Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Nov 14.
Published before final editing as: JCO Oncol Pract. 2025 Oct 9:OP2400751. doi: 10.1200/OP-24-00751

Comparing Trends in Telemedicine Use and Geographic Disparities in Utilization among Patients with Cancer in the United States

Matthew Mackwood 1,2,3, Qianfei Wang 2, Tor D Tosteson 1,2,4,5, Gabriel Brooks 2,3,5,6, Rebecca E Smith 1,2, Anna N A Tosteson 1,2,5,6
PMCID: PMC12614432  NIHMSID: NIHMS2109853  PMID: 41066715

Abstract

Purpose:

Studies have shown that rural populations were less likely than urban populations to use telemedicine during and after the COVID-19 pandemic. These trends are not well characterized nationally for patients with cancer.

Patients and Methods:

This cohort analysis, using two years of Medicare fee-for-service claims (2020-2021), identified patients with incident or prevalent lung or colon cancer in 2020. Rurality of patient residence was classified using rural-urban commuting area codes as metropolitan, micropolitan, or small town/rural. We used generalized estimating equations to model outpatient (telemedicine and in-person) visit trends in 2020 and 2021 by rurality, adjusting for local COVID-19 rates, US region, and clinical and demographic factors.

Results:

We identified 355,868 patients (66% lung, 34% colon). Median age was 76 (SD 7). 78% of patients lived in metropolitan areas, 85% were white, and 53% female. 44% had at least one instance of chemotherapy, surgery, or radiation treatment in 2020. In 2020, compared to patients residing in metropolitan areas, small town/rural-residing patients were less likely to use telemedicine (1.28 visits per year [95% CI:1.27 to 1.29] versus 2.34 [95%CI:2.33 to 2.35] for metropolitan patients), had fewer total outpatient in-person visits (12.37 [95%CI:12.34 to 12.41] versus 13.71 [95%CI:13.70 to 13.72]), and had more emergency department visits (0.85 [95%CI:0.84 to 0.86] versus 0.48 [95%CI: 0.48 to 0.48]) while inpatient utilization was similar. Similar trends were seen in 2021 and by cancer type.

Conclusion:

Lower rural telemedicine use continued beyond the pandemic onset. Relatively lower in-person and higher emergency department use suggests that telemedicine expansion did not improve overall access to care for rural cancer patients.

Introduction

The outbreak of the 2019 novel coronavirus (COVID-19) pandemic in the United States (US) led to an unprecedented and rapid increase in telemedicine visits in March 2020.1 This shift occurred in response to public health emergency measures to limit the spread of COVID-19 and was aided by policy changes that enabled a wide range of outpatient services to occur via telemedicine. These changes included payment parity for telemedicine and in-person care, and were implemented from March 2020 through at least September 20252,3 across all medical specialties, including oncology.4 Telehealth encompasses a broad scope of services5; this paper focuses on the subset provided remotely in real time between patient and provider utilizing audio (e.g., telephone) or audio-visual technology, hereafter referred to as “telemedicine”.

Previous work suggested wide variation in the use of telemedicine for cancer care during the early pandemic (estimates ranging from 24-72% of outpatient visits)4,6. Groups defined by urban residence, younger age, White non-Hispanic race/ethnicity, primary English speaking, and higher income generally used more telemedicine for cancer care6-11. Higher-acuity, more medically complex patients were noted as less likely to use telemedicine10,11. The extent to which these differences reflect the respective regional contexts of the studies, rather than broader-scale phenomena, and how “late pandemic” trends might differ from early-pandemic ones, remains poorly described.

Understanding these trends is especially important for rural populations, who are historically underserved for cancer care12, less likely to have reliable broadband internet access13, and for whom telemedicine has the highest potential impact in ameliorating burdens of travel costs and time14. Understanding is likewise important for older adults given their relatively higher incidence of cancer15 and lower telemedicine use rates to date. Furthermore, payment parity for telemedicine was approved only temporarily by the Center for Medicare & Medicaid Services (CMS)16, and with various state-level policies eroding reimbursements for telemedicine17, its future role in cancer care remains uncertain.

To improve understanding of telemedicine utilization on a national scale for the Medicare population, with a focus on differences over time by rurality, we analyzed telemedicine use in the context of overall outpatient, emergency department (ED), and hospital utilization across 2020 to 2021 Medicare fee-for-service (FFS) insurance claims data. We hypothesized that previously described rural-urban disparities in telemedicine use7-11 would be observed nationally, with significant regional variation in telemedicine use.

