Cancer Causes & Control (2021) 32:1129–1148 https://doi.org/10.1007/s10552-021-01461-x ORIGINAL PAPER Theory, methods, and operational results of the Young Women’s Health History Study: a study of young‑onset breast cancer incidence in Black and White women Ellen M. Velie1,2  · Lydia R. Marcus1,2  · Dorothy R. Pathak3  · Ann S. Hamilton4 · Ralph DiGaetano5 · Ron Klinger5 · Bibi Gollapudi5 · Richard Houang6 · Nicole Carnegie7 · L. Karl Olson8 · Amani Allen9 · Zhenzhen Zhang10 · Denise Modjesk4 · Gwendolyn Norman11 · Darek R. Lucas1,2 · Sapna Gupta12 · Hallgeir Rui13 · Kendra Schwartz14 Received: 23 November 2020 / Accepted: 11 June 2021 / Published online: 22 July 2021 © The Author(s) 2021 Abstract Purpose The etiology of young-onset breast cancer (BC) is poorly understood, despite its greater likelihood of being hormone receptor-negative with a worse prognosis and persistent racial and socioeconomic inequities. We conducted a population- based case–control study of BC among young Black and White women and here discuss the theory that informed our study, exposures collected, study methods, and operational results. Methods Cases were non-Hispanic Black (NHB) and White (NHW) women age 20–49 years with invasive BC in metro- politan Detroit and Los Angeles County SEER registries 2010–2015. Controls were identified through area-based sampling from the U.S. census and frequency matched to cases on study site, race, and age. An eco-social theory of health informed life-course exposures collected from in-person interviews, including socioeconomic, reproductive, and energy balance fac- tors. Measured anthropometry, blood (or saliva), and among cases SEER tumor characteristics and tumor tissue (from a subset of cases) were also collected. Results Of 5,309 identified potentially eligible cases, 2,720 sampled participants were screened and 1,812 completed inter- views (682 NHB, 1140 NHW; response rate (RR): 60%). Of 24,612 sampled control households 18,612 were rostered, 2,716 participants were sampled and screened, and 1,381 completed interviews (665 NHB, 716 NHW; RR: 53%). Ninety-nine% of participants completed the main interview, 82% provided blood or saliva (75% blood only), and SEER tumor characteristics (including ER, PR and HER2 status) were obtained from 96% of cases. Conclusions Results from the successfully established YWHHS should expand our understanding of young-onset BC etiol- ogy overall and by tumor type and identify sources of racial and socioeconomic inequities in BC. Keywords Breast cancer · Young-onset breast cancer · Epidemiology · Life-course · Health status disparities · Premenopause Abbreviations CSP Cancer Surveillance Program AAPOR A merican Association for Public Opinion FFQ Food Frequency Questionnaire BC Breast cancer HR H ormone receptor BIA Bioelectrical impedance analysis IRB Institutional review board CAPI Computer-assisted personal main interview MCW Medical College of Wisconsin CARE Contraceptive and Reproductive Endpoints MDCSS Metro Detroit Cancer Surveillance System CPHS California Committee for the Protection of MSU Michigan State University Human Subjects NHB Non-Hispanic Black CRIC Cancer Research Informatics Core NHW Non-Hispanic White QC Quality control RCA Rapid case ascertainment * Ellen M. Velie RR Response rate velie@uwm.edu SEER Surveillance: Epidemiology and End Results Extended author information available on the last page of the article US U nited States Vol.:(012 3456789) 1130 Cancer Causes & Control (2021) 32:1129–1148 USC University of Southern California most affluent tracts [18]. The incidence of BC overall has UWM University of Wisconsin—Milwaukee also increased among women residing in the most disad- W W eighted vantaged counties more rapidly than among women in the WSU Wayne State University most affluent counties: from 1981–1990 to 2001–2010, inci- YWHHS Young Women Health History Study dence increased by 15% in the most disadvantaged and only by 9% in the most affluent counties [24]. Black and White women residing in the most disadvantaged counties (> 20% Introduction poverty) also had a higher prevalence of HR- BC relative to women residing in wealthier counties (< 10% poverty) in In the United States (US), nearly one quarter of annual breast 2004–2007 [25]; this is most pronounced for NHB compared cancer (BC) cases occur in women under 50 years of age to NHW women < 50 years old (HR-/HR + ratio = 1.51, 95% and the incidence is increasing [1, 2]. The etiology of BC Confidence Interval (CI): 1.20, 1.90) [25]. Similar patterns varies by age [3, 4] and is poorly understood in young-onset are seen at the census-tract level: women residing in tracts BC [3, 5–8]. Breast tumors are also now recognized to have with intermediate and low compared to high socioeconomic different histopathologic and molecular characteristics with status index had 1.81 (95% CI 1.20, 2.71) and 1.95 (95% heterogeneous etiology, prognosis, and treatment [9–12]. 1.27–2.99) relative risk ratios for TNBC, respectively, in Tumors in young women are also more likely to present at a 2005–2017 [21]. later stage, have a worse prognosis, and be hormone recep- Few modifiable factors have been identified to inform tor-negative (HR-)[13–15]. Non-Hispanic White (NHW) and BC prevention strategies [26], particularly in young women non-Hispanic Black (NHB) women have the highest inci- [9, 27–31] and by tumor type [9, 13], or to explain racial dences of BC in the U.S. [2] and racial and socioeconomic and socioeconomic inequities in BC incidence [32–34]. We inequities in BC also persist [16–18]. conducted a population-based case–control study of BC risk Racial inequities exist in the U.S. in overall BC mor- among NHB and NHW women aged < 50 years old from tality and incidence, particularly in younger women, and diverse socioeconomic backgrounds in the US: The Young there are unequal distributions of tumor subtypes. Over- Women’s Health History Study (YWHHS). Our research is all BC mortality was 40% higher in NHB compared to informed by an eco-social theory of health, which situates NHW women during 2013–2017 [17] and this inequity is health outcomes—particularly those between groups— particularly pronounced among women < 50 years of age, within a complex socio-historical context; eco-social theory where mortality was 82% higher in NHB compared to seeks to identify the pathways through which that context is NHW women in 2018 [19]. Though overall incidence of embodied [35, 36]. Further, we recognize racism is a potent BC among NHB women has historically been lower than social determinant that continues to regulate differences in NHW women, rates are now nearly equal [2], and among the exposures to socioeconomic and other opportunities by race, youngest women (aged < 40 years) incidence rates have con- thereby contributing to racial health inequities in the U.S. sistently been higher among Black women [2, 20]. Among [37, 38]. We hypothesize that socio-cultural factors related to women < 50 years of age, NHB women also had a 90% race and socioeconomic position determine exposures over higher incidence of the most aggressive HR-/HER2- (i.e., the life-course (e.g., reproductive and energy balance fac- triple-negative (TNBC)) tumors compared to NHW women tors) that modify biology and, in turn, risk for young-onset in 2012–2016 [2]. Studies examining racial residential segre- BC tumor types (Fig. 1) [22, 36, 38–44]. In this paper, we gation have observed that among Black women, both a lower document details of the YWHHS study design, life-course [21] and higher [22] proportion of Black residents in census measures collected, data collection methods, response and tracts is associated with a higher odds of TNBC. Everyday cooperation rates, and provide a description of our final experiences of discrimination have also been associated with study population. increased incidence of BC among Black women, particularly among those aged < 50 years [23], potentially contributing to an explanation for observed patterns of racial residential Methods segregation and TNBC [22]. Socioeconomic inequities in BC mortality and incidence Overall study objectives also exist. Poorer women have historically had lower mor- tality from BC at all ages [18]; however, mortality from BC The primary objectives of the YWHHS were to provide has steadily increased since 1950 among women residing in insight into modifiable early life and life-course factors asso- disadvantaged census tracts and decreased among women ciated with young-onset (< 50 years) BC risk and to under- in affluent tracts[18]) such that, by 2013, BC mortality in stand racial and socioeconomic inequities in BC risk in the the most disadvantaged tracts was 6% higher than in the U.S. [40, 44–47]. We are investigating: (1) the association 1 3 Cancer Causes & Control (2021) 32:1129–1148 1131 Fig. 1 YWHHS conceptual framework: socio-historical context, life-course reproductive and energy balance factors, and breast cancer risk among young non-Hispanic Black and White women between early life and life-course factors and risk for BC the socio-historic context of race/ethnicity (hereafter “race”) overall and by tumor subtypes among young NHB and NHW and life-course socioeconomic position (SEP) on BC risk, women [9, 27–32], (2) the potentially modifying effects of and have also (3) created a bio-repository of blood (or saliva) 1 3 1132 Cancer Causes & Control (2021) 32:1129–1148 and breast tumor tissue for current and future study of the Eligibility criteria (see Table 1) contribution of biomarkers, gene-environment interactions, and gene expression on BC risk in young women. Study tracking system Overall study design A centralized computer system that tracked all correspond- BC cases were identified from the metropolitan Detroit ing study data and biospecimens was adapted and managed (Oakland, Wayne, and Macomb counties) and Los Ange- for YWHHS by the USC Cancer Research Informatics Core les County Surveillance, Epidemiology and End Results (CRIC). (SEER) registries diagnosed between 2010 and 2015. Con- trols were