Risk factors for loss to follow-up after traumatic injury: An updated view of a chronic problem

BACKGROUND

Loss to follow-up after traumatic injury occurs at rates of up to 47%. However, the most recent data are over a decade old, and recent changes in traumatic injury patterns necessitate an updated assessment of risk factors for loss to follow-up after trauma.

METHODS

We conducted a retrospective chart review of trauma admissions from January 1, 2018 to December 31, 2021. Categorical variables were compared using χ2 analyses, and continuous variables were analyzed using Mann-Whitney Wilcoxon tests. Multivariable logistic regression was used to adjust for relevant factors identified on unadjusted analysis.

RESULTS

Among 3,034 patients, overall loss to follow-up was 36.9%. Non-White patients, patients who underwent operations or non-surgical procedures, and patients discharged to rehabilitation facilities were more likely to have follow-up appointments within 30 days. Patients with substance use disorder and, among White patients, those with public insurance had higher loss to follow-up rates. Having a follow-up appointment scheduled with a primary care provider was the single most significant factor associated with attending a follow-up appointment.

CONCLUSION

Social determinants of health, such as insurance status and substance use disorder, are associated with loss of follow-up after trauma. Primary care appointments are associated with the highest attendance rates, supporting that all patients should be offered primary care appointments after traumatic injury.

Breast Cancer Screening and BI-RADS Scoring Trends Before and During the COVID-19 Pandemic in an Academic Safety-Net Hospital

BACKGROUND

Little is known about how the COVID-19 pandemic affected screening mammography rates and Breast Imaging Reporting and Data Systems (BI-RADS) categorizations within populations facing social and economic inequities. Our study seeks to compare trends in breast cancer screening and BI-RADS assessments in an academic safety-net patient population before and during the COVID-19 pandemic.

PATIENTS AND METHODS

Our single-center retrospective study evaluated women ≥ 18 years old with no known breast cancer diagnosis who received breast cancer screening from March 2019-September 2020. The screening BI-RADS score, completion of recommended diagnostic imaging, and diagnostic BI-RADS scores were compared between the pre-COVID-19 era (from 1 March 2019 to 19 March 2020) and COVID-19 era (from 20 March 2020 to 30 September 2020).

RESULTS

Among the 11,798 patients identified, screened patients were younger (median age 57 versus 59 years, p < 0.001) and more likely covered by private insurance (35.9% versus 32.3%, p < 0.001) during the COVID-19 era compared with the pre-COVID-19 era. During the pandemic, there was an increase in screening mammograms categorized as BI-RADS 0 compared with the pre-COVID-19 era (20% versus 14.5%, p < 0.0001). There was no statistically significant difference in rates of completion of diagnostic imaging (81.6% versus 85.4%, p = 0.764) or assignment of suspicious BI-RADS scores (BI-RADS 4-5; 79.9% versus 80.8%, p = 0.762) between the two eras.

CONCLUSIONS

Although more patients were recommended to undergo diagnostic imaging during the pandemic, there were no significant differences in race, completion of diagnostic imaging, or proportions of mammograms categorized as suspicious between the two time periods. These findings likely reflect efforts to maintain equitable care among diverse racial groups served by our safety-net hospital.

Disparate impact of the COVID-19 pandemic on delays in colorectal cancer treatment: A National Cancer Database study

BACKGROUND

Timely treatment for patients with colorectal cancer may have been disrupted by the COVID-19 pandemic. We evaluated the impact of the pandemic on delays to treatment with surgery or systemic therapy for patients with colorectal cancer and delineated factors predictive of delayed treatment.

METHODS

Using the National Cancer Database, patients diagnosed with colorectal cancer were categorized by year of diagnosis as COVID-19 era (2020) versus pre-COVID-19 (2018-2019). Categorical variables were compared by χ2 analysis. Multivariate logistic regression was used to assess odds ratios for delayed time to surgery or chemoimmunotherapy, defined as >60 days.

RESULTS

In total, 50,689 patients colorectal cancer were diagnosed patients who were pre-COVID-19 vs 21,331 within the COVID-19-era. Patients diagnosed with COVID-19 had a higher stage at diagnosis. There were no differences in the proportion of delayed time to surgery for patients diagnosed in 2020, but patients who were tested for COVID-19 had increased proportions of delayed time to surgery (P < .0001). In multivariate analysis, Black race (P = .0026) and uninsured/underinsured status (P = .0017) were associated with delayed time to surgery. Diagnosis during COVID-19 did not increase delayed time to chemoimmunotherapy, regardless of COVID-19 testing or positivity; however, delays were seen for Black (P < .0001), Hispanic (P < .0001), and uninsured/underinsured patients (P < .0001).

CONCLUSION

Although the pandemic did not delay treatment for colorectal cancer overall, vulnerable and underserved populations were disproportionately affected by delays to all forms of therapy. The difference in colorectal cancer outcomes in the coming years as a result of delays in treatment may be significant for these patients.

Bullet holes: A novel model to identify the most impactful gaps in the firearm homicide prevention laws of each state

BACKGROUND

Firearm homicides (FH) are a major cause of mortality in the United States (US). Firearm law implementation is variable across states, and legislative gaps may represent opportunities for FH prevention. For each state, we sought to identify which firearm law category would have been most effective if implemented and how effective it would have been.

METHODS

We conducted a retrospective cohort study examining the effects of firearm laws on FH rates in the 48 contiguous US states 2010-2019. Data were obtained from the CDC WONDER and FBI UCR databases, State Firearm Law Database, and US Census. Firearm laws were grouped into 14 categories. We assessed the association between the presence of each law category and FH rate as an incidence rate ratio (IRR) using a Poisson regression accounting for state population characteristics and laws of surrounding states. We estimated the IRR for each state that did not have a given law category present and determined which of these missing law categories would have been associated with the greatest reduction in FH rate.

RESULTS

FH rates varied widely across states and increased from a mean of 3.2 (SD = 1.7) to 4.2 (SD = 2.9) FH per 100,000. All law categories were significantly associated with decreased FH rate (p < 0.05), with IRR ranging from 0.25-0.85. The most effective missing law category differed between states but was most commonly child access prevention (34.09% of states), assault weapons and large-capacity magazines (15.91%), preemption (15.91%), and concealed carry permitting (13.64%). In total across 2010-2019, we estimated that 129,599 fewer FH would have occurred with enactment of the most effective missing law category in each state.

CONCLUSIONS

Modeling firearm law prevention of FH with regard to state legislative and population characteristics can identify the highest impact missing law categories in each state. These results can be used to inform efforts to reduce FH.

LEVEL OF EVIDENCE

Level III, Prognostic/Epidemiological.

