3 articles need to answer these questions: What are the elements of the study design, and why the authors made the choices they did? Identify the importance of this study and what it adds to the knowl

3 articles need to answer these questions: What are the elements of the study design, and why the authors made the choices they did? Identify the importance of this study and what it adds to the knowl
185 Original Report: Precision Medicine in Health Disparities Research IntroductIon Prostate cancer is one of the lead – ing causes of cancer among men in the United States. 1 In addition to African American race, established risk factors for prostate cancer include family his – tory of prostate cancer and increasing age; the average age at prostate cancer diagnosis is sixty-six years. 1 Increasing age is also a risk factor for developing chronic conditions that include hy – pertension, diabetes, high cholesterol, and cardiovascular disease. 2 Thus, as men age, they are at risk for develop – ing multiple acute and chronic condi – tions that increase their likelihood of morbidity and mortality. A substan – tial proportion of prostate cancer pa – tients have at least one co-morbidity, or a chronic condition that is distinct from their primary prostate cancer di – agnosis. 3 Previous research has shown that being diagnosed with prostate cancer and having a co-morbid con – dition (eg, diabetes, hypertension, cardiovascular disease) is associated with an increased risk of dying from causes other than prostate cancer. For example, prostate cancer patients in the Surveillance, Epidemiology, and Endpoints Registry (SEER) who had two or more chronic conditions had a 43% to 48% chance of dy – ing from any cause within five years of their prostate cancer diagnosis. 4 Because co-morbidities are com – mon among men who have a per – sonal history of prostate cancer and these other chronic conditions may C o -morbidities in a r etrospe Ctive C ohort of p rostate C an Cer p atients Melanie Jefferson, PhD 2,3; Richard R. Drake, PhD 1,2; Michael Lilly, MD 2,4; Stephen J. Savage, MD 2,5; Sarah Tucker Price, MD 6; Chanita Hughes Halbert, PhD 2,3 Objective: To characterize rates of co- morbidity among prostate cancer patients treated with radical prostatectomy and to examine the association between co-mor – bidity status and race, clinical factors, and health behaviors for cancer control. Design/Study Participants: Retrospective cohort study among prostate cancer patients treated with radical prostatectomy. Setting: Academic medical center located in the southeastern region of the United States. Main Outcome Measure: Patients with at least one of five co-morbid conditions considered were categorized as having a co-morbidity, and those without any were categorized as not having a co-morbid condition. Co-morbid conditions consid – ered were hypertension, diabetes, heart problems, stroke, and high cholesterol, which had been recorded in the electronic medical record as part of their past medical history. Results: Fifty-one percent of participants had a co-morbidity, with hypertension being the most common. The average number of co-morbidities among study participants was .87. In a multivariate logistic regression analysis, being diagnosed with prostate can – cer within the past four years was associated with an increased likelihood of having a co- morbidity (OR=4.71, 95% CI=2.69, 8.25, P=.0001) compared with diagnosis five or more years ago. Age was also associated with an increased likelihood of having a co- morbidity (OR=1.30, 95% CI=1.005, 1.68, P=.05). In this study cohort, race, stage at diagnosis, and PSA level were not statisti – cally associated with co-morbidity status. Conclusion: Better chronic disease management is needed among prostate cancer survivors through more effective survivorship care planning and interventions that promote health behaviors. Ethn Dis. 2020;30(Suppl 1):185-192; doi:10.18865/ ed.30.S1.185 Keywords: Co-morbidities, Retrospective Cohort, Prostate Cancer 1 Department of Cell and Molecular Pharmacology and Experimental Therapeutics, Medical University of South Carolina, Charleston, SC2 Hollings Cancer Center, Medical University of South Carolina, Charleston, SC 3 Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, Charleston, SC4 Department of Medicine, Medical University of South Carolina, Charleston, SC5 Department of Urology, Medical University of South Carolina, Charleston, SC6 Department of Family Medicine, Medical University of South Carolina, Charleston, SC Address correspondence to Chanita Hughes Halbert, PhD, Medical University of South Carolina, 68 President Street, Suite BE103, Charleston, SC 29425; 843.876.2421; [email protected] Ethnicity & Disease, Volume 30, Supplement 1, 2020 186 be their cause of death, co-morbidity status should be integrated and con – sidered as part of making decisions about prostate cancer treatment. 5-7 To do this, it is important to have an understanding of the distribution of co-morbidities among diverse patient populations. In population-based samples, for instance, African Ameri – cans are more likely to have hyperten – sion and cardiovascular disease com – pared with Whites. 8,9 Characterizing the distribution of co-morbidities specifically among men who have a personal history of prostate cancer is To extend previous research that examined co-morbidity in prostate cancer patients who were treated with all treatment modalities, 3 the purpose of this study was to characterize co- morbidities among prostate cancer patients treated with radical prosta – tectomy. Because of racial differences in the rates of chronic conditions (eg, hypertension, cardiovascular disease), we were also interested in determin – ing if minority and non-minority prostate cancer patients differed in terms of having co-morbidities and the types of chronic conditions with which they have been diagnosed. We hypothesized that minority pa – tients would be more likely to have at least one co-morbidity compared with non-minority patients. An ad – ditional objective of this study was to examine the association between co-morbidities and tumor char – acteristics (eg, stage of disease) to provide insight about how adverse prognostic factors for prostate can – cer are associated with the potential risk of death from chronic conditions among prostate cancer survivors. Lastly, since dietary behaviors and physical activity are behavioral risk factors for chronic conditions, 12 we also examined the relationship be – tween co-morbidity status and fruit/ vegetable intake and physical activity. M ater Ia ls a nd M ethods Study Population Participants in this study were men who had a personal history of prostate cancer and had provided a tissue sample as part of having a radi – cal prostatectomy. Prostate cancer tis – sue samples were collected by the Bio – repository and Tissue Analysis (BTA) Shared Resource at the Hollings Can – cer Center (HCC) after patients pro – vided written informed consent and privacy authorization using institu – tional guidelines at the Medical Uni – versity of South Carolina (MUSC). As part of this informed consent process, men agreed for their tissue samples to be used as a part of cancer research and agreed to be contacted about participating in future studies. More than 90% of prostatectomy pa – tients provided consent for their tis – sue sample to be stored in the BTA. The HCC Biorepository was queried to identify men who had an ICD-10-CM code of C61 (malignant neoplasm of prostate) and CPT codes of 55810 (prostatectomy, perineal radical) and 55866 (laparoscopic pro – cedures on the prostate) since 2011. The resulting study sample included 316 prostate cancer patients who had a tissue sample available in the HCC Biorepository when study recruitment was initiated in 2016. Of these 316, 83 (26%) completed a structured so – cial determinants survey that provid – ed the data to examine the association between health behaviors for can – cer control and co-morbidity status. Procedures All study procedures were ap – proved by the institutional review board at MUSC. First, information on sociodemographic characteristics (eg, race, age), prostate cancer vari – ables, and co-morbidities was ab – stracted from the electronic health record (EHR) of eligible patients who were identified from the HCC Biore – pository. Next, patients were contact – …the purpose of this study was to characterize co- morbidities among prostate cancer patients treated with radical prostatectomy. also important for survivorship care planning. This is especially true given that it was estimated that 164,690 new prostate cancer cases would occur in 2018 with only 29,430 deaths. 10 Receiving quality care for co-morbidities during and after the acute treatment phase for cancer is necessary to reduce the likelihood of morbidity and mortality from chron – ic illnesses among cancer patients. Re – cent research has shown that having a greater number of co-morbidities and being African American are associ – ated with wanting more information to help guide their follow-up care. 11 Ethnicity & Disease, Volume 30, supplement 1, 2020 Co-morbidities in Prostate Cancer Patients – Jefferson et al 187 ed by mailed invitation to complete a structured telephone interview that measured social determinants and health behaviors for cancer control. Patients could decline to participate in the social determinants survey by contacting the program manager at MUSC by telephone or email. Those who did not decline to com – plete the social determinants survey were contacted by a research assistant at MUSC to complete a 30-min – ute social determinants survey. Measures The following data elements were abstracted from the EHR for each study participant: year of birth; date of diagnosis; prostate specific anti – gen (PSA) levels at diagnosis; patho – logic stage at diagnosis (T1a; T1c; T2a; T2b; T2c; T3a; T3b); Gleason score; race (White, African Ameri – can); height; weight; and systolic and diastolic blood pressure at the time of the pre-surgical consultation visit. Co-morbidities were obtained from the patient’s problem list as recorded in the EHR. Patients were categorized as having a history of hypertension (yes or no), diabetes (yes or no), heart problems (yes or no), stroke (yes or no), or high cholesterol (yes or no). We focused on these chronic condi – tions because they are among the leading causes of morbidity and mor – tality in the United States 13 and have been associated with an increased risk of all-cause mortality among prostate cancer patients. 