Background:

Cancer screening is a complex process involving multiple steps and levels of influence (e.g., patient, provider, facility, health care system, community, or neighborhood). We describe the design, methods, and research agenda of the Population-based Research to Optimize the Screening Process (PROSPR II) consortium. PROSPR II Research Centers (PRC), and the Coordinating Center aim to identify opportunities to improve screening processes and reduce disparities through investigation of factors affecting cervical, colorectal, and lung cancer screening in U.S. community health care settings.

Methods:

We collected multilevel, longitudinal cervical, colorectal, and lung cancer screening process data from clinical and administrative sources on >9 million racially and ethnically diverse individuals across 10 heterogeneous health care systems with cohorts beginning January 1, 2010. To facilitate comparisons across organ types and highlight data breadth, we calculated frequencies of multilevel characteristics and volumes of screening and diagnostic tests/procedures and abnormalities.

Results:

Variations in patient, provider, and facility characteristics reflected the PROSPR II health care systems and differing target populations. PRCs identified incident diagnoses of invasive cancers, in situ cancers, and precancers (invasive: 372 cervical, 24,131 colorectal, 11,205 lung; in situ: 911 colorectal, 32 lung; precancers: 13,838 cervical, 554,499 colorectal).

Conclusions:

PROSPR II's research agenda aims to advance: (i) conceptualization and measurement of the cancer screening process, its multilevel factors, and quality; (ii) knowledge of cancer disparities; and (iii) evaluation of the COVID-19 pandemic's initial impacts on cancer screening. We invite researchers to collaborate with PROSPR II investigators.

Impact:

PROSPR II is a valuable data resource for cancer screening researchers.

Cancer screening is a complex process involving multiple steps (i.e., risk assessment, screening, abnormal screen evaluation, diagnosis, and treatment) in the context of multiple levels of influence (e.g., patient, provider, facility, health care system, community or neighborhood, state/national guidelines/policy/legislation; refs. 1–3). Failures or disruptions in any step can lead to suboptimal cancer outcomes. Screening process failures in the United States may be due to fragmentation of health care stemming from an array of payers, payment models, and health care settings; variable health care access and delivery; and increasingly complex, risk-based screening. Furthermore, screening in the United States is often delivered not through organized programs but opportunistically, which can lead to inconsistencies in both screening outreach and management of abnormalities detected at screening. Finally, the risk of screening process failures may be disproportionately experienced by minoritized populations (refs. 4, 5; i.e., groups that are devalued in society and given less access to its resources). The Population-based Research to Optimize the Screening Process (PROSPR II) consortium was established by the NCI in 2018 to enhance the understanding of multilevel factors that affect the delivery and quality of cancer screening in the United States to identify opportunities to improve cancer screening processes and reduce disparities (6).

PROSPR II builds on the accomplishments of the PROSPR I consortium. Guided by conceptual frameworks for the screening process (1, 7, 8), PROSPR I established the feasibility of reporting screening process measures, characterizing heterogeneity in screening performance across health care settings, and documenting disparities between populations in different settings (9–12). Despite these accomplishments, questions persisted about factors contributing to the observed variation across health care settings, largely due to limitations in the ability to conceptualize and measure characteristics beyond those of the patient (e.g., provider, facility, health care system levels; ref. 6). In addition, standards for screening test performance and follow-up varied (13), and questions remained about how to measure quality in screening delivery and how variation in quality may affect screening outcomes (14). Furthermore, PROSPR I's 5-year study period limited investigation of less prevalent screening abnormalities and longer-term outcomes. Finally, PROSPR I lacked data on the most recent cancer screening process to be endorsed with a grade “A” or “B” recommendation by the U.S. Preventive Services Task Force (USPSTF), lung cancer screening via low-dose CT (LDCT). Thus, PROSPR II was initiated with new goals, includes both new health care systems and a new lung research center, and includes additional multilevel data and follow-up beyond the PROSPR I study period.

PROSPR II investigators focus on three different screening processes that range widely in their time since implementation. Cervical cancer screening was first introduced in the 1950s in the United States, but has undergone many changes in screening modalities, screening recommendations, and management of abnormal results over the last two decades. Colorectal cancer screening became more widespread after colonoscopy was first recommended as a screening modality by the USPSTF in 2002, with accompanying changes in reimbursement. Finally, lung cancer screening received its first USPSTF recommendation only in 2013, with Centers for Medicare and Medicaid Services payment rules issued in 2015. All screening seeks an optimum balance of benefits and harms, and with improved knowledge of cancer etiology, screening processes have become increasingly risk based. Individual risk may modify screening entry, modality, interval, and exit. Common to all cancer screening processes is risk stratification based on age, with the addition of factors such as sex, race, smoking history, family history, history of abnormal test results, or cancer or precancer (e.g., cervical intraepithelial neoplasia grade 3 or colorectal adenoma) history, depending on the specific organ type. Although optimization of cancer screening is expected to lead to more benefit and less harm, it has likely made screening more complicated to communicate and deliver. Within this context, the PROSPR II research agenda aims to improve cervical, colorectal, and lung cancer screening processes by advancing: (i) conceptualization and measurement of multilevel factors and quality measures, (ii) knowledge of cancer screening disparities, and (iii) evaluation of initial impacts of the COVID-19 pandemic on cancer screening.

The PROSPR II consortium also aims to share data with the cancer screening research community. The PROSPR DataShare (PDS) website (15) will enable researchers who are not part of the consortium to request deidentified public use datasets, apply for limited datasets, and propose additional data collection activities leveraging PROSPR II infrastructure. To facilitate these activities, PDS also provides current data dictionaries and a list of PROSPR publications. For PROSPR II, this article describes (i) the design and objectives, (ii) data sources and methods used to collect screening process data, (iii) key characteristics of study populations and clinical settings, (iv) select screening process volumes in study populations, and (v) the research agenda. We aim to highlight the value of PROSPR II science and resources for the cancer screening research community.

PROSPR II design, objectives, and settings

The PROSPR II consortium includes three multisite research centers and a coordinating center (6). PROSPR II Research Centers (PRC), one each focused on cervical, colorectal, and lung cancer screening (Table 1), represent diverse health care environments and include study populations of screened and unscreened individuals. PRCs aim to identify factors that can be modified to improve organ-specific screening processes in community health care settings. The PROSPR II Coordinating Center, located at Fred Hutchinson Cancer Research Center, leads the development of measures of health care system level factors that may impact the cancer screening process and screening process quality measures that can be applied across organ types. The Coordinating Center also leads data analysis for research projects that include more than one organ type and is responsible for developing policies and processes for data sharing outside of the PROSPR II consortium.

Table 1.

Multilevel characteristics by PROSPR II Research Center (2010 through 2017 or 2019).