Social Determinants of Health Associated With the Need for Urgent Versus Elective Cholecystectomy at an Urban, Safety-Net Hospital

Background: Benign gallstone disease is the most frequent indication for cholecystectomy in the United States. Many patients present with complicated disease requiring urgent interventions, which increases morbidity and mortality. We investigated the association between individual and population-level social determinants of health (SDoH) with urgent versus elective cholecystectomy. Patients and Methods: All patients undergoing cholecystectomy (2014-2021) for benign gallstone disease were included. Demographic and clinical data were linked to population-level SDoH characteristics using census tracts. Data were analyzed using descriptive and inferential statistics. Results: A total of 3,197 patients met inclusion criteria; 1,913 (59.84%) underwent urgent cholecystectomy, 1,204 (37.66%) underwent emergent cholecystectomy, and 80 (2.5%) underwent interval cholecystectomy. On multinomial logistic regression, patients who were older (relative risk [RR], 1.010; p < 0.001), black (RR, 1.634; p = 0.008), and living in census tracts with a higher percent of poverty (RR, 0.017; p = 0.021) had a higher relative risk of presenting for urgent cholecystectomy. Patients who were female (RR, 0.462; p < 0.001), had a primary care provider (PCP; RR, 0.821; p = 0.018), and lived in census tracts with low supermarket access (RR, 0.764; p = 0.038) had a lower relative risk of presenting for urgent cholecystectomy. Only age (RR, 1.066; p < 0.001), female gender (RR, 0.227; p < 0.001), and having a PCP (RR, 1.984; p = 0.034) were associated with presentation for interval cholecystectomy. Conclusions: Patients who were older, black, and living in census tracts with high poverty levels had a higher relative risk of presenting for urgent cholecystectomy at our institution, whereas females and patients with PCPs were more likely to undergo elective cholecystectomy. Improved access to primary care and surgical clinics for all patients at safety-net hospitals may result in improved outcomes in the management of benign gallstone disease by increasing diagnosis and treatment in the elective setting.

Gun Laws, Stay-at-Home Orders, and Poverty: Surges in Pandemic Firearm Violence in Large US Cities

INTRODUCTION

The COVID-19 pandemic heralded a surge in firearm homicides (FH) in many, but not all, urban centers. We aimed to determine the relationship of firearm restrictive legislation, stay-at-home orders (SaHOs), and FH during the height of the COVID-19 pandemic in US cities.

METHODS

Demographics and socioeconomic data were captured from the 2020 US Census for large (population ≥250,000) cities. FH data were captured from the Gun Violence Archive. We retrieved firearm recovery estimates from the Bureau of Alcohol, Tobacco, and Firearms Firearms Trace Database. Firearm restrictive legislation was gathered from the State Firearm Laws Database. SaHO durations were found from press releases and government sources. Variables with P ≤ 0.200 in univariate linear regression were entered into a final multivariable model.

RESULTS

A median of 7.5 FH per 100,000 people occurred in the 85 included US cities across 32 states in 2020 (range, 0.35-69.80 per 100,000). In multivariable regression, longer SaHOs (β: 0.033, 95% confidence interval [CI]: 0.014-0.053, P = 0.001) and higher poverty (β: 0.471, 95% CI: 0.280-0.670, P < 0.001) were associated with increases in FH. Handgun-specific laws (β: -0.793, 95% CI: -1.430 to -0.160, P = 0.015) were associated with lower FH.

CONCLUSIONS

We found that poverty and longer SaHOs were associated with increased FH in large US cities during the height of the pandemic, while handgun-specific laws were associated with a decrease. Reducing poverty, mitigating the negative effects of SaHOs, and expanding handgun-specific legislation may protect from surges in FH during future crises.

Death by the Minute: Inequities in Trauma Care for Victims of Firearm Violence

BACKGROUND

Gunshot wounds (GSWs) remain a significant source of mortality in the United States. Timely delivery of trauma care is known to be critical for survival. We sought to understand the relationship of predicted transport time and death after GSW. Given large racial disparities in firearm violence we also sought to understand disparities in transport times and death by victim race, an unstudied phenomenon.

METHODS

Firearm mortality data were obtained from the Boston Police Department 2005-2023. Firearm incidents were mapped using ArcGIS. Predicted transport times for each incident to the closest trauma center were calculated in ArcGIS. Spatial autoregressive models were used to understand the relationship between victim race, transport time to a trauma center and mortality associated with the shooting incidents.

RESULTS

There were 4,545 shooting victims with 758 deaths. Among those who lived, the median transport time was 9.4 minutes (IQR 5.8, 13.8) and 10.5 minutes (IQR 6.4, 14.6, p = 0.003) for those who died. In the multivariable logistic regression, increased transport time to the nearest trauma center (OR 1.024, 95% CI 1.01-1.04) and age (OR 1.016, 95% CI 1.01-1.02) were associated with mortality. There was a modest difference in median transport time to the nearest trauma center by race with non-Hispanic Black at 10.1 minutes, Black Hispanic 9.2 minutes, white Hispanic 8.5 minutes, and non-Hispanic white 8.3 minutes (p < 0.001).

CONCLUSIONS

Our results highlight the relationship of transport time to a trauma center and death after a GSW. Non-white individuals had significantly longer transport times to a trauma center and predicted mortality would have been lower with white victim transport times. These data underscore the importance of timely trauma care for GSW victims and can be used to direct more equitable trauma systems.

LEVEL OF EVIDENCE

Level III, Prognostic/Epidemiological.

Impact of the COVID-19 Pandemic on Delays to Breast Cancer Surgery: Ripples or Waves?

BACKGROUND

Adherence to current recommendations for optimal time from diagnosis to treatment for patients with breast cancer may have been disrupted by the COVID-19 pandemic. This study aimed to evaluate the impact of the pandemic on time to surgery or systemic treatment with chemotherapy or immunotherapy for patients diagnosed with breast cancer.

METHODS

Using the National Cancer Database, patients diagnosed with breast cancer in 2020 were compared to those diagnosed from 2018-2019 (Pre-COVID). Sub-analyses were performed for patients who were tested for COVID-19 and those who had a positive result in 2020. Multivariate logistic regression was used assess odds ratios for delayed time to surgery (DTS, defined as > 90 days) or systemic therapy (defined as > 120 days).

RESULTS

In total, 230,997 patients were diagnosed with breast cancer in 2018 and 2019 compared to 102,065 in 2020. Of the 2020 cohort, 47,659 (46.7%) received COVID-19 testing; of which, 3,158 (6.6%) resulted positive. A larger proportion of COVID-tested or COVID-positive patients had higher stage at diagnosis. DTS was more likely for patients who were diagnosed in 2020, uninsured or underinsured, non-white, Hispanic, less educated, or age < 70 years. Similar factors were predictive of delay to systemic therapy (less age < 70 years); however, diagnosis in 2020 was not.

CONCLUSION

The COVID-19 pandemic was associated with significant DTS for breast cancer but spared time to systemic therapy. Delays disproportionately impacted vulnerable and underserved patient populations. The true clinical effects of these delays may yet be realized for breast cancer patients.

Delay and Disparity in Observed vs Predicted Incidence Rate of Screenable Cancer During the COVID-19 Pandemic

BACKGROUND

The COVID-19 pandemic resulted in disruption of healthcare services, including cancer screenings, yet data on this are limited. We sought to compare observed and expected cancer incidence rates for screenable cancers, quantifying potential missed diagnoses.

STUDY DESIGN

Lung, female breast, and colorectal cancer patients from 2010 to 2020 in the National Cancer Database were standardized to calculate annual incidence rates per 100,000. A linear regression model of 2010 through 2019 incidence rates (pre-COVID) was used to calculate predicted 2020 incidence compared with observed incidence in 2020 (COVID) with subanalyses for age, sex, race, ethnicity, and geographic region.