identified through area-based sampling from the Ascertainment, sampling, recruitment, and screening 2010 Census and matched to cases by study site, age, and race. Primary data collected included: (1) an in-person com- Ascertainment, sampling, recruitment, and screening activi- puter-assisted personal interview (CAPI) conducted with a ties for cases and controls are outlined in Fig. 2. life history calendar, (2) anthropometric measurements, (3) blood collection (or saliva when not available) and related Cases questionnaire, (4) SEER tumor type information, including ER, PR and HER2 status, and (5) breast tumor tissue col- Potentially eligible cases were identified by the Metropolitan lected from participants’ BC surgeries. Additional collected Detroit Cancer Surveillance System (MDCSS) SEER regis- data included: (6) an interviewer-completed built environ- try and the LA County Cancer Surveillance Program (CSP) ment survey of participants’ neighborhoods, (7) a survey SEER registry. For both sites, cases were identified through completed by participants’ primary childhood caregiver, and rapid case ascertainment (RCA), which aims to identify (8) childhood photos of body size. We also requested (9) per- cases within 3–6 months after diagnosis. mission to obtain information from the health department(s) where women gave birth and (10) where she was born, and Case sampling. We sampled from all eligible NHW (11) most recent mammogram reports from healthcare pro- 45–49  years of age due to budgetary constraints. Given viders. Participation in the main study questionnaire was that there is a paucity of studies among NHB women, the necessary for enrollment; all other study components were youngest women (< 45  years of age), and women diag- optional. This study protocol was approved by the Institu- nosed with estrogen receptor-negative tumors, we retained tional Review Boards at the University of Wisconsin—Mil- all NHB women diagnosed 20–49 years of age, all NHW waukee (UWM); Michigan State University (MSU); Wayne women 20–44 years of age, and among NHW women aged State University (WSU); the Michigan Department of Com- 45–49  years, oversampled women with estrogen receptor- munity Health; University of Southern California (USC); the negative tumors. Thus, all eligible NHB cases 20–49 years California Committee for the Protection of Human Subjects of age and NHW cases 20–44 years of age were included, (CPHS); and for the Medical College of Wisconsin (MCW), and a sample of NHW cases aged 45–49 years (n = 829 of IRB oversight was deferred to UWM. The California Cancer 2,527 Detroit; n = 883 of 2,782 LA), sampled as follows: Registry also approved the study. between 09/01/2010 and 08/31/2012 30.5% of all NHW 45–49 year old cases; between 08/31/2012 and 08/31/2015 Study organization 84.5% of ER- cases and 40.8% of ER + tumors. The YWHHS Coordinating Center (initially hosted at MSU, Case screener interview. All sampled cases were screened moved to UWM in 2014) were responsible for study design, to determine final eligibility status. Cases not success- development, and oversight of the study tracking system. fully screened by a study site team were checked against Westat, a research services corporation, and study collabora- the updated SEER Registry to determine eligibility status. tors developed the control sampling design, oversaw identi- Cases initially sampled were considered ineligible for the fication and recruitment of control participants, and created following reasons: not U.S.-born (n = 373), self-identified final study sample weights. Final recruitment, in-person as neither White nor Black (n = 153), self-identified as His- interviews, and biospecimen collection were conducted at panic (n = 151), previous cancer diagnosis (n = 117), resided two field sites: Los Angeles County (at USC) and metro- outside of the study areas at reference date (see definition politan Detroit (at WSU). A community advisory panel was of reference date in Table 1; n = 50), tumor had ineligible assembled and consulted about data collection materials and histology (n = 44), did not speak English (n = 29), updated study methodologies. age or reference date was out-of-range (n = 17), physically or mentally unable to complete the interview (n = 14), or 1 3 Cancer Causes & Control (2021) 32:1129–1148 1133 Table 1 Eligibility criteria for cases of breast cancer and controls, Young Women’s Health History Study Cases Controls Eligibility criteria: Eligibility criteria: 1. Identified as female by SEER Registry 1. Identified as female by household roster 2. 20–49 years of age at reference date 2. 20–49 years of age at reference date 3. Race/ethnicity: self-reported Non-Hispanic Black or non-Hispanic 3. Race/ethnicity: self-reported Non-Hispanic Black or non-Hispanic White1 White2 4. Resident of metropolitan Detroit or Los Angeles County at reference 4. Resident of metropolitan Detroit or Los Angeles County at reference date date 5. Born in the U.S 5. Born in the U.S 6. No previous diagnosis of in situ or invasive breast cancer 6. No previous diagnosis of in situ or invasive breast cancer 7. No previous cancer diagnosis except for cervical in situ or common 7. No previous cancer diagnosis except for cervical in situ or common skin cancer skin cancer 8. Not residing in an institution (e.g., prison, shelter, nursing home) at 8. Not residing in an institution (e.g., prison, shelter, nursing home) at reference date reference date 9. Physically and mentally able to complete the interview 9. Physically and mentally able to complete the interview 10. Able to complete interview in English 10. Able to complete interview in English 11. Diagnosed with histologically confirmed invasive BC by SEER between 1 September 2010 and 31 August 2015 (ICD-9-CM code C50.0-C50.9, excluded breast lymphoma, Paget’s disease, mesen- chymal tumors, including sarcomas, and hemangiosarcoma’s of the breast: 8800–8805, 8540/3, 8541/3, 8542/3, 8543/3, 9000–9805, 9820–9989) Reference date: Reference date: Date of first microscopic cytologic/histologic BC diagnosis Four months prior to screening 1 For cases, race/ethnicity was initially determined by SEER– derived from medical report or hospital admissions. Participants with “Hispanic” or “Arab American” last names based on SEER last name lists [74] at both study sites and participants with “Asian” last names based on SEER lists in LA County were considered ineligible 2 For controls, race/ethnicity was initially reported on the household roster (potentially by proxy) based on Census 2010 as “Hispanic or Latina origin” and as many races as applied: “Black/African American, White, Asian, Native Hawaiian/other Pacific Islander, American Indian/Alaska Native, or Other [51]. Westat also applied SEER Hispanic surname lists in LA. Final race/ethnicity determination was self-reported on the screener. Participants were asked to report their ethnicity as “Hispanic or Latina origin,” and then to select the race they identified with most: “Black or African American; White; American Indian or Native American or Alaska Native; Arab American or Chaldean; East Asian or South- east Asian; Asian Indian or South Asian; Native Hawaiian or other Pacific Islander; Some Other Group; Refused; Don’t know.” Participants who did not identify as “Hispanic or Latina origin” and those who identified as “Black or African American” or “White” were considered eligible NOTE: We use the terms Black and African American interchangeably [75] institutionalized at reference date (n = 7). Two percent of Coordinating Center to be entered into the study tracking cases were ineligible for screening for one or more of these database for recruitment. reasons. In Detroit, a letter was sent to each eligible case’s physician before cases were contacted; if the physician did Control sampling. A three-stage area probability sample not respond within three weeks the case could be contacted, was conducted to provide coverage of metropolitan Detroit except for a few Detroit hospitals that required active physi- and LA County from which YWHHS case participants were cian approval. identified (see Supplemental Materials). The first stage of sample selection was that of PSUs (Primary Sampling Controls Units) consisting of one or more Census blocks as identi- fied in the U.S. Census conducted in 2010. Within sampled YWHHS investigators and the Westat team developed the PSUs, the second stage was the sampling of approximately area-based control sampling strategy and Westat developed 24,000 + addresses from listings based on addresses served the statistical sampling methodology [48, 49]. Westat also by the U.S. Postal Service. Households within occupied oversaw control identification and recruitment, household sampled addresses were rostered to identify members rostering, screener interviews, and initiated control recruit- who were potential controls for the study. The third stage ment efforts. Once potentially eligible controls were identi- of sample selection involved randomly selecting women fied, their contact information was provided to the YWHHS from among those potentially eligible. The sampling rates 1 3 1 134 Cancer Causes & Control (2021) 32:1129–1148 Fig. 2 Control and case sampling, eligibility, and recruitment: Young Women’s Health History Study employed were designed to obtain a set of controls that were and control women. After sending the introductory letter, frequency matched to the expected case distribution within study staff (WSU/USC) telephoned women to determine study site by race (NHB/NHW) and 5-year age intervals. (cases) or confirm (controls) eligibility, answer questions, and identify a location and time for an in-person interview. Control household roster. A total of 24,612 households were Women not reached by phone were sent follow-up letters sampled (Table  2) and 21,668 were determined eligible and reminder postcards, and, in some cases, in-person visits. for roster. An introductory letter, brief roster, and a $2 bill Women who declined to participate were asked to complete were mailed to