Cervical Cancer Screening Research Center 2010–2017Colorectal Cancer Screening Research Center 2010–2017Lung Cancer Screening Research Center 2010-September 2019
Patient-level characteristics 
Total population N = 862,524 N = 6,384,599 N = 2,130,716 
 n (%) n (%) n (%) 
Age (years) at cohort entrya 
 18–29 251,596 29.2 — — — — 
 30–39 176,587 20.5 — — 474,854 22.3 
 40–49 147,602 17.1 2,844,872 44.6 536,993 25.2 
 50–59 141,182 16.4 1,754,898 27.5 521,831 24.5 
 60–69 91,399 10.6 1,109,135 17.4 352,580 16.6 
 70–89 54,158 6.3 675,694 10.6 244,458 11.5 
Sexb 
 Female 862,524 100.0 3,327,355 52.1 1,144,001 53.7 
 Male — — 3,056,760 47.9 986,584 46.3 
 Other — — 36 0.0 0.0 
 Unknown — — 448 — 131 — 
Race and ethnicityb 
 Hispanic 195,526 24.9 1,524,831 26.6 128,196 6.8 
 White 403,907 51.4 2,808,356 48.9 1,237,366 65.3 
 Black 100,894 12.9 504,900 8.8 270,391 14.3 
 Asian 58,592 7.5 779,288 13.6 134,708 7.1 
 Native American or Alaskan Native 3,066 0.4 18,842 0.3 5,527 0.3 
 Native Hawaiian or Other Pacific Islander 3,654 0.5 38,199 0.7 27,714 1.5 
 Other 7,827 1.0 3,976 0.1 43,494 2.3 
 Multiple races 11,688 1.5 59,189 1.0 46,341 2.4 
 Unknown 77,370 — 647,018 — 236,979 — 
Health plan payer or insurance type in calendar year of cohort entryc 
 Medicaid 138,502 16.1 158,336 2.5 89,551 4.4 
 Medicare 89,575 10.4 1,241,034 19.5 261,826 12.9 
 Commercial/private 480,679 55.7 4,807,771 75.5 1,657,685 81.4 
 Other 53,587 6.2 66,355 1.0 26,504 1.3 
 Uninsured/medical financial assistance 100,181 11.6 97,644 1.5 0.0 
 Unknown — 13,459 — 95,150 — 
Body mass index (kg/m2) closest to the end of the cohort entry yeard 
 <18.5 21,349 3.0 45,549 1.0 21,292 1.2 
 18.5–24.9 239,688 34.2 1,162,036 26.4 471,216 27.0 
 25–29.9 199,235 28.4 1,591,135 36.1 596,966 34.3 
 30–34.9 123,578 17.6 944,546 21.4 365,433 21.0 
 35+ 117,318 16.7 661,683 15.0 287,802 16.5 
 Unknown 161,356 — 1,979,650 — 388,007 — 
Charlson comorbidity index score closest to the end of the first full year of enrollment/observationd 
 0 535,353 75.9 3,845,062 73.7 1438,880 75.7 
 1 97,974 13.9 714,513 13.7 254,768 13.4 
 2 41,552 5.9 327,152 6.3 109,520 5.8 
 3+ 30,854 4.4 329,304 6.3 96,522 5.1 
 Unknown 156,791 — 1,168,568 — 231,026 — 
Smoking status closest to the end of the cohort exit yeard 
 Never smoker 516,609 71.3 3,590,861 68.3 990,472 54.9 
 Former smoker — — — — 573,693 31.8 
 Current smoker — — — — 241,308 13.4 
 Ever smoker 207,802 28.7 1,669,254 31.7 — — 
 Unknown 138,113 — 1,124,484 — 325,243 — 
Number of primary care encounters during first full year of enrollment/observatione 
 No encounters 49,793 7.1 1,084,281 20.9 238,026 11.2 
 (minimum, 25%ile, median, 75%ile, maximum) (1, 1, 2, 4, 100) (1, 1, 3, 4, 267) (1, 3, 4, 7, 290) 
Provider-level characteristicsf 
Total providers delivering primary care or organ-specific cancer screening care in PROSPR data n = 28,879 n = 51,268 n = 16,772 
 n (%) n (%) n (%) 
Provider type 
 MD/DO 18,506 74.0 27,905 56.9 9,140 54.5 
 Physician's assistant 1128 4.5 1,666 3.4 546 3.3 
 NP/LPN/RN 4,277 17.1 7,057 14.4 2,407 14.4 
 Medical assistant 527 2.1 1,194 2.4 760 4.5 
 Other provider 567 2.3 11,241 22.9 1,153 6.9 
 Unknown 3,874 — 2,205 — 2,766 — 
Provider specialty 
 Family medicine 6,583 26.8 7,795 16.8 2,103 20.1 
 General internal medicine/internal medicine, other 6,770 27.6 20,915 45.1 2,649 25.3 
 Gastroenterology — — 893 1.9 237 2.3 
 Obstetrics/Gynecology 2,635 10.7 — — 197 1.9 
 Oncology 319 1.3 303 0.7 341 3.3 
 Pathology 500 2.0 137 0.3 199 1.9 
 Pulmonary — — — — 390 3.7 
 Radiology 139 0.6 481 1.2 691 6.6 
 Nursing 1,935 7.9 1,569 3.4 951 9.1 
 Multiple/other specialties 5,658 23.1 14,294 30.8 2,723 26.0 
 Unknown 4,340 — 4,881 — 6,291 — 
Facility-level characteristicsg 
 n n n 
Internal hospitals 39 20 
Internal facilities providing primary care 71 258 272 
Internal facilities exclusively providing other types of care 51 167 
External facilities 3,679 1,379 Not reported 
PROSPR Research Center-level characteristics 
Health care systems 
  • 1. Kaiser Permanente Washington

  • 2. Mass General Brigham

  • 3. Parkland Health & Hospital System

 
  • 1. Kaiser Permanente Northern California

  • 2. Kaiser Permanente Southern California

  • 3. Kaiser Permanente Washington

  • 4. Parkland Health & Hospital System

 
  • 1. Henry Ford Health System

  • 2. Kaiser Permanente Colorado

  • 3. Kaiser Permanente Hawaii

  • 4. Marshfield Clinic Health System

  • 5. University of Pennsylvania Health System

 
Type of care delivery systemsh Mixed model, integrated, integrated safety-net Integrated, mixed model, integrated safety-net Mixed model and integrated 
Geographic regions represented Eastern Massachusetts; Washington; Dallas County, Texas Northern California; Southern California; Washington; Dallas County, Texas Metropolitan Detroit, Michigan; Denver/Boulder Front Range, Colorado; Oahu, Hawaii Island, Maui, and Kauai; central, north central, northwestern Wisconsin; Philadelphia, Pennsylvania and Greater Delaware Valley (including sites in New Jersey and Delaware) 
Primary screening strategies offered Cervical cytology (with reflex HPV testing), co-testing (cervical cytology and HPV testing), primary HPV screening Fecal immunochemical test, colonoscopy Low-dose chest CT 
Systems with centralized population health teams 2/3 3/4 4/5 
Systems with organization-wide policies or guidelines for organ-specific cancer screening 3/3 4/4 2/5 
Systems with incentive-based programs/policies for organ-specific cancer screening for providers and/or facilities 3/3 4/4 0/5 
Systems with performance measures for organ-specific cancer screening policies 3/3 4/4 0/5 
Cervical Cancer Screening Research Center 2010–2017Colorectal Cancer Screening Research Center 2010–2017Lung Cancer Screening Research Center 2010-September 2019
Patient-level characteristics 
Total population N = 862,524 N = 6,384,599 N = 2,130,716 
 n (%) n (%) n (%) 
Age (years) at cohort entrya 
 18–29 251,596 29.2 — — — — 
 30–39 176,587 20.5 — — 474,854 22.3 
 40–49 147,602 17.1 2,844,872 44.6 536,993 25.2 
 50–59 141,182 16.4 1,754,898 27.5 521,831 24.5 
 60–69 91,399 10.6 1,109,135 17.4 352,580 16.6 
 70–89 54,158 6.3 675,694 10.6 244,458 11.5 
Sexb 
 Female 862,524 100.0 3,327,355 52.1 1,144,001 53.7 
 Male — — 3,056,760 47.9 986,584 46.3 
 Other — — 36 0.0 0.0 
 Unknown — — 448 — 131 — 
Race and ethnicityb 
 Hispanic 195,526 24.9 1,524,831 26.6 128,196 6.8 
 White 403,907 51.4 2,808,356 48.9 1,237,366 65.3 
 Black 100,894 12.9 504,900 8.8 270,391 14.3 
 Asian 58,592 7.5 779,288 13.6 134,708 7.1 
 Native American or Alaskan Native 3,066 0.4 18,842 0.3 5,527 0.3 
 Native Hawaiian or Other Pacific Islander 3,654 0.5 38,199 0.7 27,714 1.5 
 Other 7,827 1.0 3,976 0.1 43,494 2.3 
 Multiple races 11,688 1.5 59,189 1.0 46,341 2.4 
 Unknown 77,370 — 647,018 — 236,979 — 
Health plan payer or insurance type in calendar year of cohort entryc 
 Medicaid 138,502 16.1 158,336 2.5 89,551 4.4 
 Medicare 89,575 10.4 1,241,034 19.5 261,826 12.9 
 Commercial/private 480,679 55.7 4,807,771 75.5 1,657,685 81.4 
 Other 53,587 6.2 66,355 1.0 26,504 1.3 
 Uninsured/medical financial assistance 100,181 11.6 97,644 1.5 0.0 
 Unknown — 13,459 — 95,150 — 
Body mass index (kg/m2) closest to the end of the cohort entry yeard 
 <18.5 21,349 3.0 45,549 1.0 21,292 1.2 
 18.5–24.9 239,688 34.2 1,162,036 26.4 471,216 27.0 
 25–29.9 199,235 28.4 1,591,135 36.1 596,966 34.3 
 30–34.9 123,578 17.6 944,546 21.4 365,433 21.0 
 35+ 117,318 16.7 661,683 15.0 287,802 16.5 
 Unknown 161,356 — 1,979,650 — 388,007 — 
Charlson comorbidity index score closest to the end of the first full year of enrollment/observationd 
 0 535,353 75.9 3,845,062 73.7 1438,880 75.7 
 1 97,974 13.9 714,513 13.7 254,768 13.4 
 2 41,552 5.9 327,152 6.3 109,520 5.8 
 3+ 30,854 4.4 329,304 6.3 96,522 5.1 
 Unknown 156,791 — 1,168,568 — 231,026 — 
Smoking status closest to the end of the cohort exit yeard 
 Never smoker 516,609 71.3 3,590,861 68.3 990,472 54.9 
 Former smoker — — — — 573,693 31.8 
 Current smoker — — — — 241,308 13.4 
 Ever smoker 207,802 28.7 1,669,254 31.7 — — 
 Unknown 138,113 — 1,124,484 — 325,243 — 
Number of primary care encounters during first full year of enrollment/observatione 
 No encounters 49,793 7.1 1,084,281 20.9 238,026 11.2 
 (minimum, 25%ile, median, 75%ile, maximum) (1, 1, 2, 4, 100) (1, 1, 3, 4, 267) (1, 3, 4, 7, 290) 
Provider-level characteristicsf 
Total providers delivering primary care or organ-specific cancer screening care in PROSPR data n = 28,879 n = 51,268 n = 16,772 
 n (%) n (%) n (%) 
Provider type 
 MD/DO 18,506 74.0 27,905 56.9 9,140 54.5 
 Physician's assistant 1128 4.5 1,666 3.4 546 3.3 
 NP/LPN/RN 4,277 17.1 7,057 14.4 2,407 14.4 
 Medical assistant 527 2.1 1,194 2.4 760 4.5 
 Other provider 567 2.3 11,241 22.9 1,153 6.9 
 Unknown 3,874 — 2,205 — 2,766 — 
Provider specialty 
 Family medicine 6,583 26.8 7,795 16.8 2,103 20.1 
 General internal medicine/internal medicine, other 6,770 27.6 20,915 45.1 2,649 25.3 
 Gastroenterology — — 893 1.9 237 2.3 
 Obstetrics/Gynecology 2,635 10.7 — — 197 1.9 
 Oncology 319 1.3 303 0.7 341 3.3 
 Pathology 500 2.0 137 0.3 199 1.9 
 Pulmonary — — — — 390 3.7 
 Radiology 139 0.6 481 1.2 691 6.6 
 Nursing 1,935 7.9 1,569 3.4 951 9.1 
 Multiple/other specialties 5,658 23.1 14,294 30.8 2,723 26.0 
 Unknown 4,340 — 4,881 — 6,291 — 
Facility-level characteristicsg 
 n n n 
Internal hospitals 39 20 
Internal facilities providing primary care 71 258 272 
Internal facilities exclusively providing other types of care 51 167 
External facilities 3,679 1,379 Not reported 
PROSPR Research Center-level characteristics 
Health care systems 
  • 1. Kaiser Permanente Washington

