Khera, Nandita

Abstract Background: A cancer diagnosis results in a multitude of health care decisions that require adequate patient understanding to navigate the health care system. Understanding the type and degree of information needs of cancer patients is imperative to instructing effective and pertinent patient education. This study sought to evaluate the impact of a patient centered cancer symposium on knowledge level, knowledge needs and symptom burden in a cancer patient population. Methods: Surveys were distributed to attendees of the Mayo Clinic "Living with and Surviving Cancer" patient symposium in January 2016. A follow up survey was sent to participants 3 months after the symposium. Surveys included demographic data, questionnaire evaluating disease comprehension and information assessment, symptom burden, desired information and quality of life assessment. Responses were compared between patients with hematologic and solid tumor malignancies by chi-square test. Pre and post comparisons were done in patients completing both surveys by McNemar test. Results: Demographics: 150 patients completed the pre-intervention survey. Average age of participants 67.5 years. Some patients reported more than 1 malignancy. Disease types represented included hematologic malignancies (N=60), specifically multiple myeloma (N=23), lymphoma (N=26), leukemia (N=8) and MPNs (N=3) as well as solid tumor malignancies (N=90). Most patients were diagnosed with cancer at least 1 year prior to the intervention (N=137). Patients with hematologic malignancies were more likely to be on active treatment at the time of the convention (N=35, 58%) compared to solid tumor malignancies (N=37,41%). Baseline Knowledge: Most patients reported an adequate level of knowledge defined as the proportion of participants reporting "quite a bit" or "very much understanding" on the topic of disease understanding (76%). Lowest levels of knowledge were legal issues (18%), pain management (31%), disease relapse symptoms(46%), fatigue management (48%), end of life decision making (50%) and financial management (50%). Patients with hematologic malignancies were more likely to report understanding of fatigue management (N=36, 60%) but less like to report understanding disease risk factors (N=20, 33%), or increased nutritional interest following diagnosis (N=32, 53%) compared to solid tumor malignancy (N=33,37%) (N=55,61%)(N=70,78%), respectively (all p <0.05). The EORTC-INFO25 at baseline revealed hematologic patients had received less information on sexual side effects of treatments (N=7, 14%) compared to solid tumor patients (N=30, 35%; p=0.004), and reported a lower score of treatment information (50 vs 59; p=0.04). Pre and Post Analysis: Of the 150 initial participants, 81 completed the pre and post-convention surveys (54%). Post-symposium, significant improvement defined as the proportion of participants reporting "quite a bit" or "very much" understanding was noted of disease (pre 75% vs post 89%), stress management (pre 56% vs post 69%), maintain relationships (pre 60% vs post 75%) and pain management (pre 33% vs post 51%)(p-value<0.05). The desire for increased information decreased post-symposium regarding all 9 topics, with significant decreases noted in disease, treatment options and side effects, legal issues and stress management. The EORTC INFO25 analysis revealed a decrease in the reported desire for more information (mean; pre 70 vs post 53, p-value <0.05) Discussion: Cancer patients have a variety of information needs. A patient centered educational conference was able to produce significant improvement in understanding in 4 of 17 topics (24%), including general disease, stress and pain management, and relationship maintenance that was durable at 3 months. At least some improvement in understanding was noted in 10 of 17(59%) additional topics. Knowledge deficit improved in all 9 topics assessed, significantly in five. Cancer patient knowledge needs are varied, and likely evolve over time as a patient progress through stages of disease and treatment. Education should re-evaluate and redistribute information to cancer patients over the course of treatment and survivorship. Disclosures Mikhael: Abbvie: Research Funding; Celgene: Research Funding; Sanofi: Research Funding; Onyx: Research Funding. Dueck:Bayer: Honoraria. Mesa:Novartis: Consultancy; Ariad: Consultancy; Promedior: Research Funding; Galena: Consultancy; Incyte: Research Funding; Celgene: Research Funding; CTI Biopharma: Research Funding; Gilead: Research Funding.