4,14 Height and weight values were used to calculate body mass index using the Centers for Dis – ease Control and Prevention BMI calculator. 15 Stage was recoded into a binary variable of early vs later stage disease (T1/T2 vs T3). The amount of time since diagnosis was calculated based on the date of diagnosis. We recoded time since diagnosis into a binary variable of within the past four years or five or more years from the date study recruitment was initiated. The social determinants survey as – sessed self-reported race and ethnic – ity, self-report on the co-morbidities in the study, and fruit/vegetable in – take and physical activity using items from the Health Information Nation – al Trends Survey (HINTS). 16 Specifi – cally, men were asked how many cups of fruit and vegetables they eat each day (1=none/don’t know, 2=½ cup or less, 3=½ to 1 cup, 4=1 to 2 cups, 5=2 to 3 cups, 6=3 to 4 cups, 7=more than four cups). Men who reported eating at least 2 to 3 cups were categorized as meeting recommended guidelines for each dietary behavior variable. Next, men were asked if they had par – ticipated in any physical activities or exercises during the past month (yes or no). Those who reported yes were asked how many days they were phys – ically active or exercised of at least moderate intensity and on these days, how long they typically performed these behaviors. The total number of minutes for moderate intensity physi – cal activity per week was calculated by multiplying the number of days by the minutes reported. Men who reported no physical activity during the past month, those who reported that they had been physically active during the past month, but had not been active during the past week, and those who reported physical activity, but did not meet the physical activ – ity guidelines (eg, less than 150 min – utes/week) were coded as not meeting guidelines. 17 The remaining partici – pants were coded as having met the guidelines for physical activity. 17 Data Analysis Descriptive statistics were gener – ated first to characterize the study patients (n=316). Next, frequencies were generated to identify the co- morbidities that were most and least common among participants. Chi square tests of association and t-tests were performed to examine the asso – ciation between co-morbidity status, race, and clinical variables. Variables that had a bivariate association of P<.25 with co-morbidity status were included in the multivariate logistic regression analysis. Lastly, multivari - ate logistic regression analysis was used to identify factors having signifi - cant independent associations with co-morbidity status. Race was also included in the regression model re - gardless of the significance of the bi - variate association with co-morbidity status because of disparities in chron - ic diseases. This same approach was used to examine the relationship be - tween co-morbidity status and fruit/ vegetable intake and physical activ - ity in the sub-set of men who com - pleted the social determinants survey. r esults Table 1 shows the characteristics of the study patients. Thirty-two per - cent of patients were racial minori - ties (eg, African American) and 68% where non-minorities. The mean (SD) age was 65.2 (6.7). With respect to prostate cancer variables, the mean PSA was 9.2 (SD=10.9) and 77% of Ethnicity & Disease, Volume 30, Supplement 1, 2020 Co-morbidities in Prostate Cancer Patients - Jefferson et al 188 men had been diagnosed with stage T2 disease and 23% had been diag - nosed with stage T3 disease. In addi - tion, 72% of men had a Gleason score of 3+4 or 3+3. Sixty-nine percent of men were diagnosed within the past four years and 31% were diagnosed more than five years ago. In the sub- set of men who completed the social determinants survey, 31% met guide - lines for fruit intake, 25% met guide - lines for vegetable intake, and 30% met guidelines for physical activity. With respect to co-morbidity status, 51% of men had at least one co-morbid condition; men were most likely to have high blood pres - sure (42%), high cholesterol (24%), diabetes (12%), heart problems (9%), and stroke (.63%). Among those who had at least one co-morbidity, the mean (SD) number was .87 (1.02). Table 2 shows the results of the bivariate analyses of co-morbidity. Time since diagnosis had significant bivariate association with co-morbid - ity status. Men who were diagnosed within the past four years were more likely to have a co-morbidity com - pared with those who were diagnosed more than four years ago (61% vs 30%, chi square=25.4, P=.0001). The mean PSA was also higher among men who had a co-morbid - ity (mean=10.2, SD=13.6) com - pared with those who did not have a co-morbidity (mean=8.2, SD=6.6) (t=-1.70, P=.09). Sixty-one percent of men who had stage T3 disease had a co-morbidity compared with 49% of men who were diagnosed with stage T2 disease (chi square=3.04, P=.08). There were also small mean differ - ences in age based on co-morbidity status. For instance, the average (SD) age was 65.7 (6.5) among men who had a co-morbidity compared with 64.8 (7.0) among those without any co-morbidities (t=-1.20, P=.23). Fifty-five percent of men from racial minority groups had a co-morbidity compared with 49% of non-minor - ity men (chi square=1.04, P=.31). BMI was similar between men who had at least one co-morbid condi - tion (mean=29.5, SD=5.6) and those who did not have any co-morbidities (Mean=29.4, SD=4.4) (t-value=-.20, P=.84). Among the sub-set of men who completed the social determi - nants survey, none of the behavioral risk factors (fruit/vegetable intake or physical activity) were associated sig - nificantly with having a co-morbidity. The results of the multivariate lo - gistic regression model of co-morbidi - ty status are provided in Table 3. Only time since diagnosis had significant independent association with having a co-morbid condition. Men who were diagnosed within the past four years had a greater likelihood of hav - ing a co-morbid condition compared with those who were diagnosed more than four years ago (OR=4.71, 95% CI=2.69, 8.25, P=.0001). The likeli - hood of having a co-morbid condition increased with older age (OR=1.30, 95% CI=1.005, 1.68, P=.05). d Iscuss Ion The purpose of this study was to examine co-morbidity rates among prostate cancer survivors who were treated with radical prostatectomy. Overall, 51% of men had one co- morbidity, the average number of co-morbidities was .87, and high blood pressure was the most com - mon co-morbid condition. Forty- two percent of men had high blood pressure, but less than 1% had a his - tory of stroke. Chronic disease man - agement among cancer patients and survivors is an important priority 18; our findings underscore the need for greater chronic disease management among prostate cancer patients, re - gardless of their racial background, especially when they are within the first four years of being diagnosed. The overall rates for co-morbidi - ties in our sample were higher than those reported in previous research, 3 but there was some consistency in the Table 1. Study patient characteristics Variable Level n (%) Race Minority 101 (32%) Non-minority 215 (68%) Gleason score 4+3/4+4/4+5/5+3/5+5 79 (28%) 3+3/3+4 201 (72%) Stage T3 69 (23%) T2 229 (77%) Time since diagnosis Within past four years 217 (69%) Five or more year 99 (31%) Age Mean (SD) 65.2 (6.7) PSA Mean (SD) 9.2 (10.9) BMI Mean (SD) 29.4 (5.0) Ethnicity & Disease, Volume 30, supplement 1, 2020 Co-morbidities in Prostate Cancer Patients - Jefferson et al 189 rates for individual co-morbidities in our sample and other studies. For in - stance, Edwards et al 3 found that 13% of prostate cancer patients in a na - tional sample had a history of diabetes whereas 11% of patients in our study had a history of this disease. Further, 56% of prostate cancer patients who had a radical prostatectomy reported a history of hypertension 18 and 42% patients in our sample had hyperten - sion. However, 30.5% of prostate cancer patients in a national sample had at least one co-morbidity. 3 Our higher overall rates of co-morbidity may be due to our including hyper - tension, whereas other studies, in - cluding the study by Edwards and colleagues, based co-morbidity on conditions included in the Charlson Co-Morbidity Index, which does not include hypertension. 3,19,20 Together with the findings from previous stud - ies demonstrating that hypertension is associated with an increased risk of biochemical recurrence among men treated with radical prostatectomy, 21,22 the exclusion of hypertension in stud - ies that examine prostate cancer out - comes may be a significant omission. In addition to being a risk factor for all-cause mortality and death from cardiovascular disease, 9 hypertension was associated with an increased risk of biochemical recurrence among prostate cancer patients. 21,22 There continues to be significant racial dis - parities in prostate cancer incidence and mortality. 1 African American men have the greatest incidence of pros - tate cancer among men in the United States and are about twice as likely as White men to die from this disease. 