  • 2. Mass General Brigham

  • 3. Parkland Health & Hospital System

 
  • 1. Kaiser Permanente Northern California

  • 2. Kaiser Permanente Southern California

  • 3. Kaiser Permanente Washington

  • 4. Parkland Health & Hospital System

 
  • 1. Henry Ford Health System

  • 2. Kaiser Permanente Colorado

  • 3. Kaiser Permanente Hawaii

  • 4. Marshfield Clinic Health System

  • 5. University of Pennsylvania Health System

 
Type of care delivery systemsh Mixed model, integrated, integrated safety-net Integrated, mixed model, integrated safety-net Mixed model and integrated 
Geographic regions represented Eastern Massachusetts; Washington; Dallas County, Texas Northern California; Southern California; Washington; Dallas County, Texas Metropolitan Detroit, Michigan; Denver/Boulder Front Range, Colorado; Oahu, Hawaii Island, Maui, and Kauai; central, north central, northwestern Wisconsin; Philadelphia, Pennsylvania and Greater Delaware Valley (including sites in New Jersey and Delaware) 
Primary screening strategies offered Cervical cytology (with reflex HPV testing), co-testing (cervical cytology and HPV testing), primary HPV screening Fecal immunochemical test, colonoscopy Low-dose chest CT 
Systems with centralized population health teams 2/3 3/4 4/5 
Systems with organization-wide policies or guidelines for organ-specific cancer screening 3/3 4/4 2/5 
Systems with incentive-based programs/policies for organ-specific cancer screening for providers and/or facilities 3/3 4/4 0/5 
Systems with performance measures for organ-specific cancer screening policies 3/3 4/4 0/5 

Abbreviations: DO, doctor of osteopathic medicine; HFHS, Henry Ford Health System; HPV, human papillomavirus; KPNC, Kaiser Permanente Northern California; KPSC, Kaiser Permanente Southern California; KPWA, Kaiser Permanente Washington; LDCT, low-dose chest computed tomography; LPN, licensed practical nurse; MD, doctor of medicine; MGB, Mass General Brigham; NP, nurse practitioner; Parkland, Parkland Health & Hospital System; PRC, PROSPR Research Center; PROSPR, Population-based Research to Optimize the Screening Process; RN, registered nurse; UPHS, University of Pennsylvania Health System.

aFor the lung PRC, cohort entry begins at 35 years of age.

bSex, race, and ethnicity are reported as indicated in the electronic health record and may not be reflective of individual gender identity or racial and ethnic identity. For race and ethnicity, Hispanic individuals are reported as Hispanic regardless of race classifications. Multiple races include two or more of the following categories: White, Black, Asian, Native American or Alaskan Native, Native Hawaiian or Other Pacific Islander, and Other. KPNC and KPSC do not capture “other” race data.

cThe following hierarchy was used for health insurance: Medicaid, Medicare, Commercial/private insurance, Other, Uninsured. For the cervical PRC, those with “Other” health insurance included state-subsidized coverage (excluding Medicaid) and government programs (e.g., Ryan White, Title V, XX, and BCCS programs). Medical financial assistance at Parkland is based on financial and geographic eligibility. Patients receive annual or bi-annual financial counseling to determine payer eligibility. For the lung PRC, those with an integrated care health plan membership due to Medicare advantage or Medicaid were considered commercially insured.

dFor the cervical PRC Parkland and MGB sites, weight and smoking status were calculated using values closest to the beginning of the cohort entry year within the same calendar year as cohort entry, and if no value was available during the calendar year, the closest prior recorded value was used. MGB had the following variations from Charlson Comorbidity Index specifications: comorbidities were identified from diagnosis codes as well as problem lists, and HIV status was confirmed on the basis of the date that a third primary HIV diagnosis code was observed in the health record. Parkland had the following variation from Charlson Comorbidity Index score specifications: HIV status was ascertained on the basis of a combination of diagnosis codes, HIV clinic visit, and/or disease-staging laboratory tests. For the lung PRC, UPHS did not include HIV status in the Charlson Comorbidity Index calculation due to state-level restrictions in data use.

eEnrollment-based sites did not require a primary care visit for cohort entry. The cervical PRC has one enrollment-based site (KPWA) and two primary care utilization-based sites that require a primary care visit to enter the cohort (MGB and Parkland). The colorectal PRC has three enrollment-based sites (KPWA, KPNC, KPSC) and one primary care utilization-based site (Parkland). The lung PRC has three enrollment-based sites (KPCO, KPHI, MCHS); one site with two cohorts (HFHS, comprising one enrollment-based cohort contributing approximately 30% of the total cohort with the rest from a primary care utilization-based cohort); and one primary care utilization-based site (UPHS).

fTotal number of providers reflects those providing primary care or organ-specific cancer screening process care for PROSPR cohort members over the study period and does not reflect all providers within each healthcare system. The lung PRC collected data about primary care providers, providers associated with an LDCT, and providers performing potential procedures to follow-up LDCT findings based on the procedures in Zhao, et al. (24), not all provider types. Provider specialty classification was not restricted to providers with MDs/DOs. Multiple/other specialties included the following specialties if they were captured by PROSPR sites: cervical PRC (gastroenterology, pulmonary, multiple specialties, other specialties not listed); colorectal PRC (obstetrics/gynecology, pulmonary, multiple specialties, other specialties not listed); lung PRC (multiple specialties, other specialties not listed). Obstetrics/gynecology includes midwifery.

gTotal number of facilities reflects those providing primary care or organ-specific cancer screening process care for PROSPR cohort members over the study period and may not reflect all facilities within each health care system. Internal hospitals include only those operated by the health care system. Internal hospitals are counted in the internal hospitals row and in rows for facility type (e.g., internal facilities providing primary care). External facilities include contracted hospitals and facilities where the types of care provided are unknown. The cervical PRC interpreted facilities as “outpatient care centers” (i.e., distinct medical campus or building) that could include multiple clinics and/or medical offices. The cervical PRC collects more granular facility-level data than the other PROSPR II PRCs, including department- and clinic-level data. Counts of the lung PRC internal facilities are based on administrative data on internal facilities providing primary care and exclusively other types of care during the study period. UPHS data are not included in counts of internal facilities exclusively providing other types of care. Most of the facilities within the lung PRC that provide primary care services, also provide specialty care services. The lung PRC facilities data collection and cleaning prioritized the capture of data on internal facilities. Future work within the lung PRC will capture additional information on external facilities, but these data are currently unavailable.

hMixed model is defined as a health care system that has both integrated and fee-for-service models or accepts multiple payers; integrated is defined as a health care system coordinating care across primary, specialty, and inpatient services; and integrated safety-net is defined as a health care system providing integrated care to uninsured and under-insured populations.