Abstract Background: Patient (Pt) outcomes have improved significantly in multiple myeloma (MM) due to widespread use of novel therapeutic agents and stem cell transplant (SCT) butoutcome disparities among racial/ethnic groups exist. Differences in access to and utilization of novel therapeutics are likely contributing factors. We performed an analysis of the SEER-Medicare database to explore such differences. Methods: Pts with a confirmed MM diagnosis between 2007-2009 and continuous Medicare coverage from 1 year prior to diagnosis through end of 2012 were included. Medicare Part D data (oral medications) is not available prior to 2007 so utilization analysis of lenalidomide (LEN), thalidomide (THAL) and bortezomib (BORT) was performed from 2007-2009. Demographic/survival data (PEDSAF files) and drug/SCT utilization (NCH, OUTSAF, PDESAF files) were obtained (HCPCS/NCH codes). Categorical and continuous variables were compared using /Fisher tests and Wilcoxon tests, respectively. Gender, age, and year-adjusted p-values were calculated using logistic regression. A proportional odds model was used to detect time-dependent trends in drug usage within 12 months of MM diagnosis across year of diagnosis by race. Separate multivariate models were fit for each drug adjusting for gender, age, drug use and race. Associations between drug use and overall survival (OS) were examined using a time dependent proportional hazards model. Results: A total of 5358 MM pts were included (any Medicare claims) with 2733 (51%) males. Mutually exclusive race categories were White (W; 67%), African-American (AA; 18%), Hispanic (H; 10%) and Asian (A; 5%). Median follow up was 2.4 yrs. Baseline pt characteristics were significantly different for age with AA being the youngest at diagnosis (p<0.001) and for gender (p<0.001) with W having the largest (53.4%) and AA the smallest (45.2%) proportion of males. Medicare claims capture only pts without any supplemental insurance and so drug utilization in this database was much lower than expected for all subgroups. Year-wise comparison showed that LEN use within the first year of MM diagnosis increased from 2007 to 2009 (16.5% to 27.7%; p<0.001) and there was a significant decrease in median days to first dose (p=0.002). Over the same period, use of THAL decreased (26.6% to 14.6%; p<0.001) while use of BORT (7.8% to 15.4%; p<0.001), SCT (3.3% to 6.2%; p<0.001) and more than 1 novel agent (2.5% to 5.7%; p<0.001) increased significantly. Utilization patterns within first year of MM diagnosis were significantly different among racial subgroups for LEN, THAL, BORT, SCT and >1 novel agent use. (Figure 1) No significant difference was seen for days to first dose of LEN/THAL for any race but for BORT, H got first dose in median 117 days after diagnosis as compared to median 46-51 days for other races (p=0.025). During the first year after diagnosis, W and AA had higher BORT only usage while H and A had a higher LEN/THAL only usage (p<0.001). For drug use by year and race, a significant increase in LEN use was seen from 2007 to 2009 in W (p<0.001), AA (p=0.001) and A (p<0.001) but not H (p=0.26) while THAL use decreased significantly for all races. BORT use increased significantly for all except A (p=0.602) and SCT use increased significantly for all except AA (p=0.068). Inferior OS was significantly associated with older age (p<0.001) and males (p<0.001). LEN/THAL use within the first year of diagnosis did not affect OS while BORT use within the first year of diagnosis increased mortality risk (HR 1.18, p=0.021). SCT use improved OS (HR 0.52, p<0.001). Conclusions: In this comprehensive analysis of MM therapeutics utilization, we noted significant variability among different racial groups. The use of LEN, BORT and SCT has increased while THAL has decreased over time but disparities exist in utilization among different races, with lower use of LEN in AA, lower use of BORT in A and significantly delayed use of BORT in H, which are all notable since racial/ethnic minorities are fast growing in the US population mix and seem to have inequitable utilization and/or access. Inferior survival with early BORT use may represent selection bias of pts with more advanced/aggressive disease. These trends need to be examined in larger sample sizes especially from commercial payers to eliminate drug access and utilization disparities and achieve equitable benefit of therapeutic advances across all racial subgroups. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures Ailawadhi: Pharmacyclics: Consultancy; Novartis: Consultancy; Amgen Inc: Consultancy; Takeda Oncology: Consultancy.