RESULTS

In total, 1,707,395 lung, 2,200,505 breast, and 1,066,138 colorectal cancer patients were analyzed. After standardizing, the observed 2020 incidence was 66.888, 152.059, and 36.522 per 100,000 compared with the predicted 2020 incidence of 81.650, 178.124, and 44.837 per 100,000, resulting in an observed incidence decrease of -18.1%, -14.6%, and -18.6% for lung, breast, and colorectal cancer, respectively. The difference was amplified on subanalysis for lung (female, 65 or more years old, non-White, Hispanic, Northeastern and Western region), breast (65 or more years old, non-Black, Hispanic, Northeastern and Western region), and colorectal (male, less than 65 years old, non-White, Hispanic, and Western region) cancer patients.

CONCLUSIONS

The reported incidence of screenable cancers significantly decreased during the COVID-19 pandemic (2020), suggesting that many patients currently harbor undiagnosed cancers. In addition to the human toll, this will further burden the healthcare system and increase future healthcare costs. It is imperative that providers empower patients to schedule cancer screenings to flatten this pending oncologic wave.

The Impact of the COVID-19 Pandemic on Hepatocellular Carcinoma Time to Treatment Initiation: A National Cancer Database Study

BACKGROUND

The COVID-19 pandemic strained oncologic care access and delivery, yet little is known about how it impacted hepatocellular carcinoma (HCC) management. Our study sought to evaluate the annual effect of the COVID-19 pandemic on time to treatment initiation (TTI) for HCC.

METHODS

The National Cancer Database was queried for patients diagnosed with clinical stages I-IV HCC (2017-2020). Patients were categorized based on their year of diagnosis as "Pre-COVID" (2017-2019) and "COVID" (2020). TTI based on stage and type of treatment first received was compared by the Mann-Whitney U test. A logistic regression model was used to evaluate factors of increased TTI and treatment delay (> 90 days).

RESULTS

In total, 18,673 patients were diagnosed during Pre-COVID, whereas 5249 were diagnosed during COVID. Median TTI for any first-line treatment modality was slightly shorter during the COVID year compared with Pre-COVID (49 vs. 51 days; p < 0.0001), notably in time to ablation (52 vs. 55 days; p = 0.0238), systemic therapy (42 vs. 47 days; p < 0.0001), and radiation (60 vs. 62 days; p = 0.0177), but not surgery (41 vs. 41 days; p = 0.6887). In a multivariate analysis, patients of Black race, Hispanic ethnicity, and uninsured/Medicaid/Other Government insurance status were associated with increased TTI by factors of 1.057 (95% CI: 1.022-1.093; p = 0.0013), 1.045 (95% CI: 1.010-1.081; p = 0.0104), and 1.088 (95% CI: 1.053-1.123; p < 0.0001), respectively. Similarly, these patient populations were associated with delayed treatment times.

CONCLUSIONS

For patients diagnosed during COVID, TTI for HCC, while statistically significant, had no clinically significant differences. However, vulnerable patients were more likely to have increased TTI.

Mitigating disparities in breast cancer treatment at an academic safety-net hospital

PURPOSE

Among women with non-metastatic breast cancer, marked disparities in stage at presentation, receipt of guideline-concordant treatment and stage-specific survival have been shown in national cohorts based on race, ethnicity, insurance and language. Little is published on the performance of safety-net hospitals to achieve equitable care. We evaluate differences in treatment and survival by race, ethnicity, language and insurance status among women with non-metastatic invasive breast cancer at a single, urban academic safety-net hospital.

METHODS

We conducted a retrospective study of patients with invasive ductal or lobular breast cancer, diagnosed and treated between 2009 and 2014 at an urban, academic safety-net hospital. Demographic, tumor and treatment characteristics were obtained. Stage at presentation, stage-specific overall survival, and receipt of guideline-concordant surgical and adjuvant therapies were analyzed. Chi-square analysis and ANOVA were used for statistical analysis. Unadjusted survival analysis was conducted by Kaplan-Meier method using log-rank test; adjusted 5 year survival analysis was completed stratified by early and late stage, using flexible parametric survival models incorporating age, race, primary language and insurance status.

RESULTS

520 women with stage 1-3 invasive breast cancer were identified. Median age was 58.5 years, 56.1% were non-white, 31.7% were non-English-speaking, 16.4% were Hispanic, and 50.1% were Medicaid/uninsured patients. There were no statistically significant differences in stage at presentation between age group, race, ethnicity, language or insurance. The rate of breast conserving surgery (BCS) among stage 1-2 patients did not vary by race, insurance or language. Among patients indicated for adjuvant therapies, the rates of recommendation and completion of therapy did not vary by race, ethnicity, insurance or language. Unadjusted survival at 5 years was 93.7% for stage 1-2 and 73.5% for stage 3. Adjusting for age, race, insurance status and primary language, overall survival at 5 years was 93.8% (95% CI 86.3-97.2%) for stage 1-2 and 83.4% (95% CI 35.5-96.9%) for stage 3 disease. Independently, for patients with early- and late-stage disease, age, race, language and insurance were not associated with survival at 5-years.

CONCLUSION

Among patients diagnosed and treated at an academic safety-net hospital, there were no differences in the stage at presentation or receipt of guideline-concordant treatment by race, ethnicity, insurance or language. Overall survival did not vary by race, insurance or language. Additional research is needed to assess how hospitals and healthcare systems mitigate breast cancer disparities.

Persistent Child Poverty and Mortality in a Cohort of Children with Cancer in Alabama

BACKGROUND

One-fifth of U.S. counties are designated persistent child poverty counties (≥20% of children in poverty since 1980). The association between a persistent child poverty environment and mortality in children with cancer is unknown.

METHODS

Our cohort includes 2,089 children with cancer (2000-2016) in Alabama. We used multivariable Cox proportional hazards modeling (adjusted for sociodemographics/clinical characteristics) to assess mortality by persistent child poverty designation at 1, 5, and 10 years from diagnosis. Distance to treatment was subsequently explored.

RESULTS

Forty-two percent of the cohort lived in a persistent child poverty county; they were more likely to be African American (P < 0.0001), have public/no insurance (P = 0.0009), and live >100 miles to treatment (P < 0.0001). Children in persistent child poverty counties were 30% more likely to die by 5 years [95% confidence interval (CI) = 1.06-1.59; P = 0.012]. Distance (per 20-mile increase) to treatment was associated with a 9% increased mortality risk (P < 0.0001). Children with both exposures (distance >100 miles and persistent child poverty) faced the highest mortality risk at 5 years (HR = 1.80; 95% CI = 1.39-2.33; P < 0.0001). In subanalysis, children exposed to persistent child poverty were at higher risk for cancer-related mortality. However, the risk of health-related mortality did not differ.

CONCLUSIONS

Among children with cancer from the Deep South, persistent child poverty was a prevalent exposure associated with inferior overall survival. Distance to treatment was independently associated with inferior survival. Children with both exposures had the highest risk of mortality.

IMPACT

Persistent child poverty is associated with inferior survival among children with cancer; mechanisms underlying this disparity warrant investigation. See related commentary by Orjuela-Grimm and Beauchemin, p. 295.

Using the TIDieR checklist to describe development and integration of a web-based intervention promoting healthy eating and regular exercise among older cancer survivors

Objective

To facilitate replication and future intervention design of web-based multibehavior lifestyle interventions, we describe the rationale, development, and content of the AiM, Plan, and act on LIFestYles (AMPLIFY) Survivor Health intervention which provides healthy eating and exercise behavior change support for older cancer survivors. The intervention promotes weight loss, improvements in diet quality, and meeting exercise recommendations.