The cervical PRC includes three sites—Kaiser Permanente Washington (KPWA), Parkland Health & Hospital System/University of Texas Southwestern (Parkland), and Mass General Brigham (MGB). KPWA is a mixed-model health care system providing health insurance and care for Washington state residents. Parkland Health & Hospital System is an integrated safety-net health care system for underinsured and uninsured residents of Dallas County, Texas, with data reported by their academic partner, the University of Texas Southwestern. The health care settings and patient populations for KPWA and Parkland have been described previously (16). MGB is an integrated health care system in the greater Boston, Massachusetts area that includes two academic medical centers—Brigham and Women's Hospital and Massachusetts General Hospital—and their affiliated primary care networks, as well as community and specialty hospitals. Study cohort eligibility for the cervical PRC varied slightly by site. KPWA included female health plan enrollees who selected, were assigned, or were attributed to a KPWA primary care provider, and resided in the Seattle-Puget Sound Surveillance Epidemiology and End Results (SEER) registry (17) catchment area. Parkland and MGB included female patients with at least one visit at a primary care or women's health clinic during the study period. The cervical PRC aims to evaluate the levels at which variation in cervical cancer screening and follow-up may occur, investigate how multilevel factors influence guideline adherence for those at average risk, and generate evidence to inform screening strategies for those at altered risk [e.g., decreased risk due to human papillomavirus (HPV) vaccination].

The colorectal PRC includes KPWA and Parkland along with Kaiser Permanente Northern California (KPNC) and Kaiser Permanente Southern California (KPSC). KPNC and KPSC are integrated health care systems providing health insurance and care to California residents. Details of these four health care systems, particularly with respect to the colorectal cancer screening process, were published previously (8). The colorectal PRC had slight study cohort eligibility differences across sites. KPWA, KPNC, and KPSC included members enrolled in the respective health plans. KPWA additionally restricted eligibility to individuals who selected, were assigned, or were attributed to a KPWA primary care provider, and resided in the Seattle-Puget Sound SEER registry catchment area. Parkland included individuals with at least one primary care visit during the study period. The colorectal PRC aims to fill evidence gaps about who should receive colorectal cancer screening and surveillance and when, evaluate why screening process failures occur, and understand how to improve colorectal cancer test effectiveness.

The lung PRC includes five sites—Kaiser Permanente Colorado (KPCO), Kaiser Permanente Hawaii (KPHI), Henry Ford Health System (HFHS), Marshfield Clinic Health System (MCHS), and University of Pennsylvania Health System (UPHS). KPCO, KPHI, HFHS, MCHS, and UPHS provide health care, and in many cases health insurance, for individuals in, respectively, Colorado, Hawaii, Michigan, Wisconsin, and Pennsylvania and the Greater Delaware Valley. Patient populations and lung cancer screening programs in these health care systems have been described previously (18). The lung PRC study cohort included enrollees in KPCO, KPHI, HFHS, and MCHS health plans and individuals with at least one primary care visit at HFHS or UPHS during the study period. The lung PRC aims to evaluate lung cancer screening uptake and outcomes, determine harms and costs of lung cancer screening, assess use of tobacco cessation strategies (e.g., tobacco cessation medication use) in the context of lung cancer screening, develop and validate imaging biomarkers to improve risk stratification of LDCT findings, and ultimately, to address disparities in lung cancer screening.

Data collection

PROSPR II collected longitudinal cervical, colorectal, and lung cancer screening process data, enabling research across more than one organ type. Data began on January 1, 2010 and collection is planned through at least December 31, 2020. Each PRC collected patient-level information on cohort sociodemographics; screening process data specific to each organ type (risk factors, screening tests, diagnostic evaluations, excisional treatment procedures, and outcomes); cancer screening history prior to cohort entry; cancer diagnoses; and cancer deaths. PRC cohorts purposely included a wide age range to evaluate cancer screening occurring outside of guideline-recommended ages. PRCs also obtained data about provider and facility characteristics (e.g., provider type, provider specialty, facility type). Patient, provider, and facility data were extracted from clinical and administrative data sources, including the electronic health record, within each health care system (definitions are provided in Supplementary Table S1). Cancer data were obtained from linkages to hospital and central cancer registries. In addition, all sites completed (i) the PROSPR II Organizational Factors Survey, which captured characteristics about each health care system (e.g., structure, having centralized population health teams, screening and follow-up policies, incentives for cancer screening) and (ii) the International Cancer Screening Network COVID-19 Survey that assessed the initial effects of the COVID-19 pandemic on cancer screening in different health care settings (e.g., communication, impact on resources, patient follow-up; ref. 19). The PROSPR II consortium built on the extensive harmonization, quality assurance, and documentation efforts developed in PROSPR I (16, 20) to ensure data accuracy and completeness.

Cervical PRC screening process data included HPV vaccinations, cytology and HPV tests/results, diagnostic evaluations (i.e., colposcopies and biopsies) and results, and ablative/excisional treatments (i.e., cryotherapy, laser therapy, conization, loop electrosurgical excision procedure, and surgery) and associated histopathologic diagnoses (16). Screening process data unique to the colorectal PRC included family history of colorectal cancer, stool-based (e.g., fecal immunochemical) tests and results, colonoscopy procedures and results, and colorectal surgeries. Lung PRC screening process data included smoking history, completed LDCT procedures, interpretation of LDCTs (Lung-RADS assessment), non-LDCT imaging procedures of the chest, and diagnostic procedures (18).

Analysis

To facilitate comparisons across organ types and highlight the breadth of available PROSPR II data, we calculated frequencies of multilevel characteristics by PRC during 2010 through 2017 or 2019 (abbreviated as 2010–2017/2019) for this analysis. The different study period end dates reflect the different data collection timelines at the PRCs. At the patient level, we tabulated sociodemographic characteristics of the cervical, colorectal, and lung cancer screening cohorts. Time-varying characteristics were collected annually and were calculated in the 12 months after cohort entry or at cohort exit to balance data availability and consistency across PRCs. At the provider level, we calculated the distribution of provider types and specialties for those delivering primary or cancer screening care to cohort members, combining unrelated provider types/specialties into a separate category. At the facility level, we enumerated the number of internal hospitals, internal facilities providing primary care, internal facilities exclusively providing nonprimary care, and any external facilities (i.e., not operated by the health care system) serving as care delivery settings for cohort members during the study period. At the PRC level, we described care delivery models, enumerated select organizational characteristics of the health care systems within each PRC, and described the effects of the COVID-19 pandemic on health care systems. We also calculated the volumes of screening and diagnostic tests/procedures performed and abnormalities detected in the cervical, colorectal, and lung cohorts during 2010–2017/2019. Institutional Review Boards at the Coordinating Center and across health care systems within each PRC approved all research activities. We obtained data with waivers of consent and conducted research activities in accordance with the U.S. Common Rule.

Data availability statement

PROSPR II data are available for collaboration and sharing after appropriate approvals and agreements are completed. Additional details are provided at: https://healthcaredelivery.cancer.gov/prospr/datashare/.

Study cohorts and clinical settings of PROSPR II research centers

From 2010 to 2017/2019, PRC study populations included over 9 million screened and unscreened individuals (Table 1). The cervical PRC cohort (female patients ages 18–89 years) was younger than both the colorectal PRC cohort (individuals ages 40–89 years) and the lung PRC cohort (individuals ages 35–89 years). PRC cohorts were racially and ethnically diverse with substantial proportions of Black, Asian, and Hispanic individuals (cervical PRC: 13% Black, 8% Asian, 25% Hispanic; colorectal PRC: 9% Black, 14% Asian, 27% Hispanic; lung PRC: 14% Black, 7% Asian, 7% Hispanic). Although most individuals across all PRCs had commercial/private health insurance coverage, 28% of female patients in the cervical PRC were uninsured, had Medicaid coverage, or received medical financial assistance. The cervical PRC cohort had the greatest proportion of individuals classified in the 18.5–24.9 kg/m2 range for body mass index (34%), followed by the lung PRC (27%) and colorectal PRC (26%). Most individuals (74%–76%) had a score of 0 per the Charlson comorbidity index (21, 22) [or site-specific comorbidity score (23) using prevalent diagnoses], indicating a low burden of illness. The proportion of individuals with the greatest illness burden varied only slightly by PRC (comorbidity score category ≥3: 6%, 5%, and 4%, respectively, among the colorectal, lung, and cervical PRCs). Approximately 45% of the lung PRC cohort had ever smoked compared with 32% of the colorectal PRC cohort and 29% of the cervical PRC cohort. Cohorts based on health plan enrollment (i.e., “enrollment-based sites” comprising all KP sites, MCHS, and part of HFHS) included individuals who did not have a primary care encounter during their first year in the cohort; that proportion varied across enrollment-based sites and PRCs (cervical PRC: 7%; colorectal PRC: 21%; lung PRC: 11%), as did the median number of primary care encounters among individuals who completed at least one visit. Variation in primary care encounters across PRCs was expected, as the colorectal PRC included primarily enrollment-based sites. Length of enrollment during 2010–2017/19 for individuals from enrollment-based sites is shown in Supplementary Tables S2–S4, as is duration of engagement with primary care utilization-based sites with the terminus defined as not utilizing primary care services in 37 months.