Methods

Dataset and cohort definitions

We analyzed 2020-2021 Medicare FFS claims data via the CMS Virtual Research Data Center, focused on beneficiaries who were age 65 or older on January 1st, 2020 with continuous enrollment in Medicare Parts A and B during the two-year study period. We obtained patient demographics from the Master Beneficiary Summary File, and utilization from the MEDPAR, Carrier and Outpatient files. We used a previously described, claims-based algorithm to identify patients with incident or prevalent lung or colorectal cancer (version #4 described by Bronson et al.)18 at any point during 2020. This cohort was followed through December 31, 2021. Lung and colorectal cancer populations were chosen because they are identifiable via claims algorithms and among the surgically-treated cancers we have studied by rurality18,19.

Data definitions and measures

Data were analyzed using SAS (version 9.4). The primary outcome of interest was the proportion of outpatient visits provided via telemedicine, compared by rurality. In order to describe the full continuum of care for these patients, we included all outpatient visits (telemedicine and in-person/office) in the analysis regardless of whether they were explicitly coded for cancer, reasoning that the impact of cancer extends beyond visits where it was specifically billed. As secondary outcomes, we also included ED and hospital utilization to further contextualize differences in utilization by rurality. Rurality was classified at the zip code-level according to rural-urban commuting area (RUCA) codes (defined on the basis of population density and commuting patterns); we grouped codes by increasing rurality according to WWAMI rurality classification20 as metropolitan, micropolitan, and small town/rural. We used CPT codes to create indicators for receipt of cancer surgery, chemotherapy, or radiation in 2020. A count of Hierarchical Condition Category (HCC) codes were used as a proxy for patients’ medical complexity21. Dual-eligibility for Medicare and Medicaid, which is closely tied to low income (threshold varies by state), was used as the best available proxy for individual socio-economic status22. We included US Census region to explore regional variation. We also included patient demographics including age and race/ethnicity. Given known issues with the accuracy of Medicare race/ethnicity coding, to minimize invalid inferences we limited groupings for analysis to White, Black, Hispanic, and all others (consolidating the categories of unknown, Asian/Pacific Islander, American Indian/Alaska Native, and other), using Research Triangle Institute codes, which improved accuracy utilizing name and geographic data23. To examine potential relationships between care integration and telemedicine use, we also categorized patients based upon whether the preponderance of their primary care in 2020 was delivered in an integrated (i.e., owned by a health care system or multi-specialty group providing services beyond primary care) or independent practice, leveraging the IQVIA OneKey database24. To account for the association between COVID prevalence and telemedicine utilization25, we included a geographic measure of COVID prevalence at the hospital referral region level26,27.

Statistical Analysis

We estimated weekly rates of telemedicine and in-person outpatient visits, as well as ED and hospitalization rates, using generalized estimating equations.28 Equations included the aforementioned clinical and demographic characteristics as covariates, as well as an indicator for week to model the time trend. The correlation between repeated measurements within a subject was addressed by specifying a compound symmetry correlation structure. We modeled outcome trends overall, and then examined results by rurality. To examine hypotheses that regional differences and factors related to race might be differentially associated with rural telemedicine use patterns, we also repeated our analyses with interactions for region and race.

To enhance understanding of whether telemedicine use results in higher subsequent utilization compared to office visits, we also performed secondary analyses of the relationship between outpatient visit method and subsequent care utilization. We identified all office or telemedicine visits, excluded those with a visit in the preceding two weeks to avoid double counting, and then noted whether hospitalization, ED, office, or telemedicine visits occurred in the ensuing two weeks. We also explored the potential effect modification of telemedicine use on utilization by modeling ED visits with and without a patient-level indicator of telemedicine use. This study was approved by the Dartmouth College Committee for the Protections of Human Subjects. Analyses were conducted from fall 2022 through summer 2024.

Results

Patient characteristics

We identified 355,868 patients with claims data consistent with lung (241,118) or colorectal (125,788) cancer in 2020. Both groups had similar utilization trends, and subsequent data are presented for the combined cohort. Demographics are reported in Table 1 (combined) and Appendix Tables 1 and 2 (individual cohorts and by telemedicine use, respectively). Across the cohort, metropolitan patients were older (14% over age 85, versus 11% for both micropolitan and small town/rural), more likely to be female (54% versus 50% versus 49%), less likely to be white (84% versus 91% versus 92%), and less likely to have undergone active cancer treatment with chemotherapy, surgery or radiation in 2020 (57% versus 53% versus 52%). The large sample resulted in very small p-values (<0.001) for these comparisons. There were not substantial differences by rurality in the prevalence of colon versus lung cancer, the proportion of patients attributed to a health care system, or HCC count.