Abstract Background: A diagnosis of cancer leads to significant information requirements to facilitate health care decision making, understand management options, and health care system navigation. Many patients demonstrate a knowledge deficit that can lead to decreased quality of life and poor health care compliance. This study sought to evaluate the impact of a patient centered cancer symposium on knowledge level, symptom burden, and desired information in a broad population of cancer patients. Methods: Surveys were distributed to attendees of the Mayo Clinic "Living with and Surviving Cancer" patient symposium in January 2015. A follow up survey was sent to participants 3 months after the symposium. Surveys included demographic data, questionnaire evaluating disease comprehension, symptom burden, desired information and quality of life. Results: Demographics: 113 patients completed the pre-intervention survey. Average age of participants was 64.7 years. Disease types represented included hematologic malignancies (N=50), specifically multiple myeloma (N=24), lymphoma (N=17), leukemia (N=6), MPN/MDS (n=3), as well as solid tumor malignancies (N=77). Highest knowledge topics were disease (80%), screening tests (74%), and monitoring tools (72%). Lowest knowledge topics were financial management (13%), legal issues (13%), and pain management (35%). Pre and Post Analysis: Of the 113 initial participants, 79 patients completed the pre- and post-symposium surveys. In the post-symposium setting significant improvement was noted in understanding of disease (pre 80% vs post 92%), treatment options (pre 60% vs post 76%), nutrition (pre 68% vs post 84%) and legal issues (pre 15% vs post 32%). Non-significant knowledge improvement was noted in areas of risk factors and symptoms, screening/monitoring tools, disease side effects, managing stress/fatigue, end of life decision making and health care navigation. Most patients desired increased understanding regarding disease, screening tests, nutrition, and stress and fatigue management. The level of patient reported desire for information of these topics decreased in the post-symposium setting, statistically significant decrease noted in 4 of 5 topics assessed. In hematology patients who completed pre and post surveys (N=36), a higher level of baseline knowledge was reported regarding disease understanding (89%), tools monitoring response (83%) and recurrence (75%), symptoms (83%), relapse symptoms (64%) and treatment options (72%). In this group the decrease in desire for information was less prominent in the post-symposium setting. Discussion: This data demonstrates improved patient reported understanding on a variety of topics with durable response at 3 months. A majority of patients reported adequate baseline knowledge of many topics prior to the symposium but also a desire of increased cancer specific information. Knowledge deficit in cancer care includes disease specific topics, social stressors and health care navigation. The percentage of participants reporting a high desire for more information remained high but did decrease, suggesting decreased knowledge deficit in the post-symposium setting. Patient education should strive to be accessible, all-inclusive and reactive to patient needs. Table 1.Reported "quite a bit" or "very much" understanding and desired information regarding cancer carePrePostP-valueN=79(%)N=79(%)I understand…How to manage financial considerations of my disease/treatment39(51)41(53)0.85How to manage my pain24(34)23(32)0.81Risk factors that can lead to my disease37 (49)42 (56)0.56How to navigate the health care system43(55)47(60)0.56How to manage fatigue38(50)43(56)0.54Process of making end of life decisions33(47)41(53)0.52Screening tests60 (76)63 (81)0.5Symptoms associated with relapse35(45)42(54)0.27Disease symptoms54(70)60(76)0.2Tools monitoring recurrence49(62)56 (72)0.19Tools monitoring response56(71)63(81)0.15How to manage stress46(60)56(71)0.12Legal issues of my disease11(15)24(32)<0.05Treatment options47(60)60(76)<0.05What to eat49(68)66(84)<0.05My disease62(80)71(92)<0.05I desire increased understanding…Of my disease59(77)53(70)0.4Of managing stress50(65)35(51)<0.05Regarding nutrition64(84)54(70)<0.05Of the screening tests62(79)43(57)<0.05How to manage fatigue54(72)42(54)<0.05 Disclosures Tibes: TetraLogic Pharmaceuticals: Research Funding. Mesa:Novartis. Research- incyte, Gilead, cti, Genentech, promedior, NS Pharma: Consultancy.