all sampled residential addresses. The same a brief questionnaire about demographic characteristics to follow-up household contact recruitment protocol was then characterize non-respondents. used as the National Health and Nutrition Examination Sur- vey [50]. A total of 18,612 household were rostered. The In‑person interview scheduling and informed consent. Study roster asked the initials/name, age, and race/ethnicity of all participants were interviewed at their selected location. adult women 20–50 years old in the household (see Supple- Prior to interview, participants were mailed a confirmation mentary Materials for additional details). letter and their interviewer’s business card with a photo- graph. Before the interview, the participant was asked to Control screener interview. An in-person screener interview read and sign a consent form that described the study and was conducted to determine the final eligibility of poten- participant rights and safeguards; it also requested permis- tially eligible women identified and sampled from the house- sion to conduct the interview and each component of the hold roster. Those who completed the screener received $5. study. Women were informed they could refuse any ques- Respondents willing to participate or interested in learning tions and terminate the interview at any time. Women who more were asked to provide their contact information for a had a mammogram were asked to complete a separate con- study site (WSU/USC) interviewer to contact them. sent form that requested permission to obtain information from her healthcare provider about her last mammogram Data collection before reference date. Additionally, case participants were asked to provide consent to obtain tumor tissue sampled In‑home case–control interview recruitment. An introduc- at the time of diagnosis or thereafter. A thank you gift of tory letter and study brochure were sent to all sampled case 1 3 Cancer Causes & Control (2021) 32:1129–1148 1135 1 3 Table 2 Overall ascertainment numbers by race and site, Young Women’s Health History Study Detroit LA Both Sites NHB NHW Other Total NHB NHW Other Total NHB NHW Other Total N N N N N N N N N N N N Cases All potentially eligible 738 1,721 68 2,527 589 2,040 153 2,782 1,327 3,761 221 5,309 cases identified by S EERa Sampled from potentially 738 1,318 61 2,117 589 1,428 135 2,152 1,327 2,746 196 4,269 eligible cases Not screened or incom- 311 665 14 990 163 378 18 559 474 1,043 32 1,549 plete screener SEER Registry Deter- 11 55 14 80 22 116 18 156 33 171 32 236 mined Ineligibleb Hospital/physician refusal 34 48 0 82 – – – – 34 48 0 82 or active physician approval not receivedc Participant not reached 94 407 0 501 123 232 0 355 217 639 0 856 Incomplete screener (eligi- 172 155 0 327 18 30 0 48 190 185 0 375 ble/potentially eligible) Completed screener 427 653 47 1,127 426 1,050 117 1,593 853 1,703 164 2,720 Eligible 379 564 0 943 364 706 0 1,070 743 1,270 0 2,013 Ineligible 48 89 47 184 62 344 117 523 110 433 164 707 Eligible/potentially 679 1,174 0 1,853 505 968 0 1,473 1,184 2,142 0 3,326 eligible (screened & not screened) Not interviewed 296 638 0 934 206 374 0 580 502 1,012 0 1,514 Not screened—hospital/ 34 48 0 82 – – – – 34 48 0 82 physician refusal or active physician approval not receivedc Died before interview 26 34 0 60 40 24 0 64 66 58 0 124 Too ill to conduct inter- 4 4 0 8 6 9 0 15 10 13 0 23 view Participant refused 122 225 0 347 84 168 0 252 206 393 0 599 Proxy said participant 4 8 0 12 3 9 0 12 7 17 0 24 refused Moved away from study 6 15 0 21 11 38 0 49 17 53 0 70 area Unavailable for interview 100 304 0 404 62 126 0 188 162 430 0 592 Eligible and interviewed 383 536 0 919 299 594 0 893 682 1,130 0 1,812 1136 Cancer Causes & Control (2021) 32:1129–1148 1 3 Table 2 (continued) Detroit LA Both Sites NHB NHW Other Total NHB NHW Other Total NHB NHW Other Total N N N N N N N N N N N N Controls  Households sampled – – – 9,994 – – – 14,618 – – – 24,612  Total households ineligi- – – – 1,418 – – – 1,526 – – – 2,944 ble for roster  Vacant households – – – 1,284 – – – 709 – – – 1,993  Non-dwelling unit, e.g., – – – 127 – – – 278 – – – 405 institutional quarters  Excluded 60% of – – – – – – – 485 – – – 485 potential Hispanic householdsd  Language barrier – – – 7 – – – 54 – – – 61  Total households eligible – – – 8,576 – – – 13,092 – – – 21,668 for roster  Households with no – – – 774 – – – 2,282 – – – 3,056 responsee  Locked buildings/gated – – – 91 – – – 1,197 – – – 1,288 communities  Households rostered – – – 7,802 – – – 10,810 – – – 18,612  By mail – – – 2,802 – – – 3,333 – – – 6,135  In personf – – – 5,000 – – – 7,427 – – – 12,477  Households with par- – – – 1,709 – – – 1,512 – – – 3,221 ticipants sampled for s creeningg  Households not screened – – – 310 – – – 292 – – – 602  Participants potentially 643 1,038 73 1,761 624 897 125 1,653 1,929 1,261 198 3,414 eligible & sampledh,i  Potentially eligible par- 98 251 7 362 127 190 13 336 222 441 20 698 ticipants—incomplete screeneri,j  Completed potential 545 787 66 1,399 497 707 112 1,317 1,039 1,488 178 2,716 participant screeneri,k  Ineligible participants 72 150 66 288 89 238 112 440 161 388 178 728  Eligible or potentially 473 638 0 1,111 408 469 0 877 881 1,107 0 1,988 eligible participants  Didn’t agree to be con- 12 18 0 30 10 15 0 25 22 33 0 55 tacted  Eligible 12 17 0 29 10 15 0 25 22 32 0 54 Cancer Causes & Control (2021) 32:1129–1148 1137 1 3 Table 2 (continued) Detroit LA Both Sites NHB NHW Other Total NHB NHW Other Total NHB NHW Other Total N N N N N N N N N N N N  Unknown eligibility 0 1 0 1 0 0 0 0 0 1 0 1 status  Agreed to be contacted 461 620 0 1,081 398 454 0 852 859 1,074 0 1,933  Not interviewed 109 255 0 364 85 103 0 188 194 358 0 552  Found to be ineligible 1 3 0 4 5 3 0 8 6 6 0 12  Died before interview 0 2 0 2 0 0 0 0 0 2 0 2  Too ill to conduct 0 1 0 1 0 1 0 1 0 2 0 2 interview  Participant refused 51 147 0 198 48 67 0 115 99 214 0 313  Proxy said participant 2 4 0 6 3 5 0 8 5 9 0 14 refused  Moved away from study 8 10 0 18 3 9 0 12 11 19 0 30 area  Unavailable for interviewl 47 88 0 135 26 18 0 44 73 106 0 179  Could not be located 1 2 0 3 1 2 0 3 3 3 0 6  Eligible & interviewedm 352 365 0 717 313 351 0 664 665 716 0 1,381 a Excludes NHB and NHW cases < 45 years of age identified by the SEER Registry post the study recruitment period (n = 146; DT: n = 74; LA: n = 72) b Sampled/potentially eligible cases who did not complete a telephone screener and were determined to be ineligible based on SEER information c Active physician approval required by specific hospitals among a subset of case participants in Detroit d For efficiency, 60% of households identified by the Westat address list vendor as likely to include at least one “Hispanic” adult were randomly excluded. Information from the other 40% was used to impute adjusted sampling values e Non-response households include those that refused, that were not reached after maximal contacts, that were locked buildings staff were unable to enter, or where language barriers existed f 7% and 18% of in-person rosters were completed by neighbors in Detroit and Los Angeles, respectively g Households containing more than one potentially eligible and sampled woman (Detroit: 49; LA: 121) h Of sampled potentially eligible participants, five lacked “race” values, 2 reported not knowing their self-selected “race,” and seven refused to report a “race” value i If participants lacked a self-reported “race” value at the screener level, their reported “race” value from the household roster was used instead j Of potentially eligible participants who did not complete a screener, 4 were missing “race” values, 1 reported not knowing their self-selected “race,” and 6 refused to report a “race” value k Of participants who completed a screener, 1 reported not knowing their self-selected “race” and 1 refused to report a “race” value l Includes participants lost (n = 6) or unable to contact to schedule an interview (n = 171) m Households containing more than one eligible and interviewed participant (Detroit: 20; LA: 43) 1138 Cancer Causes & Control (2021) 32:1129–1148 $75, which was later increased to $100, was provided for identified by digitally masking the participant’s eyes/face, if the main interview. requested. Built environment survey. Interviewers conducted a survey of neighborhood characteristics, primarily at the Main questionnaire. The YWHHS questionnaire captured time of the interview [59, 60]. Surveys not completed by the information about energy balance factors (e.g., childhood end of study recruitment (6.5%) were conducted remotely and adult diet, physical activity, and adult body size), fac- via Google Maps Street View using photos collected at the tors known to affect life-course energy balance (e.g., food date closest to the interview date [61]. Primary caregiver security, sleep patterns, built environment), known risk fac- survey Participants were asked to mail their primary child- tors for BC (e.g., reproductive and family history), as well hood caregiver a brief survey. Caregivers were given $10. as race/ethnicity and life-course socioeconomic indicators. The survey included respondent’s demographics, biologic Collected information related to race/ethnicity includes self- mother’s pregnancy with the participant, and the study par- reported race and Hispanic ethnicity, as well as the race/ ticipant’s childhood body size, physical activity, and SEP. ethnicity others typically ascribe to the participant. We also asked about early life discrimination, experiences of every- Biospecimen collection day discrimination and the source of discrimination. Life- course socioeconomic indicators include residential history, Blood. All study participants were asked to provide a blood household percent poverty (HPP), educational attainment, sample. Samples were collected by a phlebotomist, gener- and occupational status [51, 52]. HPP was calculated using ally at the second visit (96%, 4% at first visit). Phlebotomists household net income adjusted for household size. Other