Table 1.

Demographics of Medicare Fee-for-Service Patients with Prevalent or Incident Lung or Colon Cancer in 2020, by Rurality

Characteristic Overall, No. (%) Metropolitan, No.
(%)		 Micropolitan, No.
(%)		 Small Town /
Rural, No. (%)		 P-value for
difference across
ruralities(a)
(N=355868) (n=277967) (n=40925) (n=36976)
Age
 65-74 167286 (47) 128528 (46) 20359 (50) 18399 (50) <0.001
 75-84 141771 (40) 111194 (40) 16040 (39) 14537 (39)
 85 or older 46811 (13) 38245 (14) 4526 (11) 4040 (11)
Race(b)
 White 303184 (85) 232238 (84) 37072 (91) 33874 (92) <0.001
 Black 23317 (7) 20153 (7) 1840 (4) 1324 (4)
 Hispanic 11715 (3) 10264 (4) 913 (2) 538 (1)
 Asian / Pacific Islander 8762 (2) 8417 (3) 238 (0.6) 107 (0.3)
 American Indian / Alaska Native 1538 (0.4) 604 (0.2) 324 (0.8) 610 (2)
 Other 7352 (2) 6291 (2) 538 (1) 523 (1)
Sex
 Female 187646 (53) 148803 (54) 20562 (50) 18282 (49) <0.001
 Male 168222 (47) 129164 (46) 20363 (50) 18694 (51)
Cancer Type
 Colon 120269 (34) 93558 (34) 13981 (34) 12731 (34) 0.003
 Lung 235599 (66) 184410 (66) 26944 (66) 24245 (66)
Cancer Treatment in 2020
 Chemotherapy 98130 (28) 74785 (27) 12118 (30) 11227 (30) <0.001
 Surgery 48934 (14) 37440 (13) 5905 (14) 5589 (15) <0.001
 Radiation 55784 (16) 42467 (15) 6961 (17) 6356 (17) <0.001
 None of the above 198953 (56) 158036 (57) 21745 (53) 19172 (52) <0.001
HCC Count(c)
 0 or 1 64900 (18) 50893 (18) 7224 (18) 6783 (18) <0.001
 2 64358 (18) 49855 (18) 7607 (19) 6896 (19)
 3 54789 (15) 42510 (15) 6474 (16) 5805 (16)
 4 or 5 76501 (21) 59253 (21) 9111 (22) 8137 (22)
 6 or more 95320 (27) 75456 (27) 10509 (26) 9355 (25)
Medicaid Eligibility in addition to Medicare
 Yes 36530 (10) 28281 (10) 4150 (10) 4099 (11) <0.001
 No 319338 (90) 249686 (90) 36775 (90) 32877 (89)
Received Primary Care in a Health System-associated Practice
 Yes 265890 (75) 207742 (75) 30560 (75) 27588 (75) 0.8532
 No 89978 (25) 70225 (25) 10365 (25) 9388 (25)
Open in a new tab
(a)

calculated using chi-squared, with P-values less than 0.05 considered statistically significant.

(b)

from RTI codes in Medicare Beneficiary files.

(c)

HCC = Hierarchical Condition Category, a measure of chronic / complex medical conditions where higher numbers indicate more conditions

Overall Telemedicine Visit Trends

Adjusted overall trends in telemedicine utilization are shown in Figure 1, with US national COVID-19 case rates superimposed. Initial trends showed a massive increase in telemedicine utilization at the onset of the COVID-19 pandemic to a peak of 43% of outpatient visits in the week of April 5th, subsiding after 3 months to a range between 11-12%, increasing in late 2020/early 2021 to 13-14%, then diminishing to 6-8% of outpatient visits through the end of 2021. Overall, 62% of patients used telemedicine in 2020 and 48% in 2021; outpatient visits with a cancer diagnosis represented 75% of visits in 2020 and 72% in 2021, slightly higher for telemedicine in both years (Appendix Tables 3 and 4).

Figure 1.

Figure 1.

Open in a new tab

Adjusted weekly telehealth use rates by patients with lung or colon cancer in 2020–2021, compared against new COVID-19 case rates.

