other/unknown | 761 | 251 (33.0%) | 510 (67.0%) | 1.30 [1.12–1.52] |
| Ethnicity | ||||
| Non-Hispanic | 37 629 | 14 408 (38.3%) | 23 221 (61.7%) | 1.00 [Ref] |
| Hispanic | 1862 | 600 (32.2%) | 1262 (67.8%) | 1.30 [1.18–1.44] |
| Unknown | 673 | 255 (37.9%) | 418 (62.1%) | 1.01 [0.86–1.19] |
| Insurance status | ||||
| Private insurance | 1140 | 344 (30.2%) | 796 (69.8%) | 1.00 [Ref] |
| Not insured | 17 418 | 7529 (43.2%) | 9889 (56.8%) | 1.76 [1.54–2.00] |
| Medicaid | 4591 | 1216 (26.5%) | 3375 (73.5%) | 2.11 [1.96–2.27] |
| Medicare | 15 551 | 5644 (36.3%) | 9907 (63.7%) | 1.33 [1.27–1.39] |
| Other government | 983 | 316 (32.1%) | 667 (67.9%) | 1.60 [1.40–1.84] |
| Unknown | 481 | 214 (44.5%) | 267 (55.5%) | 0.95 [0.79–1.14] |
| Area of residencea | ||||
| Metropolitan | 32 187 | 12 150 (37.7%) | 20 037 (62.3%) | 1.00 [Ref] |
| Urban | 5899 | 2317 (39.3%) | 3582 (60.7%) | 0.93 [0.87–0.99] |
| Rural | 686 | 264 (38.5%) | 422 (61.5%) | 0.93 [0.78–1.10] |
| Educational attainmenta | ||||
| Highest quartile | 8386 | 3571 (42.6%) | 4815 (57.4%) | 1.00 [Ref] |
| Second highest quartile | 9818 | 3782 (38.5%) | 6036 (61.5%) | 1.11 [1.03–1.20] |
| Second lowest quartile | 9057 | 3252 (35.9%) | 5805 (64.1%) | 1.27 [1.18–1.37] |
| Lowest quartile | 6711 | 2221 (33.1%) | 4490 (66.9%) | 1.54 [1.43–1.67] |
| Median household incomea | ||||
| Less than $40227 | 6225 | 2088 (33.5%) | 4137 (66.5%) | 1.00 [Ref] |
| $40227–$50353 | 7660 | 2809 (36.7%) | 4851 (63.3%) | 0.87 [0.81–0.93] |
| $50354–$63332 | 7750 | 2963 (38.2%) | 4787 (61.8%) | 0.81 [0.76–0.87] |
| $63333+ | 12 260 | 4941 (40.3%) | 7319 (59.7%) | 0.74 [0.70–0.79] |
| Distance to hospital (miles) | ||||
| ≤10 | 19 724 | 7686 (39.0%) | 12 038 (61.0%) | 1.00 [Ref] |
| 11–20 | 6669 | 2593 (38.9%) | 4076 (61.1%) | 1.00 [0.94–1.06] |
| 21–50 | 7640 | 2810 (36.8%) | 4830 (63.2%) | 1.09 [1.03–1.15] |
| 51–10 | 3683 | 1232 (33.5%) | 2451 (66.5%) | 1.27 [1.17–1.36] |
| >100 | 2448 | 942 (38.5%) | 1506 (61.5%) | 1.02 [0.93–1.11] |
Abbreviations: CI, confidence interval; OR, odds ratio; PORT, postoperative radiation therapy; Ref, reference category.
Certain columns may not sum to the total due to missing observations.
In this database, 62.0% (n = 24 901) of patients began PORT beyond 42 days (6 weeks) of surgery. By 8 and 10 weeks, 31.7% and 15.9% of patients had not yet initiated PORT, respectively, with the complete breakdown presented in Table 2. Bivariate comparisons of demographics, clinical and surgical characteristics, and hospital characteristics of those treated with PORT within or beyond 6 weeks are presented in Tables 1, 3, and 4, respectively (as well as Tables S1–S3, which are the same tables presented with percentages out of all patients).
TABLE 2.
NCDB and TriNetX time to PORT breakdown
| Commencement of PORT | NCDB (n = 40 164) | NCDB cumulative frequency | TriNetX (n = 3165) | TriNetX cumulative frequency |
|---|---|---|---|---|
| ≤6 weeks | 15 263 (38.0%) | 15 263 (38.0%) | 1142 (36.1%) | 1142 (36.1%) |
| 6–8 weeks | 12 169 (30.3%) | 27 432 (68.3%) | 931 (29.4%) | 2073 (65.5%) |
| 8–10 weeks | 6327 (15.8%) | 33 759 (84.1%) | 491 (15.5%) | 2564 (81.0%) |
| 10–12 weeks | 2993 (7.5%) | 36 752 (91.5%) | 244 (7.7%) | 2808 (88.7%) |
| >12 weeks | 3412 (8.5%) | 40 164 (100%) | 357 (11.3%) | 3165 (100%) |
Abbreviations: NCDB, National Cancer Database; PORT, postoperative radiation therapy.
TABLE 3.
