Examining disparities in large‐scale patient‐reported data capture using digital tools among cancer patients at clinical intake

Abstract Background Patient‐reported data can improve quality of healthcare delivery and patient outcomes. Moffitt Cancer Center (“Moffitt”) administers the Electronic Patient Questionnaire (EPQ) to collect data on demographics, including sexual orientation and gender identity (SOGI), medical history, cancer risk factors, and quality of life. Here we investigated differences in EPQ completion by demographic and cancer characteristics. Methods An analysis including 146,142 new adult patients at Moffitt in 2009–2020 was conducted using scheduling, EPQ and cancer registry data. EPQ completion was described by calendar year and demographics. Logistic regression was used to estimate associations between demographic/cancer characteristics and EPQ completion. More recently collected information on SOGI were described. Results Patient portal usage (81%) and EPQ completion rates (79%) were consistently high since 2014. Among patients in the cancer registry, females were more likely to complete the EPQ than males (odds ratio [OR] = 1.17, 95% confidence interval [CI] = 1.14–1.20). Patients ages 18–64 years were more likely to complete the EPQ than patients aged ≥65. Lower EPQ completion rates were observed among Black or African American patients (OR = 0.59, 95% CI = 0.56–0.63) as compared to Whites and among patients whose preferred language was Spanish (OR = 0.40, 95% CI = 0.36–0.44) or another language as compared to English. Furthermore, patients with localized (OR = 1.16, 95% CI = 1.12–1.19) or regional (OR = 1.16, 95% CI = 1.12–1.20) cancer were more likely to complete the EPQ compared to those with metastatic disease. Less than 3% of patients self‐identified as being lesbian, gay, or bisexual and <0.1% self‐identified as transgender, genderqueer, or other. Conclusions EPQ completion rates differed across demographics highlighting opportunities for targeted process improvement. Healthcare organizations should evaluate data acquisition methods to identify potential disparities in data completeness that can impact quality of clinical care and generalizability of self‐reported data.


| BACKGROUND
2][3] Cancer patients often have unmet needs, such as unmanaged symptoms or barriers to healthcare, that can be identified through these questionnaires.
5][6] Specifically, some cancer centers have developed enterprise-wide [7][8][9] or clinic-specific 10 intake forms to collect pertinent patient-reported data prior to a patient's first visit.Equitable participation in these data collection efforts is essential for ensuring that all patients receive patient-centered care.However, to the best of our knowledge, no information is available on how the completeness of these data vary demographic factors among cancer patients.
H. Lee Moffitt Cancer Center & Research Institute ("Moffitt"), a National Cancer Institute-Designated Comprehensive Cancer Center in Tampa, Florida delivers multidisciplinary cancer care through disease-based clinics that historically administered paper questionnaires specific to each clinic.Paper-based questionnaires included overlapping questions across clinics, contributing to patient burden and inefficiencies, and limited use of patient-reported data for clinical decision support, quality improvement and research.In 2009, Moffitt began administering, to all NPs, the Electronic Patient Questionnaire (EPQ), a clinical intake form for the ascertainment of baseline demographic and clinical data.The goal of the EPQ was to scale data acquisition across thousands of patients simultaneously, ensure EPQ data were available to healthcare providers directly in the electronic medical record, reduce patient burden through harmonization of questions across clinics and more rapidly introduce changes to accommodate evolving clinical and operational needs.Primarily deployed through a patient portal which most patients access days before their first visit, the EPQ can be completed in multiple sessions, giving patients an opportunity to assemble relevant medical information and ensuring privacy when completing sensitive questions, 11,12 such as past medical history, cancer risk factors, sexual orientation, and gender identity.
Despite the benefits of an electronic approach, important questions remain about the representativeness of selfreported data, particularly through the lens of diversity and equity, given the potential risk of perpetuating disparities in access to high-quality oncology care through the deployment of digital tools.Previous research has shown disparities in access to patient portals by race, ethnicity, Results: Patient portal usage (81%) and EPQ completion rates (79%) were consistently high since 2014.Among patients in the cancer registry, females were more likely to complete the EPQ than males (odds ratio [OR] = 1.17, 95% confidence interval [CI] = 1.14-1.20).Patients ages 18-64 years were more likely to complete the EPQ than patients aged ≥65.Lower EPQ completion rates were observed among Black or African American patients (OR = 0.59, 95% CI = 0.56-0.63)as compared to Whites and among patients whose preferred language was Spanish (OR = 0.40, 95% CI = 0.36-0.44)or another language as compared to English.Furthermore, patients with localized (OR = 1.16, 95% CI = 1.12-1.19)or regional (OR = 1.16, 95% CI = 1.12-1.20)cancer were more likely to complete the EPQ compared to those with metastatic disease.Less than 3% of patients selfidentified as being lesbian, gay, or bisexual and <0.1% self-identified as transgender, genderqueer, or other.
Conclusions: EPQ completion rates differed across demographics highlighting opportunities for targeted process improvement.Healthcare organizations should evaluate data acquisition methods to identify potential disparities in data completeness that can impact quality of clinical care and generalizability of selfreported data.