attempted to obtain 30 mL (approximately 2 tablespoons) factors associated with social context collected include life- collected in four 10-mL vacutainers: two with no additive course experiences of adversity (including childhood expe- and two with EDTA. For cases, our protocol indicated sam- riences), financial status and use of governmental subsidies, ples should not be collected until at least two months after food insecurity, occupational status, and health insurance last treatment (average days post treatment = 376 days; 95% status. Other information on factors potentially associated CI 353.9, 398.6). Participants who provided blood samples with BC risk include prenatal exposures, medical history, were originally given a $20 thank you gift, which was later non-steroidal anti-inflammatory medication use, contracep- increased to $25. Samples were processed at the MSU Cyto- tive use, hormone medication use, fertility history, and life- genics laboratory and MCW Tissue Bank. course personal and secondhand tobacco exposure, as well as alcohol use. Study questions were developed based on Blood Questionnaire. Phlebotomists administered a ques- existing questionnaires [53–57]. tionnaire to each participant at the time of blood draw. Multiple tools were used throughout the questionnaire to Questions addressed recent medication use; medical his- assist participants with recall, including a life history calen- tory; menstrual, pregnancy, and lactation status; and recent dar of key life events [58], showcards, which also provided food, beverage, alcohol, and tobacco consumption. a non-verbal method of responding to sensitive questions, and a photobook of oral contraceptive, hormone, and thyroid Menstrual calendar. During the main interview, if a partici- medications [58]. pant reported menstruating within the past year and if they consented to have their blood drawn, they were asked to Additional components of  the  in‑person interview: anthro‑ complete a menstrual calendar that indicated each day they pometric assessment. Height, weight, waist circumference, experienced menstrual bleeding until the date their blood and body composition (assessed by Tanita bioelectrical was drawn. If participants had not completed this calendar impedance analysis (BIA)) were measured. Diet. A modi- at the time of blood draw, the phlebotomist completed it fied version of the full 100-item Block Food Frequency with the participant for the preceding two months. Questionnaire (FFQ) was developed by NutritionQuest (Berkeley, CA) with the study PI (Velie) to capture total diet Menstrual postcard. At the end of the blood draw, menstru- and foods suspected to be associated with BC risk (e.g., cru- ating participants were given a pre-addressed stamped post- ciferous vegetables) in the 12 months prior to reference date. card, and asked to record the date of the first day of their The FFQ was administered on paper or verbally during the next menstrual cycle and mail it; this information was used interview; those who did not complete it at the interview to determine the participant’s menstrual phase at the time returned it via mail or at the phlebotomy visit. Childhood her blood was drawn. diet was assessed with a food list. Childhood photographs. Participants provided photos from “head to toe” at ages 6, Saliva. Participants unwilling or unable to provide a blood 9, 12, 15, and 18  years to validate recalled relative body sample were asked to provide a saliva sample with the size (assessed by somatotype); photos were scanned and de- Oragene OG-500 DNA kit. Saliva samples were collected 1 3 Cancer Causes & Control (2021) 32:1129–1148 1139 immediately after administration of the main questionnaire, Study response and cooperation rate calculations by the phlebotomist at the second visit, or mailed to the par- ticipant after the first visit and returned by mail. Response and cooperation rates were calculated using impu- tation methods in accordance with the American Association Tumor SEER Information. Tumors were characterized by ER, for Public Opinion (AAPOR) guidelines [62] (see Supple- PR, and HER2 molecular subtypes, and histological grade mental Tables 1 and 2). to differentiate luminal A and luminal B tumors using data from SEER registries [11]. SEER reports also included Sample weights ICD-O codes, tumor size, laterality, lymph node involve- ment, and initial treatment and surgical history. Sample weights were created for both cases and controls to account for sampling design and non-response. Weights Tumor Tissue. To evaluate other tumor characteristics, e.g., reflect probabilities of selection and adjustments for non- Ki-67 status [11], tumor tissue from consenting cases was response. Adjustments for non-response were done at the requested from hospitals or clinics where they were stored; screener and main interview levels. To achieve the frequency when possible, tumor samples were taken before treatment. matching of controls to cases, a weighted distribution of When adequate tissue was provided, tumor microarrays cases for each study site was established across cells of age (TMAs) were created. and race. The sample weights of controls were then post- stratified to the weighted totals within each of these cells Biospecimen storage. All blood, saliva, and tumor tissue [63]. Additionally, replicate weights were created to develop biospecimens are stored at the MCW Tissue Bank as part of estimates of variability, including standard errors. Demo- the YWHHS Biorepository. Separate biomarker studies will graphic characteristics were obtained for 86% of sampled be conducted with all collected biospecimens. controls (complete roster information), and 100% of sam- pled case participants (age, race, site, county, ER status) to Interviewer Training and Quality Control Measures inform non-response weights. Replicate weights were cre- ated for case–control analyses and case-only analyses. A sec- Control recruitment interviewer training. Control field inter- ond set of weights was created for control-only analyses, to viewers were employees of Westat. Interviewers from both weight controls to the source population. Replicate weights study sites were trained together to synchronize data col- were also created for blood sample analyses. lection. Once they demonstrated adherence to all protocols they were certified for data collection. Statistical analyses Study site interviewer and  phlebotomist training. Training Primary analyses are conducted using survey weighted was conducted by the YWHHS Coordinating Center to syn- multiple logistic regression to account for study design chronize data collection. All field staff completed appropri- and potential confounding. Where appropriate, potential ate IRB-mandated training and field safety training and were effect modification by study site, race and/or socioeco- certified by the YWHHS Coordinating Center once they nomic position are being evaluated. For some analyses, demonstrated adherence to all protocols and competence in structural equation modeling (SEM) with latent variables is a complete study interview. being conducted to evaluate exposures over the life-course [64]. Additionally, for some analyses we are using survey Main interview and  phlebotomy quality control. Interviews weighted polytomous logistic regression to assess hetero- and phlebotomy visits of consenting participants were audio geneity in risk by tumor subtypes. recorded for quality control. The first five recorded inter- views completed by each interviewer and additional inter- views as needed based on performance (4.8% in Detroit; Operational results 2.6% in LA of completed interviews) were reviewed by a trained evaluator. The evaluator documented discrepancies Case participation in recorded responses, deviations from protocol, and appro- priate probing, and provided detailed feedback to each inter- A total of 5,309 potentially eligible women were identified viewer. through the Detroit (n = 2,527) and LA (n = 2,782) SEER registries (Table 2). Of these, 80% were sampled (see Case Sampling), and 3,326 were determined to be eligible or potentially eligible (Table 2). Among sampled cases, 124 women died before they could be interviewed and 82 could 1 3 1140 Cancer Causes & Control (2021) 32:1129–1148 not be contacted because physician or hospital permission higher for NHB women (57.9%) compared to NHW women was not obtained. Other reasons for non-interview included: (48.3%), and for LA (58.5%) compared to Detroit (49.3%) 177 could not be located, 70 moved away from the study (Supplemental Table 1) but did not vary significantly by age area, 23 were too ill, and 415 did not respond after maximum (Supplemental Table 2). contact attempts. Of the 3,326 sampled and potentially eli- gible participants, study staff had the opportunity to recruit Main interview 2,435 participants. Of these, 623 declined to participate, and 1,812 women were interviewed (ER + n = 1,310; ER- Location of completed interviews n = 437). The overall cooperation rate was 74.4% (Detroit: 71.9%, LA: 77.2%) and response rate was 59.8% (Detroit: A total of 73.2% and 80.8% of interviews were conducted in- 53.1%, LA: 66.4%) (Supplemental Table 1). Response rates home, 3.4% and 3.0% were conducted at a study site office, were higher for NHB women (60.2%) than NHW women and 23.5% and 16.2% were conducted at other locations (59.8%), and for LA (66.4%) than Detroit (53.1%) (Supple- (e.g., a coffee shop, local library, or healthcare provider’s mental Table 1), but did not vary significantly by age (Sup- office) for cases and controls, respectively. Distributions of plemental Table 2). interview locations were similar across study sites. Control participation Interview timing A total of 24,612 households were sampled in Detroit The median period between reference date and interview (n = 9,994) and LA (n = 14,618) (Table 2). Of these, 21,668 date was 153 days for controls and 378 days for cases (Sup- were eligible or potentially eligible and 18,612 households plemental Table 3). completed a roster (86% response