1 Previous research has shown that Af - rican Americans are more likely than Whites to have high blood pressure. 9 Specifically, hypertension was associ - ated with a two-fold increase in bio - chemical recurrence among African American and White men who were treated with radical prostatectomy. 22 While Post and colleagues 22 found that African American prostate can - cer patients were significantly more likely to have hypertension compared with White patients, there were non - significant racial differences in overall rates of hypertension in our study and mean levels of systolic and diastolic blood pressure did not differ between Table 2. Bivariate analysis of comorbidity status Variable Level % Comorbidity Chi Square P Race Minority 55% 1.04 .31 Non-minority 49% Gleason score 4+3/4+4/4+5/5+3/5+5 52% .03 .86 3+3/3+4 51% Stage T3 61% 3.04 .08 T2 49% Time since diagnosis Within past four years 61% 25.4 .0001 Five or more year 30% Co-MorbidityMean (SD) No-Comorbidity Mean (SD) T-Value P Age 65.7 (6.5) 64.8 (7.0) -1.20 .23 BMI 29.5 (5.6) 29.4 (4.4) -.20 .84 PSA 10.2 (13.6) 8.2 (6.6) -1.70 .09 Table 3. Logistic regression model of comorbidity Variable Level Odds ratio 95% CI P Race Minority 1.56 .90, 2.69 .11 Non-minority Age a 1.30 1.005, 1.68 .05 Stage T3 1.20 .65, 2.21 .56 T2 Time since diagnosis Within past four years 4.71 2.69, 8.25 .0001 Five or more years PSA a 1.17 .85, 1.61 .33 a. ORs for continuous variables reflect the OR for a 1-SD unit change in the covariate. Ethnicity & Disease, Volume 30, Supplement 1, 2020 Co-morbidities in Prostate Cancer Patients - Jefferson et al 190 minorities and non-minorities in our sample (data not shown). However, our sample showed higher blood pressures and higher BMI measures overall. The average (SD) systolic and diastolic blood pressures were 142.1 (17.5) and 82.3 (9.3), respec - tively, the average (SD) BMI was high (29.4, 5.0), and 37% of men in our sample were obese. This may explain why there were no racial differences in co-morbidity status in our study. We found that men who had been diagnosed with prostate cancer within the past four years had a sig - nificantly increased likelihood of hav - ing a co-morbid condition compared with those who had been diagnosed five or more years ago. This may be due to temporal changes in the ex - tent to which co-morbidities are re - corded in electronic medical records. All health care providers and systems were required to implement and dem - onstrate meaningful use of electronic medical records in January 2014 23; there may be greater documentation of co-morbidities in electronic medi - cal records as information systems were introduced and expanded to meet federal requirements. However, recent research has shown high agree - ment between patient self-reported co-morbidities and documentation of these conditions in the medical re - cord. 24 Further, a similar proportion of men had specific co-morbidities based on self-report and electronic medical record. For instance, 44% of men self-reported hypertension and 44% of men had hypertension according to the electronic health record. Similarly, 8% of men self-re - ported a personal history of diabetes. Additional research is needed to de - termine why men who have a shorter time from prostate cancer diagno - sis are more likely than longer-term survivors to have a co-morbidity. Interestingly, none of the behav - ioral risk factors (eg, diet, physical activity) for co-morbidities were as - sociated significantly with having a chronic disease among men who completed the social determinants survey. This may be due to the small number of men who were included in this analysis; however, it is important to note that a minority of these partic - for chronic disease management 25; and our findings emphasize the im - portance of developing behavioral interventions to enhance these behav - iors in prostate cancer survivors, espe - cially those who have a co-morbidity. Study Limitations In considering the results of this study, some limitations should be noted. First, co-morbidities were ex - amined among men who were treated with radical prostatectomy at one academic health center, and men who have several comorbid condi - tions do not receive surgery as their primary treatment. Therefore, our study may reflect the lowest percent - age of co-morbidity for prostate can - cer diagnosis. Co-morbidities should be examined among men who have been treated with different modali - ties at diverse academic and commu - nity oncology clinical settings. Sec - ondly, co-morbidity was determined based on the presence of the leading causes of death in the United States (eg, cardiovascular disease, stroke, hypertension, diabetes) at the time of medical abstraction in a retrospective cohort of prostate cancer patients. Other co-morbidity indices include a more extensive list of conditions 20; however, these measures may not ask about chronic diseases that are com - mon in minority populations. No - tably, the inclusion of hypertension in our measure of co-morbidity may be a better reflection of the chronic disease burden in diverse samples of prostate cancer patients. Lastly, it is also important to determine co-mor - bidity status prospectively at the time of diagnosis to be able to examine the association between chronic condi - Overall, 51% of men had one co-morbidity, the average number of co- morbidities was .87, and high blood pressure was the most common co-morbid condition. ipants met recommended guidelines for fruit/vegetable intake and physical activity. Recommendations for physi - cal activity cancer survivors include 150 minutes of moderate intensity exercise weekly 25; however, only 30% of men in our study met this recom - mendation. Similarly, only 31% and 25% met recommended guidelines for fruit and vegetable intake, respec - tively. Diet behaviors and physical activity are important strategies for cancer control among prostate cancer survivors and are also recommended Ethnicity & Disease, Volume 30, Supplement 1, 2020 Co-morbidities in Prostate Cancer Patients - Jefferson et al 191 tions and prostate cancer outcomes. Detailed information on when prostate cancer patients were diag - nosed with chronic diseases should be captured as part of prospective studies to understand the trajectory of co-morbidity in these patients. Study Implications Despite these potential limita - tions, the results of our study have important implications for prostate cancer survivorship. First, our find - ings demonstrate that chronic disease management is needed among pros - tate cancer patients. Survivorship care plans are now being implemented at the conclusion of cancer treatment to facilitate the patient’s transition back to primary care by summariz - ing their cancer diagnosis, treatment, and follow-up care 26,27; however, our findings suggest that efforts may also be needed to manage diseases such as hypertension and to promote cancer control behaviors at all phases of can - cer survivorship. A little more than 50% of the patients in our study had at least one co-morbidity, 37% were obese, and only about one third of men met recommended guidelines for diet and physical activity. Although we were not able to determine the specific age and date at which men were diagnosed with co-morbid con - ditions such as diabetes and hyper - tension because this information was not recorded in the electronic medi - cal record, blood pressure and obe - sity were measured at the time of the pre-surgical consultation visit. Fur - ther, the mean levels for systolic and diastolic blood pressure were above normal ranges and 87% and 57% of participants in our study had values that were above 120 mm Hg and 80 mm Hg, respectively, regardless of hy - pertension status. Blood pressure may have been elevated due to anxiety at the time of the pre-surgical consul - tation visit. However, national data show that only about 50% of individ - uals who have hypertension have this condition under control, men were less likely than women to have con - trolled hypertension, and there are racial differences in the rates of con - trolled disease in national samples. 28 Research in breast cancer patients has shown that adherence to noncan - cer medications for chronic condi - tions decreases during the first year after treatment; potential reasons for reduced adherence to noncancer medications include greater priori - tization of cancer treatment and fi - nancial toxicity. 29 To our knowledge, financial toxicity and adherence to noncancer medications has not been examined specifically among prostate cancer patients and these are impor - tant areas for future research. Because of the high burden of chronic disease and the potential for cancer patients to reduce their adherence to noncan - cer medications following their diag - nosis and treatment, 29 primary care services may need to be integrated into oncology care. Primary care on - cology is emerging as a cancer care service in which providers focus on the medical and psychological impact of cancer treatment, with the manage - ment of co-morbidities as one com - ponent of this specialty. 30,31 Primary care oncologists could play an impor - tant role in managing co-morbidities at diagnosis, through treatment, and during short- and long-term survi - vorship among prostate cancer pa - tients. Future studies are needed to evaluate the impact of primary care oncology on the management of co- morbidities and prostate cancer out - comes in diverse patient populations. acknowledgeMents This study was supported by National Institute of Minority Health and Health Dis - parities grant #U54MD010706. This study was also supported in part by the Bioreposi - tory and Tissue Analysis Shared Resource at the Hollings Cancer Center at the Medical University of South Carolina through grant #P30CA138313 from the National Cancer Institute. We are very appreciative to Tamara Dobson-Brown for assistance with data col - lection and management. We would like to thank the men who participated in this study. Conflict of Interest No conflicts of interest to report. Author Contributions Research concept and design: Jefferson, Drake, Savage; Acquisition of data: Jefferson, Lilly, Savage; Data analysis and interpreta - tion: Jefferson, Lilly, Savage, Tucker Price; Manuscript draft: Jefferson, Drake, Lilly, Tucker Price; Acquisition of funding: Drake; Administrative: Jefferson, Lilly, Savage, Tucker Price; Supervision: Jefferson, Savage References1. American Cancer Society. Cancer Facts and Figures, 2018 . Atlanta, GA: American Cancer Society; 2018. 