Abstract Background: The association of clinical trial participation with outcomes in cancer patients is controversial, with some studies reporting better outcomes for trial participants while others do not. It is possible that the difference in outcomes is due to differences in baseline characteristics of participants and non-participants. The Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 0201 trial was a large randomized, multi-center study comparing peripheral blood (PB) with bone marrow (BM) for hematopoietic cell transplantation (HCT) from HLA-matched or 1 allele mismatched unrelated donors (URD) in patients with acute leukemia, chronic myeloid leukemia and myelodysplastic syndromes. The trial reported no significant survival differences between PB and BM; there was decreased risk of graft failure but increased risk of chronic graft-versus-host disease (GVHD) with PB grafts (Anasetti et al NEJM 2012). Methods: We compared characteristics of BMT CTN 0201 study participants (n=494) with HCT recipients who did not participate on the trial (n=1384) but were potentially eligible by virtue of known characteristics. These patients received similar conditioning and GVHD prophylaxis as trial participants at 38 United States trial centers during the study time-period. Centers that had no eligible non-trial participants were excluded. Outcomes between patients on and off trial were compared using Cox proportional hazards regression models. Data were obtained from the Center for International Blood and Marrow Transplant Research (CIBMTR) that collects patient data on all trial and non-trial allogeneic HCT recipients through its registry function. Results: Approximately 29% [494/1878] of apparently eligible patients were randomized and analyzed as part of the trial. There were no significant differences between study participants and non-participants in age or sex distribution, disease types, HLA or sex matching, comorbidities, or interval from diagnosis to HCT. Higher proportions of non-participants were Hispanic, had a lower performance status and lower disease risk, and received myeloablative conditioning and antithymocyte globulin (ATG) or Campath than the study participants. PB was used more commonly as the graft source in the non-participants as compared to the study participants (65% vs. 50%, p<0.001). In multivariate models, overall mortality (HR 1.09, 95% CI 0.95-1.25, p=0.22), transplant-related mortality (HR 1.08, 95% CI 0.89-1.32, p=0.42) and risks of acute grade II-IV (HR 0.96, 95% CI 0.82-1.12, p=0.57) and chronic GVHD (HR 1.06, 95% CI 0.92-1.23, p=0.41) were comparable between study non-participants and participants. Non-participants were at higher risk for relapse than participants (HR 1.24, 95%CI 1.04-1.49, p=0.02) in the multivariate analysis adjusted for disease risk and other clinical variables. Center effect was examined and was not found to be significant (p=0.10). There was no significant interaction between graft source and trial participation (P=0.22) and the risk of mortality was comparable for BM (HR 0.99, 95% CI 0.80-1.22, p=0.90) and PB recipients (HR 1.17, 95% CI 0.97-1.41, p=0.09) whether on trial or off trial. (Figure) Conclusion: Approximately 29% of patients who appeared eligible and were treated with conditioning and GVHD prophylaxis regimens allowed on the trial were enrolled. Despite some differences in clinical characteristics, most of the outcomes of non-trial participants were comparable to trial participants, suggesting that there is no definitive evidence that the outcomes of patients in a clinical trial are superior to those of the non-participants treated in a similar fashion. The comparable outcomes also indicate that the results of this trial were not seriously affected by selection bias and are therefore representative of the larger population. Figure 1: Adjusted overall survival in BMT CTN 0201 trial participants and non-participants Figure 1:. Adjusted overall survival in BMT CTN 0201 trial participants and non-participants Figure 2 Figure 2. Disclosures No relevant conflicts of interest to declare.