rate) (Supplemental Table 1). Households not rostered because they were in Length of main questionnaire an inaccessible gated community included in LA 9% and Detroit 1% of potentially eligible households. Of households The questionnaire included 639 questions (excluding prob- that completed rosters, 3,414 participants were sampled and ing questions and repeat questions about exposures over 2,720 completed screeners (88% response rate, Supplemental the life-course). The median administration time of the Table 1). Reasons that screeners were not obtained were the questionnaire was 130 and 120 min for cases and controls, following: resided outside the study area (n = 24), was too ill respectively (Supplemental Table 4). The median duration (n = 2), was not reached after maximum attempts (n = 132), of the measured anthropometry section was 11 min  for or sampled in error (n = 9). Of the 3,247 participants sam- both cases and controls (Supplemental Table 4). Interview pled for screening that interviewers had the opportunity to time for study participants was longer for NHB women screen, 83.6% were screened. Of these, 1,988 were eligible (141 min) compared to NHW women (119 min) and for or potentially eligible and 97.2% agreed to be contacted by poorer women (HHP < 150; 132 min) compared to wealthier study site staff. Thus, Westat provided control participant women (HHP ≥ 300; 120 min). information for 1,933 women. Of these, study site staff had no opportunity to interview 223 women for the following Description of interviewed study population reasons: 12 were ineligible, 2 died before interview, 6 could not be located, 30 moved away from the study area, 2 were Table 3 shows the weighted demographic characteristics of too ill, and 171 were not reached after the maximum num- interviewed study participants. Cases were more likely to be ber attempts. Thus, 1,708 participants were confirmed to be wealthier than controls (52.0% vs. 46.3% HHP ≥ 300) and eligible and agreed to be contacted by the study site staff. Of less likely to be unemployed (17.9% vs. 25.9%). Participants these, 327 women refused to participate in the study (4% via were similar across study sites, although both NHB and proxy) and 1,381 completed the main interview (Table 2). NHW women were more likely to be poor (HHP < 150%) in Accounting for the household roster cooperation rate (94%), Detroit than LA. NHB women across both study sites were screener cooperation rate (84%), and study site recruitment also significantly more likely to be poor (35.1% cases; 49.1% cooperation rate (81%), the overall study cooperation rate controls) compared to NHW women (12.3% cases; 15.8% was 65% (Supplemental Table 1). Similarly, taking into controls) (Table 3). account the household roster response rate (86%), the par- ticipant screener response rate (88%), Westat agreed to be Completion of study components contacted response rate (98%), and the study site recruit- ment response rate (72%) led to an overall control response Response rates for all ancillary data collection efforts and rate of 53% (supplemental Table 1). Response rates were for biospecimen collection are reported in Table 4. Nearly 1 3 Cancer Causes & Control (2021) 32:1129–1148 1141 1 3 Table 3 Weighted demographic characteristics of interviewed participants by site and case–control status, Young Women’s Health History Study (N = 3,193) Detroit Los Angeles Total Case Control Case Control Case Control N W N W N W N W N W N W Percenta Percenta Percenta Percenta Percenta Percenta Total 919 100.0 717 100.0 893 100.0 664 100.0 1812 100.0 1381 100.0 Age at reference date  20–29 years 23 2.3 94 2.2 45 3.6 152 3.6 68 2.9 246 2.9  30–34 years 63 5.7 99 5.6 77 6.5 132 6.4 140 6.0 231 6.0  35–39 years 164 13.8 123 13.8 143 12.6 128 12.7 307 13.2 251 13.3  40–44 years 337 30.6 201 31.3 337 29.4 110 29.4 674 30.0 311 30.4  45–49 yearsb 332 47.7 200 47.0 292 47.9 142 47.9 623 47.8 342 47.4 Race  Non-Hispanic Black 383 32.5 352 32.5 299 27.6 313 27.6 682 30.2 665 30.2  Non-Hispanic White 536 67.5 365 67.5 594 72.4 351 72.4 1130 69.8 716 69.8 Household poverty  < 150 percent 246 24.5 275 30.7 122 13.0 194 20.3 368 19.2 469 25.9  150 to < 300 percent 258 27.7 196 26.4 211 21.8 168 24.2 469 25.0 364 25.4  ≥ 300 percent 371 43.5 220 39.5 531 61.8 285 54.0 902 52.0 505 46.3  Refused/don’t know/missing 44 4.3 26 3.3 29 3.4 17 1.5 73 3.8 43 2.4 Household poverty by race/ethnicity  Non-Hispanic Black  < 150 percent 168 43.1 200 54.2 73 24.4 159 42.3 241 35.1 359 49.1  150 to < 300% 99 25.7 86 25.8 97 32.9 70 25.3 196 28.8 156 25.6  ≥ 300 percent 94 25.4 47 13.9 118 39.2 73 30.8 212 31.3 120 21.1 Non-Hispanic White  < 150 percent 78 15.6 75 19.5 49 8.7 35 12.0 127 12.3 110 15.8  150 to < 300 percent 159 28.6 110 26.7 114 17.7 98 23.8 273 23.3 208 25.3  ≥ 300 percent 277 52.2 173 51.9 413 70.4 212 62.8 690 61.0 385 57.2 Refused/Don’t Know/Missing 44 4.3 26 3.3 29 3.4 17 1.5 73 3.8 43 2.4 Employment (outside the home)  None 168 17.2 189 25.4 162 18.7 152 26.5 330 17.9 341 25.9  Part-time (< 35 h a week) 144 16.8 101 15.6 134 15.8 106 18.4 278 16.3 207 16.9  Full-time (≥ 35 h a week) 568 62.8 368 54.4 549 60.4 338 50.4 1,117 61.3 706 52.5  Full-time student 30 2.9 56 4.2 33 3.6 67 4.4 63 3.2 123 4.3  Refused/don’t know/missing 9 1.0 3 0.3 15 1.5 1 0.4 24 1.3 4 0.3 Educational attainment  Less than high school diploma 195 20.0 185 22.9 83 9.2 106 11.9 278 14.9 291 17.8  High school graduate or GED 353 36.8 300 40.0 294 32.2 261 41.6 647 34.7 561 40.7 1142 Cancer Causes & Control (2021) 32:1129–1148 1 3 Table 3 (continued) Detroit Los Angeles Total Case Control Case Control Case Control N W N W N W N W N W N W Percenta Percenta Percenta Percenta Percenta Percenta  Vocational school, associate’s degree, or some college 204 24.1 144 22.1 313 34.5 210 29.2 517 28.9 354 35.4  Bachelor’s degree or higher 167 19.1 88 15.0 200 23.9 87 17.3 367 21.3 175 16.1  Refused/don’t know/missing 0 0 0 0 3 0.2 0 0 3 0.1 0 0 Primary caregiver’s educational attainment  Less than high school diploma 193 19.7 147 20.5 93 10.4 78 12.4 286 15.4 225 16.7  High school graduate or GED 356 41.5 268 38.7 263 32.8 191 30.0 619 37.5 459 34.6  Vocational school, associate’s degree, or some college 211 21.5 192 25.2 310 32.9 242 35.8 521 26.8 434 30.1  Bachelor’s degree or higher 107 12.4 72 10.8 204 21.3 139 18.9 311 16.5 211 14.6  Refused/don’t know/missing 52 5.0 38 4.8 22 2.5 14 2.9 74 3.7 52 3.9 Age at first birth Nulliparous 171 19.1 138 16.1 326 36.1 265 29.2 497 27.1 403 22.2  First birth age < 20 years 190 17.9 194 22.8 101 10.4 112 17.8 291 14.4 306 20.5  First birth age 20–29 years 385 42.4 278 41.5 234 25.5 194 29.6 619 34.5 472 35.9  First birth age ≥ 30 years 173 20.5 107 19.7 231 27.8 93 23.4 404 23.9 200 21.4  Missing 0 0 0 0 1 0.1 0 0 1 0.1 0 0 Birth cohort  Born 1961–1969 459 59.2 287 60.3 428 59.9 173 56.1 887 59.5 460 58.4  Born 1970–1979 407 35.8 274 33.9 372 32.5 246 35.8 779 34.3 520 34.8  Born 1980–1989 51 4.8 130 5.1 87 7.3 197 7.6 138 6.0 327 6.3  Born 1990–1995 2 0.1 26 0.7 6 0.4 48 0.5 8 0.2 74 0.6 a Weighted percentages incorporate post-stratification sample weights b Among 45–49 year old Non-Hispanic White women diagnosed with invasive BC, 48.7% of Detroit and 35.3% of LA cases were sampled; cases identified post the study recruitment (N = 258 (n = 66 Black; n = 192 White)) were considered not sampled; all other case subgroups were sampled at 100% Cancer Causes & Control (2021) 32:1129–1148 1143 Table 4 Completion rates of study materials by case–control status and race, Young Women’s Health History Study Cases Controls Non-Hispanic Non-Hispanic Total Non-Hispanic Non-Hispanic Total Black White Black White Percent Percent Percent Percent Percent Percent Study material Main interview (all eligible) (N): (682) (1130) (1812) (665) (716) (1381) Completed all sections 98 99 99 99 99 99 Life history calendar 99 98 98 98 100 99 Anthropometry measurements  Height, weight, waist/hip circumference 96 94 95 98 94 96  Bioelectric impedance a ssaya 81 88 85 82 87 85  Photographs of body size 18 55 41 19 52 36  Food Frequency Questionnaire 70 84 79 57 79 69  Neighborhood notes 88 91 90 92 96 94  Main interview audio consent 96 99 98 98 98 98  Main interview audio (of consented) 53 84 72 52 83 68 Residential census block information  12 Months before reference date 95 96 96 96 96 96  Age 12 82 86 85 82 86 84  Permission to obtain health department information about 90 94 92 89 94 92 participant’s birth  Permission to contact in future 98 99 99 99 99 99 Blood/saliva:  Blood sample/saliva kit for DNA analyses 79 87 84 78 83 81  Blood sample 70 77 75 74 75 75  Blood Questionnaireb 99 99 99 99 99 99  Menstrual calendarc 88 93 92 90 94 92  Given and returned menstrual p ostcardd 43 69 61 42 63 53 Breast tumor:  Tumor ER/PR/HER2 status from SEER 95 97 96 – – –  Tumor tissue consent received 96 97 96 – – –  Tumor tissue available of consented e 78 50 60 – – –  Tumor tissue collected of available (as of 10/1/2020) e 47 68 58 – – –  Among women who had mammogram (N): (606) (1045) (1651) (285) (357) (642)  Permission to obtain last mammogram 98 99 99 94 95 95  Among gravid women (N): (548) (766) (1314) (506) (472) (978)  Permission to obtain health department information about 90 96 94 87 95 91 participant’s pregnancies  Among participants with eligible c aregiversF (N): (500) (973) (1473) (513) (633) (1146)  Caregiver survey 48 70 63 37 68 54 a Percentages based on number of participants who were not pregnant at the time of interview, N = 1659 cases and N = 1256 controls b Percentages based on participants who completed blood draws c Percentages based on participants who completed blood draws and were pre- or perimenopausal at time of blood draw d Percentages based on participants who completed blood draws and were given menstrual postcards e Tumor percent available based on number of participants with tumor material considered available of those who consented. Tumor availability determined by Slide Retrieval Program in LA and Epidemiology Research Core in Detroit. Primary reasons tumor not available were that there was not enough tumor tissue available for analysis and the hospital at which the specimen is stored does not allow researchers to take samples f Percentages based on number of participants who completed interviews and who didn’t report mother is “deceased” or “not in contact.” 