At the provider and facility levels, PROSPR II data reflect only providers and facilities that delivered primary care or cancer screening care (e.g., ordering, performing, or evaluating a cancer screening test or procedure) to individuals across PRC cohorts from 2010 to 2017/2019. Among over 96,000 providers meeting this definition, the majority were physicians (both allopathic and osteopathic) at all PRCs [Table 1 (24)]. The fraction of other provider types differed across PRCs (e.g., nurses were a greater fraction of providers in the cervical PRC compared with the colorectal and lung PRCs, and other provider types were most common at the colorectal PRC). Providers with family medicine or internal medicine specialties predominated at all PRCs, followed by providers with multiple/other specialties. PROSPR II data included hospitals operated by PRC health care systems (cervical PRC: n = 8, colorectal PRC n = 39, lung PRC n = 20). The number of facilities varied across PRCs, with a mix of internal facilities (i.e., those operated by the health care system) providing primary care, internal facilities exclusively providing care types other than primary care, and external facilities. Patient, provider, and facility characteristics varied across sites within the cervical, colorectal, and lung PRCs [Supplementary Tables S2–S4 (24)].

Each PRC included three to five health care systems representing diverse care delivery models across numerous geographic regions (Table 1). On the basis of the PROSPR II Organizational Factors Survey data, all health care systems within the cervical and colorectal PRCs had organization-wide policies or guidelines for cancer screening, reported accompanying performance measures, and offered provider- and/or facility-level incentives for achieving specific measures—and most systems had centralized population health teams. In contrast, while most health care systems within the lung PRC used centralized population health teams, only two had organization-wide lung cancer screening policies or guidelines, and none reported performance measures or had provider or facility incentives for lung cancer screening, likely reflecting the lack of national quality measures for lung cancer screening in the United States.

Tests, procedures, and abnormalities in PROSPR II cohorts by research center

The total volume of tests, procedures, abnormal test/procedure results, diagnostic evaluations, and cancer diagnoses by PRC reflects organ-specific screening processes and differences in PRC cohorts [Table 2 (24)]. PRCs identified incident invasive cancers (372 cervical; 24,131 colorectal; 11,205 lung), in situ cancers (911 colorectal; 32 lung), and precancers (13,838 cervical; 554,499 colorectal). The volume of cancer screening process data differed by site within PRCs [Supplementary Tables S5–S7 (24)].

Table 2.

Tests, procedures, and abnormalities among PROSPR cohort members by PROSPR Research Center (2010 through 2017 or 2019).

Cervical Cancer Screening Research CenterColorectal Cancer Screening Research CenterLung Cancer Screening Research Center
2010–20172010–20172010-September 2019
(n)(n)(n)
Tests and procedures (regardless of indication)a 
 Cervical cytology alone 598,707 — — 
 Cervical cytology with a reflex HPV test 21,005   
 Co-test (cervical cytology and HPV test) 218,053 — — 
 HPV test alone 5,251 — — 
 Other cervical cytology and HPV test/unknown cervical testing modality 10,416 — — 
 FIT/gFOBT — 7,503,537 — 
 Flexible sigmoidoscopy — 177,072 — 
 Colonoscopy — 1,694,317 — 
 Low-dose chest CT for LCS (LDCT) — — 27,231 
 Chest CT following an LDCT — — 7,787 
Abnormal/positive tests and proceduresb 
 Abnormal cervical cytology and/or HPV+ 87,263 — — 
 FIT+/gFOBT+ — 353,110 — 
 Colorectal adenoma(s) detected on colonoscopy — 554,499 — 
 Lung-RADS 3 or 4 — — 3,581 
Diagnostic evaluations following abnormal/positive tests or proceduresc 
 Colposcopy 77,303 — — 
 Excisional treatments 6,387 — — 
 Colonoscopy — 332,866 — 
 Lung procedures — — 2,945 
Invasive cancer, in situ cancer, and pre-cancerous diagnosesd 
 Incident invasive cancer diagnoses 372 24,131 11,205 
 Incident in situ cancer diagnoses — 911 32 
 Incident precancerous diagnoses 13,838 554,499 — 
Cervical Cancer Screening Research CenterColorectal Cancer Screening Research CenterLung Cancer Screening Research Center
2010–20172010–20172010-September 2019
(n)(n)(n)
Tests and procedures (regardless of indication)a 
 Cervical cytology alone 598,707 — — 
 Cervical cytology with a reflex HPV test 21,005   
 Co-test (cervical cytology and HPV test) 218,053 — — 
 HPV test alone 5,251 — — 
 Other cervical cytology and HPV test/unknown cervical testing modality 10,416 — — 
 FIT/gFOBT — 7,503,537 — 
 Flexible sigmoidoscopy — 177,072 — 
 Colonoscopy — 1,694,317 — 
 Low-dose chest CT for LCS (LDCT) — — 27,231 
 Chest CT following an LDCT — — 7,787 
Abnormal/positive tests and proceduresb 
 Abnormal cervical cytology and/or HPV+ 87,263 — — 
 FIT+/gFOBT+ — 353,110 — 
 Colorectal adenoma(s) detected on colonoscopy — 554,499 — 
 Lung-RADS 3 or 4 — — 3,581 
Diagnostic evaluations following abnormal/positive tests or proceduresc 
 Colposcopy 77,303 — — 
 Excisional treatments 6,387 — — 
 Colonoscopy — 332,866 — 
 Lung procedures — — 2,945 
Invasive cancer, in situ cancer, and pre-cancerous diagnosesd 
 Incident invasive cancer diagnoses 372 24,131 11,205 
 Incident in situ cancer diagnoses — 911 32 
 Incident precancerous diagnoses 13,838 554,499 — 

Abbreviations: CIN, cervical intraepithelial neoplasia; CT, computerized tomography; FIT, fecal immunochemical test; gFOBT, guaiac fecal occult blood test; HPV, human papillomavirus; LCS, lung cancer screening; LDCT, low-dose chest CT; Lung-RADS, lung imaging reporting and data system; PROSPR, Population-based Research to Optimize the Screening Process.

aHPV test alone reflects different clinical practices over time including follow-up of prior abnormalities and primary HPV screening. LDCT includes codes G0297 and S8032 (and local codes mapped to these codes prior to data transfer). CT following an LDCT includes codes 71250 and 71260.

bLung-RADS 3 or 4 includes codes G0297 and S8032 (and local codes mapped to these codes prior to data transfer).

cColposcopy category includes colposcopies only, biopsies, and endocervical curettage. Excisional treatment is defined as loop electrosurgical excision procedure, conization, and/or excisional treatments not otherwise specified. Colonoscopies include only those following gFOBT+/FIT+ results. Lung procedures include the first procedure following a Lung-RADS 3 or 4 assessment and within 12 months of a screening LDCT and the codes are aligned with Zhao, et al (24).

dInvasive cancer, in situ cancer, and precancerous diagnoses are calculated at the diagnosis level, not the person level, for the entire study population. Colorectal precancerous lesions are defined as colonoscopies where adenoma(s) were detected. Cervical precancerous lesions are defined as CIN II/moderate dysplasia, CIN II-III, CIN III/severe dysplasia/carcinoma in situ (stage 0), high-grade squamous intraepithelial lesion, and adenocarcinoma in situ of the cervix.

Initial changes in cancer screening care at PROSPR II health care systems in response to the COVID-19 pandemic

All 10 PROSPR II healthcare systems completed the International Cancer Screening Network COVID-19 Survey between May 2020 and August 2021. Survey data demonstrated that these health care systems initially experienced a range of disruptions due to the COVID-19 pandemic that impacted cancer screening services (19). All PROSPR II health care systems largely deferred cancer screening at the onset of the pandemic with few exceptions such as continued mailing of home-based fecal immunochemical tests for colorectal cancer screening and, in some systems, opportunistic screening for patients who were due/overdue at the time of an in-person visit for other reasons. Most PROSPR II health care systems also deferred follow-up visits after an abnormal cancer screening result, although there were variations in policies by organ type. In addition, some health care systems observed patients choosing to postpone follow-up visits. Cancer screening research and pilot programs requiring patient contact were halted at all systems except one. All systems retrained or redeployed cancer screening professionals to aid in COVID-19 response efforts, and some systems repurposed their cancer screening infrastructure.

Rich, multilevel data from the PROSPR II consortium will enable investigations of cervical, colorectal, and lung cancer screening processes in community settings to evaluate quality and identify disparities. PROSPR data are unique in both breadth and depth, documenting not only the initial screening step, but also downstream follow-up and treatment, and including detailed test results and histopathologic diagnoses. As expected, we observed differences in patient, provider, and facility characteristics across the cervical, colorectal, and lung PRCs. These differences reflect organ-specific nuances, varying cohort age ranges, and characteristics and sizes of the health care systems in each PRC. Within PRCs, variations in multilevel characteristics across sites are likely due to different geographic catchment areas, types/structures of care delivery systems, and varying populations served across health care systems. However, these detailed and comprehensive data offer a unique opportunity to evaluate patient, provider, facility, and health care system factors impacting the cancer screening care continuum across 10 diverse health care systems.