Abstract Background: Outcomes in multiple myeloma (MM) have significantly improved, leading to higher prevalence of this disease. Increasing efforts have been made by various agencies to educate patients (pts), manage quality of life (QoL) and survivorship. A global impact of these efforts including pt awareness, disease knowledge and participation in their care, as well as the physical and emotional impact of MM and its treatments (Rx) has not been assessed. We undertook an international, pt-reported survey to investigate these factors and understand any underlying disparities. Methods: A 61-question patient-reported questionnaire (MM-Q), addressing global aspects of MM pts care (components in Figure 1) was used. MM-Q was administered in collaboration with the International Myeloma Foundation (IMF). To limit responses by MM pt only, “SmartPatient” MM community members and IMF Support Group Leader list were invited to complete the survey. Some of the questions were mandatory, and those with missing responses were withdrawn from the final analysis of that pts MM-Q. Categorical and continuous variables were compared among groups using the Chi-square and Kruskal-Wallis tests, respectively using SAS software (v9.3) with a two-sided significance level of 0.05. Results: Survey responses are available from 894/1240 (72%) individuals thus far. Average survey completion time was 16 minutes. Data from 746 MM-Q was included in this analysis with 466 (62.5%) complete responses. Pts from United States (US) represented 85.7% of the responders (majority from New York State). Ethnicity (n=590) was reported as non-Latino by 97%, race (n=580) was reported as White by 95%, family history of MM was reported by 19 pts and 81 reported prior history of a primary malignancy (most common, 37% non-melanoma skin cancer). Living situation questions (n=538) revealed majority were in a household of 2 (62%), were unemployed/on public assistance (53.5%), travelled alone to medical appointments (74%), with travel times varying from <30 min (55%) to >1 hr (17%). Majority (83%) had no prior knowledge of MM, 63% were patient support group members, and 92% did MM-related research on internet. With respect to MM Rx (n=524), 70% had been on Rx within 3 months and 12% reported non-compliance. Prior Rx included thalidomide (30%), lenalidomide (66%), pomalidomide (11.5%), bortezomib (63.5%), carfilzomib (11%), steroids (100%), and clinical trial participation in 25% pts. Side effects from Rx/QoL in prior 3 months (n=403) are summarized in figure 2. Pts satisfaction with Rx/treating staff (n=378) revealed majority felt their Rx to be effective (92%), right for them (87%) and side effects were as expected (84%). In general, pts acknowledged positive attributes of the Rx staff (worked towards same goal; 91%, participated in decisions; 81%, honesty 89%, confidence; 92%), and their doctor's explanations regarding Rx (benefits; 84%, side effects; 74%, test results; 83%), with the exception of emotional effects of Rx (48%). 94% pts rated their care as good/excellent and 95% were willing to participate in a follow-up survey. Stratification of statistically significant responses by geographical categories, age and time since diagnosis are shown in table 1 and analysis by occupation and educational status is ongoing. Conclusions: We conducted a large patient-reported survey to understand the current state of awareness among MM pts and a global impact of the diagnosis and its Rx on them. We showed that majority of pts who accessed the survey responded (72%), which may be a product of literacy, education and access to technology. Majority of the pts were White and non-Latino, highlighting the possible role of access and awareness in ethnic minorities. We report the large prevalence of psychophysical burden of MM and its Rx in MM pts. In general, MM pts had a high rate of satisfaction related to the Rx and their healthcare providers, but the frequency of discussion regarding emotional effects of Rx between doctors and pts was conspicuously low (48%) and the more recently diagnosed pts were less likely to be satisfied with their treating staff. We also demonstrated significant behavioral pattern differences in geographical, age and time-since-diagnosis subgroups, some of which may be sociocultural, while others may be a result of more patient awareness and understanding of the disease. This knowledge will be very helpful in devising MM survivorship strategies. Disclosures No relevant conflicts of interest to declare.

Abstract Abstract 602 Although, MM remains the most common indication for ASCT in the US, a controversy exists about the timing of ASCT. Since early ASCT (eASCT) versus delayed ASCT (dASCT) results in similar progression free survival (PFS) and overall survival (OS) in the novel agent era, economic assessment of one strategy over the other may help both clinicians and policy makers in establishing guidelines in the MM treatment paradigm. Here we designed a study to model a decision tree of both the strategies and estimate their cost-effectiveness. A decision tree was developed to compare the outcomes of ASCT in newly diagnosed transplant eligible MM patients after chemotherapy with novel drug lenalidomide which formed the base case. The survival data was obtained from our published data on eASCT versus dASCT in MM patients who underwent initial therapy with dexamethasone + immunomodulatory agents (IMiDs). The initial decision node on the decision tree model was