1 3 1144 Cancer Causes & Control (2021) 32:1129–1148 all participants completed the main interview (99%) and training and oversight, and through the study’s central- provided anthropometry measurements (95% of cases and ized tracking system. Other strengths include its in-depth 96% of controls). Most also provided blood samples (75% assessment of social context, including residential history of cases and controls), or if blood was not provided, saliva and current built environment. Additionally, biomarkers and (84% of cases and 81% of controls provided blood or saliva). both inherited genetic factors associated with BC and gene In addition, 60% of women with BC who consented to allow expression changes can be evaluated in this population-based us to retrieve tumor tissue had tissue available for analysis study of young-onset BC—all of which are understudied. and thus far, of available participant tumor tissue, 58% has been retrieved (n = 660). Nearly all interviewed participants Limitations (97%) agreed to be contacted in the future. Limitations of this study include potential residual recall bias for exposures that could not be validated. The study, Discussion however, used methods such as a life calendar, to minimize these issues [65]; life-course exposures were collected We successfully conducted the YWHHS: a large popula- with recall aids, and YWHHS was able to validate recalled tion-based case–control epidemiologic study based on the responses for key exposures, e.g., using measured adult and eco-social theory of disease etiology [42] to identify poten- childhood photos to validate recalled anthropometry. The tially modifiable factors associated with young-onset BC study sample size also limits our ability to examine young- overall and by molecular tumor subtypes, and to investigate onset BC risk by some rarer tumor subtypes and within some racial and socioeconomic inequities in BC among NHB and population subgroups for small effect sizes and more rare NHW young women. For the extensive in-person interview exposures; data from this study can be pooled with other (median time 120–130 min), we achieved a 60% response studies to evaluate these questions. The timing of blood sam- rate among cases and 53% response rate among controls, and ple collection also prohibits examination of factors poten- the cooperation rate, among those we had the opportunity to tially affected by treatment or “case” status, though exten- interview was 74% among cases and 65% among controls. sive information was collected to allow the study of these This was achieved through extensive follow-up efforts with potential influences. Additionally, information on “race” the use of a centralized computer tracking system. Subse- is ultimately self-reported but was originally based on the quently we achieved a high response rate to our request for SEER registry for cases. SEER registry reports of “race” and blood (75%) or saliva samples when blood was not available “Hispanic ethnicity,” however, are highly correlated with (82%). With linkage to NCI SEER cancer registry data, we self-report [66, 67]. have valid information on the definition of a breast cancer An additional limitation could be the study response case and detailed information on tumor subtype. With survey rates; however, complete enumeration of cases in the SEER data linked to biospecimen information, we have collected registry and 86% enumeration of sampled control house- comprehensive data to address this study’s research ques- holds enabled us to incorporate non-response sample tions, as well as future studies of breast cancer. This is one weights to mitigate this limitation. Declining response rates the largest, population-based case–control studies of young- for national-level surveys, particularly telephone surveys, onset BC. Additionally, to our knowledge, this is the largest are well documented over the course of the survey period, population-based case–control study of BC in young NHB and the challenges that caused this decline in rates also women and the largest where extensive life-course individ- contributed to reduced response rates for YWHHS cases ual-level socioeconomic measures were collected to evaluate and controls [68]. Study response rates are, however, well racial and socioeconomic inequities in BC risk. within ranges reported in the literature [53, 69, 70], par- ticularly for the data collection time period, participants’ Strengths ages, and the well-recognized challenges in enrolling disad- vantaged populations [71, 72]. We found that women were Strengths of this study include its exclusive focus on young more willing to participate when interviewers were similar women (aged < 50  years) incorporating information on in race and age (data not shown) [71, 73] and that response tumor subtypes [9], and that it is designed to shed light on rates may have been lower among White women in Detroit inequities in risk in young NHB compared to NHW women due to interviewer-participant age incongruence. Recruit- by life-course SEP. Other strengths include its population- ment and scheduling challenges included that women who based ascertainment of cases and controls and availability of were juggling childcare, work, other family responsibilities created sample weights. The centralized YWHHS Coordi- or challenging cancer treatment regimens often resched- nating Center synchronized data collection across study sites uled interviews. To address these obstacles exclusive tel- through conduct of all study interviewer and recruitment ephone schedulers were hired, targeted letters were mailed 1 3 Cancer Causes & Control (2021) 32:1129–1148 1145 to address concerns regarding confidentiality and time con- Tumor Subtyping Consultants: Dr. Howard Chang; Dr. Sandra Haslam; straints, in-person follow-up visits were attempted with con- Dr. Melissa Troester; Dr. Mark Sherman; YWHHS Biospecimen Labo- trols in Detroit and cases and controls in LA, and the study ratory Staff (Michigan State University): Dr. Rachel Schiffman; Alice Schehr; Melanie Adkins; Dr. Sainan Wei; Genetic Biostatistician Con- incentive was increased. sultant: Dr. Goncalo Abecasis; Nutritional Assessment Consultants (NutritionQuest): Tory Block; Dr. Jean Norris; Kinesiology Consult- Future directions ants and Interviewer Trainers: Dr. Emily Guseman (OSU); Dr. Kimbo Yee; YWHHS Interviewer Training and Quality Control (Michigan State University/University of Wisconsin—Milwaukee): Dr. Jeanne Analyses using collected YWHHS data are in progress. Meier; Scientific Advisors: Dr. Otis Brawley; Dr. Lawrence Brody; Dr. Additional supplemental projects are possible, including Larry Kushi; Dr. Camara Jones; Dr. Julie Palmer; Dr. Mark Sherman; pooling of data, particularly to study rarer tumor subtypes, and Dr. Anne Sumner; and our past YWHHS Central Coordinating studies to evaluate risk for other BC tumor subtypes, to study Center Research team members (Michigan State University/University of Wisconsin, Milwaukee): Kara Mannor; Steven Larmore; Marielle factors associated with mammograms and BC survival, to Gagnier; James Dodge; Andrew Jessmore; Olga Prushinskaya; The- study biomarkers, e.g., gene expression, to integrate exter- resa Kowalaski; Kevin Petersen; Stephan Diljak; Hannah Selig; Cristin nal data with data on geocoded life-course residential his- McArdle; Dr. Julie Schuppie; Amy Parry; Beneet Pandey; Bethany tories, and/or to evaluate intermediate biomarkers and BC Canales; Cory Steinmetz; David Strong; Sofia Haile; James Groh; Jenn Woo; Brian Thayer; Dan Sanfelippo; Nicole Carlson and Anamarie risk. Results from YWHHS will expand our understanding LeDuc. Additionally, we would like to thank Drs. Leslie Bernstein of potentially modifiable factors associated with BC risk and Katie Henderson, City of Hope, for assistance with the develop- overall and by subtype and should identify sources of racial ment of the study design and obtaining funding, as well as Dr. Karen and socioeconomic inequities in young-onset BC. Klomparens, Dean of the Graduate School, Michigan State University, for her support. The authors assume full responsibility for analyses and interpretation of these data. Supplementary Information The online version of this article con- tains supplementary material available (https:// doi. org/ 10. 