K E Y W O R D S
4][15][16] This study aimed to describe patient portal usage and EPQ completion at a high-volume comprehensive cancer center and to identify differences by patient clinical characteristics and sociodemographic variables.

| Study population and study design
Moffitt serves a large primary catchment area composed of 15 Florida counties encompassing 30% of the state's population, 17 and serves patients from all 50 states and more than 140 countries. 18Patients seek care at Moffitt for services across the cancer continuum, including prevention, screening, diagnosis, treatment, survivorship care, second opinions, and therapy for select benign or premalignant conditions.To better understand the characteristics of patients who complete or do not complete the EPQ and assess the generalizability of these self-reported data for research, we conducted a comprehensive analysis of all new adult patients from 2009 to 2020.

| Data collection
Data were obtained via various data fields from electronic data capture systems, with different data types available for different years, as described below.

| Patient-reported data
In 2009, Moffitt sought to convert existing paper-based NP intake forms into a single EPQ administered to all NPs.To harmonize existing paper-based forms developed by individual clinics, a multidisciplinary team of clinical faculty, staff, and researchers was convened and has governed EPQ content to this day.Questions were included in the EPQ only if there was clear clinical utility.The specific formatting of questions was determined based on existing standards and potential downstream uses of the data, including clinical decision support, quality improvement, and research.Moffitt's Patient Family Advisory Council was consulted during the EPQ revision process.
The EPQ was designed to be completed by patients once, as a baseline patient intake form, and includes 12 core modules with questions pertaining to demographics, visit history, personal and family history of cancer, other family history, women's reproductive history, cancer risk factors, past medical history, current symptoms, surgical history, psychosocial assessment, and quality of life (Table S1).Based on different skip patterns, patients who received the core EPQ may answer anywhere between 87 and 230 questions.In addition to these core EPQ modules, disease-specific "spokes" include questions uniquely relevant to subsets of cancer patients, such as questions pertaining to the diseases themselves, relevant risk factors, and quality of life considerations; these disease-specific questions were interspersed with "core" questions to optimize question flow.Spokes are currently used for patients with thoracic, cutaneous, prostate, bladder, colorectal, pancreas, liver, malignant hematology, ovarian or breast cancers.The EPQ incorporated previously validated tools to examine patient cancer risk, physical health, and quality of life.Examples of such tools include the International Physical Activity Questionnaires, Fagerstorm Test for Nicotine Dependence, Semiquantitative Food Frequency Questionnaire, and Short Form 12 (SF-12) Health Survey.To enhance the quality of clinical care and research for sexual and gender minority populations, questions to ascertain patients' sexual orientation and gender identity (SOGI) were added to the EPQ in 2016.Questions, associated definitions, and response options were modeled on the Fenway Institute's "Do Ask, Do Tell" guidelines. 19he questionnaire has historically been available in English with shorter, paper versions available in Spanish; development of a Spanish-language EPQ is underway.
To optimize accessibility, the EPQ was integrated into Moffitt's electronic patient portal.Once a patient is scheduled for a NP appointment a link to register for the portal and complete the EPQ is emailed to the patient, a process which has been consistently implemented over time.Patients who have not completed the EPQ at the time of their initial appointment are asked to complete it on a tablet in the clinic, although only a small percentage of patients use the tablet (12.8% in 2022).

| Clinic appointments and patient portal utilization
The Soarian scheduling system, a Cerner product, was used to identify the number of patients with NP appointments, select demographic characteristics as recorded by scheduling staff, preferred language, and the subset who accessed the patient portal at least once prior to the date of the NP appointment.This information was available starting in 2014.