dichotomized based on the cutoff of 12 months for eASCT versus dASCT. The probabilities of events during the length of follow up (60 months) period which formed our base case were: (a) eASCT cohort: Induction chemotherapy for median of 6 months - ASCT - median 30 months to relapse (drug free interval) followed by median therapy of 24 months for relapsed MM; (b) dASCT cohort: Median initial therapy for 36 months - ASCT - median 24 months to relapse (drug free interval). The cost analysis was done from a US third party payer perspective. The cost data for MM ASCT was obtained from the published data in which direct hospital costs were obtained from Nationwide Inpatient Sample and were converted to Medicare cost-to-charge ratio for urban hospitals. Chemotherapy cost consisted of actual drug costs that calculated based on average wholesale prices AWP (2008), costs of medical oncology visits, and the costs of managing adverse events. Post-ASCT maintenance costs were not included in the analysis. TreeAge Pro 2012 software was used to perform a static decision analysis. The outcomes were expressed in the units of health utility values obtained from phase III HOVON trial which measured quality of life (QoL) using EORTC QLQ-C30 & EuroQoL-5D. All scales and items were linearly transformed to range from 0 to 100. The utilities measured by the EuroQol-5D ranged from 0.40 before induction therapy to 0.70 after ASCT and translated to QALYs using area under the curve method. Using a Consumer Price Index (CPI), the costs of ASCT and chemotherapy were converted into 2012 US$. The robustness of the decision model was tested through univariate analyses and tornado diagram to identify those variables which if varied across clinically relevant measures may result in difference in outcomes. The expected cost of providing treatment to newly diagnosed MM patients for a period of 5 years when they undergo eASCT after induction with IMiDs was calculated to be $211,869. This was $52,454 less than the expected cost of undergoing dASCT. Patients who underwent eASCT had an expected benefit of 2.06 QALYs following treatment, which was 0.19 QALYs more than the patients who received dASCT. This implies that eASCT is preferred over (dominates) dASCT. One-way sensitivity analysis was undertaken to identify factors that were found to be most influential and included OS, overall response rate (ORR) and 1 year ASCT related mortality rate (TRM). Specifically, eASCT showed dominance even if the probability of OS was lowered to 52.1% from the actual case base of 65%, or the probability of ORR was lowered to ≥ 75.8% from the actual case base of 92%, or the probability of 1 year TRM was increased to 11.5% from actual case base of 2%, indicating the robustness of the model favoring cost effectiveness of eASCT. In absence of a survival benefit, the cost savings of early ASCT as measured by ICER would favor this approach, and it would help clinicians, patients, policy makers and third party payers in decision making. Future prospective studies in MM ASCT should incorporate economic and QOL components in the analysis and also focus on the added cost of post-ASCT maintenance therapies with respect to ICER and QALYs. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 1940 Background: Chronic graft-versus-host disease (GVHD) is a significant source of morbidity, mortality, impaired patient-reported quality of life (QOL), greater symptom burden, and prolonged duration of immune suppressive therapy following allogeneic hematopoietic cell transplantation (HCT). While available data support the adverse prognosis of overlap subtype of chronic GVHD, the relative importance of site of gastrointestinal (GI) and type of hepatic involvement is not known. Methods: We analyzed prospectively acquired observational cohort data to examine whether the site of GI involvement (esophageal, upper GI, lower GI) and type of hepatic laboratory test abnormality (bilirubin, alkaline phosphatase (AP), alanine aminotransferase (ALT) elevation over the upper limit of normal based on study site-specific laboratory reference ranges) among chronic GVHD-affected patients are associated with overall survival (OS) and non-relapse mortality (NRM), symptoms (Lee Symptom Scale), and quality of life (QOL) per SF-36 and FACT-BMT instruments. Results: This analysis included 567 individual subjects with baseline and 1548 follow up visits. The majority were incident cases (59%), adults (98%), and predominantly White/Non-Hispanic. Median time from HCT to cohort enrollment was 11.9 months (range 3–294 months). Overall NIH chronic GVHD severity was mild in 10%, moderate in 52%, and severe in 38% at enrollment. Prior acute GVHD was noted in 66%, KPS < 80% in 17%, and platelet count at chronic GVHD onset of < 100 K/uL in 23%. At enrollment, GI involvement was seen in 40% (upper GI 20%, esophagus 16%, lower GI 13%), and liver involvement was seen in 52% (AP 38%, ALT 38%, bilirubin 10%). Time from transplant to cGVHD onset was not different for patients with or without GI involvement or hepatic involvement. In multivariate analysis utilizing data from enrollment visits only and adjusting for patient and transplant variables, lower GI involvement (HR 1.7, p=0.03) and elevated bilirubin (HR 2.36, p=0.001) were associated with OS; both were also associated with NRM. In multivariable analysis using all visits (time-dependent covariates), GI score greater than zero (HR 1.7, p=0.01) and elevated bilirubin (HR 3.7, p<0.001) were