2. World Health Organization. Chronic Diseases and their Common Risk Factors . Available at: https://www.who.int/chp/chronic_disease_re - port/media/Factsheet1.pdf . 3. Edwards BK, Noone AM, Mariotto AB, et al. Annual Report to the Nation on the status of cancer, 1975-2010, featuring prevalence of comorbidity and impact on survival among persons with lung, colorectal, breast, or prostate cancer. Cancer . 2014;120(9):1290- 1314. https://doi.org/10.1002/cncr.28509 PMID:24343171 4. Albertsen PC, Moore DF, Shih W, Lin Y, Li H, Lu-Yao GL. Impact of comorbidity on survival among men with localized prostate cancer. J Clin Oncol . 2011;29(10):1335-1341. https://doi.org/10.1200/JCO.2010.31.2330 PMID:21357791 5. American Urological Association. Clinically Localized Prostate Cancer: AUA/ASTRO/SUO Guideline. Available at: https://www.auanet. Ethnicity & Disease, Volume 30, Supplement 1, 2020 Co-morbidities in Prostate Cancer Patients - Jefferson et al 192 org/guidelines/prostate-cancer-clinically-localized-(2017) . 6. Thompson I, Thrasher JB, Aus G, et al; AUA Prostate Cancer Clinical Guideline Update Panel. Guideline for the management of clinically localized prostate cancer: 2007 update. J Urol . 2007;177(6):2106-2131. https://doi.org/10.1016/j.juro.2007.03.003 PMID:17509297 7. Hoffman KE. Management of older men with clinically localized prostate cancer: the signifi - cance of advanced age and comorbidity. Semin Radiat Oncol . 2012;22(4):284-294. https:// doi.org/10.1016/j.semradonc.2012.05.005 PMID:22985811 8. Davis J, Penha J, Mbowe O, Taira DA. Prevalence of single and multiple leading causes of death by race/ethnicity among US adults aged 60 to 79 years. Prev Chronic Dis . 2017;14:E101. https://doi.org/10.5888/ pcd14.160241 PMID:29049018 9. Fryar CD, Ostchega Y, Hales CM, Zhang G, Kruszon-Moran D. Hypertension prevalence and control among adults: United States, 2015-2016. NCHS Data Brief . 2017;Oct(289):1-8. PMID:29155682 10. Siegel RL, Miller KD, Jemal A. Cancer statis - tics, 2018. CA Cancer J Clin . 2018;68(1):7- 30. https://doi.org/10.3322/caac.21442 PMID:29313949 11. O’Malley DM, Hudson SV, Ohman-Strick - land PA, et al. Follow-up care education and information: identifying cancer survivors in need of more guidance. J Cancer Educ . 2016;31(1):63-69. https://doi.org/10.1007/ s13187-014-0775-y PMID:25524391 12. Adams ML, Grandpre J, Katz DL, Shenson D. Linear association between number of modifiable risk factors and multiple chronic conditions: Results from the Behavioral Risk Factor Surveillance System. Prev Med . 2017;105:169-175. https://doi.org/10.1016/j. ypmed.2017.09.013 PMID:28917949 13. Heron M. Deaths: Leading causes for 2016. National Vital Statistics Reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics . National Vital Statistics System. 2018;67:1-77. 14. Daskivich TJ, Fan KH, Koyama T, et al. Effect of age, tumor risk, and comorbidity on competing risks for survival in a U.S. population-based cohort of men with prostate cancer. Ann Intern Med . 2013;158(10):709- 717. https://doi.org/10.7326/0003-4819-158- 10-201305210-00005 PMID:23689764 15. Centers for Disease Control and Prevention. Body Mass Calculator . Available at: https:// www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/english_bmi_calculator/bmi_calcu - lator.html . 16. National Cancer Institute. Health Information National Trends Survey . Available at: http:// hints.cancer.gov/ . 17. Halbert CH, Bellamy S, Briggs V, et al. A comparative effectiveness education trial for lifestyle health behavior change in African Americans. Health Educ Res . 2017;32(3):207- 218. https://doi.org/10.1093/her/cyx039 PMID:28335038 18. Snyder CF, Frick KD, Herbert RJ, et al. Co - morbid condition care quality in cancer survi - vors: role of primary care and specialty provid - ers and care coordination. J Cancer Surviv . 2015;9(4):641-649. https://doi.org/10.1007/ s11764-015-0440-4 PMID:25716644 19. National Cancer Institute. NCI Co-morbidity Index Overview . Available at: https://health - caredelivery.cancer.gov/seermedicare/consider - ations/comorbidity.html . 20. Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbid - ity index using physician claims data. J Clin Epidemiol . 2000;53(12):1258-1267. https:// doi.org/10.1016/S0895-4356(00)00256-0 PMID:11146273 21. Asmar R, Beebe-Dimmer JL, Korgavkar K, Keele GR, Cooney KA. Hypertension, obesity and prostate cancer biochemical recurrence after radical prostatectomy. Prostate Cancer Prostatic Dis . 2013;16(1):62-66. https://doi. org/10.1038/pcan.2012.32 PMID:22907512 22. Post JM, Beebe-Dimmer JL, Morgenstern H, et al. The metabolic syndrome and biochemi - cal recurrence following radical prostatec - tomy. Prostate Cancer . 2011;2011:245642. https://doi.org/10.1155/2011/245642 PMID:22096652 23. US Department of Health and Human Services. HITECH Act Enforcement Interim Final Rule. Available at: https://www.hhs.gov/ hipaa/for-professionals/special-topics/hitech-act-enforcement-interim-final-rule/index.html . 24. Ye F, Moon DH, Carpenter WR, et al. Com - parison of patient report and medical records of comorbidities: results from a population-based cohort of patients with prostate cancer. JAMA Oncol . 2017;3(8):1035-1042. https:// doi.org/10.1001/jamaoncol.2016.6744 PMID:28208186 25. Rock CL, Doyle C, Demark-Wahnefried W, et al. Nutrition and physical activity guide - lines for cancer survivors. CA Cancer J Clin . 2012;62(4):243-274. https://doi.org/10.3322/ caac.21142 PMID:22539238 26. Skolarus TA, Wolf AM, Erb NL, et al. American Cancer Society prostate cancer survivorship care guidelines. CA Cancer J Clin . 2014;64(4):225-249. https://doi.org/10.3322/ caac.21234 PMID:24916760 27. Song L, Dunlap KL, Tan X, et al. Enhancing survivorship care planning for patients with localized prostate cancer using a couple-focused mHealth symptom self-management program: protocol for a feasibility study. JMIR Res Protoc . 2018;7(2):e51. https://doi. org/10.2196/resprot.9118 PMID:29483070 28. Centers for Disease Control and Prevention. High Blood Pressure Fact Sheet . Available at: https://www.cdc.gov/dhdsp/data_statistics/fact_sheets/fs_bloodpressure.htm . 29. Yang J, Neugut AI, Wright JD, Accordino M, Hershman DL. Nonadherence to oral medica - tions for chronic conditions in breast cancer survivors. J Oncol Pract . 2016;12(8):e800- e809. https://doi.org/10.1200/ JOP.2016.011742 PMID:27407167 30. Allam O, Gray A, Bailey H, Morrey D. Pri - mary care oncology: addressing the challenges. Inform Prim Care . 2006;14(3):167-173. PMID:17288702 31. Shaw A. What Is Primary Care Oncol - ogy? Unique Issues Facing Primary Care Provid - ers. Available at: https://www.peacehealth. org/sites/default/files/6_shaw_unique-issues-facing-pcps.pdf . Ethnicity & Disease, Volume 30, Supplement 1, 2020 Co-morbidities in Prostate Cancer Patients - Jefferson et al 3 articles need to answer these questions: What are the elements of the study design, and why the authors made the choices they did? Identify the importance of this study and what it adds to the knowl American Journal of Epidemiology Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health 2020. This work is written by (a) US Government employee(s) and is in the public domain in the US. Vol. 189, No. 11 DOI: 10.1093/aje/kwaa073 Advance Access publication: May 30, 2020 Study Design A Prospective Cohort Study to Evaluate the Impact of Diet, Exercise, and Lifestyle on Fertility: Design and Baseline Characteristics Sunni L. Mumford∗ , Erica Johnstone, Keewan Kim, Mudsar Ahmad, Shanna Salmon, Karen Summers, Kayla Chaney, Ginny Ryan, James M. Hotaling, Alexandra C. Purdue-Smithe, Zhen Chen, and Traci Clemons ∗ Correspondence to Dr. Sunni L. Mumford, Epidemiology Branch, Division of Intramural Population Health Research,Eunice Kennedy ShriverNational Institute of Child Health and Human Development, National Institutes of Health, 6710B Rockledge Drive, MSC 7004, Bethesda, MD 20892 (e-mail: [email protected]). Initially submitted July 7, 2019; accepted for publication April 24, 2020. Diet, lifestyle, and psychosocial factors might inf luence fertility for men and women, although evidence is mixed, and couple-based approaches are needed for assessing associations with reproductive outcomes. The Impact of Diet, Exercise, and Lifestyle (IDEAL) on Fertility Study is a prospective cohort with contemporaneous detailed follow-up of female partners of men enrolled in the Folic Acid and Zinc Supplementation Trial studying couples seeking infertility treatment (2016–2019). Follow-up of men continued for 6 months, while female partners were followed for 9 months while attempting pregnancy and throughout any resulting pregnancy (up to 18 months). Longitudinal data on diet, physical activity (including measurement via wearable device), sleep, and stress were captured at multiple study visits during this follow-up. A subset of women (IDEALplus) also completed daily journals and a body fat assessment via dual-energy x-ray absorptiometry. IDEAL enrolled 920 women, and IDEALPlus enrolled 218. We demonstrated the ability to enroll women in a prospective cohort study contemporaneous to a partner-enrolled randomized trial. In combination with data collected on male partners, IDEAL data facilitates a couple-based approach to