NCDB analysis of clinical and surgical characteristics
| Patient variable | Total patients (n = 40 164) | Initiation of PORT ≤6 weeks (n = 15 263) | Initiation of PORT >6 weeks (n = 24 901) | OR [95% CI] |
|---|---|---|---|---|
| Cancer primary site | ||||
| Oropharynx | 15 099 | 7198 (47.7%) | 7901 (52.3%) | 1.00 [Ref] |
| Oral cavity | 16 415 | 4297 (26.2%) | 12 118 (73.8%) | 2.56 [2.45–2.69] |
| Hypopharynx | 896 | 304 (33.9%) | 592 (66.1%) | 1.77 [1.53–2.04] |
| Larynx | 7754 | 3464 (44.7%) | 4290 (55.3%) | 1.12 [1.06–1.19] |
| AJCC clinical stage groupa | ||||
| I | 2719 | 1354 (49.8%) | 1365 (50.2%) | 1.00 [Ref] |
| II | 2766 | 985 (35.6%) | 1781 (64.4%) | 1.79 [1.61–1.99] |
| III | 3677 | 1457 (39.6%) | 2220 (60.4%) | 1.51 [1.36–1.67] |
| IV | 10 468 | 3886 (37.1%) | 6582 (62.9%) | 1.68 [1.54–1.82] |
| Unknown | 126 | 48 (38.1%) | 78 (61.9%) | 1.61 [1.11–2.32] |
| AJCC pathological stage groupa | ||||
| I | 1126 | 485 (43.1%) | 641 (56.9%) | 1.00 [Ref] |
| II | 1406 | 468 (33.3%) | 938 (66.7%) | 1.51 [1.29–1.78] |
| III | 3044 | 1083 (35.6%) | 1961 (64.4%) | 1.37 [1.19–1.57] |
| IV | 29 015 | 10 084 (34.8%) | 18 931 (65.2%) | 1.42 [1.25–1.60] |
| Surgical margin statusa | ||||
| Negative | 26 698 | 8873 (33.2%) | 17 825 (66.8%) | 1.00 [Ref] |
| Positive | 10 596 | 4811 (45.4%) | 5785 (54.6%) | 0.59 [0.57–0.62] |
| Unknown | 2870 | 1579 (55.0%) | 1291 (45.0%) | 0.40 [0.37–0.44] |
| Postoperative length of stay (days) | ||||
| 0–3 | 18 596 | 9205 (49.5%) | 9391 (50.5%) | 1.00 [Ref] |
| 4–7 | 7012 | 2273 (32.4%) | 4739 (67.6%) | 1.78 [1.67–1.90] |
| 8–14 | 7664 | 1928 (25.2%) | 5736 (74.8%) | 2.04 [1.92–2.16] |
| 15–21 | 1186 | 155 (13.1%) | 1031 (86.9%) | 6.51 [5.49–7.73] |
| >21 | 1042 | 132 (12.7%) | 910 (87.3%) | 6.75 [5.61–8.12] |
| Unknown | 4664 | 1570 (33.7%) | 3094 (66.3%) | 1.93 [1.80–2.06] |
| Readmission within 30 days of discharge | ||||
| None | 37 474 | 14 428 (38.5%) | 23 046 (61.5%) | 1.00 [Ref] |
| Unplanned | 1078 | 278 (25.8%) | 800 (74.2%) | 1.80 [1.56–2.06] |
| Planned | 726 | 269 (37.1%) | 457 (62.9%) | 1.06 [0.91–1.23] |
| Unknown | 886 | 288 (32.5%) | 598 (67.5%) | 1.30 [1.12–1.49] |
| Radiation modalitya | ||||
| External beam | 2005 | 1007 (50.2%) | 998 (49.8%) | 1.00 [Ref] |
| Conformal or 3D therapy | 11 774 | 4767 (40.5%) | 7007 (59.5%) | 0.67 [0.61–0.74] |
| IMRT | 25 843 | 9296 (36.0%) | 16 547 (64.0%) | 1.21 [1.15–1.26] |
| Concurrent chemoradiation | ||||
| No | 30 062 | 10 946 (36.4%) | 19 116 (63.6%) | 1.00 [Ref] |
| Yes | 10 102 | 4317 (42.7%) | 5785 (57.3%) | 0.76 [0.73–0.80] |
| Charlson-Deyo comorbidity score | ||||
| 0 | 29 888 | 11 641 (38.9%) | 18 247 (61.1%) | 1.00 [Ref] |
| 1 | 6765 | 2390 (35.3%) | 4375 (64.7%) | 1.16 [1.10–1.23] |
| ≥2 | 3511 | 1232 (35.1%) | 2279 (64.9%) | 1.18 [1.09–1.27] |
Abbreviations: AJCC, American Joint Committee on Cancer; CI, confidence interval; IMRT, intensity-modulated radiation therapy; OR, odds ratio; PORT, postoperative radiation therapy; Ref, reference category; 3D, 3-dimensional.
Certain columns may not sum to the total due to missing observations.
TABLE 4.
NCDB analysis of hospital characteristics
| Patient variable | Total patients (n = 40 164) | Initiation of PORT ≤6 weeks (n = 15 263) | Initiation of PORT >6 weeks (n = 24 901) | OR [95% CI] |
|---|---|---|---|---|
| Treatment facility typea | ||||
| Community Cancer Program | 1909 | 797 (41.7%) | 1112 (58.3%) | 1.00 [Ref] |
| Comprehensive Community Cancer Program | 10 436 | 4524 (43.3%) | 5912 (56.7%) | 0.93 [0.84–1.03] |
| Academic/Research Program | 20 557 | 6905 (33.6%) | 13 652 (66.4%) | 1.41 [1.28–1.55] |
| Integrated Network Cancer Program | 6165 | 2606 (42.3%) | 3559 (57.7%) | 0.97 [0.88–1.08] |
| Number of facilities involved in treatment | ||||
| All treatment at 1 CoC Facility | 30 255 | 11 823 (39.1%) | 18 432 (60.9%) | 1.00 [Ref] |
| Treatment at >1 CoC Facility | 9909 | 3440 (34.7%) | 6469 (65.3%) | 1.20 [1.15–1.26] |
| Surgery and radiation at same facility | ||||
| Yes | 20 312 | 7024 (34.6%) | 13 288 (65.4%) | 1.00 [Ref] |
| No | 19 852 | 8239 (41.5%) | 11 613 (58.5%) | 1.34 [1.28–1.39] |
| Region of carea | ||||
| Northeast | 8070 | 2702 (33.5%) | 5368 (66.5%) | 1.00 [Ref] |
| Midwest | 11 014 | 4604 (41.8%) | 6410 (48.2%) | 0.70 [0.66–0.74] |
| South | 14 025 | 5296 (37.8%) | 8729 (62.2%) | 0.83 [0.78–0.87] |
| West | 5958 | 2230 (37.4%) | 3728 (62.6%) | 0.84 [0.78–0.90] |
Abbreviations: CI, confidence interval; CoC, commission on cancer; OR, odds ratio; PORT, postoperative radiation therapy; Ref, reference category.
Certain columns may not sum to the total due to missing observations.
On multivariable analysis, significant demographic predictors of timeliness of PORT included (OR >1 indicative of delay) age 50–59 years (OR, 95% CI) (1.18, 1.09–1.29) and 60–69 years (1.12, 1.03–1.22) compared to <50 years, female sex (1.12, 1.06–1.19) compared to male, black race (1.28, 1.16–1.41) versus white, Medicaid (1.64, 1.50–1.79), Medicare (1.24, 1.17–1.32), other government insurance (1.56, 1.33–1.82), and uninsured status (1.46, 1.25–1.70) compared to private insurance, belonging to the lowest quartile of educational attainment (1.28, 1.19–1.39) compared to the highest, and distance to the treatment center >100 miles (0.74, 0.67–0.82) compared to ≤10 miles. Clinical and surgical factors that remained significant included oral cavity cancer site (1.70, 1.60–1.81) compared to oropharynx, positive surgical margin status (0.79, 0.74–0.83) compared to negative margins, postoperative length of stay 4–7 days (1.63, 1.52–1.75), 8–14 days (2.02, 1.88–2.18), 15–21 days (4.44, 3.65–5.41), and >21 days (4.69, 3.80–5.80) compared to 0–3 days, unplanned hospital readmission (1.53, 1.30–1.80) compared to no readmission, IMRT radiation modality (1.18, 1.12–1.25) compared to external beam, and concurrent chemoradiation (0.78, 0.74–0.83) versus not receiving it. Significant hospital characteristics included treatment at an academic center (1.18, 1.05–1.32) compared to community cancer program, surgery and radiation at the different facilities (1.50, 1.42–1.57) compared to same facility, and Midwest (0.74, 0.69–0.80) and South (0.82, 0.77–0.88) regions of care compared to Northeast. The specific subcategories associated with delay are presented with their respective effect sizes in Figure 1. Factors that were not associated with delay on multivariable analysis included ethnicity, urban/rural area of residence, household income, Charlson–Deyo comorbidity score, clinical and pathological stage, and number of CoC facilities involved in treatment.