| Cancer characteristics
The Moffitt Cancer Registry is a highly curated database used to record all malignancies diagnosed and/or treated at Moffitt that are defined as reportable to Florida's statewide cancer registry. 20Certified tumor registrars and abstractors document information relevant for each cancer diagnosis.In this analysis, we examined EPQ completion by demographic/clinical characteristics among all NPs to Moffitt, as well as within the subset who were included in the Cancer Registry.Variables extracted from the Cancer Registry from 2009 to 2020 included: sex, age at presentation/diagnosis, race/ethnicity, primary cancer site, and stage at diagnosis/presentation.

| Statistical analysis
Numbers of patients scheduled for NP appointments were tabulated by year of visit and demographic characteristics, as were the numbers and percentages of those who accessed the patient portal.Similarly, the number of patients who completed the EPQ were tabulated by year, as were the subset of those patients entered in the Cancer Registry with a date of first contact within 1 year of EPQ completion.Cancer Registry-recorded demographics were described for patients who completed the EPQ, overall and among the subset of patients entered in the Cancer Registry within 1 year of their EPQ completion; chi-squared and Cochran-Armitage tests were used to test statistical significance of observed differences.
Demographic and clinical characteristics for all NPs entered into the Cancer Registry were described.Patient characteristics were described separately among those who did and did not complete the EPQ.Unadjusted logistic regression was used to estimate the association between the odds of completing the EPQ and each of the described variables.Multivariable logistic regression was used to control for potential confounding.Associations between EPQ completion and stage at presentation and stage at diagnosis were adjusted for all demographic characteristics.Multivariable logistic models were used to examine possible interactions between tumor stage and each of the demographic factors including race, ethnicity, and language preference.Percentages of patients who completed and did not complete the EPQ were presented by cancer characteristics.Using data captured in 2016-2020, sex assigned at birth, relationship status, gender identity, and sexual orientation were described for all patients who completed EPQ, and separately, among those with a Cancer Registry-recorded date of first contact within 1 year of EPQ completion.Lastly, among patients in the Cancer Registry, logistic regression was used to examine the association between demographics and whether a patient had complete data for questions on SOGI.
De-identified data were used and therefore institutional review board approval was not required.Analysis were conducted using R version 4.02 (RRID:SCR_001905) 21 with tidyverse (RRID:SCR_019186) 22 and DescTools. 23wo sided p ≤ 0.05 were considered statistically significant.

| Patient portal use and EPQ completion
Patient portal usage increased over time from 76.3% in 2014 to 83.3% in 2020 (Table 1).The number of patients with NP appointments completing the EPQ increased from 743 patients in 2009 to over 17,000 in 2019 and over 15,000 patients in 2020.EPQ completion rates steadily increased from 71.1% in 2014, peaked at 84.2% in 2017, and decreased to 77.7% in 2020, the first year of the COVID-19 pandemic.During 2014-2020, 51% of patients who completed the EPQ were recorded in the Cancer Registry as having been diagnosed and/or treated for cancer within 1 year of EPQ completion.The remaining patients likely received other services at Moffitt, such as second opinions, cancer prevention and screening, and treatment for nonmalignant conditions.
Of the approximately 131,000 patients who were scheduled for NP appointments in 2014-2020, 54% were female, 56% were aged ≥65 years, 86% were White, and 89% were non-Hispanic (Table 2).Statistically significant differences of small magnitude were observed in the proportions of patients with NP appointments who accessed the patient portal across categories of sex and age.Patient portal usage varied by race (82%, 67%, 81%, and 71% for patients of White, Black, Asian, and Other races, respectively, p < 0.0001) and ethnicity (70% and 82% for Hispanic and non-Hispanic patients, respectively, p < 0.0001).EPQ completion rates were >95% for all age groups within the age range of 18-55 years and decreased to 88% and 68% for patients ages 55-64 and ≥65 years, respectively.Higher EPQ completion rates were observed for White and Asian patients as compared with Black and Other races.Similarly, regarding ethnicity, 81% of non-Hispanic patients completed the EPQ compared to 68% of Hispanic patients.The percentage of EPQ completers who were diagnosed and/or treated for cancer within 1 year of EPQ completion was greater for males (57% male vs. 46% female, p < 0.0001) and older patients (57% of patients aged ≥65 vs. 46% aged 18-64, p-trend < 0.0001).different from White patients with respect to EPQ completion.However, compared to White patients, Black or African American patients (OR = 0.59, 95% CI = 0.56-0.63)and patients categorized as other race (OR = 0.56, 95% CI = 0.52-0.61)were less likely to complete the EPQ.Similarly, Hispanic patients were less likely to complete the EPQ than non-Hispanic patients, although this difference was not statistically significant in the multivariable model that included language preference, where patients who prefer Spanish (OR = 0.40, 95% CI = 0.36-0.44)or other languages (OR = 0.60, 95% CI = 0.50-0.71)had significantly lower EPQ completion rate as compared to patients who prefer English.Notably, a minimal proportion (0.2%) of patients identified as being both Hispanic and Black.EPQ completion rates ranged from 59% of patients with lung cancer to 78% of patients with breast cancer or uterine cancer (Figure 1).Compared to patients with distant disease at the time of presentation to Moffitt, the EPQ was more often completed by patients with less advanced disease (Figure 2).For example, patients presented at Moffitt with a localized (OR = 1.16, 95% CI = 1.12-1.19) or a regional disease (OR = 1.16, 95% CI = 1.12-1.20)were 16% more likely to complete the survey than those who had metastatic disease.No statistically significant interaction effects were found between stage at presentation and race, ethnicity, or language preference, respectively (p-interaction = 0.162-0.894).Similar results were observed for stage at diagnosis (Figure 2).