understanding associations between lifestyle factors and infertility treatment outcomes. We describe in detail the study design, recruitment, data collection, lessons learned, and baseline characteristics. diet; fertility; infertility treatment; lifestyle; pregnancy Abbreviations: DXA, dual-energy x-ray absorptiometry; FAZST, Folic Acid and Zinc Supplementation Trial; IDEAL, Impact of Diet, Exercise, and Lifestyle; IVF, in vitro fertilization; SD, standard deviation. Infertility affects approximately 16% of couples in the United States (1,2). It is estimated that roughly one-third of infertility is caused by male disorders, one-third by female disorders, and one-third by combined male and female disorders (3). Many of the urological and/or gynecolog- ical disorders that lead to infertility are idiopathic, and unexplained infertility comprises a significant proportion of cases. At the same time, many of the potential identifiable “causes” of infertility, including oligo-ovulation, unilateral tubal obstruction, and suboptimal semen parameters, are not absolute blocks to pregnancy, and thus in many couples, multiple factors are likely at play. Psychosocial stressors, diet, and other potentially modifiable lifestyle factors inboth men and women have been shown to influence the hypothalamic pituitary gonadal axis, reproductive hormone concentrations, anovulation (4–8), inflammation (9–14), and other endocrine and metabolic pathways critical for reproduction (15), and could therefore have important downstream impacts on fertility (16–28). A couple-based definition of infertility highlights the importance of a couple-based approach to assessing lifestyle factors and re- productive success. Associations between lifestyle and psy- chosocial factors and ovulation, conception, implantation, and embryonic and fetal development remain largely un- explored, but they offer potential low-cost strategies to improve fertility. As yet, few well-conducted prospective 1254Am J Epidemiol.2020;189(11):1254–1265 Design of the IDEAL Fertility Study1255 studies have evaluated how these couple-level and individual factors might influence fertility in couples conceiving spon- taneously, as well as in those seeking infertility treatment. In order to address these important research questions, we designed a prospective cohort study called Impact of Diet, Exercise, and Lifestyle (IDEAL) on Fertility by expanding follow-up of female partners of male participants in the Folic Acid and Zinc Supplementation Trial (FAZST). The FAZST trial is a multisite double-blind block-randomized placebo-controlled clinical trial to evaluate the effect of fo lic acid and zinc sulfate supplementation on semen quality and infertility treatment outcomes among male partners of couples seeking infertility treatment (29). The design of IDEAL allows for the analysis of preconception measures as well as longitudinal measures across critical sensitive windows of development throughout pregnancy, with the ability to combine data on both male and female partners for a couple-based approach. Furthermore, in this cohort of women recruited prior to conception, IDEAL is also able to evaluate associations between lifestyle factors and early pregnancy losses that are often not captured in studies that recruit pregnant women given that more than half of preg- nancy losses occur prior to 9 weeks of gestation (30). IDEAL study participants have presented seeking fertility treatment and, as such, are encouraged to test for pregnancy at the time of a missed period, and undergo serum pregnancy testing at this time, thus allowing for capture of very early pregnancy losses. The purpose of this paper is to describe the design of the IDEAL study, including recruitment and methodol- ogy, lessons learned, and baseline characteristics of enrolled participants. METHODS Study objectives The primary aim of the IDEAL study is to evaluate the impact of dietary and other potentially modifiable lifestyle factors on prospectively measured pregnancy outcomes among couples seeking infertility treatment, including live birth, pregnancy, pregnancy loss, and specific pregnancy complications. Specific aims include the evaluation of preg- nancy-related outcomes in relation to: 1) dietary intake (par- ticularly dietary fiber, fat, phytoestrogens, caffeine, and vitamin D intake); 2) diet patterns (e.g., Mediterranean diet and low-carbohydrate diet); 3) adiposity and fat distribution (measured by dual-energy x-ray absorptiometry (DXA)) and the metabolic influences of adiposity, including leptin, insulin resistance, lipids, and inflammation; and 4) other lifestyle factors, such as physical activity, sleep, and psycho- social stress. Study design and target population IDEAL is a prospective cohort study of the female part- ners in couples attempting to conceive and seeking infertility treatment. Women were recruited from couples whose male partners were enrolled in FAZST (29). Randomization was stratified by site and intended infertility treatment (in vitro fertilization (IVF), non-IVF/study site, and non-IVF/outsideclinic). Follow-up of men continued for 6 months, while fe- male partners were followed for 9 months while attempting to conceive and throughout any resulting pregnancy (up to 18 months). The IDEAL study specifically includes detailed follow-up through questionnaires, biospecimen collection, and activity monitoring during this time for a subset of female partners of FAZST participants, including up to 2 visits during infertility treatment and up to 3 pregnancy visits if they conceived. This study was funded by the Intramural Research Pro- gram of theEunice Kennedy ShriverNational Institute of Child Health and Development and conducted at the Uni- versity of Utah and the University of Iowa. Institutional review board approval was obtained at each site and at the data coordinating center. All participants provided written informed consent. The FAZST trial was registered with clinicaltrials.gov(NCT01857310). The IDEAL study was overseen by an independent Data Safety and Monitoring Board, also responsible for the FAZST trial, that regularly reviewed the safety of all study participants, as well as data quality and timeliness, participant recruitment, accrual, and retention. Eligibility criteria Eligible participants included heterosexual couples (fe- male partner 18–45 years old and male partner aged 18 years or older) in a committed relationship, actively attempt- ing to conceive, and seeking infertility treatment. Specific eligibility criteria for the male partners have been detailed previously (29). Only female partners of male participants enrolled in FAZST at the University of Utah and the Uni- versity of Iowa were eligible for inclusion in IDEAL. In addition, women had to be willing to complete the additional questionnaires and biospecimen collection (Figure 1). Fur- ther eligibility for IDEALplus was based on willingness to undergo a DXA scan and to complete daily journals, and male partner willingness to undergo a DXA scan. Recruit- ment for FAZST started on June 3, 2013, with recruitment for IDEAL and IDEALplus starting on February 26, 2016. Based on these criteria, 920 women were enrolled in IDEAL and 218 in IDEALplus (note that after the recruitment goal for IDEALplus was met, the enrollment for this portion was closed). Baseline study visit The baseline enrollment visit was coordinated to occur at the same time as the FAZST enrollment visit. Data col- lection for female partners in FAZST at baseline consisted of a basic questionnaire regarding diet, physical activity, medication use, and reproductive history, as well as blood and urine collection and measurement of blood pressure and pulse. The baseline study visit for IDEAL included a questionnaire assessing pelvic pain and sexual health, as well as saliva collection. Participants were provided with a wrist-based fitness tracker (Fitbit Charge HR; Fitbit Inc., San Francisco, California) to track daily physical activity, heart rate, and sleep for 9 months while attempting to conceive, and throughout any resulting pregnancy (up to 18 months). Am J Epidemiol.2020;189(11):1254–1265 1256Mumford et al. Figure 1.Recruitment and eligibility criteria for the Impact of Diet, Exercise, and Lifestyle (IDEAL) on Fertility Study, Folic Acid and Zinc Supplementation Trial (FAZST), and IDEALplus, United States, 2016–2019. CBAVD, congenital bilateral absence of the vas deferens; DXA, dual-energy x-ray absorptiometry; HIV/AIDS, human immunodeficiency virus/acquired immune deficiency syndrome. Participants were instructed on how to use the fitness tracker and to sync their data to a smart phone application. The sub- set of women enrolled in IDEALplus received a DXA scan to assess body fat composition, as did their male partners. All women completed a urine pregnancy test at their baseline visit to confirm nonpregnant status. Women scheduled for DXA scans separate from their baseline visit were provided an additional urine pregnancy test to confirm nonpregnant status on the day of their scheduled DXA scan. DXA scans were conducted within 1 week of the baseline visit. Women in IDEALplus were also given access to an online journal to complete daily entries assessing dietary intake, sleep patterns, medication use, physical activity, and stress. Follow-up visits As part of their participation in FAZST, female par- ticipants completed brief monthly online questionnaires regarding their pregnancy status and progress of infertilitytreatments during the 9 month follow-up period and through- out any pregnancy conceived during that window. Medical record abstraction by trained staff assessed infertility testing and treatments, prenatal care, obstetrical and neonatal out- comes using inpatient hospital records, and outpatient clinic records. As part of the IDEAL protocol, participants also com- pleted up to 2 follow-up