FIGURE 1.
Predictors of PORT delay on multivariable analysis
Analysis of temporal trends indicated a gradual increase in the annual percentage of patients who failed to initiate PORT within 6 weeks from 2015 to 2019 (Figure 2).
FIGURE 2.
Temporal trends in the failure to initiate PORT within 6 weeks of surgery according to year of diagnosis
Analysis of risk factors for prolonged LOS and readmissions is presented in Tables 5 and 6 (and Tables S4 and S5). Among patients with oral cavity or hypopharynx tumor sites, 43.6% and 46.2% had postoperative LOS at least 8 days, compared to just 8.9% and 28.8% for oropharynx and larynx, respectively. Additionally, 3.2% and 4.2% of patients with oral cavity and hypopharynx tumor sites had unplanned hospital readmissions, compared to 2.1% and 2.3% of patients with oropharynx and larynx tumor sites, respectively. Additional patients more likely to experience a hospital stay ≥8 days included those of black race (42.9% vs. 25.9% of white patients), Hispanic ethnicity (35.5% vs. 27.5% of non-Hispanic patients), who were uninsured or had Medicaid (40.5% and 43.8%, respectively, compared to 21.4% for private insurance), who belonged to the lowest educational quartile (34.8% vs. 23.0% for highest educational quartile), who earned an income <$40227 (35% vs. 24.1% for income ≥$63333), who lived ≥100 miles from the hospital (34.4% vs. 25.9% for ≤10 miles), with a Charleson–Deyo comorbidity score of ≥2 (37.0% vs. 25.3% for score of 0), with clinical and pathological disease stage of 4 (34.1% and 32.8%, respectively, compared to 8.5% and 11.1% for clinical and pathological stage of 1), who had negative surgical margins (33.4% vs. 19.6% for positive margins), who had unplanned readmissions (43.7% vs. 27.4% for no readmission), and who underwent external beam or IMRT (26.1% and 29.8%, respectively vs. 14.3% for conformal or 3D therapy). Additional patients more likely to experience an unplanned hospital readmission included those of black race (3.8% vs. 2.6% for white), with Medicaid insurance (3.3% vs. 2.4% for private), who lived in rural settings (3.4% vs. 2.5% for urban), belonged to the lowest educational quartile (3.1% vs. 2.6% for highest quartile), who made <$40227 (3.3% vs. 2.6% for income ≥$63333), who lived >100 miles from the treatment center (3.2% vs. 2.6% for ≤10 miles), with a Charleson–Deyo comorbidity score of ≥2 (4.0% vs. 2.4% for a score of 0), with clinical or pathological disease stage of 4 (2.9% and 3.0%, respectively, vs. 1.2% and 1.5% for disease stage of 1), with negative surgical margins (3.0% vs. 2.5% for positive margins), with increased postoperative LOS (5.9% for LOS > 21 days vs. 1.6% for LOS 0–3 days), and who underwent IMRT (2.9% vs. 2.0% conformal or 3D therapy).
TABLE 5.
Risk factors for prolonged length of stay
| Postoperative length of stay (days) |
||||||
|---|---|---|---|---|---|---|
| Patient variable | 0–3 (n = 18 596) | 4–7 (n = 7012) | 8–14 (n = 7664) | 15–21 (n = 1186) | >21 (n = 1042) | p-value |
| Age | ||||||
| <50 | 2392 (51.4%) | 1006 (21.6%) | 1016 (21.8%) | 137 (2.9%) | 106 (2.3%) | <0.001a |
| 50–59 | 5381 (53.0%) | 2018 (19.9%) | 2091 (20.6%) | 336 (3.3%) | 321 (3.2%) | |
| 60–69 | 6062 (51.6%) | 2391 (20.3%) | 2567 (21.8%) | 396 (3.4%) | 340 (2.9%) | |
| >70 | 4761 (52.3%) | 1597 (17.9%) | 1990 (22.3%) | 317 (3.5%) | 275 (3.1%) | |
| Sex | ||||||
| Male | 14 291 (54.5%) | 5069 (19.3%) | 5295 (20.2%) | 817 (3.1%) | 739 (2.8%) | <0.001a |
| Female | 4305 (46.3%) | 1943 (20.9%) | 2369 (25.5%) | 369 (4.0%) | 303 (3.3%) | |
| Race | ||||||
| White | 16 856 (54.4%) | 6117 (19.7%) | 6243 (20.2%) | 931 (3.0%) | 826 (2.7%) | <0.001a |
| Black | 1072 (38.4%) | 523 (18.7%) | 878 (31.5%) | 176 (6.3%) | 142 (5.1%) | |
| Asian/Pacific Islander | 372 (34.7%) | 228 (21.3%) | 367 (34.2%) | 62 (5.8%) | 43 (4.0%) | |
| Other/unknown | 296 (44.6%) | 144 (21.7%) | 176 (26.5%) | 17 (2.6%) | 31 (4.7%) | |
| Ethnicity | ||||||
| Non-Hispanic | 17 538 (52.7%) | 6566 (19.7%) | 7095 (21.3%) | 1106 (3.3%) | 957 (2.9%) | <0.001a |
| Hispanic | 737 (43.8%) | 348 (20.7%) | 464 (27.6%) | 63 (3.7%) | 71 (4.2%) | |
| Unknown | 321 (57.8%) | 98 (17.7%) | 105 (18.9%) | 17 (3.1%) | 14 (2.5%) | |
| Insurance status | ||||||
| Not insured | 383 (37.4%) | 227 (22.1%) | 323 (31.5%) | 46 (4.5%) | 46 (4.5%) | <0.001a |
| Private insurance | 9049 (58.5%) | 3123 (20.2%) | 2700 (17.5%) | 334 (2.2%) | 259 (1.7%) | |
| Medicaid | 1424 (34.7%) | 875 (21.4%) | 1305 (31.8%) | 239 (5.8%) | 255 (6.2%) | |