associated with OS; results were similar for NRM. Separate models were constructed to examine the relationship of severity of GI and hepatic involvement (0–3 severity score per NIH consensus) and mortality: Increasing lower GI severity at enrollment was significantly associated with non-relapse mortality, with progressively increasing HR for greater severity levels (lower GI score 2/3 vs. 0: HR 5.5, 95% CI 2.3–13.1, p = 0.0001). Increasing bilirubin was also associated with NRM (bilirubin score 2/3 vs. 0: HR 3.3, 95% CI 1.3–8.4, p=0.01). A similar association was observed for bilirubin elevation and NRM in the time-dependent model (bilirubin 2/3 vs. 0: HR 8.9, 95% CI 3.3–24, p < 0.0001). When analyzing all visits, any esophageal involvement and GI score greater than zero were associated with both symptoms and QOL while elevated bilirubin was associated with QOL. We found no consistent evidence that upper GI involvement, AP, ALT, or NIH liver score add prognostic value for survival, overall symptom burden, or quality of life. Conclusion: These data support important differences in patient-reported outcomes according to GI and hepatic involvement among chronic GVHD affected patients, and identify those with elevated bilirubin or higher GI score at any time, or lower GI involvement at cohort enrollment, as patients at greater risk for mortality under current treatment approaches. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 1954 Introduction: Vitamin D (VD) deficiency can cause osteomalacia, aching bone pain, muscle weakness, fatigue, increase risk of fracture and precipitate or exacerbate osteopenia/osteoporosis. Allogeneic hematopoietic stem cell transplant (HCT) patients are susceptible to low VD level secondary to poor oral intake, decreased exposure to sunlight and treatment related malabsorption. VD level has been correlated to cancer incidence and VD metabolites have been used in the treatment of myeloid leukemia. VD shows promising immunomodulatory properties and correction of low VD level may mitigate manifestations of graft versus host disease (GVHD). Reports have suggested that low VD level appears to increase the incidence of GVHD. We hypothesize that there is a relationship between low VD level and morbidity (specifically incidence of acute GVHD), mortality and relapse incidence after HCT. We therefore studied VD levels pre- and post-HCT to determine if VD level impacts these outcomes. Patients and Methods: 241 patients underwent myeloblative or non-myeloablative HCT between January 1, 2009 and January 31, 2011 at our institution. Baseline demographic data, disease characteristics, transplant variables and outcomes data were obtained from the transplant database. These data were supplemented by retrospective chart review for pre- and post-transplant VD level. Categories for VD level included normal (>30 ng/ml) or abnormal (<30 ng/ml). Results: 131 (54%) of patients who underwent myeloblative or non-myeloablative HCT had their VD level evaluated either pre-transplant, post-transplant or both. Pre-HCT 57 (56%) patients, 100 days post-HCT 55 (59%) patients and 365 days post-HCT 12 (32%) patients had a low VD level. 51 (38%) of patients had a reduced intensity HCT and 80 patients (62%) had an ablative conditioning regimen. 11 (8.4%) patients had acute GVHD. Comparison of VD level among those with acute GVHD did show a higher incidence of acute GVHD between those with a low or normal VD level (HR=3.14, 95% CI: 0.35–28.33) however this association was not statistically significant (p=0.3079). Survival analysis in those with a low VD level pre-HCT showed there was not a higher risk of mortality (HR=1.14, 95% CI: 0.18–7.38) after adjusting for post- VD level, and this association was not statistically significant (p=0.8921). Survival analysis in those with a low VD level post-HCT did show a higher risk of mortality (HR=2.59, 95% CI: 0.26–25) after adjusting for pre-HCT VD level, however, this association was not statistically significant (p=0.4155). The relationship between VD level and relapse at 1 year post-HCT was not able to be examined because the VD level for patients who relapsed was not available. Conclusions: Half of patients undergoing HCT had VD testing pre- or post- HCT. Just over half of the patients tested had a low VD level pre- and 100 days post-HCT. It is notable that 365 days post HCT the number of patients with low VD level had decreased. This could be attributable to less time in the hospital thus increasing sun exposure, increasing performance status allowing better ingestion and absorption of VD in the gastrointestinal tract or proper supplementation of a low level noted previously. There was no significant difference in incidence of acute GVHD by VD level though there was a trend for increased risk. Pre- and post-HCT VD level did not significantly impact mortality but there was a trend toward higher risk in those with a low VD level post-HCT. This is the first study, to our knowledge, to evaluate the impact of VD level on mortality post-HCT. Our study confirms that many patients have a low VD level pre-and post-HCT. These findings, especially the trend towards a higher mortality risk and higher incidence of acute GVHD in those who have a low VD level post-HCT, warrant further prospective investigation. VD supplementation may be a low cost, easy to implement addition to routine post HCT care that might reduce HCT associated mortality. Disclosures: Reeder: Celgene: Research Funding.