at-home data-collection visits while attempting to conceive, at months 2 and 4 after enrollment, that included online questionnaires and at-home biospeci- men collections (Figure 2). Specifically, participants com- pleted online questionnaires regarding dietary intake via the Automated Self-Administered 24-Hour Dietary Assess- ment (ASA24) (31) as well as other questionnaires, includ- ing the International Physical Activity Questionnaire (32), Pittsburgh Sleep Quality Index (33), Berlin Questionnaire for sleep apnea (34), the Perceived Stress Scale (35), and the Center for Epidemiologic Studies Depression Scale (36), to assess physical activity, sleep, stress, and depression. Am J Epidemiol.2020;189(11):1254–1265 Design of the IDEAL Fertility Study1257 Figure 2. Data collection according to study visit in the Impact of Diet, Exercise, and Lifestyle (IDEAL) on Fertility Study, United States, 2016–2019. DXA, dual-energy x-ray absorptiometry; FFQ, food frequency questionnaire. Circled plus sign indicates IDEALplus study components. Am J Epidemiol.2020;189(11):1254–1265 1258Mumford et al. Ta b l e 1.Baseline Characteristics of Participants in the Impact of Diet, Exercise, and Lifestyle on Fertility Study, IDEALplus, and Folic Acid and Zinc Supplementation Trial, United States, 2016–2019 IDEAL (n= 920)IDEALplus (n=218)FAZST Non-IDEAL (n= 1,450)FA Z S T (n= 2,370) Characteristic No. % No. % No. % No. % Age, years a 30.7 (5.2) 30.3 (5.0) 30.7 (5.0) 30.7 (5.1) BMI a,b 29.2 (8.4) 28.7 (27.4) 28.0 (8.0) 28.5 (8.2) Randomization strata IVF 118 12.8 26 11.9 255 17.6 373 15.7 Non-IVF study site 701 76.2 153 70.2 957 66.0 1,658 70.0 Non-IVF outside clinic 101 11.0 39 17.9 238 16.4 339 14.3 Race/ethnicity White 765 83.2 184 84.4 1,191 82.1 1956 82.5 Black 12 1.3 4 1.8 31 2.1 43 1.8 Asian 41 4.5 9 4.1 90 6.2 131 5.5 Hispanic/Latino 52 5.7 11 5.0 75 5.2 127 5.4 Other race/ethnic groups 45 4.9 9 4.1 56 3.9 101 4.3 Do not wish to provide 5 0.5 1 0.5 7 0.5 12 0.5 Education High school or less 126 13.7 22 10.1 149 10.3 275 11.6 Some college 307 33.4 87 39.9 480 33.1 787 33.2 Bachelor’s degree 320 34.8 77 35.3 529 36.5 849 35.8 Master’s degree or higher 158 17.2 31 14.2 268 18.5 426 18.0 Do not wish to provide 9 1.0 1 0.5 24 1.7 33 1.4 Annual income, $ <40,000 118 12.8 29 13.3 215 14.8 333 14.1 40,000–74,999 325 35.3 83 38.1 552 38.1 877 37.0 75,000–99,999 198 21.5 44 20.2 297 20.5 495 20.9 ≥100,000 219 23.8 45 20.6 311 21.4 530 22.4 Do not wish to provide 60 6.5 17 7.8 75 5.2 135 5.7 Health insurance No 23 2.5 6 2.8 46 3.2 69 2.9 Yes 894 97.2 212 97.2 1,384 95.4 2,278 96.1 Do not wish to provide 3 0.3 0 0.0 20 1.4 23 1.0 Insurance cover infertility treatment No 378 41.1 97 44.5 636 43.9 1,014 42.8 Yes 287 31.2 59 27.1 395 27.2 682 28.8 Don’t know/do not wish to provide 229 24.9 56 25.7 353 24.3 582 24.6 Missing 26 2.8 6 2.8 66 4.6 92 3.9 Smoking Smoking in the past 3 months Never 798 86.7 198 90.8 1,257 86.7 2055 86.7 Rarely (1–4 times per month) 29 3.2 4 1.8 31 2.1 60 2.5 Sometimes (2–6 times per week) 11 1.2 1 0.5 10 0.7 21 0.9 Daily 20 2.2 4 1.8 28 1.9 48 2.0 Missing 62 6.7 11 5.0 124 8.6 186 7.8 Table continues Am J Epidemiol.2020;189(11):1254–1265 Design of the IDEAL Fertility Study1259 Ta b l e 1.Continued IDEAL (n= 920)IDEALplus (n=218)FAZST Non-IDEAL (n= 1,450)FA Z S T (n= 2,370) Characteristic No. % No. % No. % No. % Alcohol consumption in past 3 months Never 474 51.5 137 62.8 798 55.0 1,272 53.7 Rarely (1–4 times per month) 308 33.5 56 25.7 373 25.7 681 28.7 Sometimes (2–6 times per week) 72 7.8 12 5.5 150 10.3 222 9.4 Daily 4 0.4 2 0.9 4 0.3 8 0.3 Missing 62 6.7 11 5.0 125 8.6 187 7.9 Partner randomization arm Active 460 50.0 116 53.2 725 50.0 1,185 50.0 Placebo 460 50.0 102 46.8 725 50.0 1,185 50.0 Abbreviations: BMI, body mass index; FAZST, Folic Acid and Zinc Supplementation Trial; IDEAL, Impact of Diet, Exercise, and Lifestyle; IVF, in vitro fertilization. aValues are expressed as mean (standard deviation).bWeight (kg)/height (m) 2. Participants were asked to collect a saliva sample at months 2 and 4, and toenails at month 4, to be returned to the clinic at the male partner’s visits. For couples undergoing IVF, serum was collected at the time of oocyte retrieval, as was follicular fluid and granulosa cells pooled from multiple follicles. Participants who became pregnant during the 9-month follow-up period had 3 additional pregnancy follow-up clinic visits (1 per trimester) that included questionnaires, biospecimen collection, anthropometric measurements, and vital signs. All biospecimens were stored at−80 ◦C. The first trimester visit was scheduled at 6–8 weeks’ gestational age, the second trimester visit at 18–22 weeks’ gestational age, and the third trimester visit at 30–36 weeks’ gestational age. Throughout follow-up, participants were asked about daily fitness tracker use to ensure appropriate syncing and recording of information. Outcome measures The primary outcome of interest in the IDEAL study was live birth. Secondary outcomes included human chori- onic gonadotropin–recognized pregnancy, clinical intrauter- ine pregnancy, ectopic pregnancy, early pregnancy loss, spe- cific pregnancy complications (including cesarean delivery, preeclampsia, gestational diabetes, preterm birth, and infant being small for gestational age), and early embryonic devel- opment parameters among couples using assisted reproduc- tive technologies such as in IVF (29). RESULTS From February 26, 2016, to December 30, 2017, a total of 920 women were enrolled in the IDEAL Study, withthe last third-trimester visit completed on January 31, 2019. This represents 39% (920/2,370) of all female partners of FAZST participants and 97% (920/948) of those eligible for enrollment at the Utah and Iowa sites during the IDEAL recruitment period. From these 920, 218 women were addi- tionally enrolled in IDEALplus (24% of women enrolled in the IDEAL Study). Of those enrolled, 394 became pregnant over follow-up (360 with one pregnancy and 34 with two pregnancies), with 311 live births, 114 pregnancy losses, and 3 with unknown pregnancy outcome (Web Figure 1, available athttps://academic.oup.com/aje). Baseline characteristics of IDEAL participants (n= 920), all female partners of FAZST participants (n= 2,370), female partners not included in the IDEAL Study (n= 1,450), and women enrolled in IDEALplus (n= 218) are summa- rized inTa b l e 1. The average age and body mass index of women enrolled in IDEAL were similar to female partners of FAZST participants overall (age: in IDEAL, 30.7 (standard deviation (SD), 5.2) years vs., in FAZST, 30.7 (SD, 5.1) years; body mass index: in IDEAL, 29.2 (SD, 8.4) vs., in FAZST, 28.5 (SD, 8.2)), although the average body mass index was slightly higher when compared with female partners not enrolled in IDEAL (IDEAL, 29.2 (SD, 8.4) vs., in non-IDEAL, 28.0 (SD, 8.0)). Most couples enrolled in IDEAL were planning non-IVF treatments at a study site upon enrollment, although a lower proportion of couples enrolled in IDEAL were planning to undergo IVF as compared with all couples in FAZST (12.8% vs. 15.7%). Women enrolled in IDEAL and IDEALplus tended to have a higher annual income and were more likely to have health insurance and insurance coverage for infertility treatment compared with female partners of FAZST participants not enrolled in IDEAL. Furthermore, women in IDEALplus were more likely never smokers and never drinkers compared with women enrolled in IDEAL Am J Epidemiol.2020;189(11):1254–1265 1260Mumford et al. Ta b l e 2 .Biospecimen Collections and Questionnaire Completions in the Impact of Diet, Exercise, and Lifestyle on Fertility Study, United States, 2016–2019 Infertility Treatment Pregnancy Follow-up Baseline (n= 920) a Month 2 (n= 828)Month 4 (n= 734)Oocyte Retrieval (IVF) b (n= 107)First Trimester (n= 302)Second Trimester (n= 290)Third Trimester (n= 267) Collection Component No. % No. % No. % No. % No. % No. % No. % Biospecimen collection Blood 773 84 44 41 205 68 176 61 154 57 Urine 918 99242 80 211 73 191 71 Saliva 874 95 649 78 573 78 243 80 214 74 192 71 Toenail 616 84 Follicular f luid60 56 Cord blood Placenta Questionnaires Lifestyle 815 89 738 89 631 86 259 86 198 68 177 66 ASA24 611 74 467 64 258 85 223 77 Anthropometry assessment Anthropometry 918 99254 84 218 75 193 71 DXA scan c 173 80 Abbreviations: ASA24, Automated Self-Administered 24-Hour Dietary Assessment; DXA, dual-energy x-ray absorptiometry; IVF, in vitro fertilization.aExpectedn. Denominators at each time point for calculation of % completion are based on the number of women expected to complete a given visit; some women withdrewfrom study participation, became pregnant during follow-up and switched to the pregnancy track, or had a pregnancy loss (and were not expected at subsequent pregnancy questionnaires and returned to infertility treatment follow-up). Of 920 women enrolled, 103 became pregnant, 5 withdrew, and 16 had a pregnancy loss before the month-2 visit (n= 828 expected at month 2); of 828 expected at month 2, 122 became pregnant, 6 withdrew, and 34 had a pregnancy loss before the month-4 visit (n= 734 expected at month 4); of 734 expected at month 4, 203 became pregnant, 5 withdrew, and 64 had a pregnancy loss through the end of follow-up. A total of 428 pregnancies were observed (from 394 women); of these, 101 were early pregnancy losses and 25 withdrew prior to the first trimester visit (n= 302 expected at first-trimester visit); of 302 expected at the first trimester, 12 experienced a loss during the second trimester (n= 290 expected at second-trimester visit); of 290 expected at the second trimester, 1 had a late loss, 1 withdrew, and 21 delivered preterm prior to their third-trimester visit (n= 267 expected at third-trimester visit). See Web Figure 2 for a detailed timeline. bNexpected = 107.cNexpected = 218. Am J Epidemiol.2020;189(11):1254–1265 Design of the IDEAL Fertility