Study design and participants
This cross-sectional, mixed methods study was designed to: 1) obtain general information about AA and non-Hispanic white MM survivors in two large urban areas of the Midwest (Milwaukee and Chicago), and 2) identify optimal methodologies to capture specific topic areas of interest, perhaps unique to population group or geographic areas. By design, most assessment tools and data points across the two sites were identical; however, differences were permitted due to varying resources and to allow within and across site comparisons. Participants were recruited from two university hospitals with active stem cell transplant programs, reflecting cities with diverse patient populations. Patients were recruited from hospital oncology clinics, through recruitment letters and/or phone calls using contact information provided by medical oncologists (July, 2019 – July, 2020). We intentionally sought representation of AA participants since MM impacts this population disproportionately and a racially diverse perspective on intervention needs and interests would be critical to informing future studies. Thus, recruitment aimed to achieve 50% AA representation. Ethical approval was granted from the respective institutions prior to study initiation.
Eligible participants were required to be 18 years of age or older, at least 100 days post-ASCT, with a BMI ≥20 kg/m2, a performance status of ECOG 0-1 or Karnofsky >70, and life expectancy of at least 6 months. Prior to recruitment, participant eligibility was ascertained via the electronic health record or clearance from medical oncologists verifying clinical stability and appropriateness of inclusion.
Once deemed eligible, participants scheduled a time to complete informed consent, questionnaires, and a qualitative interview conducted either in-person or over the phone. For feasibility purposes, body composition and physical performance were only collected at the Chicago site. At this site, the in-person visit was scheduled no more than 30 days after initial study eligibility and entailed ~90 minutes to complete questionnaires and/or physical measures. In March 2020, all in-person methodologies were stopped due to the impact of COVID19. Thereafter, recruitment and consent were conducted over the phone and data collection was restricted to phone or mailed questionnaires. Participants received a $20 gift card for their participation.
Demographics and Medical Information
Demographic data included age, gender, race/ethnicity, marital status, employment status, annual household income, household size, and education. Clinical data on co-morbid conditions and treatments were collected. Height and weight were measured and used to classify normal weight, overweight, and obese, based on calculated BMI.
Body composition and physical performance (Chicago only)
Body composition was non-invasively measured using dual energy x-ray absorptiometry (DXA). A certified radiation technologist performed and analyzed participants using a Discovery W (Hologic Inc.) device, which was calibrated daily with manufacturer’s phantom. Applying the methods of Kaul et al. , android fat was automatically defined and measurements of abdominal and visceral fat were obtained from the android region. Sarcopenia was explored using appendicular skeletal mass (ASM)/ height2 using the cut-points of Baumgartner et al.  (7.26 for men and 5.45 for women). Cardiorespiratory fitness was assessed using a sub-maximum 6-minute walk test on an indoor track. Total distance was recorded and used to determine fitness level. Upper body muscle strength was measured using Jamar Hand Dynamometer following standard procedures . Measures were obtained twice for each participant’s non-dominant and dominant hands. Scores of the dominant hand were averaged and the cut-point of <27 kg and <20 kg for men and women, respectively, were used to define compromised functioning . Lower body muscle strength was measured by quantifying the number of ‘Chairs Stands’ in 30 seconds; <5 rises in 15 seconds was used to indicate compromised strength.
Two different dietary screeners were used to assess ease and comparability of results across sites. Both allowed for the evaluation of general trends in intake with minimal burden. Participants at the Milwaukee site completed the validated Block Fat/Sugar/Fruit/Vegetable Screener . This tool takes ~20 min to complete and queries about usual consumption and portion sizes of 55 food items. Participants at the Chicago site completed the Block Fruit/Vegetable/Fiber Screener ; a 10-item on-line fruit, vegetable and fiber screener that ranks individuals with regard to their usual intake of fruits and vegetables . Estimates of total fat, saturated fat, added sugars, fruit, vegetables, and/or dietary fiber were generated.
The Godin Leisure Physical Activity Index asks about time spent engaged in light, moderate, and strenuous activities over the past 7-day period . Results are used to classify participants as sufficiently active (meets current physical activity guidelines of 150 mins of moderate activity or 75 mins of vigorous activity per week) or insufficiently active. Because current cancer survivorship guidelines recommend a minimum of twice weekly resistance exercise training (RET) , participants were also asked: 1) if they engage in RET and 2) if yes, how many times per week.
Quality of life (QOL), symptom burden and social support
Symptom burden and functioning were measured using the Patient-Reported Outcomes Measurement Information System (PROMIS) . The PROMIS-29 profile assesses seven health domains (Physical Function, Anxiety, Depression, Fatigue, Sleep Disturbance, Ability to Participate in Social Roles and Activities, Pain Interference). To further assess social relationships, PROMIS short forms measuring social isolation, companionship, and various domains of support, such as informational and emotional support, were also administered. Short forms measuring self-efficacy in managing symptoms, cognitive function, and sexual function were administered given these are frequent concerns in the MM population .
Using open-ended questions in interviews (Milwaukee) and written prompts (Chicago) on the survey, information regarding (1) post diagnosis changes in health behaviors, (2) post diagnosis changes in general health, (3) unmet needs, and (4) interests related to a lifestyle intervention were gathered.
Data analyses were performed for quantitative and qualitative data. For quantitative data, categorical data were described using percentages, while continuous data were presented using means and standard deviations. PROMIS scores were reported on the T-distribution, which has been normalized to the U.S. population, so that 50 corresponds to the U.S. average with a standard deviation of 10. Higher scores represent more of that domain. Based on recent work by Jensen et al., a 3-point difference in T-score was considered clinically significant between our study population compared to the U.S. population not affected by cancer . Qualitative data from interviews relied on thematic coding. Several authors independently reviewed participant responses to open-ended questions to identify key topics areas and codes within each topic area. Subsequently, coders met to group codes into themes and to resolve discrepancies. Quantitative and qualitative data was stratified to examine racial, gender, and site differences.