Utilization Trends by Rurality

Adjusted time trends of telemedicine visits stratified by rurality are shown in Figure 2, with annual utilization rates for telemedicine alongside office visits, ED visits and hospitalizations are visualized in Figure 3. Small town/rural cancer patients had nearly half as many telemedicine visits as metropolitan patients in both years (in 2020, 1.28 visits per year [95% CI 1.27 to 1.29] versus 2.34 [2.33 to 2.35]; in 2021, 0.88 [0.87 to 0.89] versus 1.60 [1.59 to 1.60]). Small town/rural patients had lower office visit rates as well (in 2020, 12.37 [12.34-12.41] versus 13.71 [13.70-13.72]; in 2021, 12.16 [12.12-12.20] versus 14.43 [14.41-14.45]). While the decline across 2020-2021 in metropolitan telemedicine use was offset by a corresponding increase in office visits, micropolitan and small town/rural use of both visit types decreased. Hospitalization rates were similar across rurality, while ED utilization was higher in more rural environments.

Figure 2. Weekly Proportion of Outpatient Care via Telemedicine by Patients Within Each Rurality, 2020 - 2021, Among Medicare Fee-For-Service Beneficiaries with Lung or Colon Cancer.

Figure 2.

Open in a new tab

Figure 3. Annual Utilization Rates per Medicare Fee-for-Service Patient with Prevalent or Incident Lung or Colon Cancer, by Ruralityab.

Figure 3.

Open in a new tab

aWith metropolitan as referent, all comparisons by rurality within year are significant at p<0.001 level, except for hospitalizations in 2020 for metro vs small town/rural, p = 0.726

bFor comparisons between years within each category of rurality, P < 0.001, except for outpatient in-person visits for micropolitan 2020 vs 2021 where p = 0.031

Figure 4 displays telemedicine use by rurality, stratified by region and race (data in Appendix Table 5). When examined by U. S. region, the difference by rurality persists though the extent of use varies, with the smallest relative differences in the Northeast (relative rate of telemedicine in metropolitan versus small town/rural, 1.27 [1.24 to 1.30] 2020, 1.11 [1.08 to 1.15] 2021) and largest in the Midwest (1.56 [1.53 to 1.58] 2020, 1.44 [1.40 to 1.48] 2021). Telemedicine differences across rurality by race/ethnicity were most pronounced for patients identified as Black, where metropolitan use approached double that of rural (1.84 [1.74 to 1.94] 2020, 1.76 [1.62 to 1.91] 2021) and was smallest among patients identified as Hispanic (1.39 [1.30 to 1.48] 2020, 1.28 [1.17 to 1.41] 2021).

Figure 4: Proportion of Visits Via Telemedicine, Medicare Fee-for-Service Patients with Prevalent or Incident Lung or Colon Cancer by Rurality Within Groups Defined by US Region and Race/Ethnicity.

Figure 4:

Open in a new tab

aWith the South region as referent within each year and rurality, all other regions are significantly different at the p < 0.001 level except small town/rural Midwest in 2020, p =0.007

bWith White as referent within each year and rurality, all other groups are significantly different at the p < 0.001 level except metropolitan Black in 2021, p=0.27. The White, Black and Other groups are exclusive of Hispanic ethnicity.

Telemedicine Visit Trends by Medical and Demographic Indicators

Adjusted telemedicine visit rates among other subgroups of interest, including chemotherapy or radiation treatment, age, race, US region, and dual Medicare / Medicaid eligibility, are reported in Figure 5. Age was not associated with substantial differences in telemedicine use. Patients who received chemotherapy or radiation treatment had substantially lower rates of telemedicine use in 2020, with diminishing differences into 2021 (peak of 44% outpatient visits via telemedicine for no radiation versus 34% radiation; peak of 49% no chemotherapy versus 32% chemotherapy, each converging to 7-9%). Regionally, the Northeast had the highest rates of telemedicine use early in the study period, peaking at 54% of outpatient visits, with the West leading after 4 months and both higher than the South or Midwest (week of July 12th 2020: 21% for West, 20% for Northeast, 14% South and 10% Midwest). Patients with dual Medicare/Medicaid eligibility, compared to those only on Medicare, used more telemedicine after the first month of the pandemic (week of June 7th 2020: 23% dual versus 17% Medicare only, declining to 10-11% versus 6-7% during 2021). Other subgroups, including whether the patient was primarily seen within a healthcare system versus independent practice, whether the patient had cancer surgery or not, and HCC count, were associated with minimal (< 1%) differences in telemedicine use.

Figure 5. Weekly Proportion of Outpatient Care via Telemedicine by Select Characteristics, 2020 - 2021, Among Medicare Fee-For-Service Beneficiaries with Lung or Colon Cancer.