Methods

The Template for Intervention Description and Replication (TIDieR) checklist was used to provide a comprehensive description of the AMPLIFY intervention, consistent with CONSORT recommendations.

Results

A social cognitive theory web-based intervention founded on the core components of efficacious print and in-person interventions was conceptualized and developed through an iterative collaboration involving cancer survivors, web design experts, and a multidisciplinary investigative team. The intervention includes the AMPLIFY website, text and/or email messaging, and a private Facebook group. The website consists of: (1) Sessions (weekly interactive e-learning tutorials); (2) My Progress (logging current behavior, receiving feedback, setting goals); (3) Tools (additional information and resources); (4) Support (social support resources, frequently asked questions); and (5) Home page. Algorithms were used to generate fresh content daily and weekly, tailor information, and personalize goal recommendations. An a priori rubric was used to facilitate intervention delivery as healthy eating only (24 weeks), exercise only (24 weeks), or both behaviors concurrently over 48 weeks.

Conclusions

Our TIDieR-guided AMPLIFY description provides pragmatic information helpful for researchers designing multibehavior web-based interventions and enhances potential opportunities to improve such interventions.

Location of death and hospice use in children with cancer varies by type of health insurance

BACKGROUND

Disparities in end-of-life (EOL) care for children with cancer remain understudied. We addressed this gap by examining patterns of EOL care, with a focus on location of death and hospice utilization.

METHODS

We used MarketScan - a nationally representative dataset with Medicaid and commercial claims to conduct a retrospective observational study of children with cancer who died between 2011 and 2017 at age ≤21 years. We examined rates of (a) home death, (b) hospice utilization, (c) and medically intense interventions in the last 30 days (e.g., intubation).

RESULTS

Of the 1492 children in the cohort, 44% had Medicaid and 56% commercial insurance; 71% carried a solid tumor diagnosis, and 37% were between the ages of 15 and 21 years at the time of death. Forty percent died at home; children with Medicaid were less likely to die at home (relative risk [RR] = 0.82, 95% confidence interval [CI]: 0.73-0.92; reference: commercial). Forty-five percent enrolled in hospice, for a median of 2 days. Hospice enrollment rates did not vary with insurance. However, children with Medicaid spent less time enrolled (incidence rate ratio [IRR] = 0.22, 95% CI: 0.17-0.27). Among children with Medicaid, Black children were less likely to die at home (RR = 0.69, 95% CI: 0.52-0.92) and enroll on hospice (RR = 0.71, 95% CI: 0.55-0.91) than non-Hispanic White children. Medically intense interventions did not vary with insurance or race.

CONCLUSION

Only 40% of children with cancer die at home, and the duration of hospice enrollment is short. EOL care varies significantly with insurance. It is imperative that we determine if these patterns and disparities represent EOL preferences, provider biases, or differences in quality or availability of hospice.

Using CT-based body composition metrics and frailty index in predicting survival among older adults with cancer

12057

Background: Older adults with cancer are at an increased risk of treatment related toxicities and excess mortality during cancer treatment. While altered body composition and frailty are associated with worse survival among older adults with cancer, no prior study has examined their combined influence on survival prediction. Methods: Prospective study of older adults (≥60 years) undergoing geriatric assessment (GA) at initial visit with a medical oncologist at UAB from 9/2017-07/2021 with available abdominal computed tomography (CT) within 60 days of GA. Using multi-slice CT images from T12 to L5 level, volumetric skeletal muscle (SMV), visceral (VATV) and subcutaneous (SATV) adipose tissues, and skeletal muscle density (SMD), were derived. Sex-specific z-scores for each measure were determined. A 44-item frailty index was obtained, using the deficit accumulation model. Overall survival (OS) was defined as time from GA to death or last follow-up (11/8/2021). Kaplan-Meier estimates of survival rates were compared using log-rank statistics. Multivariable cox regression models were used to predict OS in a random sub-sample (1:1 split of training:validation set), sequentially adding frailty and each body composition measure and assessing improvement with likelihood ratio tests and Harrel’s C statistic. Results: 815 patients were included (median age 68 years, 61% men, and 75% non-Hispanic Whites. 73% had gastrointestinal malignancies (stage III, 25%, stage, IV 48%). 32% were frail, 31% pre-frail. There was a weak negative correlation between height-adjusted SMV and frailty (r = -0.16), particularly among men (r -0.24). Over a median follow-up of 25.7 months (range 0.3-49.6 months), 268 patients (33%) died. The 2-year survival rate was 75.9%, 68.5% and 52.2% among robust, pre-frail and frail (log-rank p <.001), respectively. In multivariable models adjusted for age, sex, race, cancer type and cancer stage, being frail (vs robust) (Hazards Ratio, HR = 2.32; 95%CI: 1.69-3.2; p <.001) and higher skeletal muscle volume (HR = 0.85; 95%CI: 0.72-0.99; p= 0.04, per SD increment) were independently associated with OS. Adding body composition and frailty to clinical variables led to significant improvement in prediction (Harrel’s C increased from 0.69 to 0.74). In the validation set, discrimination was similar (Harrel’s C = 0.72) and plots suggested good model calibration. Conclusions: CT-based body composition metrics and frailty are independent predictors of OS among older adults with cancer and improve survival prediction compared to routine clinical risk factors.

Rural-urban disparities in Geriatric Assessment (GA) impairments and mortality among older adults with cancer: Results from the Cancer and Aging Resilience Evaluation (CARE) Registry

12012

Background: Rural-urban disparities persist in cancer incidence and mortality, despite improvement in cancer screening and treatment. Older adults are at increased risk for cancer, represent the majority of cancer cases, and are more likely to reside in rural areas than younger adults are. GAs are recommended in clinical management of older adults with cancer, in part to identify aging-related impairments predictive of adverse outcomes; few studies have explored rural-urban disparities in GA impairments and mortality among older patients with cancer. Methods: We included 937 older adults (≥60y) from the CARE registry recently diagnosed with cancer who underwent GA at their first pre-chemotherapy visit to the UAB oncology clinic. Rural-urban status using Rural-Urban Commuting Area (RUCA) codes, classified the cohort by residence in metropolitan, micropolitan, and rural/small town areas. We included self-rated performance status (PS) (≥2 on Eastern Cooperative Oncology Group PS), instrumental activities of daily living (IADL) [≥1 impairment], physical and mental health-related quality of life (HRQoL) [t-score < 40 on PROMIS 10-item Global Health], and overall survival as outcomes. Logistic regression evaluated the association between rural-urban status and each outcome (except overall survival, where we used Cox regression analyses). Micropolitan residence was chosen as the reference category due to similar high risk in both rural and urban areas. Results: Median age at study participation was 69.0y (Interquartile range: 64.0-74.0); 12.4% resided in rural, 14.8% in micropolitan and 72.8% in urban areas; 22.5% were diagnosed with colorectal cancer, 19.0% with pancreatic, and 12.4% with hepatobiliary; 74.7% were Stage III/IV. Participants in rural areas were more likely to be white and less educated. After adjustment for age, sex, race, education and cancer type/stage, rural residence was associated with increased odds of impaired PS (OR = 1.93, 95% CI: 1.10-3.40), limitations in IADLs (OR = 1.79, 95% CI: 1.03-3.10), and impaired physical HRQoL (OR = 1.84, 95% CI: 1.05-3.22) compared to micropolitan residence. Urban residence was not significantly associated with any GA outcomes compared to micropolitan residence. Rural residence was associated with higher hazard of death compared to micropolitan residence (HR = 1.88, 95% CI: 1.09-3.24) as well as urban residence (HR = 1.67, 95% CI: 1.13-2.45). Conclusions: Among older adults with newly diagnosed cancer, rural residence was associated with impaired PS, limitations in IADLs, impaired physical HRQoL, and reduced overall survival. Implementation of routine GA among older adults in rural areas may aid in early identification and intervention on GA impairments to improve cancer outcomes.