| Sexual orientation, gender identity, and relationship status
Self-reported sex at birth and relationship status among the 145,486 patients completing EPQ in 2009-2020 as well as gender identity and sexual orientation among the 69,540 patients completing EPQ in 2016-2020 are described in Table 4. Fifty-four percent of EPQ respondents reported female sex at birth, 44% reported male, 4 reported "other" and 67 declined to answer.Using data collected after 2016, gender identity was reported as follows: 54.5% female, 45.2% male, 16 female-to-male transgender, 8 male-to-female transgender, 5 genderqueer, 15 identified with another gender category, and 89 declined to answer.Regarding sexual orientation, 1298 (2%) of EPQ respondents reported being lesbian, gay, or homosexual, 298 (0.4%) reported being bisexual, 360 (0.5%) reported "something else", and 259 (0.4%) responded "don't know".Notably, patients who were male, of younger age, and racial minorities were more likely to complete questions on SOGI than patients who were female, of older age, and White, respectively (Table S2).Most EPQ respondents (66%) reported being married, followed by 11% single, and 11% divorced or separated.Among the subset of EPQ respondents who were diagnosed and/or treated for cancer that same year, sex at birth and gender identity were more evenly distributed across males and females than for EPQ respondents overall, with other gender identities comprising approximately 0.1% of both groups.Sexual orientation and relationship status were similarly distributed among all EPQ respondents and the subset in the cancer registry.