Figure 5.

Open in a new tab

Utilization Subsequent to Telemedicine and Office Visits

Appendix Table 6 reports rates and relative likelihoods of hospitalization, ED, office or telemedicine visit within a 2-week period following an office or telemedicine visit without outpatient visits in the preceding 2 weeks. In 2020, ED, hospitalization, and office visits were more common after office than telemedicine visits (RR following telemedicine versus office visit: 0.84 [0.82 to 0.86], 0.85 [0.84 to 0.87], and 0.54 [0.53 to 0.54], respectively). In 2021, relative rates for ED visits were equivocal (1.03, [0.99 to 1.06]) and hospitalization was more frequent after telemedicine than office visits (1.23 [1.19 to 1.26]). Telemedicine visits were far more likely to have subsequent telemedicine visits (3.92 [3.88 to 3.96] in 2020, 3.28 [3.23 to 3.33] in 2021) and less likely to have subsequent office visits. In analysis of effect modification for telemedicine use on ED visit rates, we found minimal change in overall rates (the largest difference was among small town/rural patients in 2021: 0.71 [0.70 to 0.72] ED visits per patient per year without accounting for telemedicine use, versus 0.75 [0.74 to 0.76] when including in model; see Appendix Table 7).

Discussion

This retrospective cohort analysis leveraged Medicare FFS claims data to characterize national utilization trends for patients over 65 with lung or colorectal cancer from 2020 - 2021, and noted persistently higher rates of telemedicine use among metropolitan compared to rural patients. Rural patients had fewer outpatient visits in general, with the gap widening from 2020 to 2021 in our analysis, and also used the emergency department more than metropolitan patients, suggesting that telemedicine use did not meaningfully address rural-urban disparities in overall access to outpatient care for cancer patients. While this observational analysis cannot make causal inferences regarding lower rural telemedicine use, there are multiple likely contributing factors. Disparities in broadband and general technology accessibility (the “digital divide”) for rural populations are well described29-31. Rural cancer clinician workforce limitations may also play a role32-34. Given the relatively larger potential benefits to rural populations for telemedicine to save on cost and travel time29, the finding of persistently lower rural telemedicine use and its further decline over time should be of concern to care providers and policymakers concerned with the welfare of rural cancer patients.

Our findings align with previous reporting on rural-urban telemedicine disparities for cancer patients11,35; to a lesser extent our findings also affirm disparities by race, though differ in finding higher use by patients of Hispanic ethnicity and races other than white or Black7,10. Our findings also differ from most studies of cancer patients in showing higher telemedicine use for low-income patients (dual Medicaid/Medicare eligible)9, though studies beyond cancer populations have found similar patterns36,37. This divergence may be explained by a shift in the relative impact of early adoption among digitally-connected, higher-income populations in the early pandemic (a majority of cancer telemedicine studies to date) in favor of a more persistent desire or need for telemedicine among lower-income patients as pandemic risks subsided, given the emergence over time of persistently higher use among Medicaid-eligible patients in our study. Our findings further show that patients receiving radiation and chemotherapy (but not surgery) tended to receive a smaller proportion of care via telemedicine. Perhaps unsurprisingly, patients with office visits more often had subsequent office visits rather than telemedicine and those with telemedicine visits more often had subsequent telemedicine visits. These findings may reflect a continuation of usual practices of bundling office visits and treatment for the former,34 while the latter might reflect aspects of patient or provider preferences for telemedicine. While observational data suggest telemedicine use is non-inferior or even slightly superior for chronic condition management compared to in-person care37-40, outcomes specific to cancer and telemedicine are not well established and remain an area where further study is needed41.

For clinicians and health systems caring for cancer patients, our findings demonstrate an opportunity to expand access for rural populations via telemedicine. Effectively leveraged, telemedicine might help address disparate treatment delays for rural patients42, for example through partnerships with oncology centers in high-supply metropolitan areas assisting low-supply rural areas, or using satellite/community-based clinics to support telemedicine visits in areas where in-person care and home-based broadband are lacking43-45.