Adapting MultiPLe behavior Interventions that eFfectively Improve (AMPLIFI) cancer survivor health: program project protocols for remote lifestyle intervention and assessment in 3 inter-related randomized controlled trials among survivors of obesity-related cancers

BACKGROUND

Scalable, multiple behavior change interventions are needed to address poor diet, inactivity, and excess adiposity among the rising number of cancer survivors. Efficacy-tested diet (RENEW) and exercise (BEAT Cancer) programs were adapted for web delivery among middle-aged and older cancer survivors for the AMPLIFI study, a National Cancer Institute-funded, multi-site, program project.

METHODS

Throughout the continental U.S., survivors of several obesity-related cancers are being recruited for three interconnected randomized controlled trials (RCTs). Projects 1 and 2 test 6-month diet or exercise interventions versus a wait-list control condition. Upon completion of the 6-month study period, the intervention participants receive the next behavior change sequence (i.e., diet receives exercise, exercise receives diet) and the wait-list control arm initiates a 12-month combined diet and exercise intervention. Project 3 tests the efficacy of the sequential versus simultaneous interventions. Assessments occur at baseline and semi-annually for up to 2-years and include: body mass index, health behaviors (diet quality, accelerometry-assessed physical activity/sleep), waist circumference, D3 creatine-assessed muscle mass, physical performance, potential mediators/moderators of treatment efficacy, biomarkers of inflammation and metabolic regulation, health care utilization, cost, and overall health. Four shared resources support AMPLIFI RCTs: 1) Administrative; 2) Adaptation, Dissemination and Implementation; 3) Recruitment and Retention; and 4) Assessment and Analysis.

DISCUSSION

Representing a new generation of RCTs, AMPLIFI will exclusively use remote technologies to recruit, intervene and assess the efficacy of the newly-adapted, web-based diet and exercise interventions and determine whether sequential or combined delivery works best for at-risk (older, rural, racial minority) cancer survivors.

TRIAL REGISTRATION

ClinicalTrials.gov , NCT04000880 . Registered 27 June 2019.

The Impact of Income and Social Mobility on Colorectal Cancer Outcomes and Treatment: A Cross-sectional Study

OBJECTIVE

To determine the impact of income mobility on racial disparities in colorectal cancer.

BACKGROUND

There are well-documented disparities in colorectal cancer treatment and outcomes between Black and White patients. Socioeconomic status, insurance, and other patient-level factors have been shown important, but little has been done to show the discriminatory factors that lead to these outcomes.

METHODS

Data were obtained from the Surveillance Epidemiology and End-Results database for Black and White patients with colorectal cancer between 2005 and 2015. County level measures of Black (BIM) and White income mobility (WIM) were obtained from the Opportunity Atlas as a measure of intergenerational poverty and social mobility. Regression models were created to assess the relative risk of advanced stage at diagnosis (Stage IV), surgery for localized disease (Stage I/II), and cancer-specific mortality.

RESULTS

There was no significant association of BIM or WIM on advanced stage at diagnosis in Black or White patients. An increase of $10,000 of BIM was associated with a 9% decrease in hazards of death for both Black (hazard ratio 0.91, 95% confidence interval 0.86,0.95) and White (0.91, 95%CI 0.90,0.93) patients, while the same increase in WIM was associated with no significant difference in hazards among Black patients (hazard ratio 0.99, 95% confidence interval 0.97,1.02). There were no predicted racial differences in hazards of death at high levels of BIM.

CONCLUSIONS

Increased Black income mobility significantly improves survival for both Black and White patients. Interventions aimed at increasing economic and social mobility could significantly decrease mortality in both Black and White patients while alleviating disparities in outcomes.

Factors associated with loss to follow-up after radiation therapy for head and neck cancer

OBJECTIVES

Head and neck cancer (HNC) patients are at high risk for late occurring radiation-related morbidity and recurrence, necessitating close long-term medical surveillance. This study identified factors associated with becoming lost to follow-up (LTFU) at a comprehensive cancer center.

MATERIALS AND METHODS

Patients were drawn from survivors who received radiation for HNC at a single institution between 2001 and 2018. LTFU was defined as living patients without a clinical encounter within 2 years of the data query.

RESULTS

In total, 537 patients met the inclusion criteria and 57 (10.6%) were identified as LTFU. Individual comparisons identified time since completing radiation, non-White race and being unmarried as associated with LTFU. Multiple regression identified time since treatment and being unmarried as factors associated with LTFU. A decision tree correctly sorted 89.4% using time, distance, and marital status.

CONCLUSION

Time since radiation, distance to clinic, and being unmarried were factors associated with becoming LTFU.

Outcomes in Laparoscopic Roux-en-Y Gastric Bypass and Implications for Surgical Resident Education

OBJECTIVES

Some surgery residents feel inadequately prepared to perform advanced operations, partly due to losing operative opportunities to fellows. In turn, they are prompted to pursue fellowships. Allowing residents the opportunity to participate in advanced procedures and complex cases may alleviate this cycle, if their participation is safe. This study examined the effects of resident participation in laparoscopic Roux-en-Y gastric bypass procedures (LRYGBs).

DESIGN

Our MBSAQIP database was used to identify LRYGBs performed at our institution between 2015 and 2018. Operative notes were reviewed to determine training level of the assistant. Patient comorbidities and outcomes (duration of surgery, length of stay, post-operative complications, readmissions, and reoperations) were stratified by assistant level of training for comparison.

SETTING

Urban tertiary care hospital.

PARTICIPANTS

Trainees and attending surgeons acting as assistants during LRYGBs.

RESULTS

Among 987 total cases, the assistants for the procedures were chief residents (n = 549, 56%), fourth-year residents (n = 258, 26%), attending surgeons (n = 143, 14%), and third-year residents (n = 37, 4%). Attending surgeons assisted more often when patients had a BMI ≥ 45 (38% attendings vs. 25% residents, p = 0.007), ≥ 2 comorbidities (54% vs. 40%, p = 0.007), or had a history of prior bariatric surgery (22% vs. 3%, p < 0.0001).Post-operative complication rate was low (4%) and did not differ significantly between all training levels (p = 0.86). Average length of stay, readmission rates, and reoperation rates were not significantly different across training levels (p = 0.75, p = 0.072, and p = 0.91 respectively).

CONCLUSION

Complication rates, hospital length of stay, readmission rates, and reoperation rates were equivalent for patients regardless of the level of training of the assistant for LRYGBs. Involving residents in complex bariatric procedures such as LRYGB is a safe model of education that does not compromise patient safety or hospital outcomes. Involvement in advanced cases allows general surgery residents to more confidently move toward independent practice.