| DISCUSSION
6][27][28] The Moffitt patient portal provides patients with access to several services relevant to clinical care (e.g., scheduling) and encourages advanced completion of clinically relevant patient forms such as the EPQ, which may contribute to high completion rates.In addition, the availability of the EPQ via tablet in the clinic also serves to optimize completion rates, 7 particularly among patients accessing emergent care.
We aimed to assess the possibility that downstream analyses could be prone to selection bias and to identify subgroups to target in future efforts to encourage completion.Although absolute differences in completion rates across demographic categories were minimal (5%-10%), these differences reached statistical significance, likely due in part to the large samples sizes.Nevertheless, attention is warranted regarding the lower completion rates observed in males versus females, older ages versus younger ages, Black/African American, and other non-White/non-Asian races versus White and Asian races, and patients whose preferred language is Spanish or another language other than English. 291][32] Future research should implement and test additional targeted strategies to improve accessibility of portals and validated questionnaires in patients' native languages and targeted outreach to racial/ ethnic minority patients to address technical, language, and access barriers. 30,31ifferences in EPQ completion were observed for primary cancer site, however, the overall number of questions did not vary greatly across disease spokes.Patients with more extensive disease, including those with worse symptoms, may be less likely to complete a lengthy questionnaire.For example, patients with metastatic disease had lower completion rates than those with less advanced disease which may be due in part to the physical and/ or psychosocial burdens associated with advanced cancer and the treatment side effects.Thus, future quality improvement efforts should seek to minimize patient burden, taking into account unique needs of patients with advanced disease and tailoring content accordingly, with critical input from patients.
The EPQ provides a unique opportunity for patients to discreetly share their SOGI with their oncology providers.Of the 69,540 NPs who completed the EPQ in 2016-2020, 1596 (2.3%) self-identified as being lesbian, gay, or bisexual and 44 (<0.1%) self-identified as transgender, genderqueer, or an additional gender category.These proportions are lower than estimates for the general population. 33This may be due to the older age of cancer patients compared to the general population, as a U.S. nationwide survey indicated that the percentage of the population identifying as a sexual orientation or gender identity minority is 7.1% overall, including 2.6% of individuals ages 58-72, and 0.8% of those over age 72. 33Another explanation may be that some patients chose not to disclose their sexual orientation or gender identity.Based on our analysis, these patients are more likely to be female, of older age, and racial minorities.Future research should assess sexual orientation and gender minority individuals' preferences for self-identifying to healthcare providers and revise data collection instruments accordingly.
Moffitt was an early adopter of portal technology for patient reported information capture (2009) and collection of self-reported data on sexual and gender identity (2016), well ahead of the goals set by the Centers for Medicare and Medicaid Services' Promoting Interoperability Program. 34,35As a result, the data included in the current analysis span over a decade and include more than 100,000 patients, representing a major study strength.Some limitations should also be noted.Scheduling data were not available prior to 2014 at which time there was a change in the scheduling systems, and limited clinical information was available for patients not entered into the Cancer Registry; therefore, portal utilization and EPQ response could not be examined as functions of non-cancer diagnoses and/or procedures such as screening exams.Since EPQ data are captured only at the time of clinical intake, current study findings may not be generalizable to the electronic acquisition of patient-reported outcomes in the follow-up setting.Finally, results of this analysis may not be directly generalizable to other healthcare institutions, although the overall approach is important and generalizable, in that healthcare organizations should be encouraged to evaluate their data acquisition methods across patient subgroups to identify potential disparities in data completeness.
In summary, rates of patient portal use and EPQ completion were robust over time and among subgroups of patients defined by demographic and clinical factors.Differences in EPQ completion between patient subgroups can provide important insights for process improvement impacting downstream use of the data including clinical care, quality improvement, and research projects.Future studies are needed to implement and test targeted strategies to increase access to and use of electronic questionnaires among racial/ethnic minority patients as well as patients with more advanced disease and to extend this work into the follow-up setting.While digital tools can rapidly scale data collection in the clinical setting, it is important to minimize patient burden and maximize both clinical utility and patient satisfaction.Finally, in an era where real world data are increasingly being used in research and cancer care delivery, an intentional and purposeful examination of the representativeness of minority populations in those datasets is imperative to ensure predictive models trained on such data do not perpetuate existing healthcare disparities.

c
Asian Indian, Pakistani, and other Asian were grouped as Asian to compare with White in the logistic regression analysis.d American Indian or Alaska, native Hawaiian, and other were grouped as other race to compare with White in the logistic regression analysis.e The top 10 other languages preferred by the patients were: Arabic, Vietnamese, Creole, Russian, Polish, Portuguese, Mandarin, sign language, Korean, and Albanian.T A B L E 3 (Continued) F I G U R E 1 Electronic Patient Questionnaire (EPQ) completion by primary site of the cancer for new patients to Moffitt who received a cancer diagnosis between 2009 and 2020.The bar graph depicts the percent of patients who completed (green) and did not complete the EPQ (orange) by the primary site of the cancer among patients who received a new cancer diagnosis between 2009 and 2020.Sites that represented less than 2% of the total sample size were collapsed into the "Other" category.

F I G U R E 2
Electronic Patient Questionnaire (EPQ) completion by stage of the cancer for new patients to Moffitt who received a cancer diagnosis between 2009 and 2020.The percent of patients who completed (green) and did not complete the EPQ (orange) by (A) stage at diagnosis and (B) stage at presentation among patients who received a new cancer diagnosis between 2009 and 2020.The odds ratios presented were adjusted for sex, age at presentation, age at diagnosis, race, and ethnicity.