PROSPR II research agenda

Conceptualizing and measuring multilevel factors

Health services research has described many components of care delivery; however, many gaps remain in our understanding of how these components influence outcomes at the patient, provider, facility, health care system, and community or neighborhood levels (ref. 2; Fig. 1 includes exemplar elements or characteristics of these levels). In particular, scant attention has been given to how health care settings affect processes and outcomes of care. Health care settings influence every step of care delivery, yet effectively describing characteristics of these settings and understanding which characteristics may be associated with outcomes are inhibited by a lack of standardized definitions and measures to facilitate evaluation of care delivery settings and cross-setting comparison. We need to advance conceptualization and measurement of characteristics at higher levels of influence so we can evaluate and improve health care delivery—particularly in the United States, where care is delivered in heterogeneous settings with varying policies and processes.

Figure 1.

Multiple levels of influence on the cancer screening process and exemplar elements or characteristics of these levels. Concepts adapted from Taplin, et al. and Zapka, et al (2, 3).

Figure 1.

Multiple levels of influence on the cancer screening process and exemplar elements or characteristics of these levels. Concepts adapted from Taplin, et al. and Zapka, et al (2, 3).

Close modal

The PROSPR II consortium has begun this work at the facility, health care system, and neighborhood levels. At the facility level, PRC sites identified comparable facility categorizations across health care systems with different structures (e.g., defining analogous facilities within integrated, mixed-model, and integrated safety-net systems). At the system level, sites at each PRC collected contextual data describing organizational structure, clinical data infrastructure, cancer-specific screening policies, and incentive programs related to cancer screening. Neighborhood-level data based on individuals’ residence includes census tract-level geocoded social determinants of health (25)—education, poverty, racial and ethnic segregation indices, and others—that are calculated for all individuals at cohort entry. Evaluating factors at the facility, system, and neighborhood levels will be important to better understand drivers of variation and disparities in cancer screening processes and ultimately, to identify targets for potential intervention.

Conceptualizing and measuring quality

A key focus of PROSPR II is the conceptualization and measurement of cancer screening quality. Optimizing cancer screening by maximizing benefits while reducing harms requires delivery of high-quality care across the screening process continuum. However, quality measures that are applicable across cancer screening processes and capture multiple quality dimensions are currently lacking. Existing quality measures are limited, as they are organ specific and typically focus only on the initial screening step and effectiveness for average-risk individuals (26–32). The Crossing the Quality Chasm report (14) from the National Academy of Medicine characterized quality aims across six dimensions of health care: effectiveness, safety, timeliness, efficiency, equity, and patient centeredness. Using this framework, we aim to conceptualize and define quality measures across these dimensions in the specific context of cancer screening. This work will be further guided by the PROSPR I conceptual model of the cancer screening process to ensure a comprehensive examination of quality across all steps in the continuum (1).

We plan to focus on patient-level outcome and process quality indicators (33) using existing electronic administrative or clinical data available through the course of screening care delivery and administration. PROSPR II data may be used to evaluate the applicability and reliability of these measures across the represented populations and health care systems, but empiric application using data from other settings will also be needed. We anticipate that these quality measures will enable comparisons of cancer screening delivery and outcomes across organ types and health care settings as well as aid implementation of quality improvement strategies to optimize the screening process.

Identifying disparities across the cancer screening continuum

Minoritized populations (4, 5) in the United States experience a disproportionately high burden of cancer mortality for many organ types, including those for which screening is recommended (34–37). These groups include those who live in rural areas, have lower educational attainment, are uninsured or underinsured, experience poverty, and represent specific racial or ethnic groups. Coupled with this high burden of cancer mortality is low access to and uptake of cancer screening. Breast, cervical, colorectal, and lung cancer screening studies have demonstrated racial and ethnic disparities in cancer screening rates, screening quality, and time to follow-up testing or diagnosis among those with abnormal screening tests (11, 12, 38, 39). Cancer screening disparities are also prevalent in rural communities (40, 41), in those with lower socioeconomic status (42, 43), and among populations who lack insurance coverage (44, 45).

Cancer screening disparities are a focus of the PROSPR II research agenda; these disparities must be addressed to optimize the population-level impact of cancer screening and ensure that the benefits of screening reach those who are at the highest risks for cancer mortality. Several recent studies have looked beyond patient-level factors influencing cancer screening disparities to evaluate multilevel factors. These include provider-patient communication strategies (46), clinic-level use of evidence-based interventions (47), and the impacts of policies to expand health care coverage (48). These studies and prior health care services research suggest that multilevel cancer screening interventions have high potential to improve cancer screening outcomes (49). Thus, additional research is needed to understand how multilevel factors—including provider, facility, health care system, and neighborhood factors—impact cancer screening disparities.

Measuring the impact of the COVID-19 pandemic on cancer screening

The COVID-19 pandemic interrupted cancer screening care delivery at PROSPR II health care systems and other U.S. health care systems (50, 51), similar to health care settings worldwide (19). In response to the pandemic, the PROSPR II consortium has contributed to efforts to anticipate and understand its impacts on cancer screening (51, 52), and plans to collect newly important data elements, such as telehealth encounters. Analyses will inform both empiric and modeling studies estimating delays in screening and follow-up of screening abnormalities (52–56) and their consequences. Examples of topics requiring further research include evaluating evidence of cancer stage shift, identifying disparities in returning to screening, and optimizing strategies to address backlogs in cancer screening and diagnostic evaluations. Ultimately, PROSPR II's observational research, including illuminating the extent and impact of COVID-19-related screening reductions and delays, intends to identify points in the screening process to intervene upon and improve.

The PROSPR II consortium is innovating research on cancer prevention and control through its focus on: (i) the entire screening process continuum, (ii) multilevel factors influencing the quality of the screening process in community settings, and (iii) multiple organ types. While most organized screening programs, national surveys, and quality measures assess only the initial screening step, PROSPR II data enable evaluation of the complex screening process that begins with screening and is followed by diagnostic evaluation of abnormal screening results and initiation of treatment in those with confirmed abnormalities. PROSPR II also seeks to understand the multiple levels of influence and their interplay on cancer screening care and outcomes in diverse settings. A unique aspect and major strength of PROSPR II is the effort to understand similarities and differences across cancer screening processes. This can help inform optimization and coherence in care delivery across multiple organ types. We describe the PROSPR II data resources and research agenda to illustrate their value to the cancer screening research community. We invite researchers to collaborate with PROSPR II investigators (15) or advance this research agenda in their own settings.

E.F. Beaber reports grants from NCI during the conduct of the study. A. Kamineni reports grants from NCI during the conduct of the study. A.N. Burnett-Hartman reports grants from NCI at the NIH during the conduct of the study. M. Oliver reports grants from NIH during the conduct of the study. K.A. Rendle reports grants from NCI/NIH during the conduct of the study; grants from Pfizer and AstraZeneca outside the submitted work. R.A. Ziebell reports grants from NCI during the conduct of the study. J. Chubak reports grants from NIH during the conduct of the study; grants from Amgen, Inc. outside the submitted work. D.A. Corley reports grants from NCI during the conduct of the study. R.T. Greenlee reports grants from NCI via Kaiser during the conduct of the study. J.S. Haas reports grants from NCI during the conduct of the study. S. Honda reports grants from NIH/NCI during the conduct of the study; grants from NIH outside the submitted work. C. Neslund-Dudas reports grants from NCI during the conduct of the study. D.P. Ritzwoller reports grants from NIH/NCI during the conduct of the study. J.E. Schottinger reports grants from NIH grant to Southern California Permanente Medical Group during the conduct of the study; grants from NIH SCOLAR grant to SOuthern California Permanente Medical Group outside the submitted work. J.A. Tiro reports grants from NCI/NIH during the conduct of the study. A. Vachani reports grants from NCI/NIH during the conduct of the study; grants from MagArray, Inc., Precyte, Inc., Broncus Medical, and personal fees from Johnson and Johnson outside the submitted work; and A. Vachani is an advisory board member of the Lungevity Foundation (unpaid). No disclosures were reported by the other authors.