Features That Middle-aged and Older Cancer Survivors Want in Web-Based Healthy Lifestyle Interventions: Qualitative Descriptive Study

Background

With the increasing number of older cancer survivors, it is imperative to optimize the reach of interventions that promote healthy lifestyles. Web-based delivery holds promise for increasing the reach of such interventions with the rapid increase in internet use among older adults. However, few studies have explored the views of middle-aged and older cancer survivors on this approach and potential variations in these views by gender or rural and urban residence.

Objective

The aim of this study was to explore the views of middle-aged and older cancer survivors regarding the features of web-based healthy lifestyle programs to inform the development of a web-based diet and exercise intervention.

Methods

Using a qualitative descriptive approach, we conducted 10 focus groups with 57 cancer survivors recruited from hospital cancer registries in 1 southeastern US state. Data were analyzed using inductive thematic and content analyses with NVivo (version 12.5, QSR International).

Results

A total of 29 male and 28 female urban and rural dwelling Black and White survivors, with a mean age of 65 (SD 8.27) years, shared their views about a web-based healthy lifestyle program for cancer survivors. Five themes emerged related to program content, design, delivery, participation, technology training, and receiving feedback. Cancer survivors felt that web-based healthy lifestyle programs for cancer survivors must deliver credible, high-quality, and individually tailored information, as recommended by health care professionals or content experts. Urban survivors were more concerned about information reliability, whereas women were more likely to trust physicians’ recommendations. Male and rural survivors wanted information to be tailored to the cancer type and age group. Privacy, usability, interaction frequency, and session length were important factors for engaging cancer survivors with a web-based program. Female and rural participants liked the interactive nature and visual appeal of the e-learning sessions. Learning from experts, an attractive design, flexible schedule, and opportunity to interact with other cancer survivors in Facebook closed groups emerged as factors promoting program participation. Low computer literacy, lack of experience with web program features, and concerns about Facebook group privacy were important concerns influencing cancer survivors’ potential participation. Participants noted the importance of technology training, preferring individualized help to standardized computer classes. More rural cancer survivors acknowledged the need to learn how to use computers. The receipt of regular feedback about progress was noted as encouragement toward goal achievement, whereas women were particularly interested in receiving immediate feedback to stay motivated.

Conclusions

Important considerations for designing web-based healthy lifestyle interventions for middle-aged and older cancer survivors include program quality, participants’ privacy, ease of use, attractive design, and the prominent role of health care providers and content experts. Cancer survivors’ preferences based on gender and residence should be considered to promote program participation.

National Disparities in COVID-19 Outcomes between Black and White Americans

BACKGROUND

There is very limited comprehensive information on disparate outcomes of black and white patients with COVID-19 infection. Reports from cities and states have suggested a discordant impact on black Americans, but no nationwide study has yet been performed. We sought to understand the differential outcomes for black and white Americans infected with COVID-19.

METHODS

We obtained case-level data from the Centers for Disease Control and Prevention on 76,442 white and 48,338 non-Hispanic Black patients diagnosed with COVID-19, ages 0 to >80+, outlining information on hospitalization, ICU admission, ventilation, and death outcomes. Multivariate Poisson regressions were used to estimate the association of race, treating white as the reference group, controlling for sex, age group, and the presence of comorbidities.

RESULTS

Black patients were generally younger than white, were more often female, and had larger numbers of comorbidities. Compared to white patients with COVID-19, black patients had 1.4 times the risk of hospitalization (RR 1.42, p < 0.001), and almost twice the risk of requiring ICU care (RR 1.68, p < 0.001) or ventilatory support (RR 1.81, p < 0.001) after adjusting for covariates. Black patients saw a 1.36 times increased risk of death (RR 1.36, p < 0.001) compared to white. Disparities between black and white outcomes increased with advanced age.

CONCLUSION

Despite the initial descriptions of COVID-19 being a disease that affects all individuals, regardless of station, our data demonstrate the differential racial effects in the United States. This current pandemic reinforces the need to assess the unequal effects of crises on disadvantaged populations to promote population health.

Utilizing Data Visualization to Identify Survival and Treatment Differences Between Women With De Novo and Recurrent Metastatic Breast Cancer

INTRODUCTION

De novo stage IV metastatic breast cancer (MBC) and recurrent MBC are considered the same when determining guideline-based care, but differences in treatment patterns exist. Data visualization can be used to understand these differences and optimize treatment delivery.

PATIENTS AND METHODS

This retrospective study evaluated treatment patterns for de novo and recurrent MBC using the American Society of Clinical Oncology's CancerLinQ Discovery database. Spatiotemporal graphics depicting treatment data were generated for MBC subtype and stratified by de novo and recurrent MBC. Descriptive statistics for categorical and continuous variables were calculated.

RESULTS

We identified 1668 patients diagnosed and treated for MBC: 391 patients with HER2⁺ MBC, 767 patients with HR⁺/HER2- MBC, and 510 patients with triple-negative MBC. Median survival from MBC diagnosis for patients with de novo MBC and recurrent MBC was 1.4 years (interquartile range, 0.6-3.0) and 1.8 years (interquartile range, 0.7-4.5), respectively. Both patients with de novo and recurrent HER2⁺ MBC were often treated with continuous HER2-targeted therapy. Patients with de novo HR⁺/HER2- MBC often received chemotherapy followed by hormone therapy. This contrasted with treatment trends observed among patients with recurrent HR⁺/HER2- MBC who, after receiving chemotherapy, seldom went on to receive hormone therapy. Patients diagnosed with triple-negative MBC displayed less heterogeneous treatment trends.

CONCLUSION

There are observable differences in survival and practice patterns between de novo and recurrent MBC. Visualization techniques are effective in assessing large databases and could give researchers and clinicians a clearer understanding of how survival differs by disease subtype, diagnosis status, and practice patterns.

Depression among older adults with gastrointestinal malignancies

BACKGROUND

Depression among older adults with cancer is under recognized and under treated. This study characterizes the burden of depression in older adults with gastrointestinal (GI) malignancies prior to chemotherapy and its relationship with geriatric assessment (GA) domains, health-related quality of life (HRQOL), and self-reported healthcare utilization.

METHODS

Patients ≥60 years in GI oncology clinics at UAB were asked to complete a GA entitled the Cancer & Aging Resilience Evaluation (CARE). We examined depression using the Patient-Reported Outcomes Measurement Information System (PROMIS®) Depression four-item short form; moderate/severe depression was defined by a t-score ≥ 60. Multivariate analysis was used to examine associations between those with and without moderate/severe depression.

RESULTS

Of 355 included patients, 46 had mild depression (13%) and an additional 46 patients had moderate/severe depression (13%). After adjustment for age, sex, education, cancer type, and cancer stage, those who reported moderate/severe depression had a significantly increased odds of reporting falls (adjusted odds ratio [aOR] 4.01, 95% confidence interval [CI] 1.94-8.26), dependence in IADLs (aOR 7.06,CI 2.91-17.1), dependence in ADLs (aOR 6.23, CI 2.89-13.4), malnutrition (aOR 5.86, CI 2.40-14.3), frailty (aOR 13.7, CI 5.80-32.1), and fatigue (aOR 11.2, CI 3.31-37.6). Moderate/severe depression was also significantly associated with worse physical (aOR 7.58, CI 3.30-17.4) and mental (aOR 26.3, CI 10.1-68.8) HRQOL sub-scores, without significant differences in healthcare utilization.