On a policy level, we found meaningful variation by region yet persistently higher telemedicine use among metropolitan patients. While we do not know the extent to which our findings reflect differences in broadband accessibility, this highlights the potential for the 2021 Infrastructure Investment and Jobs Act, which allocated $65 billion for broadband expansion46, to facilitate rural telemedicine use. Of these funds, $42.5 billion were allocated to address broadband access disparities, but actual infrastructure was not yet built as of 2025 and continued support for expansion is needed47,48. The Trump administration’s “One Big Beautiful Bill” Act supported telemedicine by making copay-less visits a permanent option, though it did not address payment parity (set to expire by October 2025);49 on the other hand, Medicaid changes are expected to leave millions of Americans without coverage in the years to come, adding barriers to access among a population who are more inclined to use telemedicine50. Continued uncertainty in who might be able to use telemedicine and its expected reimbursements should be addressed to best support long-term investment in telemedicine innovation.

Limitations

We were not able to include data before 2020 or after 2021 due to limits in the data use agreement. Thus, we were unable to establish baseline utilization trends to enable difference-in-differences analyses of changes attributable to telemedicine adoption; however, others have found similar differences in rural ED and outpatient utilization pre-pandemic51,52 and <1% of Medicare beneficiaries used telemedicine before 202053. Our study of Medicare FFS beneficiaries may have limited generalizability for other populations, including younger patients and those covered by commercial plans or Medicare Advantage (an alternative Medicare coverage scheme)54. However, considering that a majority of cancer survivors are age 65 or older55 and that Medicare pays the largest share of cancer costs56, our analysis provides crucial information on US national telemedicine use for cancer patients, particularly considering rural populations where data are limited. While observational analyses are always subject to potential unmeasured confounding, our inclusion of 2021 data helps to tease apart longer-term trends in telemedicine use after the pandemic’s onset. Finally, as there remains no clear consensus on a “right rate” or best-practice use cases for telemedicine in oncology, deciding whether the observed differences represent an overuse or a disparity as related to care quality remains subjective.

Conclusion

Rural-urban differences in telemedicine use for patients with cancer persisted from the pandemic onset through late 2021 with lower usage by rural-residing patients. This trend aligns with larger disparities in access to care for rural populations. While broadband expansion and similar efforts to address the digital divide in telemedicine accessibility may provide opportunities to innovate in rural cancer care delivery and narrow gaps in accessing care, further research is needed to understand the optimal role for telemedicine in oncology.

Supplementary Material

PV Appendix Tables

Context Summary.

Key objective:

How much was telemedicine used in 2020 and 2021 by lung and colon cancer patients, and how did it vary by rurality?

Knowledge generated:

Telemedicine use declined over time, but was used by a majority of cancer patients in 2020 and nearly half in 2021. In both years, compared to urban patients, rural patients had fewer telemedicine and in-person visits, yet more emergency department visits, with no meaningful differences in hospitalization.

Relevance:

Clinicians and health systems should consider ways to better leverage telemedicine to expand access to care for rural cancer patients.

Acknowledgements

The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.

The statements, findings, conclusions, views, and opinions contained and expressed in this article are based in part on data obtained under license from IQVIA information services: OneKey subscription information services 2017-2022, IQVIA Inc. The statements, findings, conclusions, views, and opinions contained and expressed herein are not necessarily those of IQVIA Incorporated or any of its affiliated or subsidiary entities.

Funding

This study was supported by the Agency for Healthcare Research and Quality’s Comparative Health System Performance Initiative under grant No. 1U19HS024075 and by the National Cancer Institute (P30CA023108 and R01CA248470).

Footnotes

Author Contributions / CRediT

Matthew Mackwood: Conceptualization (supporting); writing – original draft (lead), methodology (supporting); visualization (lead)

Qianfei Wang: writing – review & editing (supporting); data curation (lead); formal analysis (lead); software (lead); methodology (supporting); project administration (supporting); visualization

Tor Tosteson: writing – review & editing (supporting); formal analysis (supporting), statistical methodology (lead); software (supporting); supervision (supporting)

Gabriel Brooks: writing – review & editing (supporting), methodology (supporting)

Rebecca Smith: writing – review & editing (supporting), methodology (supporting)

Anna Tosteson: Conceptualization (lead), writing – review & editing (supporting), methodology (supporting), supervision (lead), resources (lead), project administration (lead)

Conflicts of Interest

All authors have no conflicts of interest in this work.

Data Availability Statement

Data analyzed are governed by a Data Use Agreement (DUA) with the Centers for Medicare Services (CMS) and cannot be shared on request based on the terms of the DUA.

References

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

PV Appendix Tables
NIHMS2109853-supplement-PV_Appendix_Tables.xlsx (43.7KB, xlsx)

Data Availability Statement

Data analyzed are governed by a Data Use Agreement (DUA) with the Centers for Medicare Services (CMS) and cannot be shared on request based on the terms of the DUA.

RESOURCES