E.F. Beaber: Conceptualization, resources, data curation, supervision, investigation, methodology, writing–original draft, project administration, writing–review and editing. A. Kamineni: Conceptualization, resources, data curation, supervision, funding acquisition, investigation, methodology, writing–original draft, project administration, writing–review and editing. A.N. Burnett-Hartman: Resources, data curation, supervision, methodology, writing–original draft, writing–review and editing. B. Hixon: Data curation, formal analysis, writing–review and editing. S.C. Kobrin: Conceptualization, writing–original draft, writing–review and editing. C.I. Li: Resources, funding acquisition, writing–review and editing. M. Oliver: Data curation, formal analysis, methodology, writing–review and editing. K.A. Rendle: Resources, data curation, writing–review and editing. C.S. Skinner: Resources, funding acquisition, writing–review and editing. K. Todd: Data curation, formal analysis, methodology, writing–review and editing. Y. Zheng: Resources, funding acquisition, writing–original draft, writing–review and editing. R.A. Ziebell: Data curation, formal analysis, methodology, writing–review and editing. E.S. Breslau: Writing–review and editing. J. Chubak: Resources, data curation, funding acquisition, writing–review and editing. D.A. Corley: Resources, funding acquisition, writing–review and editing. R.T. Greenlee: Resources, data curation, writing–review and editing. J.S. Haas: Resources, data curation, funding acquisition, writing–review and editing. E.A. Halm: Resources, funding acquisition, writing–review and editing. S. Honda: Resources, data curation, writing–review and editing. C. Neslund-Dudas: Resources, data curation, writing–review and editing. D.P. Ritzwoller: Resources, data curation, funding acquisition, writing–review and editing. J.E. Schottinger: Funding acquisition, writing–review and editing. J.A. Tiro: Resources, data curation, funding acquisition, writing–review and editing. A. Vachani: Resources, funding acquisition, writing–review and editing. V.P. Doria-Rose: Conceptualization, investigation, methodology, writing–review and editing.

The authors thank the participating PROSPR II Research Centers for their collaboration. A list of the PROSPR II investigators and contributing research staff is provided at: http://healthcaredelivery.cancer.gov/prospr/. The authors also thank Maya A. Jackson for administrative assistance, Leigh E. Sheridan for verifying data accuracy, and Chris Tachibana, Ph.D. for editorial assistance.

This work was supported by the NCI at the NIH (UM1CA222035, to J. Chubak, D.A. Corley, E.A. Halm, A. Kamineni, J.E. Schottinger, and C. Sugg Skinner; UM1CA229140, to J.S. Haas, A. Kamineni, and J.A. Tiro; UM1CA221939, to D.P. Ritzwoller and A. Vachani; UM24CA221936, to C.I. Li and Y. Zheng). This article was written as part of the Population-based Research to Optimize the Screening Process (PROSPR II) consortium. The overall aim of PROSPR II is to conduct multisite, coordinated, transdisciplinary research to evaluate and improve cervical, colorectal, and lung cancer screening processes. The sites comprising the three PROSPR II Research Centers reflect the diversity of U.S. health care delivery system organizations.

The views expressed here are those of the authors only and do not necessarily represent the views of the NCI or NIH.