CONCLUSIONS

More than one out of eight older adults with a GI malignancy reported moderate/severe depression prior to chemotherapy, which was associated with impairments in several GA domains and HRQOL.

Intersectional Disparities Among Hispanic Groups in COVID-19 Outcomes

Previous geographically limited studies have shown differential impact of COVID-19 on Hispanic individuals. Data were obtained from the Centers for Disease Control and Prevention. We performed multivariate Poisson regression assessing risk of hospitalization and death in Hispanic White (HW), Hispanic Black (HB), and Hispanic Multiracial/Other (HM) groups compared to non-Hispanic Whites (NHW). The relative risk of hospitalization was 1.35, 1.58, and 1.50 (p < 0.001) for HW, HB, and HM individuals respectively when compared to NHW. Relative risk of death was 1.36, 1.72, 1.68 (p < 0.001) times higher in HW, HB, and HM compared to NHW. HW, HB, and HM individuals also had significantly increased risk of requiring mechanical ventilation and ICU admission when compared to NHW. Hispanic individuals are more likely to be hospitalized and die from COVID-19 infection than White, which underscores the need for more precise data and policies aimed at unique Hispanic groups to decrease disparities.

Health plan expenditures young adults with newly-diagnosed Hodgkin lymphoma (HL) by care at NCI-designated comprehensive cancer centers (CCC) vs. other treatment sites (non-CCC)

7081

Background: Patients diagnosed with HL between 22-39y have worse outcomes than younger patients (≤21y); we previously reported that treatment at a CCC mitigates these disparities [Wolfson, Leukemia 2017]. While there is general consensus that CCC care is expensive, expenditures for managing young adults with HL in CCC vs. non-CCC are not known. Methods: Cancer-related expenditures were examined in HL patients diagnosed between 2001-2014 at age 22-39y and treated at CCC and non-CCC sites using commercial insurance claims data (OptumLabs Data Warehouse). Multivariable generalized linear models with log link modeled average monthly health plan paid expenditures, adjusting for sociodemographics, stage, adverse events, pre-existing comorbidities, and diagnostic era. Results: Of the 1501 HL patients, 33% (n = 489) were treated at a CCC. Patients treated at CCC vs. non-CCC did not differ with respect to race, sex, income, diagnostic era or comorbidities (p≥0.3). Mean duration of enrollment was longer in CCC than non-CCC (25 vs. 23 mos; p < 0.001) patients. During the first year after HL diagnosis, total average monthly expenditures were higher in CCC ($9,111) than non-CCC ($7,834, p = 0.001), including those related to inpatient (CCC: $1,790 vs. non-CCC: $1,011; p = 0.001) and outpatient (CCC: $6,971 vs. non-CCC: $6,487; p = 0.001) expenditures. The higher CCC expenditures were associated with higher monthly rates of inpatient admissions (IRR = 1.3, p = 0.001) and outpatient visits (IRR = 1.1, p = 0.02) at CCC. Rates of chemotherapy-related inpatient admissions were higher (IRR = 2.3, p = 0.001) in CCC than non-CCC patients, while outpatient chemotherapy visit rates were lower (IRR = 0.9, p = 0.001) in CCC. During Years 2-3, total average monthly expenditures were higher in CCC ($19,259) than non-CCC ($4,145, p = 0.002) patients. Outpatient expenditures were higher in CCC ($10,164) vs. non-CCC ($2,901, p = 0.001), with higher monthly outpatient visit rates (IRR = 1.7, p = 0.001) at CCC. Conclusions: Inpatient and outpatient cancer-related expenditures in young adults with HL were higher at CCC than non-CCCs. Higher outpatient expenditures at CCC were associated with only higher monthly visit rates. Higher inpatient expenditures were in the setting of higher admission rates, including those related to chemotherapy. Additional work is necessary to understand whether these higher expenditures at CCC are related to supportive care and/or differences in facility structure and billing practices.

Geriatric assessment (GA) predictors of 1y mortality in older adults with gastrointestinal (GI) malignancies: Results from the CARE study

12047

Background: Chronologic age is an imperfect predictor of morbidity and mortality in older patients with newly-diagnosed GI malignancies. Identifying patients with GI malignancies that are at increased risk of mortality within the 1st year remains challenging given no prior studies have focused on this population, yet is critical to developing personalized treatment plans. To fill this gap, we examined predictors of 1y mortality using variables from a patient-reported GA in a prospective cohort of older adults with GI malignancies. Methods: Cancer and Aging Resilience Evaluation (CARE) is a prospective registry of older adults (≥60y) with cancer seen at UAB (J Geri Onc 2019; PMID 31005648). Patients with GI malignancies with GA completed within the timeframe of 3 mo. before and up to 6 mo. after diagnosis were included. Vital status (up to 12/7/2019) was ascertained by linking participants to LexisNexis. Multivariable Cox regression analysis was used to estimate associations between GA variables and 1y mortality, adjusting for age at cancer diagnosis, race, cancer stage (IV vs. I-III), cancer group (high risk: pancreatic, hepatobiliary, esophageal vs. low risk: colorectal, GIST, neuroendocrine, etc.), and planned chemotherapy (yes/no). Results: A total of 356 participants met eligibility criteria. Mean age at enrollment was 70y; 56.4% were females; 25% black; 47.1% had high-risk cancers. In unadjusted analysis, high-risk cancers, cancer stage, malnutrition, impaired performance status, limitations in social activities, impaired instrumental activities of daily living (IADL), physical health, mental health, anxiety, and ≥3 comorbidities were associated with higher 1y mortality. Our base model (demographic and clinical variables) demonstrated good discrimination (c statistic 0.758), but was improved with the addition of all significant GA variables (c-statistic 0.810). Fatigue and malnutrition were identified as the strongest predictors among the GA variables, and a model adding those to the base model retained high discrimination (c-statistic 0.804). The estimated 1yr survival was 53.1% for those with both fatigue and malnutrition compared to 88.1% in those with neither. Conclusions: Among older adults with GI malignancies, malnutrition and fatigue were the strongest GA predictors of 1yr mortality after adjusting for age and clinical factors. These findings provide evidence for developing targeted interventions in older patients with newly-diagnosed GI malignancies to reduce 1y mortality.

Use of self-rated health to identify frailty and predict mortality in older adults with cancer. Results from the care study