| Medicare | 7126 (51.8%) | 2545 (18.5%) | 3118 (22.7%) | 518 (3.8%) | 450 (3.3%) | |
| Other government | 398 (53.1%) | 152 (20.3%) | 146 (19.5%) | 33 (4.4%) | 21 (2.8%) | |
| Insurance status unknown | 216 (53.3%) | 90 (22.2%) | 72 (1.8%) | 16 (4.0%) | 11 (2.7%) | |
| Urban/rural | ||||||
| Metropolitan | 15 006 (52.7%) | 5525 (19.4%) | 6115 (21.5%) | 974 (3.4%) | 853 (3.0%) | 0.01a |
| Urban | 2679 (52.1%) | 1063 (20.7%) | 1101 (21.4%) | 149 (2.9%) | 151 (2.9%) | |
| Rural | 289 (48.3%) | 137 (22.9%) | 139 (23.2%) | 16 (2.7%) | 17 (2.8%) | |
| Education | ||||||
| Lowest quartile | 2707 (45.9%) | 1131 (19.2%) | 1530 (25.9%) | 279 (4.7%) | 250 (4.2%) | <0.001a |
| Second lowest quartile | 4010 (50.6%) | 1595 (20.1%) | 1807 (22.8%) | 295 (3.7%) | 216 (2.7%) | |
| Second highest quartile | 4661 (53.9%) | 1671 (19.3%) | 1795 (20.8%) | 267 (3.1%) | 250 (2.9%) | |
| Highest quartile | 4260 (56.9%) | 1499 (20.0%) | 1391 (18.6%) | 182 (2.4%) | 149 (2.0%) | |
| Median household income | ||||||
| Less than $40227 | 2483 (45.5%) | 1067 (19.6%) | 1432 (26.3%) | 252 (4.6%) | 221 (4.1%) | <0.001a |
| $40227–$50353 | 3380 (50.9%) | 1358 (20.5%) | 1500 (22.6%) | 220 (3.3%) | 181 (2.7%) | |
| $50354–$63332 | 3564 (52.2%) | 1336 (19.6%) | 1479 (21.7%) | 241 (3.5%) | 203 (3.0%) | |
| $63333+ | 6176 (56.4%) | 2126 (19.4%) | 2096 (19.1%) | 307 (2.8%) | 255 (2.3%) | |
| Great circle distance | ||||||
| ≤10 | 9843 (57.2%) | 29,12 (16.9%) | 3367 (19.6%) | 569 (3.3%) | 511 (3.0%) | <0.001a |
| 11–20 | 3251 (56.0%) | 1049 (18.1%) | 1158 (20.0%) | 190 (3.3%) | 153 (2.6%) | |
| 21–50 | 3327 (48.9%) | 1472 (21.6%) | 1585 (23.3%) | 221 (3.2%) | 198 (2.9%) | |
| 51–100 | 1386 (40.1%) | 896 (25.9%) | 928 (26.9%) | 124 (3.6%) | 120 (3.5%) | |
| >100 | 789 (35.2%) | 683 (30.5%) | 626 (27.9%) | 82 (3.7%) | 60 (2.7%) | |
| Charlson–Deyo score | ||||||
| 0 | 14 406 (55.1%) | 5108 (19.5%) | 5207 (19.9%) | 767 (2.9%) | 666 (2.5%) | <0.001a |
| 1 | 2816 (45.8%) | 1260 (20.5%) | 1589 (25.9%) | 263 (4.3%) | 217 (3.5%) | |
| ≥2 | 1374 (42.9%) | 644 (20.1%) | 868 (27.1%) | 156 (4.9%) | 159 (5.0%) | |
| Primary site | ||||||
| Oral cavity | 4653 (31.8%) | 3611 (24.7%) | 4933 (33.7%) | 816 (5.6%) | 630 (4.3%) | <0.001a |
| Oropharynx | 9981 (75.2%) | 2099 (15.8%) | 908 (6.8%) | 133 (1.0%) | 148 (1.1%) | |
| Hypopharynx | 253 (32.3%) | 168 (21.4%) | 270 (34.4%) | 41 (5.2%) | 52 (6.6%) | |
| Larynx | 3709 (54.5%) | 1134 (16.7%) | 1553 (22.8%) | 196 (2.9%) | 212 (3.1%) | |
| AJCC clinical stage group | ||||||
| 1 | 1963 (81.6%) | 237 (9.9%) | 168 (7.0%) | 18 (0.7%) | 20 (0.8%) | <0.001a |
| 2 | 1307 (52.7%) | 553 (22.3%) | 500 (20.2%) | 64 (2.6%) | 57 (2.3%) | |
| 3 | 1692 (52.0%) | 752 (23.1%) | 651 (20.0%) | 88 (2.7%) | 72 (2.2%) | |
| 4 | 4041 (43.6%) | 2071 (22.4%) | 2405 (26.0%) | 387 (4.2%) | 358 (3.9%) | |
| Unknown | 55 (52.9%) | 20 (19.2%) | 20 (19.2%) | 6 (5.8%) | 3 (2.9%) | |
| AJCC pathological stage group | ||||||
| 1 | 753 (75.1%) | 139 (13.9%) | 89 (8.9%) | 7 (0.7%) | 15 (1.5%) | <0.001a |
| 2 | 582 (47.0%) | 315 (25.4%) | 273 (22.0%) | 39 (3.1%) | 30 (2.4%) | |
| 3 | 1291 (47.4%) | 649 (23.8%) | 639 (23.5%) | 83 (3.0%) | 60 (2.2%) | |
| 4 | 11 614 (45.1%) | 5697 (22.1%) | 6527 (25.3%) | 1032 (4.0%) | 907 (3.5%) | |
| Surgical margin status | ||||||
| Negative | 10 154 (42.5%) | 5761 (24.1%) | 6279 (26.3%) | 927 (3.9%) | 771 (3.2%) | <0.001a |
| Positive | 6251 (68.0%) | 1145 (12.5%) | 1300 (14.1%) | 245 (2.7%) | 253 (2.8%) | |
| Unknown | 2191 (90.8%) | 106 (4.4%) | 85 (3.5%) | 14 (0.6%) | 18 (0.7%) | |
| Readmission within 30 days of discharge | ||||||
| None | 17 683 (53.0%) | 6514 (19.5%) | 7137 (21.4%) | 1080 (3.2%) | 950 (2.8%) | <0.001a |
| Unplanned | 306 (28.8%) | 291 (27.4%) | 326 (30.7%) | 77 (7.3%) | 61 (5.7%) | |
| Planned | 362 (51.6%) | 146 (20.8%) | 149 (21.2%) | 24 (3.4%) | 21 (3.0%) | |
| Unknown | 245 (65.7%) | 61 (16.4%) | 52 (13.9%) | 5 (1.3%) | 10 (2.7%) | |
| Radiation modality | ||||||
| Conformal or 3D therapy | 1302 (74.7%) | 193 (11.1%) | 185 (10.6%) | 35 (2.0%) | 29 (1.7%) | <0.001a |
| External beam | 5679 (54.4%) | 2032 (19.5%) | 2133 (20.4%) | 333 (3.2%) | 265 (2.5%) | |
| IMRT | 11 314 (49.5%) | 4715 (20.6%) | 5269 (23.1%) | 805 (3.5%) | 734 (3.2%) | |
| Concurrent chemoradiation | ||||||
| No | 13 834 (52.0%) | 5320 (20.0%) | 5760 (21.7%) | 899 (3.4%) | 788 (3.0%) | 0.14a |
| Yes | 4762 (53.5%) | 1692 (19.0%) | 1904 (21.4%) | 287 (3.2%) | 254 (2.9%) | |
Abbreviations: AJCC, American Joint Committee on Cancer; IMRT, intensity-modulated radiation therapy; 3D, 3-dimensional.
Chi-square p-value.
TABLE 6.