Table 3
presents associations between demographics and EPQ completion among NPs in 2009-2020 who were diagnosed and/or treated for cancer that same year.Females were more likely to complete the Patient portal usage and electronic patient questionnaire (EPQ) completion among patients with a new patient (NP) appointment at Moffitt in 2009-2020.
T A B L E 1

Portal usage EPQ completion EPQ completion and new diagnosis Patient characteristics a Number of patients with NP appointment Number of patients with NP appointments and used portal by the time of the NP appointment Number of patients who completed EPQ Number of patients who completed EPQ and had a date of first contact within 1 year of EPQ completion
the number of unique patients with NP appointments and used portal by the time of the NP appointment out of the Number of unique patients with NP appointments.the number of unique patients who completed the EPQ and were entered in Cancer Registry within 1 year of EPQ completion out of the number of patients who completed the EPQ.Patient characteristics, portal usage and electronic patient questionnaire (EPQ) completion among patients with a new patient (NP) appointment at Moffitt in 2014-2020.
aThe Soarian Scheduling System was used as the data source for this table.bPercentagecalculatedasc Percentage calculated as the number of unique patients who completed the EPQ out of the number of patients with NP appointments.dPercentagecalculatedas EPQ than males (odds ratio [OR] = 1.17, 95% confidence interval [CI] = 1.14-1.20).EPQ completion rates were lower among those ages 18-25 (OR = 0.70, 95% CI = 0.52-0.93)andhigheramongthose ages 56-64 (OR = 1.16, 95% CI = 1.09-1.23)ascomparedtothose aged ≥65 (Table3).Asian patients were not statisticallyT A B L E 2 Percentage calculated as the number of unique patients with NP appointments and used portal by the time of the NP appointment out of the Number of unique patients with NP appointments.cPercentagecalculated as the number of unique patients who completed the EPQ out of the number of patients with NP appointments.Percentage calculated as the number of unique patients who completed the EPQ and were entered in Cancer Registry within 1 year of EPQ completion out of the number of patients who completed the EPQ.Portal data race categorizations differ from those used in the Moffit Cancer Registry.Thus, the following race subcategories were combined as Asian: Asian Indian, Pakistani, Chinese, Filipino, Japanese, Kampuchean (including Khmer, Cambodian), Korean, Laotian, Hmong, Melanesian, Other Asian including Asian and Oriental, Thai, and Vietnamese; Native Hawaiian or Other Pacific Islander: Guamanian, Hawaiian, Samoan, Chamorran, Fiji Islander, Tahitian, Tongan, New Guinean, Micronesian, Polynesian, and Pacific Islander; and White: White and White Hispanic.Electronic Patient Questionnaire (EPQ) completion by patient demographic characteristics for new patients to Moffitt who received a cancer diagnosis between 2009 and 2020.
b d e p-Values were calculated using chi-squared tests for sex, race, and ethnicity with missing data excluded from the analysis.fp-valueswere calculated using the Cochran-Armitage trend test for age with missing data excluded from the analysis.ghFor sample size considerations, when analyzing race, American Indian, Aleutian, Eskimo, native Hawaiian or other Pacific islander, other race, or multiple races were combined as one category to compare with Asian, Black, and White.iWhenanalyzing ethnicity, multiple ethnicities were combined with Hispanic to compare with non-Hispanic.T A B L E 2 (Contiuned)T A B L E 3 Patients in this table were restricted to those whose date of first contact were between 2009 and 2020.
a The MCC Cancer Registry was used as the data source for this table.b

EPQ SOGI data (available years) All Moffitt patients who completed an EPQ a EPQ completers newly diagnosed and/ or treated for cancer at Moffitt within 1 year of EPQ completion b
Electronic Patient Questionnaire (EPQ) completion by patient sexual orientation and gender identity (SOGI) for new patients (NPs) to Moffitt between 2009 and 2020.Includes all NPs to Moffitt who completed an EPQ in 2009-2020, including screening patients, second opinions, those being treated for premalignant or nonmalignant conditions, as well as patients diagnosed with and/or treated for cancer.b Includes the subset of NPs to Moffitt who completed an EPQ and for whom a date of first contact as recorded in the cancer registry was within 1 year of EPQ completion (i.e., new Moffitt patients who were diagnosed and/or treated for a reportable malignancy the same year as EPQ completion).
T A B L E 4a