1007/ Author contributions EMV: Conceptualization, supervision, meth- s10552- 021-0 1461-x). odology, funding acquisition, data curation, writing—original draft, writing—reviewing and editing. LRM: Project administration, data Acknowledgments We would first like to extend our deep appreciation curation, formal analysis, writing—original draft, writing—reviewing to the women who contributed as participants to the Young Women’s and editing. DRP: Conceptualization, methodology, funding acquisi- Health History Study. We would also like to thank the following indi- tion, data curation, writing—reviewing and editing. ASH: Conceptual- viduals who contributed to the study design and data collection. Com- ization, supervision, methodology, funding acquisition, data curation, munity Advisors: Twyla Griffin; Kommah McDowell; Hope Bradford; writing—reviewing and editing. RD: Conceptualization, supervision, Katie Clark; Diana Dyer; Brenda Krentler; Karen Owens; Vernessa methodology, funding acquisition, data curation, writing—reviewing Patrick; Karry Samulski; Lori Wesby; Hanna Weber; the Metropolitan and editing. RK: Project administration, supervision, data curation, Detroit SEER registry and Epidemiology Research Core (Wayne State writing—reviewing and editing. BG: Project administration, supervi- University/Karmanos Cancer Institute): Dr. Jennifer Beebe-Dimmer, sion, data curation, writing—reviewing and editing. RH: Conceptu- Julie Ruterbusch and Fawn Vigneau; the Michigan State Vital Statis- alization, methodology, funding acquisition, writing—reviewing and tics Registrar: Dr. Glenn Copeland; the Los Angeles County SEER editing. NC: Conceptualization, methodology, writing—reviewing and registry: Dr. Dennis Deapen; Justin Cook; Maria Isabel Gaeta; Yaping editing. LKO: Conceptualization, methodology, funding acquisition, Wang; YWHHS Los Angeles County Data Collection and Processing writing—reviewing and editing. AA: Conceptualization, methodology, Team (University of Southern CA): Denise Modjeski; Kashonda Davis; writing—reviewing and editing. ZZ: Conceptualization, methodology, Wendy McGlothlin; Paige Rosenthal; Jennifer Zelaya; Renee Bicker- data curation, writing—reviewing and editing. DM: Project administra- staff-Magee; Elesa Maxie; Priscilla Gardner; the YWHHS Metropolitan tion, supervision, data curation, writing—reviewing and editing. GN: Detroit Data Collection and Processing Team (Wayne State University/ Project administration, supervision, data curation, writing—review- Karmanos Cancer Institute): Dr. Gwendolyn Norman; Landa Daniels; ing and editing. DRL: Project administration, formal analysis, writ- Tara Baird; Amanda Bullock; Terry Smith; Mary Beth Kolbicz; Verona ing—reviewing and editing. SG: Project administration, data curation, Ivory; Arkeshia Barnes; Heloise Glenn; Velma White; Terry Smith; writing—reviewing and editing. HR: Conceptualization, supervision, Ernestine Anthony; and Deborah Kimbrough; our Westat Team: Dr. methodology, data curation, writing—reviewing and editing. KS: Con- Jeanne Rosenthal; Giannella De Rienzo; Craig Ray; Jane Li; Sabrina ceptualization, supervision, methodology, funding acquisition, data Zhang; and the many field interviewers in Detroit and LA; YWHHS curation, writing—reviewing and editing. Computer Tracking System (University of Southern California Can- cer Informatics Core): Aarti Vaishnav; Reed Comire; Vaibhav Bora; Funding This work was directly supported by the National Institute Jeet Poonater; Waikeung Louis Lee; and Charanya Ram Kumar; Sur- of Health (NIH) National Cancer Institute (NCI) grant R01CA136861 vey Biostatistical Consultant (National Cancer Institute): Dr. Barry (E.Velie). The collection of cancer incidence data from California used Graubard; Racial Sensitivity Trainer/Field Work Consultants (Univer- in this study was supported by the California Department of Public sity of Southern California): Dr. Karen Lincoln; Dr. Rose Monteiro; Health pursuant to California Health and Safety Code Sect. 103885; Questionnaire Development Consultants: Dr. Lorraine Halinka Mal- Centers for Disease Control and Prevention’s (CDC) National Program coe; Dr. Christine Erdmann; YWHHS Biospecimen Biorepository Staff of Cancer Registries, under cooperative agreement 5NU58DP006344; (Medical College of Wisconsin Tissue Bank): Dr. Saul Suster; Mary the National Cancer Institute’s Surveillance, Epidemiology and End Rau; Janelle Lang-Piette; Ellen Schneider; Matthew Dunham; Whit- Results Program under contract HHSN261201800032I awarded to the ney Stibb; YWHHS Detroit Tumor Processing Team (Medical College University of California, San Francisco, contract HHSN261201800015I of Wisconsin): Dr. Craig MacKinnon; Dr. Zainab Basir; Kathy Stoll; awarded to the University of Southern California, and contract Los Angeles YWHHS Tumor Collection/Processing: Dr. Wendy Cozen; HHSN261201800009I awarded to the Public Health Institute. Z. Dr. Debra Hawes; Jose Aparicio; Dr. Maria Sibug-Saber; Moli Chen; 1 3 1146 Cancer Causes & Control (2021) 32:1129–1148 Zhang was supported by the NIH Office of Research on Women’s 6. Assi H et al (2013) Epidemiology and prognosis of breast cancer Health and the National Institute of Child Health and Human Devel- in young women. J Thorac Dis 5(1):S2–S8 opment K12HD043488 (Building Interdisciplinary Research Careers in 7. Nichols HB, Schoemaker MJ, Wright LB, McGowan C, Brook Women’s Health, BIRCWH). The ideas and opinions expressed herein MN, McClain KM, Jones ME, Adami HO, Agnoli C, Baglietto are those of the authors and do not necessarily reflect the opinions of L, Bernstein L (2017) The premenopausal breast cancer collabo- the funders. ration: a pooling project of studies participating in the National Cancer Institute Cohort Consortium. Cancer Epidemiol Prev Bio- Availability of data and material The datasets analyzed during the cur- marks. 26(9):1360–9 rent study are not publicly available because main study findings are 8. Johnson KC, Glantz SA (2008) Evidence secondhand smoke in process of being analyzed, but are available from the corresponding causes breast cancer in 2005 stronger than for lung cancer in 1986. author on reasonable request. Prev Med 46(6):492–496 9. Barnard M, Boeke C, Tamimi R (2015) Established breast cancer risk factors and risk of intrinsic tumor subtypes. Biochem Biophys Declarations Acta 1856(1):73–85 1 0. Perou CM et  al (2000) Molecular portraits of human breast Conflict of interest None declared. tumours. Nature 406(6797):747–752 1 1. Goldhirsch A et al (2013) Personalizing the treatment of women Ethical approval This study protocol was approved by the Institutional with early breast cancer: highlights of the St Gallen International Review Boards at the University of Wisconsin—Milwaukee (UWM); Expert Consensus on the Primary Therapy of Early Breast Cancer Michigan State University (MSU); Wayne State University (WSU); 2013. Ann Oncol 24(9):2206–2223 the Michigan Department of Community Health; University of South- 12. Balic M et al (2019) St. Gallen/Vienna 2019: a brief summary of ern California (USC); the California Committee for the Protection of the consensus discussion on the optimal primary breast cancer Human Subjects (CPHS); and for the Medical College of Wisconsin treatment. Breast Care 14(2):103–110 (MCW), IRB oversight was deferred to UWM. The California Cancer 1 3. Shoemaker ML, White MC, Wu M, Weir HK, Romieu I (2018) Registry also approved the study. Differences in breast cancer incidence among young women aged 20–49 years by stage and tumor characteristics, age, race, and Consent to participate Written informed consent was obtained from ethnicity, 2004–2013. Breast Cancer Res Treat 169(3):595–606 all participants included in the YWHHS. 1 4. Chen HL et al (2016) Effect of age on breast cancer patient prog- noses: a population-based study using the SEER 18 Database. Consent for publication All participants included in the final YWHHS PLoS ONE 11(10):e0165409 sample consented to having their data published in scientific publica- 15. Eccles SA et al (2013) Critical research gaps and translational tions. priorities for the successful prevention and treatment of breast cancer. Breast Cancer Res 15(5):R92 1 6. Bray F et al (2018) Global Cancer Statistics 2018: GLOBCAN Open Access This article is licensed under a Creative Commons Attri- estimates of incidence and mortality worldwide for 36 cancers in bution 4.0 International License, which permits use, sharing, adapta- 185 countries. Cancer 68(6):394–424 tion, distribution and reproduction in any medium or format, as long 1 7. Henley A et al (2020) Annual Report to the Nation on the Status of as you give appropriate credit to the original author(s) and the source, Cancer, Part 1: National Cancer Statistics. Cancer 126:2225–2249 provide a link to the Creative Commons licence, and indicate if changes 18. Singh GK, Jemal A (2017) Socioeconomic and racial/ethnic dis- were made. The images or other third party material in this article are parities in cancer mortality, incidence, and survival in the United included in the article’s Creative Commons licence, unless indicated States, 1950–2014: over six decades of changing patterns and otherwise in a credit line to the material. If material is not included in widening inequalities. J Environ Public Health 2017:2819372 the article’s Creative Commons licence and your intended use is not 1 9. SEER*Explorer Application: Breast Cancer Recent Trends in permitted by statutory regulation or exceeds the permitted use, you will SEER Age-Adjusted Mortality Rates, 2000–2018 by Race/Eth- need to obtain permission directly from the copyright holder. To view a nicity, Female, Ages <50, SEER, Editor. 2020. copy of this licence, visit http://c reati vecom mons.o rg/l icens es/b y/4.0 /. 20. SEER*Explorer Application: Breast Cancer Recent Trends in SEER Age-Adjusted Incidence Rates, 2000–2017 by Race/Eth- nicity, Female, Ages 15–39, All Stages, Delay-adjusted Rates, References SEER, Editor. 2020. 21. Qin B et al (2021) Neighborhood social environmental factors and breast cancer Subtypes among Black Women. Cancer Epidemiol 1. Ward E et al (2019) Annual Report to the Nation on the Status of Biomarkers Prev 30(2):344–350 Cancer, 1999–2015, Featuring Cancer in Men and Women ages 22. Linnenbringer E et al (2020) Associations between breast cancer 20–49. J Natl Cancer Inst subtype and neighborhood socioeconomic and racial composi- 2. DeSantis C et al (2019) Breast cancer statistics, 2019. CA Cancer tion among Black and White women. Breast Cancer Res Treat J Clin 69(6):438–451 180(2):437–447 3. Warner ET et al (2013) Reproductive factors and risk of premeno- 23. Taylor TR et al (2007) Racial discrimination and breast cancer pausal breast cancer by age at diagnosis: are there differences incidence in US Black women: the Black Women’s Health Study. before and after age 40? Breast Cancer Res Treat 142(1):165–175 Am J Epidemiol 166(1):46–54 4. White AJ et al (2015) Overall and central adiposity and breast 24. Lu G et al (2018) The fluctuating incidence, improved survival cancer risk in the Sister Study. Cancer 121(20):3700–3708 of patients with breast cancer, and disparities by age, race, and 5. Chollet-Hinton L et al (2016) Breast cancer biologic and etiologic socioeconomic status by decade, 1981–2010. Cancer Manag Res heterogeneity by young age and menopausal status in the Caro- 10:4899–4914 lina Breast Cancer Study: a case-control study. Breast Cancer Res 2 5. Andaya AA et al (2012) Socioeconomic disparities and breast 18(1):79 cancer hormone receptor status. Cancer Causes Control 23(6):951–958 1 3 Cancer Causes & Control (2021) 32:1129–1148 1147 26. Colditz GA, Bohlke K, Berkey CS (2014) Breast cancer risk accu- 48. DiGaetano R, Waksberg J (2002) Commentary: trade-offs in the mulation starts early: prevention must also. Breast Cancer Res development of a sample design for case-control studies. Am J Treat 145(3):567–579 Epidemiol 155(8):771–775 2 7. Kawai M et al (2014) Height, body mass index (BMI), BMI 49. Brogan DJ et al (2001) Comparison of telephone sampling and change, and the risk of estrogen receptor-positive, HER2-positive, area sampling: response rates and within-household coverage. Am and triple-negative breast cancer among women ages 20 to 44 J Epidemiol 153(11):1119–1127 years. Cancer 120(10):1548–1556 50. National Health And Nutrition Examination Survey III: Field 28. Ma H et al (2015) Reduced risk of breast cancer associated with Operations Manual. 