Risk factors for hospital readmission
| Readmission within 30 days of discharge |
|||||
|---|---|---|---|---|---|
| Patient variable | None (n = 37 474) | Unplanned (n = 1078) | Planned (n = 726) | Unknown (n = 886) | p-value |
| Age | |||||
| <50 | 4870 (92.6%) | 164 (3.1%) | 106 (2.0%) | 118 (2.2%) | 0.14a |
| 50–59 | 10 656 (93.3%) | 320 (2.8%) | 210 (1.8%) | 238 (2.1%) | |
| 60–69 | 12 494 (93.3%) | 343 (2.6%) | 253 (1.9%) | 307 (2.3%) | |
| >70 | 9454 (93.7%) | 251 (2.5%) | 157 (1.6%) | 223 (2.2%) | |
| Sex | |||||
| Male | 27 766 (93.4%) | 789 (2.7%) | 533 (1.8%) | 651 (2.2%) | 0.86a |
| Female | 9708 (93.1%) | 289 (2.8%) | 193 (1.9%) | 235 (2.3%) | |
| Race | |||||
| White | 32 794 (93.5%) | 911 (2.6%) | 612 (1.7%) | 763 (2.2%) | <0.001a |
| Black | 2878 (91.9%) | 118 (3.8%) | 65 (2.1%) | 69 (2.2%) | |
| Asian/Pacific Islander | 1105 (92.6%) | 30 (2.5%) | 31 (2.6%) | 27 (2.3%) | |
| Other/unknown | 697 (91.6%) | 19 (2.5%) | 18 (2.4%) | 27 (3.5%) | |
| Ethnicity | |||||
| Non-Hispanic | 35 092 (93.3%) | 1028 (2.7%) | 690 (1.8%) | 819 (2.2%) | 0.10a |
| Hispanic | 1751 (94.0%) | 36 (1.9%) | 28 (1.5%) | 47 (2.5%) | |
| Unknown | 631 (93.8%) | 14 (2.1%) | 8 (1.2%) | 20 (3.0%) | |
| Insurance status | |||||
| Not insured | 1067 (93.6%) | 27 (2.4%) | 29 (2.5%) | 17 (1.5%) | <0.001a |
| Private insurance | 16 332 (93.8%) | 412 (2.4%) | 323 (1.9%) | 351 (2.0%) | |
| Medicaid | 4257 (92.7%) | 152 (3.3%) | 77 (1.7%) | 105 (2.3%) | |
| Medicare | 14 499 (93.2%) | 453 (2.9%) | 259 (1.7%) | 340 (2.2%) | |
| Other government | 895 (91.0%) | 23 (2.3%) | 17 (1.7%) | 48 (4.9%) | |
| Insurance status unknown | 424 (88.1%) | 11 (2.3%) | 21 (4.4%) | 25 (5.2%) | |
| Urban/rural | |||||
| Metropolitan | 30 038 (93.3%) | 880 (2.7%) | 607 (1.9%) | 662 (2.1%) | <0.001a |
| Urban | 5475 (92.8%) | 148 (2.5%) | 100 (1.7%) | 176 (3.0%) | |
| Rural | 633 (92.3%) | 23 (3.4%) | 8 (1.2%) | 22 (3.2%) | |
| Education | |||||
| Lowest quartile | 6213 (92.6%) | 209 (3.1%) | 125 (1.9%) | 164 (2.4%) | 0.01a |
| Second lowest quartile | 8451 (93.3%) | 238 (2.6%) | 148 (1.6%) | 220 (2.4%) | |
| Second highest quartile | 9183 (93.5%) | 239 (2.4%) | 181 (1.8%) | 215 (2.2%) | |
| Highest quartile | 7855 (93.7%) | 222 (2.6%) | 161 (1.9%) | 148 (1.8%) | |
| Median household income | |||||
| Less than $40227 | 5734 (92.1%) | 206 (3.3%) | 113 (1.8%) | 172 (2.8%) | 0.001a |
| $40227–$50353 | 7181 (93.7%) | 183 (2.4%) | 130 (1.7%) | 166 (2.2%) | |
| $50354–$63332 | 7264 (93.7%) | 195 (2.5%) | 134 (1.7%) | 157 (2.0%) | |
| $63333+ | 11 450 (93.4%) | 322 (2.6%) | 238 (1.9%) | 250 (2.0%) | |
| Great circle distance | |||||
| ≤10 | 18 298 (92.8%) | 519 (2.6%) | 362 (1.8%) | 545 (2.8%) | <0.001a |
| 11–20 | 6237 (93.5%) | 161 (2.4%) | 120 (1.8%) | 151 (2.3%) | |
| 21–50 | 7191 (94.1%) | 214 (2.8%) | 118 (1.5%) | 117 (1.5%) | |
| 51–100 | 3458 (93.9%) | 106 (2.9%) | 80 (2.2%) | 39 (1.1%) | |
| >100 | 2290 (93.5%) | 78 (3.2%) | 46 (1.9%) | 34 (1.4%) | |
| Charlson–Deyo score | |||||
| 0 | 27 911 (93.4%) | 719 (2.4%) | 529 (1.8%) | 729 (2.4%) | <0.001a |
| 1 | 6302 (93.2%) | 220 (3.3%) | 129 (1.9%) | 114 (1.7%) | |
| ≥2 | 3261 (92.9%) | 139 (4.0%) | 68 (1.9%) | 43 (1.2%) | |
| Primary site | |||||
| Oral cavity | 15 227 (92.8%) | 533 (3.2%) | 294 (1.8%) | 361 (2.2%) | <0.001a |
| Oropharynx | 14 154 (93.7%) | 323 (2.1%) | 279 (1.8%) | 343 (2.3%) | |
| Hypopharynx | 812 (90.6%) | 38 (4.2%) | 17 (1.9%) | 29 (3.2%) | |
| Larynx | 7281 (93.9%) | 184 (2.3%) | 136 (1.7%) | 153 (2.0%) | |
| AJCC clinical stage group | |||||
| 1 | 2594 (95.4%) | 32 (1.2%) | 48 (1.8%) | 45 (1.7%) | <0.001a |
| 2 | 2610 (94.4%) | 63 (2.3%) | 51 (1.8%) | 42 (1.5%) | |
| 3 | 3444 (93.7%) | 99 (2.7%) | 62 (1.7%) | 72 (2.0%) | |
| 4 | 9733 (93.0%) | 308 (2.9%) | 216 (2.1%) | 211 (2.0%) | |
| Unknown | 117 (92.9%) | 3 (2.4%) | 4 (3.2%) | 2 (1.6%) | |
| AJCC pathological stage group | |||||
| 0 | 110 (96.5%) | 1 (0.9%) | 2 (1.8%) | 1 (0.9%) | 0.02a |
| 1 | 1072 (95.2%) | 17 (1.5%) | 18 (1.6%) | 19 (1.7%) | |
| 2 | 1335 (95.0%) | 30 (2.1%) | 20 (1.4%) | 21 (1.5%) | |
| 3 | 2819 (92.6%) | 84 (2.8%) | 62 (2.0%) | 79 (2.6%) | |
| 4 | 26 946 (92.9%) | 866 (3.0%) | 547 (1.9%) | 656 (2.3%) | |
| Surgical margin status | |||||
| Negative | 24 882 (93.2%) | 790 (3.0%) | 506 (1.9%) | 520 (1.9%) | <0.001a |
| Positive | 9904 (93.5%) | 261 (2.5%) | 184 (1.7%) | 247 (2.3%) | |
| Unknown | 2688 (93.7%) | 27 (0.9%) | 36 (1.3%) | 119 (4.1%) | |
| Postoperative length of stay (days) | |||||
| 0–3 | 17 683 (95.1%) | 306 (1.6%) | 362 (1.9%) | 245 (1.3%) | <0.001a |
| 4–7 | 6514 (92.9%) | 291 (4.2%) | 146 (2.1%) | 61 (0.9%) | |
| 8–14 | 7137 (93.1%) | 326 (4.3%) | 149 (1.9%) | 52 (0.7%) | |
| 15–21 | 1080 (91.1%) | 77 (6.5%) | 24 (2.0%) | 5 (0.4%) | |
| >21 | 950 (91.2%) | 61 (5.9%) | 21 (2.0%) | 10 (1.0%) | |
| Radiation modality | |||||
| Conformal or 3D therapy | 1902 (94.9%) | 40 (2.0%) | 24 (1.2%) | 39 (1.9%) | 0.003a |
| External beam | 11 044 (93.8%) | 291 (2.5%) | 200 (1.7%) | 239 (2.0%) | |
| IMRT | 24 007 (92.9%) | 743 (2.9%) | 494 (1.9%) | 599 (2.3%) | |
| Concurrent chemoradiation | |||||
| No | 28 048 (93.3%) | 825 (2.7%) | 534 (1.8%) | 655 (2.2%) | 0.45a |
| Yes | 9426 (93.3%) | 253 (2.5%) | 192 (1.9%) | 231 (2.3%) | |
Abbreviations: AJCC, American Joint Committee on Cancer; IMRT, intensity-modulated radiation therapy; 3D, 3-dimensional.