Study1261 Ta b l e 3 .Completion of Monthly Questionnaires in the Impact of Diet, Exercise, and Lifestyle on Fertility Study, IDEALplus, and Folic Acid and Zinc Supplementation Trial, United States, 2016–2019 IDEAL IDEALplus FAZST Non-IDEAL FAZST (n= 920) (n=218) (n= 1,450) (n= 2,370) Month Average Days of CompletionNo. Complete% CompleteAverage Days of CompletionNo. Complete% CompleteAverage Days of CompletionNo. Complete% CompleteAverage Days of CompletionNo. Complete% Complete 1 4.0 734 80 6.2 163 75 4.3 953 66 4.1 1,688 71 2 4.0 707 77 6.8 157 72 4.2 908 63 4.0 1,615 68 3 4.4 613 67 7.5 141 65 4.4 813 56 4.4 1,428 60 4 4.8 632 69 7.0 130 60 4.0 810 56 4.3 1,445 61 5 4.7 532 58 6.8 109 50 4.5 712 49 4.5 1,248 53 6 3.8 548 60 7.2 108 50 4.0 706 49 3.9 1,257 53 7 4.7 485 53 7.6 90 41 3.7 592 41 4.1 1,080 46 8 4.1 433 47 6.8 78 36 4.0 494 34 4.0 934 39 9 6.0 257 28 7.6 41 19 4.1 375 26 4.8 632 27 Overall 4.4 4,941 60 7.0 1,017 52 4.1 6,363 49 4.2 11,327 53 Abbreviations: FAZST, Folic Acid and Zinc Supplementation Trial; IDEAL, Impact of Diet, Exercise, and Lifestyle. or FAZST overall. Male partners of those women enrolled in the IDEAL study were similarly randomized to folic acid and zinc supplement (treatment group) or placebo. Although slightly more male partners of IDEALplus participants were randomized to treatment group compared with placebo group (53% vs. 47%), the difference was not substantial (P= 0.18). Participants in the IDEAL study provided questionnaire data and biospecimens multiple times throughout follow up and pregnancy and were largely compliant with the study protocol, with 89% completing data collection at month 2 and 86% at month 4 (Ta b l e 2). Overall, women had higher completion rates for questionnaire components of the pro- tocol than for biospecimen collection, although the rate of questionnaire completion dropped substantially as preg- nancy progressed (86%, 68%, and 66% at first, second, and third trimester, respectively) for those with a resulting pregnancy. Monthly questionnaires designed to follow infertility treat- ment and pregnancy status were completed as part of FAZST participation. Overall, women enrolled in the IDEAL and IDEALplus studies were more likely to complete the month- ly questionnaires compared with female partners of FAZST participants not enrolled in IDEAL, although completion rates for these also declined over time and the average days of completion were similar (Ta b l e 3). The average days of completion was somewhat longer for women enrolled in IDEALplus. Overall, 94% of women had some Fitbit information available over follow-up (Ta b l e 4). Notably, 78% of women had over 3 months of data synced, and only 6% of women did not participate in the Fitbit component of the study. The median number of days of activity synced was 184 (25th percentile: 124; 75th percentile: 255); with a median number of hours per day with heart rate information of 23.5 (25th percentile: 18.25; 75th percentile: 24). The median number of preconception days of activity was 141 (25th percentile: 66; 75th percentile: 188) including women who did not become pregnant over follow-up and the time prior to pregnancy among those that conceived. The median number of days of activity synced during pregnancy was 172 (25th percentile: 66; 75th percentile: 244). Women who either did not use the Fitbit or had more than 3 months of data were more likely to be older and have higher education, and were less likely to smoke or drink alcohol (Ta b l e 4). No differences in compliance were noted by stratum, body mass index, race, or insurance coverage. DISCUSSION Dietary and lifestyle factors offer the potential for low- cost interventions to improve fertility, although well- designed prospective studies are needed to better understand associations with fertility and pregnancy outcomes. The IDEAL study was designed to answer these important questions by expanding follow-up of the female partners of FAZST participants to include more detailed questionnaire, biospecimen, and wearable fitness tracking data on the female partners. The IDEAL study results demonstrate the ability to recruit a large number of female partners for an Am J Epidemiol.2020;189(11):1254–1265 1262Mumford et al. Ta b l e 4 .Baseline Characteristics of Participants in the Impact of Diet, Exercise, and Lifestyle on Fertility Study According to Compliance With Wrist-Based Fitness Tracker aBased on Number of Months With Synced Data Available, United States, 2016–2019 No Data (n= 52)<1 Month (n= 43)1–3 Months (n= 109)≥3 Months (n=716) Characteristic No. % No. % No. % No. %PValue Age, years b 32.8 (6.7) 28.1 (4.6) 29.2 (4.9) 30.9 (5.0)<0.0001 BMI b,c 28.0 (7.3) 31.3 (9.6) 29.9 (8.5) 29.1 (8.4) 0.21 Randomization strata0.16 IVF 12 23.1 5 11.6 14 12.8 87 12.2 Non-IVF study site 35 67.3 30 69.8 80 73.4 556 77.7 Non-IVF outside clinic 5 9.6 8 18.6 15 13.8 73 10.2 Race/ethnicity0.67 White 43 82.7 34 79.1 88 80.7 600 83.8 B l a c k 1 1. 9 1 2 . 3 2 1. 8 8 1. 1 Asian 4 7.7 1 2.3 3 2.8 33 4.6 Hispanic/Latino 1 1.9 3 7.0 8 7.3 40 5.6 Other race/ethnic groups 2 3.8 3 7.0 7 6.4 33 4.6 Do not wish to provide 1 1.9 1 2.3 1 0.9 2 0.3 Education0.0005 High school or less 8 15.4 8 18.6 25 22.9 85 11.9 Some college 18 34.6 18 41.9 44 40.4 227 31.7 Bachelor’s degree 16 30.8 12 27.9 27 24.8 265 37.0 Master’s degree or higher 8 15.4 3 7.0 13 11.9 134 18.7 Do not wish to provide 2 3.8 2 4.7 0 0 5 0.7 Annual income, $0.19 <40,000 8 15.4 6 14.0 17 15.6 87 12.2 40,000—74,999 17 32.7 15 34.9 47 43.1 246 34.4 75,000—99,999 6 11.5 9 20.9 16 14.7 167 23.3 ≥100,000 16 30.8 8 18.6 20 18.3 175 24.4 Do not wish to provide 5 9.6 5 11.6 9 8.3 41 5.7 Health insurance0.23 No 3 5.8 1 2.3 4 3.7 15 2.1 Yes 48 92.3 42 97.7 105 96.3 699 97.6 Do not wish to provide 1 1.9 0 0 0 0 2 0.3 Insurance cover infertility treatment0.38 No 22 42.3 22 51.2 47 43.1 287 40.1 Yes 16 30.8 9 20.9 35 32.1 227 31.7 Don’t know/do not wish to provide 10 19.2 11 25.6 23 21.1 185 25.8 Missing 4 7.7 1 2.3 4 3.7 17 2.4 Smoking Smoking in the past 3 months<0.0001 Never 39 75.0 22 51.2 80 73.4 657 91.8 Rarely (1–4 times per month) 2 3.8 3 7.0 6 5.5 18 2.5 Sometimes (2–6 times per week) 0 0 2 4.7 1 0.9 8 1.1 Daily 0 0 2 4.7 4 3.7 14 2.0 Missing 11 21.2 14 32.6 18 16.5 19 2.7 Table continues Am J Epidemiol.2020;189(11):1254–1265 Design of the IDEAL Fertility Study1263 Ta b l e 4 .Continued No Data (n= 52)<1 Month (n= 43)1–3 Months (n= 109)≥3 Months (n=716) Characteristic No. % No. % No. % No. %PValue Ever smoked≥100 cigarettes<0.0001 No 38 73.1 23 53.5 82 75.2 608 84.9 Yes 3 5.8 5 11.6 9 8.3 89 12.4 Missing 11 21.2 15 34.9 18 16.5 19 2.7 Alcohol consumption in past 3 months<0.0001 Never 26 50.0 15 34.9 46 42.2 387 54.1 Rarely (1–4 times per month) 10 19.2 13 30.2 37 33.9 248 34.6 Sometimes (2–6 times per week) 5 9.6 1 2.3 8 7.3 58 8.1 Daily 0 0 0 0 0 0 4 0.6 Missing 11 21.2 14 32.6 18 16.5 19 2.7 Partner randomization arm0.38 Active 20 38.5 22 51.2 57 52.3 361 50.4 Placebo 32 61.5 21 48.8 52 47.7 355 49.6 Abbreviations: BMI, body mass index; IVF, in vitro fertilization.aFitbit Charge HR; Fitbit Inc., San Francisco, California.bValues are expressed as mean (standard deviation).cWeight (kg)/height (m) 2. intensive study during infertility treatment (Figure 3)and provide a rich data source that can be linked to data on the male partners in FAZST for a couple-based approach to understanding lifestyle influences on infertility. The design of the IDEAL study allows for the analysis of preconception samples as well as longitudinal measure- ments across critical windows of development throughout pregnancy. Furthermore, because this cohort of women was recruited prior to conception and in the infertility clinic setting, the IDEAL study is positioned to capture the timing of key pregnancy-related events such as fertilization (in the setting of IVF), implantation, clinical pregnancy recogni- tion, early pregnancy loss, and gestational age. Considering the stress and disruption of daily-life activities that infertility treatment imposes upon couples, this study was designed to be comprehensive while minimizing burden. At-home ques- tionnaires and biospecimen collection were implemented during infertility treatment to allow for schedule flexibility preferred by participants. Indeed, higher compliance was observed for the at-home questionnaires than for compo- nents that required a clinic visit. Compliance also improved with frequent reminders and the availability of staff support for completion of any outstanding items when they came into the clinic. Further, structuring study visits to align with clinical appointments, particularly during pregnancy, as well as offering flexible hours for visits improved retention and study visit completion. Despite these efforts, monthly questionnaire compliance was low, perhaps due to the par- ticular sensitivity of the questions to confirm pregnancyand infertility treatment status during this stressful time for participants. Importantly, we had planned for detailed chart abstraction from the outset of the FAZST trial so that impor- tant outcomes, such as pregnancy, loss, and live birth, would be captured regardless of the self-report. The low response rates were observed prior to the initiation of IDEAL and so additional tools to improve staff tracking of pregnancies were implemented in IDEAL to ensure that pregnancies were reported in a timely manner so that the IDEAL preg- nancy visits could be scheduled accordingly. Specifically, real-time tracking of electronic medical records to determine pregnancy and infertility treatment status proved to be a useful strategy because compliance with monthly question- naires was low despite these improvements. The IDEAL study included innovative approaches for data collection, including a comprehensive assessment of body fat (using DXA combined with detailed anthropo- metric measurements), time-varying