Abstract Background A better understanding of social determinants of health and psychosocial and economic sequelae of cancer for an ethnoculturally diverse population is needed. A Patient-Reported Outcomes Quality of Life (PROQOL) instrument has been developed to capture less commonly discussed patient concerns. We sought to adapt this instrument and pilot it in underserved, ethnic minority cancer patients at Maricopa Integrated Health System (MIHS) in Phoenix, AZ. Methods Two focus groups (FG) were conducted to understand patient perspectives of an adapted Spanish version of PROQOL. A revised version based on FG data was then piloted among patients on active cancer treatment. Patients also completed a Linear Analog Scale Assessment (from 1-10) for quality of life (QOL) and its domains. Results All participants in both FG (11/12 female) were Hispanic, with a median age of 53 years. Participants agreed that the domains covered in the PROQOL represented the most important psychosocial needs in their cancer continuum. No difficulty in understanding the individual domains or questions for delving deeper into each domain was noted. Median age of 34 pilot study participants (74% female, 79% Hispanic) was 48 years. 28/34 completed the Spanish version. The most common concerns were ‘Money’ (32%) and ‘Cancer Diagnosis, Treatment and Survivorship’ (32%). Median overall QOL was 8 (range 3-10). Conclusions PROQOL is feasible for systematic capture of patient concerns in underserved, ethnic minority cancer patients. A better understanding of these concerns can help design interventions to improve psychosocial outcomes for the vulnerable groups and decrease disparities in care delivery.

Abstract HLA-haploidentical hematopoietic cell transplantation (Haplo-HCT) using posttransplantation cyclophosphamide (PT-Cy) has improved donor availability. However, a matched sibling donor (MSD) is still considered the optimal donor. Using the Center for International Blood and Marrow Transplant Research database, we compared outcomes after Haplo-HCT vs MSD in patients with acute myeloid leukemia (AML) in first complete remission (CR1). Data from 1205 adult CR1 AML patients (2008-2015) were analyzed. A total of 336 patients underwent PT-Cy–based Haplo-HCT and 869 underwent MSD using calcineurin inhibitor–based graft-versus-host disease (GVHD) prophylaxis. The Haplo-HCT group included more reduced-intensity conditioning (65% vs 30%) and bone marrow grafts (62% vs 7%), consistent with current practice. In multivariable analysis, Haplo-HCT and MSD groups were not different with regard to overall survival (P = .15), leukemia-free survival (P = .50), nonrelapse mortality (P = .16), relapse (P = .90), or grade II-IV acute GVHD (P = .98). However, the Haplo-HCT group had a significantly lower rate of chronic GVHD (hazard ratio, 0.38; 95% confidence interval, 0.30-0.48; P &lt; .001). Results of subgroup analyses by conditioning intensity and graft source suggested that the reduced incidence of chronic GVHD in Haplo-HCT is not limited to a specific graft source or conditioning intensity. Center effect and minimal residual disease–donor type interaction were not predictors of outcome. Our results indicate a lower rate of chronic GVHD after PT-Cy–based Haplo-HCT vs MSD using calcineurin inhibitor–based GVHD prophylaxis, but similar other outcomes, in patients with AML in CR1. Haplo-HCT is a viable alternative to MSD in these patients.