Chi-square p-value.
5.2 |. TriNetX analysis
In TriNetX, there were 3651 patients with head and neck cancer who were treated with surgery and PORT during 2015 through 2021. Of this cohort, 486 patients were excluded for the following reasons: received stereotactic radiation or brachytherapy (n = 14), received induction chemotherapy (n = 35), and underwent surgery beyond 6 months after initial diagnosis (n = 115) or PORT beyond 6 months of surgery (n = 322). This left 3165 patients who met criteria for analysis. Demographics are described in Tables 7 and S6.
TABLE 7.
TriNetX analysis of demographic characteristics
| Patient variable | Total patients (n = 3165) | Initiation of PORT ≤6 weeks (n = 1142) | Initiation of PORT >6 weeks (n = 2023) | OR [95% CI] |
|---|---|---|---|---|
| Age | ||||
| <65 years | 1570 | 564 (35.9%) | 1006 (64.1%) | 1 [Ref] |
| ≥65 years | 1595 | 578 (36.2%) | 1017 (63.8%) | 0.99 [0.85–1.01] |
| Sex | ||||
| Male | 2331 | 874 (37.5%) | 1457 (62.5%) | 1 [Ref] |
| Female | 834 | 268 (32.1%) | 566 (67.9%) | 1.27 [1.08–1.49] |
| Race | ||||
| White | 2554 | 937 (36.7%) | 1617 (63.3%) | 1 [Ref] |
| Black | 387 | 131 (33.9%) | 256 (66.1%) | 1.14 [0.91–1.43] |
| Asian | 57 | 21 (36.8%) | 36 (63.2%) | 0.99 [0.58–1.72] |
| Other/unknown | 167 | 53 (31.7%) | 114 (68.3%) | 1.25 [0.89–1.75] |
| Ethnicity | ||||
| Non-Hispanic or Latino | 2622 | 969 (37.0%) | 1653 (63.0%) | 1 [Ref] |
| Hispanic or Latino | 179 | 62 (34.6%) | 117 (65.4%) | 1.11 [0.81–1.52] |
| Other/unknown | 364 | 111 (30.5%) | 253 (69.5%) | 1.33 [1.05–1.69] |
| Marital status | ||||
| Married | 966 | 422 (43.7%) | 544 (56.3%) | 1 [Ref] |
| Never married | 418 | 142 (34.0%) | 276 (66.0%) | 1.51 [1.19–1.92] |
| Divorced | 194 | 69 (35.6%) | 125 (64.4%) | 1.41 [1.02–1.92] |
| Widowed | 133 | 46 (34.6%) | 87 (65.4%) | 1.47 [1.01–2.13] |
| Other/unknown | 1454 | 463 (31.8%) | 991 (68.2%) | 1.67 [1.41–1.96] |
| Region of care | ||||
| East | 2041 | 731 (35.8%) | 1310 (64.2%) | 1 [Ref] |
| West | 1112 | 408 (36.7%) | 704 (63.3%) | 0.96 [0.83–1.12] |
| Unknown | 12 | 3 (25.0%) | 9 (75.0%) | 1.67 [0.45–6.25] |
Abbreviations: CI, confidence interval; OR, odds ratio; PORT, postoperative radiation therapy; Ref, reference category.
Similar to NCDB, 63.9% (n = 2023) of patients in TriNetX initiated PORT greater than 6 weeks after surgery. By 8 and 10 weeks, 35.5% and 19.0%, of patients still had not initiated PORT, respectively (Table 2). Additional unique predictors of delay (OR, 95% CI) included never married (1.51, 1.19–1.92), divorced (1.41, 1.02–1.92), or widowed (1.47, 1.01–2.13) marital status compared to married, major surgical procedures including neck dissection (1.55, 1.34–1.79), laryngectomy (2.01, 1.54–2.64), and free osteocutaneous (1.68, 1.30–2.18), myocutaneous (2.59, 1.72–3.91), or fascial flaps (2.27, 1.32–2.88) compared to those who did not undergo these operations, and gastrostomy (1.30, 1.11–1.51) or tracheostomy (1.67, 1.41–1.96) dependence compared to those without (Tables 7 and 8 and S6 and S7).
TABLE 8.