1.
Beaber
EF
,
Kim
JJ
,
Schapira
MM
,
Tosteson
AN
,
Zauber
AG
,
Geiger
AM
, et al
.
Unifying screening processes within the PROSPR consortium: a conceptual model for breast, cervical, and colorectal cancer screening
.
J Natl Cancer Inst
2015
;
107
:
djv120
.
2.
Taplin
SH
,
Anhang Price
R
,
Edwards
HM
,
Foster
MK
,
Breslau
ES
,
Chollette
V
, et al
.
Introduction: understanding and influencing multilevel factors across the cancer care continuum
.
J Natl Cancer Inst Monogr
2012
;
2012
:
2
10
.
3.
Zapka
JG
,
Taplin
SH
,
Solberg
LI
,
Manos
MM
.
A framework for improving the quality of cancer care: the case of breast and cervical cancer screening
.
Cancer Epidemiol Biomarkers Prev
2003
;
12
:
4
13
.
4.
Sensoy
Ö
,
DiAngelo
R
.
Is everyone really equal?: an introduction to key concepts in social justice education
.
New York, NY
:
Teacher's College Press
;
2012
.
5.
American Medical Association [Internet] Chicago: The Association
;
c2022
.
Available from
: https://www.ama-assn.org/about/ama-center-health-equity/advancing-health-equity-guide-language-narrative-and-concepts-0.
6.
NCI
.
PROSPR
;
c2022
.
Available from
: https://healthcaredelivery.cancer.gov/prospr/.
7.
Onega
T
,
Beaber
EF
,
Sprague
BL
,
Barlow
WE
,
Haas
JS
,
Tosteson
AN
, et al
.
Breast cancer screening in an era of personalized regimens: a conceptual model and National Cancer Institute initiative for risk-based and preference-based approaches at a population level
.
Cancer
2014
;
120
:
2955
64
.
8.
Tiro
JA
,
Kamineni
A
,
Levin
TR
,
Zheng
Y
,
Schottinger
JS
,
Rutter
CM
, et al
.
The colorectal cancer screening process in community settings: a conceptual model for the population-based research optimizing screening through personalized regimens consortium
.
Cancer Epidemiol Biomarkers Prev
2014
;
23
:
1147
58
.
9.
Barlow
WE
,
Beaber
EF
,
Geller
BM
,
Kamineni
A
,
Zheng
Y
,
Haas
JS
, et al
.
Evaluating screening participation, follow-up, and outcomes for breast, cervical, and colorectal cancer in the PROSPR Consortium
.
J Natl Cancer Inst
2020
;
112
:
238
46
.
10.
Tosteson
AN
,
Beaber
EF
,
Tiro
J
,
Kim
J
,
McCarthy
AM
,
Quinn
VP
, et al
.
Variation in screening abnormality rates and follow-up of breast, cervical and colorectal cancer screening within the PROSPR Consortium
.
J Gen Intern Med
2016
;
31
:
372
9
.
11.
McCarthy
AM
,
Kim
JJ
,
Beaber
EF
,
Zheng
Y
,
Burnett-Hartman
A
,
Chubak
J
, et al
.
Follow-up of abnormal breast and colorectal cancer screening by race/ethnicity
.
Am J Prev Med
2016
;
51
:
507
12
.
12.
Burnett-Hartman
AN
,
Mehta
SJ
,
Zheng
Y
,
Ghai
NR
,
McLerran
DF
,
Chubak
J
, et al
.
Racial/ethnic disparities in colorectal cancer screening across healthcare systems
.
Am J Prev Med
2016
;
51
:
e107
15
.
13.
Corley
DA
,
Haas
JS
,
Kobrin
S
.
Reducing variation in the “Standard of Care” for cancer screening: recommendations from the PROSPR Consortium
.
JAMA
2016
;
315
:
2067
8
.
14.
Institute of Medicine (US) Committee on Quality Health Care in America
.
Crossing the quality chasm: a new health system for the 21st century
.
Washington, DC
:
National Academy Press
;
2001
.
15.
NCI
.
PROSPR
;
c2021
.
Available from
: https://healthcaredelivery.cancer.gov/prospr/datashare/.
16.
Kamineni
A
,
Tiro
JA
,
Beaber
EF
,
Silverberg
MJ
,
Wheeler
CM
,
Chao
CR
, et al
.
Cervical cancer screening research in the PROSPR I consortium: rationale, methods and baseline findings from a US cohort
.
Int J Cancer
2019
;
144
:
1460
73
.
17.
NCI Department of Health and Human Services, Surveillance, epidemiology, and end results program
.
SEER incidence data, 1973–2019
.
Bethesda, MD
:
NCI
;
2019
. Available from: http://seer.cancer.gov/data/.
18.
Rendle
KA
,
Burnett-Hartman
AN
,
Neslund-Dudas
C
,
Greenlee
RT
,
Honda
S
,
Elston Lafata
J
, et al
.
Evaluating lung cancer screening across diverse healthcare systems: a process model from the lung PROSPR Consortium
.
Cancer Prev Res
2020
;
13
:
129
36
.
19.
Puricelli Perin
DM
,
Elfstrom
KM
,
Bulliard
JL
,
Buron
A
,
Campbell
C
,
Flugelman
AA
, et al
.
Early assessment of the first wave of the COVID-19 pandemic on cancer screening services: The International Cancer Screening Network COVID-19 survey
.
Prev Med
2021
;
151
:
106642
.
20.
Rolland
B
,
Reid
S
,
Stelling
D
,
Warnick
G
,
Thornquist
M
,
Feng
Z
, et al
.
Toward rigorous data harmonization in cancer epidemiology research: one approach
.
Am J Epidemiol
2015
;
182
:
1033
8
.
21.
Charlson
ME
,
Pompei
P
,
Ales
KL
,
MacKenzie
CR
.
A new method of classifying prognostic comorbidity in longitudinal studies: development and validation
.
J Chronic Dis
1987
;
40
:
373
83
.
22.
Deyo
RA
,
Cherkin
DC
,
Ciol
MA
.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases
.
J Clin Epidemiol
1992
;
45
:
613
9
.
23.
Quan
H
,
Sundararajan
V
,
Halfon
P
,
Fong
A
,
Burnand
B
,
Luthi
JC
, et al
.
Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data
.
Med Care
2005
;
43
:
1130
9
.
24.
Zhao
H
,
Xu
Y
,
Huo
J
,
Burks
AC
,
Ost
DE
,
Shih
YT
.
Updated analysis of complication rates associated with invasive diagnostic procedures after lung cancer screening
.
JAMA Netw Open
2020
;
3
:
e2029874
.
25.
NCI Cancer Control & Population Sciences
.
Social Determinants of Health by US Census Tract
;
c2021
.
Available from
: https://healthcaredelivery.cancer.gov/social-determinants/.
26.
Yancy
B
,
Royalty
JE
,
Marroulis
S
,
Mattingly
C
,
Benard
VB
,
DeGroff
A
.
Using data to effectively manage a national screening program
.
Cancer
2014
;
120
:
2575
83
.
27.
NCI Centers for Disease Control and Prevention
;
c2021
.
Available from
: https://www.cdc.gov/cancer/nbccedp/pdf/measuring-cancer-screening-rates-508.pdf.
28.
D'Orsi
CJ
,
Sickles
EA
,
Mendelson
EB
,
Morris
EA
.
ACR BI-RADS Atlas: Breast Imaging Re-porting and Data System
.
Reston, VA, USA
:
American College of Radiology
;
2013
.
29.
Rex
DK
,
Boland
CR
,
Dominitz
JA
,
Giardiello
FM
,
Johnson
DA
,
Kaltenbach
T
, et al
.
Colorectal cancer screening: recommendations for physicians and patients from the U.S. Multi-Society Task Force on Colorectal Cancer
.
Gastrointest Endosc
2017
;
86
:
18
33
.
30.
Perkins
RB
,
Guido
RS
,
Castle
PE
,
Chelmow
D
,
Einstein
MH
,
Garcia
F
, et al
.
2019 ASCCP risk-based management consensus guidelines for abnormal cervical cancer screening tests and cancer precursors
.
J Low Genit Tract Dis
2020
;
24
:
102
31
.
31.
National Committee for Quality Assurance
;
c2021
.
Available from
: https://www.ncqa.org/hedis/measures/.
32.
Mazzone
PJ
,
White
CS
,
Kazerooni
EA
,
Smith
RA
,
Thomson
CC
.
Proposed quality metrics for lung cancer screening programs: a national lung cancer roundtable project
.
Chest
2021
;
160
:
368
78
.
33.
Donabedian
A
.
An introduction to quality assurance in health care
.
New York, NY
:
Ocxford University Press, Inc
;
2003
.
34.
Ward
E
,
Jemal
A
,
Cokkinides
V
,
Singh
GK
,
Cardinez
C
,
Ghafoor
A
, et al
.
Cancer disparities by race/ethnicity and socioeconomic status
.
CA Cancer J Clin
2004
;
54
:
78
93
.
35.
Singh
GK
,
Williams
SD
,
Siahpush
M
,
Mulhollen
A
.
Socioeconomic, rural-urban, and racial inequalities in US cancer mortality: part I-all cancers and lung cancer and part II-colorectal, prostate, breast, and cervical cancers
.
J Cancer Epidemiol
2011
;
2011
:
107497
.
36.
Siegel
RL
,
Miller
KD
,
Fuchs
HE
,
Jemal
A
.
Cancer statistics, 2021
.
CA Cancer J Clin
2021
;
71
:
7
33
.
37.
Zavala
VA
,
Bracci
PM
,
Carethers
JM
,
Carvajal-Carmona
L
,
Coggins
NB
,
Cruz-Correa
MR
, et al
.
Cancer health disparities in racial/ethnic minorities in the United States
.
Br J Cancer
2021
;
124
:
315
32
.
38.
Kasting
ML
,
Wilson
S
,
Zollinger
TW
,
Dixon
BE
,
Stupiansky
NW
,
Zimet
GD
.
Differences in cervical cancer screening knowledge, practices, and beliefs: an examination of survey responses
.
Prev Med Rep
2017
;
5
:
169
74
.
39.
Lake
M
,
Shusted
CS
,
Juon
HS
,
McIntire
RK
,
Zeigler-Johnson
C
,
Evans
NR
, et al
.
Black patients referred to a lung cancer screening program experience lower rates of screening and longer time to follow-up
.
BMC Cancer
2020
;
20
:
561
.
40.
Kurani
SS
,
McCoy
RG
,
Lampman
MA
,
Doubeni
CA
,
Finney Rutten
LJ
,
Inselman
JW
, et al
.
Association of neighborhood measures of social determinants of health with breast, cervical, and colorectal cancer screening rates in the US Midwest
.
JAMA Netw Open
2020
;
3
:
e200618
.
41.
Rohatgi
KW
,
Marx
CM
,
Lewis-Thames
MW
,
Liu
J
,
Colditz
GA
,
James
AS
.
Urban-rural disparities in access to low-dose computed tomography lung cancer screening in missouri and illinois
.
Prev Chronic Dis
2020
;
17
:
E140
.
42.
Lian
M
,
Schootman
M
,
Yun
S
.
Geographic variation and effect of area-level poverty rate on colorectal cancer screening
.
BMC Public Health
2008
;
8
:
358
.
43.
Goding Sauer
A
,
Siegel
RL
,
Jemal
A
,
Fedewa
SA
.
Current prevalence of major cancer risk factors and screening test use in the United States: disparities by education and race/ethnicity
.
Cancer Epidemiol Biomarkers Prev
2019
;
28
:
629
42
.
44.
Fedewa
SA
,
Sauer
AG
,
DeSantis
C
,
Siegel
RL
,
Jemal
A
.
Disparities in cancer screening by occupational characteristics
.
Prev Med
2017
;
105
:
311
8
.
45.
Sabatino
SA
,
Thompson
TD
,
White
MC
,
Shapiro
JA
,
de Moor
J
,
Doria-Rose
VP
, et al
.
Cancer screening test receipt - United States, 2018
.
MMWR Morb Mortal Wkly Rep
2021
;
70
:
29
35
.
46.
Kindratt
TB
,
Dallo
FJ
,
Allicock
M
,
Atem
F
,
Balasubramanian
BA
.
The influence of patient-provider communication on cancer screenings differs among racial and ethnic groups
.
Prev Med Rep
2020
;
18
:
101086
.
47.
Sharma
KP
,
DeGroff
A
,
Scott
L
,
Shrestha
S
,
Melillo
S
,
Sabatino
SA
.
Correlates of colorectal cancer screening rates in primary care clinics serving low income, medically underserved populations
.
Prev Med
2019
;
126
:
105774
.
48.
Lam
MB
,
Phelan
J
,
Orav
EJ
,
Jha
AK
,
Keating
NL
.
Medicaid expansion and mortality among patients with breast, lung, and colorectal cancer
.
JAMA Netw Open
2020
;
3
:
e2024366
.
49.
Zapka
J
,
Taplin
SH
,
Ganz
P
,
Grunfeld
E
,
Sterba
K
.
Multilevel factors affecting quality: examples from the cancer care continuum
.
J Natl Cancer Inst Monogr
2012
;
2012
:
11
9
.
50.
Mast
C
,
Munoz del Rio
A
.
Delayed cancer screenings - a second look
.
Epic Health Research Network
;
2020
.
51.
National Cancer Institute's Prospr Consortium
,
Corley
DA
,
Sedki
M
,
Ritzwoller
DP
,
Greenlee
RT
,
Neslund-Dudas
C
, et al
.
Cancer screening during the coronavirus disease-2019 pandemic: a perspective from the National Cancer Institute's PROSPR Consortium
.
Gastroenterology
2021
;
160
:
999
1002
.
52.
Croswell
JM
,
Corley
DA
,
Lafata
JE
,
Haas
JS
,
Inadomi
JM
,
Kamineni
A
, et al
.
Cancer screening in the U.S. through the COVID-19 pandemic, recovery, and beyond
.
Prev Med
2021
;
151
:
106595
.
53.
Burger
EA
,
Jansen
EE
,
Killen
J
,
Kok
IM
,
Smith
MA
,
Sy
S
, et al
.
Impact of COVID-19-related care disruptions on cervical cancer screening in the United States
.
J Med Screen
2021
;
28
:
213
6
.
54.
Smith
MA
,
Burger
EA
,
Castanon
A
,
de Kok
I
,
Hanley
SJB
,
Rebolj
M
, et al
.
Impact of disruptions and recovery for established cervical screening programs across a range of high-income country program designs, using COVID-19 as an example: a modelled analysis
.
Prev Med
2021
;
151
:
106623
.
55.
de Jonge
L
,
Worthington
J
,
van Wifferen
F
,
Iragorri
N
,
Peterse
EFP
,
Lew
JB
, et al
.
Impact of the COVID-19 pandemic on faecal immunochemical test-based colorectal cancer screening programmes in Australia, Canada, and the Netherlands: a comparative modelling study
.
Lancet Gastroenterol Hepatol
2021
;
6
:
304
14
.
56.
Rutter
CM
,
Kim
JJ
,
Meester
RGS
,
Sprague
BL
,
Burger
EA
,
Zauber
AG
, et al
.
Effect of time to diagnostic testing for breast, cervical, and colorectal cancer screening abnormalities on screening efficacy: a modeling study
.
Cancer Epidemiol Biomarkers Prev
2018
;
27
:
158
64
.

Supplementary data