1991. recreational physical activity varies by HER2 status. Cancer Med 51. Mesenbourg T et al (2010) Census Summary File 1 4(7):1122–1135 52. Jones CP et al (2008) Using “socially assigned race” to probe 29. Premenopausal Breast Cancer Collaborative, G et al (2018) Asso- white advantages in health status. Ethn Dis 18(4):496–504 ciation of Body Mass Index and age with subsequent breast cancer 53. Marchbanks PA et al (2002) The NICHD Women’s Contraceptive risk in premenopausal women. JAMA Oncol 4(11):e181771 and Reproductive Experiences Study: methods and operational 3 0. Robinson WR et al (2014) Body size across the life course and results. Ann Epidemiol 12(4):213–221 risk of premenopausal and postmenopausal breast cancer in Black 5 4. Gammon MD et al (2002) The Long Island Breast Cancer Study women, the Carolina Breast Cancer Study, 1993–2001. Cancer Project: description of a multi-institutional collaboration to iden- Causes Control 25(9):1101–1117 tify environmental risk factors for breast cancer. Breast Cancer 31. Xue F et al (2016) Body fatness throughout the life course and Res Treat 74(3):235–254 the incidence of premenopausal breast cancer. Int J Epidemiol 55. Brinton LA et al (1995) Oral contraceptives and breast cancer risk 45(4):1103–1112 among younger women. J Natl Cancer Inst 87(11):827–835 3 2. Millikan RC et al (2008) Epidemiology of basal-like breast cancer. 56. Resnick M, Bearman P, Blum R (1997) Protecting adolescents Breast Cancer Res Treat 109(1):123–139 from harm. Findings from the National Longitudinal Study on 33. Chollet-Hinton L et al (2017) Biology and Etiology of Young- Adolescent Health. JAMA 278(10):823–832 Onset Breast Cancers among Premenopausal African American 57. Hamilton C, Strader L, Pratt J (2011) The PhenX Toolkit: get the Women: Results from the AMBER Consortium. Cancer Epide- most from your measures. Am J Epidemiol 174(3):253–260 miol Biomarkers Prev 26(12):1722–1729 58. Wingo P et al (1988) The evaluation of the data collection process 34. Bandera EV et al (2015) Obesity, body fat distribution, and risk for a multicenter, population-based, case-control design. Am J of breast cancer subtypes in African American women partici- Epidemiol 128(1):206–217 pating in the AMBER Consortium. Breast Cancer Res Treat 59. Pebley NSA (2003) Neighborhood and family effects on children’s 150(3):655–666 health in Los Angeles. RAND, Santa Monica 35. Krieger N (1994) Epidemiology and the web of causation: has 6 0. Weir SS Healthy environments partnership neighborhood obser- anyone seen the spider? Soc Sci Med 39(7):887–903 vational checklist. University of Michigan, Ann Arobor, pp 1–12. 36. Krieger N (2020) Measures of racism, sexism, heterosexism, and 6 1. Kelly C et al (2013) Using Google Street View to audit the built gender binarism for health equity research: from structural injus- environment: inter-rater reliability results. Ann Behav Med tice to embodied harm-an ecosocial analysis. Annu Rev Public 45(Suppl_1):108–12 Health 41:37–62 6 2. AAPOR (2016) Standard definitions: final dispositions of case 37. Omi M, Winant H (1994) Racial formation in the United States: codes and outcome rates for surveys. AAPOR, Oakbrook Terrace from the 1960’s to the 1990’s. Routledge, New York 6 3. Li Y, Graubard B, Digaetano R (2011) Weighting methods for 3 8. Jones C (2000) Levels of racism: a theoretic framework and a population-based case-control studies with complex sampling. J gardener’s tale. Am J Public Health 90:1212–1215 R Stat Soc Ser C 60(2):165–185 3 9. Duster T (2005) MEDICINE: enhanced: race and reification in 6 4. Bollen K (1989) Structural equations with latent variables. Wiley, science. Science 307(5712):1050–1051 New York 40. Williams DR, Mohammed SA, Shields AE (2016) Understand- 6 5. Mahabir S et al (2012) Challenges and opportunities in research ing and effectively addressing breast cancer in African American on early-life events/exposures and cancer development later in life. women: Unpacking the social context. Cancer 122(14):2138–2149 Cancer Causes Control 23(6):983–990 41. Ford CL, Harawa NT (2010) A new conceptualization of ethnicity 66. Gomez SL, Glaser SL (2006) Misclassification of race/ethnicity in for social epidemiologic and health equity research. Soc Sci Med a population-based cancer registry (United States). Cancer Causes 71(2):251–258 Control 17(6):771–781 4 2. Krieger, N., Ecosocial Theory of Disease Distribution: Embody- 67. Hamilton A et al (2009) Latinas and breast cancer outcomes: ing Societal & Ecologic Context, in Epidemiology and the Peo- population-based sampling, ethnic identity, and acculturation ple’s Health. Theory and Context. 2013. p. 202–235. assessment. Cancer Epidemiol Biomarkers Prev 18(7):2022–2029 43. Williams DR, Mohammed SA (2009) Discrimination and racial 6 8. Tourangeau R, Plewes (2013) Nonresponse in social science sur- disparities in health: evidence and needed research. J Behav Med veys: a research agenda. National Academy of Sciences, Wash- 32(1):20–47 ington, DC 44. Linnenbringer E, Gehlert S, Geronimus AT (2017) Black-White 6 9. Xu M et al (2018) Response rates in case-control studies of cancer disparities in breast cancer subtype: the intersection of socially by era of fieldwork and by characteristics of study design. Ann patterned stress and genetic expression. AIMS Public Health Epidemiol 28(6):385–391 4(5):526–556 70. Palmer JR, Ambrosone CB, Olshan AF (2014) A collaborative 4 5. Jones CP (2002) Confronting institutionalized racism. Phylon study of the etiology of breast cancer subtypes in African Ameri- 50(1/2):7–22 can women: the AMBER consortium. Cancer Causes Control 4 6. Jones CP (2001) Invited commentary: "race," racism, and the 25(3):309–319 practice of epidemiology. Am J Epidemiol 154(4):299–304; dis- 71. Pinn V et al (2003) Outreach Notebook: For the inclusion, recruit- cussion 305–6. ment and retention of women and minority subjects in clinical 4 7. Krieger N (2013) History, biology, and health inequities: emergent research, U.S.D.o.H.a.H. Services, Editor. National Institutes of embodied phenotypes and the illustrative case of the breast cancer Health estrogen receptor. Am J Public Health 103(1):22–27 1 3 1148 Cancer Causes & Control (2021) 32:1129–1148 72. Bartlett DWR (2013) Recruitment and retention of African Ameri- 76. Folsom R, otter F, Williams SRTI (1987) Notes on a composite can and Hispanic girls and women in research. Public Health Nurs size measure for self-weighting samples in multiple domains. 30(2):159–166 American Statistical Association Meeting, pp 792–796 73. Moorman PG et al (1999) Participation rates in a case-control study: the impact of age, race, and race of interviewer. Ann Epi- Publisher’s Note Springer Nature remains neutral with regard to demiol 9(3):188–195 jurisdictional claims in published maps and institutional affiliations. 7 4. Schwartz K et al (2013) Enhancement and validation of an Arab surname database. J Registry Manag 40(4):176–179 75. Williams DR (1997) Race and health: basic questions, emerging directions. Ann Epidemiol 7:322–333 Authors and Affiliations Ellen M. Velie1,2  · Lydia R. Marcus1,2  · Dorothy R. Pathak3  · Ann S. Hamilton4 · Ralph DiGaetano5 · Ron Klinger5 · Bibi Gollapudi5 · Richard Houang6 · Nicole Carnegie7 · L. Karl Olson8 · Amani Allen9 · Zhenzhen Zhang10 · Denise Modjesk4 · Gwendolyn Norman11 · Darek R. Lucas1,2 · Sapna Gupta12 · Hallgeir Rui13 · Kendra Schwartz14 1 Zilber School of Public Health, University of WI – 9 Departments of Community Health Sciences Milwaukee, 1240 N. 10th Street, Milwaukee, WI 53201, and Epidemiology, School of Public Health, University USA of California Berkeley, 2121 Berkeley Way, Berkeley, 2 Departments of Medicine and Pathology, Medical College CA 94720, USA of Wisconsin, 9200 W. Wisconsin Ave, Milwaukee, 10 Division of Oncological Sciences, Knight Cancer Institute, WI 53226, USA Oregon Health & Science University, 3181 SW Sam Jackson 3 Department of Epidemiology and Biostatistics, College Park Rd, Mail Code: KCRB-PROS, Portland, OR 97239, of Human Medicine, Michigan State University, 909 Wilson USA Road Room B601, East Lansing, MI 48824, USA 11 College of Liberal Arts and Sciences, Wayne State 4 Department of Preventive Medicine, Keck School University, 4841 Cass Avenue, Detroit, MI 48201, USA of Medicine, University of Southern California, 2001 N. Soto 12 Cancer Research Informatics Core, University of Southern St, Los Angeles, CA 90089-9239, USA California Norris Cancer Center, NRT LG507, 1450 Biggy 5 Westat Inc., 1650 Research Blvd, Rockville, MD 20850, St, Los Angeles, CA 90033, USA USA 13 Department of Pathology, Medical College of Wisconsin, 6 Department of Education, Michigan State University, 620 8701 Watertown Plank RD., Milwaukee, WI 53226, USA Farm Ln, East Lancing, MI 48824, USA 14 Department of Family Medicine and Public Health Sciences, 7 Department of Mathematics, Montana State University, 732 Wayne State University, 3939 Woodward Ave, Detroit, Grant St, Bozeman, MT 59717, USA MI 48201, USA 8 Department of Physiology, Michigan State University, 567 Wilson Rd, East Lansing, MI 48824, USA 1 3