TriNetX analysis of clinical and surgical characteristics
| Patient variable | Total patients (n = 3165) | Initiation of PORT ≤6 weeks (n = 1142) | Initiation of PORT >6 weeks (n = 2023) | OR [95% CI] |
|---|---|---|---|---|
| Cancer site | ||||
| Oropharynx | 1116 | 485 (43.5%) | 631 (56.5%) | 1 [Ref] |
| Oral cavity | 1376 | 412 (29.9%) | 964 (70.1%) | 1.79 [1.52–2.12] |
| Hypopharynx | 98 | 46 (46.9%) | 52 (53.1%) | 0.87 [0.57–1.31] |
| Larynx | 575 | 199 (34.6%) | 376 (65.4%) | 1.45 [1.18–1.79] |
| Concurrent chemoradiation | ||||
| No | 1947 | 660 (33.9%) | 1287 (66.1%) | 1 [Ref] |
| Yes | 1218 | 482 (39.6%) | 736 (60.4%) | 0.78 [0.68–0.91] |
| Neck dissection | ||||
| No | 1387 | 579 (41.7%) | 808 (58.3%) | 1 [Ref] |
| Yes | 1778 | 563 (31.7%) | 1215 (68.3%) | 1.55 [1.34–1.79] |
| Free skin flap | ||||
| No | 2755 | 997 (36.2%) | 1758 (63.8%) | 1 [Ref] |
| Yes | 410 | 145 (35.4%) | 265 (64.6%) | 1.04 [0.83–1.29] |
| Free osteocutaneous flap | ||||
| No | 2839 | 1057 (37.2%) | 1782 (62.8%) | 1 [Ref] |
| Yes | 326 | 85 (26.1%) | 241 (73.9%) | 1.68 [1.30–2.18] |
| Free muscle or myocutaneous flap | ||||
| No | 3008 | 1113 (37.0%) | 1895 (63.0%) | 1 [Ref] |
| Yes | 157 | 29 (18.5%) | 128 (81.5%) | 2.59 [1.72–3.91] |
| Free fascial flap | ||||
| No | 3081 | 1125 (37.0%) | 1956 (63.0%) | 1 [Ref] |
| Yes | 84 | 17 (20.2%) | 67 (79.8%) | 2.27 [1.32–3.88] |
| Laryngectomy | ||||
| No | 2839 | 1067 (37.6%) | 1772 (62.4%) | 1 [Ref] |
| Yes | 326 | 75 (23.0%) | 251 (77.0%) | 2.01 [1.54–2.64] |
| Gastrostomy status | ||||
| No | 2030 | 776 (38.2%) | 1254 (61.8%) | 1 [Ref] |
| Yes | 1135 | 366 (32.2%) | 769 (67.8%) | 1.30 [1.11–1.51] |
| Tracheostomy status | ||||
| No | 2234 | 880 (39.4%) | 1354 (60.6%) | 1 [Ref] |
| Yes | 931 | 262 (28.1%) | 669 (71.9%) | 1.67 [1.41–1.96] |
Abbreviations: CI, confidence interval; OR, odds ratio; PORT, postoperative radiation therapy; Ref, reference category.
6 |. DISCUSSION
Substantial strides in cancer care resulted from the advent of novel treatments, adoption of the multidisciplinary approach to patient-centered care, and implementation of evidence-based practice guidelines. However, variability in adherence to guidelines results in suboptimal outcomes across various malignancies.10 In the setting of HNSCC, initiating PORT within 6 weeks of surgical resection is associated with improved locoregional control and overall survival.11–17 Graboyes et al. reported that time to PORT deviated from this guideline for approximately half of this patient cohort during 2006–2014.3 Our updated analysis demonstrated a persistence of this trend, with nearly two-thirds of patients initiating PORT beyond the 6-week timepoint.
In our study, a variety of demographic, clinical, surgical, and hospital factors predicted treatment nonadherence, with many variables sharing commonality with the prior NCDB analysis.3 Persistent associations with treatment delay included black race, public insurance, lower educational attainment, oral cavity tumor site, increased postoperative length of stay or unplanned readmissions, treatment at an academic center, and fragmented surgery and PORT at different locations. The parallels of our findings 5 years after the initial report underscores inadequate interventions to rectify this disparity despite expanding evidence dedicated to understanding patient and system-level barriers to timely PORT.3–6 Additionally, temporal trends are suggestive of worsening adherence over time. Variables with the strongest associations with PORT delay included increased length of hospitalization, unplanned readmission, oral cavity cancer site, and insurance status. When designing interventions, targeting these areas may provide the greatest opportunity for improvement.
There are reports of a few isolated but promising quality improvement endeavors resulting in reduced PORT delay. One institution reported that the percentage of free flap patients receiving timely PORT increased from 10.5% to 50% due to three interventions: standardized care-pathway order sets, prompt referrals to radiation oncology, and most importantly, assigning a patient navigator to coordinate care and serve as the liaison between patients, physicians, and schedulers.18 Another project reduced preventable delays from 24% to 9% and improved overall timeliness of care from 62% to 73% by initiating dental evaluation at the time of the new patient visit with scheduled extraction during surgery, placing radiation oncology referrals at the time of identified indication for PORT, setting automated reminders to review surgical pathology if there was a clinical concern for PORT, and creating a standardized head and neck oncology clinic visit checklist to recap key aspects of the care timeline.19 The most successful intervention to date, Navigation for Disparities and Untimely Radiation thErapy (NDURE), involves a series of three appointments with a patient navigator throughout the treatment course to coordinate care, provide patient education, and identify and resolve potential modifiable barriers to PORT. This resulted in 86% of patients receiving PORT within 6 weeks; however, it should be noted that there were only 15 patients in total in the study.20 Additionally, nomograms and machine learning algorithms are available to identify patients at highest risk for PORT delay.21,22
Prolonged length of stay and unplanned hospital readmissions are often due to surgical complications. Therefore, efforts to minimize complications via emphasis on measures such as adequate preoperative nutrition, medical optimization, and thromboprophylaxis may be useful. Importantly, multidisciplinary communication is critical in advance of surgery. Patients should either see their radiation oncologist preoperatively or have a postoperative appointment scheduled in advance. This may help avoid last minute care coordination and delays if a patient desires surgical and radiation care at different facilities for example, a variable that was significantly associated with delay in the current study. It is essential that institutions managing HNSCC patients evaluate and implement similar interventions to improve adherence to national guidelines.
A distinct advantage of TriNetX is analysis of additional factors that were not studied previously. Specifically, patients who underwent a neck dissection, certain free flaps, or laryngectomy were more likely to experience PORT delay. This also applied to those with a tracheostomy or gastrostomy tube. This suggests patients with more extensive surgeries or in need of additional ancillary services are at increased risk for PORT delay. Additionally, patients in TriNetX who were never married, divorced, or widowed were significantly more likely to receive PORT beyond 6 weeks. A possible explanation for this could be that partners were likely to provide tangible services and informal caregiving, such as transportation and care coordination that resulted in increased adherence to guidelines. The improved timeliness of treatment in the married cohort may also be extrapolated to any unmarried patients with caring partners who will provide various levels of support after surgery. Additionally, other family members, friends, neighbors, or patient navigators may provide a similar positive influence. Initiatives aimed at providing these conveniences may be worth pursuing in patients lacking support.
Now that time to PORT is officially a CoC national quality metric,7 head and neck oncologists and other members of the multidisciplinary care team will prioritize similar efforts to improve adherence. Our results provide a performance baseline, with future research necessary to continue trending outcomes moving forward. While our study provides a recent update regarding adherence to treatment guidelines and is the first to identify certain factors as predictive of delay, it is not without limitations. First, it is limited by the quality of data entry in the NCDB and TriNetX, which is a recognized weakness in performing a retrospective review and using a large de-identified database. However, the similar rate of treatment delay in two databases increased our confidence in our findings, and the large sample size of patients strengthened the generalizability. Furthermore, TriNetX was only able to capture those patients who received both surgery and radiation within the same HCO, which likely excluded a significant number of patients, but these patients with fragmented care were able to be included in the analysis of the NCDB. It is possible that there could be some patient overlap between databases, however the second database was useful in that it allowed for the presentation of more granular data. A limitation affecting our analysis of both databases is that we were unable to determine whether patients were subject to circumstances that justifiably delayed adjuvant radiation, which may have inflated our reported rate of delay. Components that are recognized as leading to prolonged time to PORT, such as delays in dental evaluation, necessary treatment preradiation, and transportation issues, were also not able to be captured through these databases.