dietary questionnaires, and sleep and physical activity assessments using Fitbit. Fitbits offer a noninvasive approach to prospectively collect detailed data on physical activity and sleep, and we found that participants in the study were highly compliant with the study protocol, with over 75% of women providing daily data for at least 3 months during follow-up. Staff noted that frequent reminders for syncing were needed. Compliance with wearing the fitness trackers was related to certain demo- graphic characteristics, including age, education, smoking, and alcohol intake. Interestingly, we observed similar pat- terns among those with high compliance as those who did Am J Epidemiol.2020;189(11):1254–1265 1264Mumford et al. not wear the fitness trackers. To our knowledge, this is the first large-scale trial to use Fitbit or an equivalent tracker to obtain real-time data on a large cohort of women trying to conceive that demonstrates feasibility of having participants wear the tracker over a relatively long period of follow-up time. This study provided a unique opportunity during critical periconception windows to explore the patterns of lifestyle factors and how these factors are associated with pregnancy outcomes. Although the results will be generalizable only to women seeking infertility treatment, these findings will provide much-needed evidence to inform clinical recom- mendations on low-cost strategies to optimize fecundability in an environment where infertility treatments are typically expensive and rarely covered by insurance. Figure 3.Lessons learned during the design and recruitment of the Impact of Diet, Exercise, and Lifestyle (IDEAL) on Fertility Study, United States, 2016–2019. ACKNOWLEDGMENTS Author affiliations: Epidemiology Branch, Division of Intramural Population Health Research,Eunice Kennedy ShriverNational Institute of Child Health and Human Development, Bethesda, Maryland (Sunni L. Mumford, Keewan Kim, Alexandra C. Purdue-Smithe); Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, University of Utah, Salt Lake City, Utah (Erica Johnstone, Mudsar Ahmad, Shanna Salmon); Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, University of Iowa Carver College of Medicine, Iowa City, Iowa (Karen Summers, Ginny Ryan); The Emmes Company LLC, Rockville, Maryland (Kayla Chaney, Traci Clemons); Department of Surgery (Urology), Center for Reconstructive Urology and Men’s Health, University of Utah School of Medicine, Salt Lake City, Utah(James M. Hotaling); Department of Obstetrics and Gynecology, Center for Reconstructive Urology and Men’s Health, University of Utah School of Medicine, Salt Lake City, Utah (James M. Hotaling); and Biostatistics and Bioinformatics Branch, Division of Intramural Population Health Research,Eunice Kennedy ShriverNational Institute of Child Health and Human Development, Bethesda, Maryland (Zhen Chen). This research was supported by the Intramural Research Program of theEunice Kennedy ShriverNational Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland (contracts HHSN275201500001C, HHSN275201300026I/ HHSN27500008, and HHSN275201300026I/ HHSN27500018). We thank all of the Folic Acid and Zinc Supplementation Trial and Impact of Diet, Exercise, and Lifestyle investigators, fellows, and staff who devoted their time and energy to the success of these studies, and the Data Safety and Monitoring Board members for ongoing oversight, constant support, and advice throughout the trial. J.M.H. reports equity in early stage startups (streamDX, Nanonc, Andro360), grants from Endo Pharmaceuticals, and grants from Boston Scientific, that are unrelated to this work. The other authors report no conflicts. REFERENCES 1. Evers JL. Female subfertility.Lancet. 2002;360(9327): 151–159. 2. Thoma ME, McLain AC, Louis JF, et al. Prevalence of infertility in the United States as estimated by the current duration approach and a traditional constructed approach. Fertil Steril. 2013;99(5):1324–1331. 3. Kumar N, Singh AK. Trends of male factor infertility, an important cause of infertility: a review of literature.J Hum Reprod Sci. 2015;8(4):191–196. 4. Schliep KC, Schisterman EF, Wactawski-Wende J, et al. Serum caffeine and paraxanthine concentrations and menstrual cycle function: correlations with beverage intakes and associations with race, reproductive hormones, and anovulation in the BioCycle study.Am J Clin Nutr. 2016; 104(1):155–163. 5. Mumford SL, Chavarro JE, Zhang C, et al. Dietary fat intake and reproductive hormone concentrations and ovulation in regularly menstruating women.Am J Clin Nutr. 2016;103(3): 868–877. 6. Mumford SL, Browne RW, Schliep KC, et al. Serum antioxidants are associated with serum reproductive hormones and ovulation among healthy women.J Nutr. 2016; 146(1):98–106. 7. Ahrens KA, Vladutiu CJ, Mumford SL, et al. The effect of physical activity across the menstrual cycle on reproductive function.Ann Epidemiol. 2014;24(2):127–134. 8. Schliep KC, Mumford SL, Vladutiu CJ, et al. Perceived stress, reproductive hormones, and ovulatory function: a prospective cohort study.Epidemiology. 2015;26(2):177–184. 9. Levine LD, Holland TL, Kim K, et al. The role of aspirin and inflammation on reproduction: the EAGeR trial (1).Can J Physiol Pharmacol. 2019;97(3):187–192. Am J Epidemiol.2020;189(11):1254–1265 Design of the IDEAL Fertility Study1265 10. Radin RG, Sjaarda LA, Silver RM, et al. C-reactive protein in relation to fecundability and anovulation among eumenorrheic women.Fertil Steril. 2018;109(2):232–239. 11. Minihane AM, Vinoy S, Russell WR, et al. Low-grade inflammation, diet composition and health: current research evidence and its translation.Br J Nutr. 2015;114(7): 999–1012. 12. Gaskins AJ, Mumford SL, Rovner AJ, et al. Whole grains are associated with serum concentrations of high sensitivity C-reactive protein among premenopausal women.J Nutr. 2010;140(9):1669–1676. 13. Fijak M, Pilatz A, Hedger MP, et al. Infectious, inflammatory and ‘autoimmune’ male factor infertility: how do rodent models inform clinical practice?Hum Reprod Update. 2018; 24(4):416–441. 14. Sarkar O, Bahrainwala J, Chandrasekaran S, et al. Impact of inflammation on male fertility.Front Biosci (Elite Ed). 2011; 3:89–95. 15. Seli E, Babayev E, Collins SC, et al. Minireview: metabolism of female reproduction: regulatory mechanisms and clinical implications.Mol Endocrinol. 2014;28(6):790–804. 16. Gaskins AJ, Chavarro JE. Diet and fertility: a review.Am J Obstet Gynecol. 2018;218(4):379–389. 17. Buck Louis GM, Sapra KJ, Schisterman EF, et al. Lifestyle and pregnancy loss in a contemporary cohort of women recruited before conception: the LIFE study.Fertil Steril. 2016;106(1):180–188. 18. Salas-Huetos A, Bulló M, Salas-Salvadó J. Dietary patterns, foods and nutrients in male fertility parameters and fecundability: a systematic review of observational studies. Hum Reprod Update. 2017;23(4):371–389. 19. Showell MG, Mackenzie-Proctor R, Jordan V, et al. Antioxidants for female subfertility.Cochrane Database Syst Rev. 2017;7:CD007807. 20. Smits RM, Mackenzie-Proctor R, Yazdani A, et al. Antioxidants for male subfertility.Cochrane Database Syst Rev. 2019;3:CD007411. 21. Wise LA, Wesselink AK, Mikkelsen EM, et al. Dairy intake and fecundability in 2 preconception cohort studies.Am J Clin Nutr. 2017;105(1):100–110. 22. Mumford SL, Sundaram R, Schisterman EF, et al. Higher urinary lignan concentrations in women but not men are positively associated with shorter time to pregnancy.J Nutr. 2014;144(3):352–358. 23. Chavarro JE, Rich-Edwards JW, Rosner BA, et al. Diet and lifestyle in the prevention of ovulatory disorder infertility. Obstet Gynecol. 2007;110(5):1050–1058.24. Silvestris E, Lovero D, Palmirotta R. Nutrition and female fertility: an interdependent correlation.Front Endocrinol (Lausanne). 2019;10:346. 25. Mumford SL, Garbose RA, Kim K, et al. Association of preconception serum 25-hydroxyvitamin D concentrations with livebirth and pregnancy loss: a prospective cohort study.Lancet Diabetes Endocrinol. 2018;6(9):725–732. 26. Kim K, Schisterman EF, Silver RM, et al. Shorter time to pregnancy with increasing preconception carotene concentrations among women with 1–2 previous pregnancy losses.Am J Epidemiol. 2018;187(9):1907–1915. 27. Schliep KC, Mumford SL, Silver RM, et al. Preconception perceived stress is associated with reproductive hormone levels and longer time to pregnancy.Epidemiology. 2019; 30(suppl 2):S76–S84. 28. Russo LM, Whitcomb BW, Mumford SL, et al. A prospective study of physical activity and fecundability in women with a history of pregnancy loss.Hum Reprod. 2018;33(7): 1291–1298. 29. Schisterman EF, Clemons T, Peterson CM, et al. A randomized trial to evaluate the effects of folic acid and zinc supplementation on male fertility and live birth: design and baseline characteristics.Am J Epidemiol. 2020; 189(1):8–26. 30. Mumford SL, Silver RM, Sjaarda LA, et al. Expanded findings from a randomized controlled trial of preconception low-dose aspirin and pregnancy loss.Hum Reprod. 2016; 31(3):657–665. 31. National Cancer Institute. Automated Self-Administered 4-Hour (ASA24) Dietary Assessment Tool.https://epi.grants. cancer.gov/asa24/. Accessed April 17, 2019. 32. Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity.Med Sci Sports Exerc. 2003;35(8):1381–1395. 33. Buysse DJ, Reynolds CF 3rd, Monk TH, et al. The Pittsburgh sleep quality index: a new instrument for psychiatric practice and research.Psychiatry Res. 1989;28(2):193–213. 34. Netzer NC, Stoohs RA, Netzer CM, et al. Using the Berlin questionnaire to identify patients at risk for the sleep apnea syndrome.Ann Intern Med. 1999;131(7):485–491. 35. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress.J Health Soc Behav. 1983;24(4): 385–396. 36. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population.Appl Psychol Measur. 1977;1:385–401. Am J Epidemiol.2020;189(11):1254–1265