In conclusion, timely initiation of PORT continues to be a challenge for nearly two-thirds of patients with HNSCC. With the introduction of a head and neck specific CoC quality measure evaluating this benchmark, we anticipate renewed institutional focus on vulnerable patients and care processes associated with delay. A lack of family support, more complex surgical care, and demographic factors associated with poorer access to care continue to be risk factors for delayed adjuvant radiation.
Supplementary Material
ACKNOWLEDGMENTS
The project described was supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant UL1 TR002014. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Footnotes
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
The research discussed was accepted for an oral presentation at the AAO-HNSF 2022 Annual Meeting and OTO experience, September 10–14, in Philadelphia, Pennsylvania.
ETHICS STATEMENT
This study was exempted by the Penn State Institutional Review Board review (STUDY00018629 for TriNetX and STUDY00018678 for NCDB).
SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.
DATA AVAILABILITY STATEMENT
The study data is available from F. Jeffrey Lorenz on reasonable request.
REFERENCES
- 1.Institute of Medicine (US) Committee on Quality of Health Care in America. Crossing the Quality Chasm. National Academies Press; 2001. doi: 10.17226/10027 [DOI] [Google Scholar]
- 2.Pfister DG, Spencer S, Adelstein D, et al. Head and neck cancers, version 2.2020, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2020;18(7):873–898. doi: 10.6004/jnccn.2020.0031 [DOI] [PubMed] [Google Scholar]
- 3.Graboyes EM, Garrett-Mayer E, Sharma AK, Lentsch EJ, Day TA. Adherence to National Comprehensive Cancer Network guidelines for time to initiation of postoperative radiation therapy for patients with head and neck cancer. Cancer. 2017;123(14):2651–2660. doi: 10.1002/cncr.30651 [DOI] [PubMed] [Google Scholar]
- 4.Noyes EA, Burks CA, Larson AR, Deschler DG. An equity-based narrative review of barriers to timely postoperative radiation therapy for patients with head and neck squamous cell carcinoma. Laryngosc Investig Otolaryngol. 2021;6(6):1358–1366. doi: 10.1002/lio2.692 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Janz TA, Kim J, Hill EG, et al. Association of care processes with timely, equitable postoperative radiotherapy in patients with surgically treated head and neck squamous cell carcinoma. JAMA Otolaryngol Neck Surg. 2018;144(12):1105–1114. doi: 10.1001/jamaoto.2018.2225 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Graboyes EM, Halbert CH, Li H, et al. Barriers to the delivery of timely, guideline-adherent adjuvant therapy among patients with head and neck cancer. JCO Oncol Pract. 2020;16(12):e1417–e1432. doi: 10.1200/OP.20.00271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Graboyes EM, Divi V, Moore BA. Head and neck oncology is on the national quality sidelines no longer—put me in coach. JAMA Otolaryngol Neck Surg. 2022;148(8):715–716. doi: 10.1001/jamaoto.2022.1389 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Ann Surg Oncol. 2008;15(3):683–690. doi: 10.1245/s10434-007-9747-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Topaloglu U, Palchuk MB. Using a federated network of real-world data to optimize clinical trials operations. JCO Clin Cancer Informatics. 2018;2:1–10. doi: 10.1200/CCI.17.00067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Institute of Medicine (US) and National Research Council (US) National Cancer Policy Board. Ensuring Quality Cancer Care. National Academies Press; 1999. doi: 10.17226/6467 [DOI] [PubMed] [Google Scholar]
- 11.Graboyes EM, Garrett-Mayer E, Ellis MA, et al. Effect of time to initiation of postoperative radiation therapy on survival in surgically managed head and neck cancer. Cancer. 2017;123(24):4841–4850. doi: 10.1002/cncr.30939 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ang KK, Trotti A, Brown BW, et al. Randomized trial addressing risk features and time factors of surgery plus radiotherapy in advanced head-and-neck cancer. Int J Radiat Oncol. 2001;51(3):571–578. doi: 10.1016/S0360-3016(01)01690-X [DOI] [PubMed] [Google Scholar]
- 13.Harris JP, Chen MM, Orosco RK, Sirjani D, Divi V, Hara W. Association of survival with shorter time to radiation therapy after surgery for US patients with head and neck cancer. JAMA Otolaryngol Neck Surg. 2018;144(4):349–359. doi: 10.1001/jamaoto.2017.3406 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Huang J, Barbera L, Brouwers M, Browman G, Mackillop WJ. Does delay in starting treatment affect the outcomes of radiotherapy? A systematic review. J Clin Oncol. 2003;21(3):555–563. doi: 10.1200/JCO.2003.04.171 [DOI] [PubMed] [Google Scholar]
- 15.Ho AS, Kim S, Tighiouart M, et al. Quantitative survival impact of composite treatment delays in head and neck cancer. Cancer. 2018;124(15):3154–3162. doi: 10.1002/cncr.31533 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tam M, Wu SP, Gerber NK, et al. The impact of adjuvant chemoradiotherapy timing on survival of head and neck cancers. Laryngoscope. 2018;128(10):2326–2332. doi: 10.1002/lary.27152 [DOI] [PubMed] [Google Scholar]
- 17.Cramer JD, Speedy SE, Ferris RL, Rademaker AW, Patel UA, Samant S. National evaluation of multidisciplinary quality metrics for head and neck cancer. Cancer. 2017;123(22):4372–4381. doi: 10.1002/cncr.30902 [DOI] [PubMed] [Google Scholar]
- 18.Voora RS, Stramiello JA, Sumner WA, et al. Quality improvement intervention to reduce time to postoperative radiation in head and neck free flap patients. Head Neck. 2021;43(11):3530–3539. doi: 10.1002/hed.26852 [DOI] [PubMed] [Google Scholar]
- 19.Divi V, Chen MM, Hara W, et al. Reducing the time from surgery to adjuvant radiation therapy: an institutional quality improvement project. Otolaryngol Neck Surg. 2018;159(1):158–165. doi: 10.1177/0194599818768254 [DOI] [PubMed] [Google Scholar]
- 20.Graboyes EM, Sterba KR, Li H, et al. Development and evaluation of a navigation-based, multilevel intervention to improve the delivery of timely, guideline-adherent adjuvant therapy for patients with head and neck cancer. JCO Oncol Pract. 2021;17(10):e1512–e1523. doi: 10.1200/OP.20.00943 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Levy DA, Li H, Sterba KR, et al. Development and validation of nomograms for predicting delayed postoperative radiotherapy initiation in head and neck squamous cell carcinoma. JAMA Otolaryngol Neck Surg. 2020;146(5):455–464. doi: 10.1001/jamaoto.2020.0222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Shew M, New J, Bur AM. Machine learning to predict delays in adjuvant radiation following surgery for head and neck cancer. Otolaryngol Neck Surg. 2019;160(6):1058–1064. doi: 10.1177/0194599818823200 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The study data is available from F. Jeffrey Lorenz on reasonable request.