12046

Background: Poor self-rated health (SRH) is a known predictor of mortality in the general adult population, but little is known about its use in older adults with cancer. The purpose of this study was to examine the association and ability of SRH to identify frail older adults and assess its ability to predict mortality in older adults with cancer. Methods: Using participants from the Cancer & Aging Resilience Evaluation (CARE) Registry who had undergone a geriatric assessment, we examined SRH using a single-item from the Patient-Reported Outcomes Measurement Information System (PROMIS) global health scale. SRH scores were dichotomized into Poor (poor and fair) and Good (good, very good, and excellent). Multivariable logistic regression analyses were used to examine associations between SRH and frailty (based on frailty index) and specific geriatric impairments adjusting for age, sex, comorbidity, cancer type and stage. Finally, the impact of SRH on all-cause mortality was assessed with a multivariable cox regression model. Results: A total of 708 participants with malignancy were included, median age was 68y, 41.5% male, and 74.6% White. Colorectal cancer was the most common cancer (27.1%) and 48.2% of the participants had Stage IV disease. Poor SRH was reported by 42% of participants and was associated with significantly higher odds of frailty (adjusted Odds Ratio [aOR] = 21.8; 95%CI 13.7-34.8). Similarly, poor SRH was independently associated with higher odds of impairments in Activities of Daily Living (ADL) (aOR = 5.6, 95%CI, 3.6-8.9), independent ADL (aOR = 8.4, 95%CI, 5.8-12.4), cognition (aOR = 4.6, 95%CI 2.3-9.3), malnutrition (aOR = 4.5, 95%CI 3.2-6.4), falls (aOR = 3.6, 95%CI 2.4-5.4), anxiety (aOR = 4.6, 95%CI 2.9-7.3), and depression (aOR = 5.4, 95%CI 3.0-9.7). The SRH demonstrated high sensitivity (84.3%) and specificity (78.4%) for identifying frailty, with a positive predictive value of 67% and negative predictive value of 90.6%. The 1y survival rate in those with Poor SRH was significantly worse (64.7% vs 84.3%, log rank p value < 0.001). In a multivariate cox regression analysis, poor SRH remained an independent predictor of worse survival (adjusted Hazard Ratio 2.29 [1.6-3.2], p< 0.01) after adjusting for age, sex, race, cancer type, stage, comorbidity, and planned treatment. Conclusions: Poor SRH is highly associated with frailty and could be a simple tool to identify frail older patients with cancer at risk for adverse events and increased mortality.

Association between chronological age and geriatric assessment (GA) to identify deficits in elderly adults with cancer: Findings from the Care Registry

12048

Background: Although ASCO and NCCN guidelines recommend that adults with cancer diagnosed at age ≥65y undergo a GA, the association between chronologic age and GA identified deficits remains understudied, and thus, the appropriate age cut-off for employing GA in clinical settings remains unknown. We addressed this gap by examining the association between chronologic age and GA deficits in older adults with cancer. Methods: The Cancer and Aging Resilience Evaluation (CARE) is an ongoing prospective registry of older adults (≥60y) with cancer at a single site. Eligible patients underwent a patient-reported GA adapted from the Cancer and Aging Research Group. The association between age categories (10y increments) and presence of GA deficits was tested using chi-squared tests of trend. Linear association between age and GA deficits was examined using Pearson correlation. Results: The median age at enrollment was 70y (60-96) for 08 participants; 58% were male. Most common cancer types were colorectal (27%), pancreatic (17%), and hepatobiliary (12%). No significant correlation was found between chronologic age and the number of GA deficits (r = 0.03). There was no association between the youngest (60-70y) vs. the oldest age groups (≥80y) with respect to the prevalence of GA deficits: frailty (33% vs. 33%, p= 0.97); impairment of activities of daily living (ADL) (20% vs. 16%, p= 0.7);impairment of instrumental ADL (50% vs 60%, p= 0.3); malnutrition (42% vs. 33%, p =0.4), cognitive impairment (8% vs. 6%, p= 0.6), falls (19% vs. 30%, p 0.1), anxiety (19% vs. 11%, p= 0.1) and depression (13.4% vs. 13.7%, p= 0.2) (Table). Prevalence of 3+ comorbidities was higher in the older patients (45% vs. 59%, p= 0.03). Conclusions: In our cohort of older adults with mostly gastrointestinal malignancies, age was not associated with GA identified deficits and the prevalence of most impairments was similar across age-groups. The use of chronologic age alone to identify which patients may benefit from GA is problematic, and adults 60yrs and above, or perhaps even younger, may derive benefits from a GA. [Table: see text]

Hospice use at end of life in children with cancer: The effect of insurance

10558

Background: The National Quality Forum has endorsed hospice care as a metric of high quality end-of-life (EOL) care for adults with cancer. Specific hospice-related quality metrics include hospice enrollment, hospice enrollment for ≥3d, and death outside of the acute care setting. These metrics have been examined extensively in adults and disparities related to a number of clinical and sociodemographic factors, including insurance, have been identified. However, for children with cancer, hospice utilization data is lacking. We addressed this gap by examining location of death and hospice utilization at EOL for children with cancer and determining whether these metrics varied with insurance status. Methods: We used national insurance claims data (Truven) to conduct a population-based analysis of patients with cancer who died between 2011 and 2017 at age 0-21y. The dataset was queried for hospice claims, inpatient claims, and location of death. The association between insurance (private vs. Medicaid) and 1) location of death, 2) hospice enrollment, and 3) days between first hospice claim and death was examined using multivariable regression analysis, adjusting for age at death, gender, and cancer diagnosis (hematologic malignancies vs. solid tumor). Results: A total of 1,492 children died at a mean age of 11y (SD: 6y); 56% were privately-insured, 56% were male, and 30% had hematologic malignancies. Overall, 58% died in the hospital (privately-insured: 54% vs. Medicaid: 63%). Forty-five percent enrolled in hospice (privately-insured: 46% vs. Medicaid: 43%) with 2% returning to the hospital to die after enrolling in hospice. The average time between first hospice claim and death was 3d (privately-insured: 10d vs. Medicaid: 2d, p = < 0.001). When compared to privately-insured children, children on Medicaid had similar likelihood of hospice enrollment (RR = 1.0, 95%Cl = 0.6-1.8). However, children on Medicaid were more likely to die in the hospital (RR = 1.3, 95%CI = 1.1-1.4) and have fewer days between hospice enrollment and death if enrolled in hospice (IRR: 0.5, 95%CI = 0.3-0.8). Conclusions: In this first study to examine national hospice utilization in children with cancer, care varies significantly with insurance status. Children on Medicaid are more likely to die in the hospital and have shorter hospice enrollment duration than children with private insurance. Whether this variation represents EOL care preferences, provider biases, differences in quality and availability of hospice or home care to different insurers, or other barriers needs to be examined.

Association of geriatric conditions with survival and health care use in older adults with colon cancer living in long-term care facilities

12045

Background: Older adults with colon cancer residing in nursing homes are at risk for experiencing geriatric conditions such as cognitive decline, limitations in activities of daily living (ADLs), needing pain medications, and incontinence, due to cancer and its treatment. We sought to investigate these factors pre- and post-diagnosis and explored their relationship with health care use and survival. Methods: We identified 483 patients age 65+ with colon cancer from 2011-2015 in SEER-Medicare with linked quarterly nursing home assessments from the Minimum Data Set both pre- and post-cancer diagnosis. We determined the number of geriatric conditions (cognitive functioning, limitation in any ADL, pain medication use, bowel/urinary incontinence) at the pre- and post-cancer diagnosis assessment. We created four groups based on changes in these factors from pre- to post- assessment: improved (n = 105), worsened (n = 25), remained limited (n = 240), never limited (n = 113). Regression models estimated how changes from pre- to post-cancer diagnosis were associated with number of emergency department (ED) visits, hospitalizations, and survival, adjusted for age, sex, race/ethnicity, insurance status, cancer stage, number of pre-cancer comorbidities, urban/rural status, and time from diagnosis. Results: Overall, 55.3% of patients were age > 80 at diagnosis, with 64.8% female; 73.3% non-Hispanic white; and 9.9% Stage IV. Pre- versus post-diagnosis, 20.7% vs. 34.8% of patients were limited in cognitive functioning, and 75.4% vs. 77.8% were limited in ADLs. About a third of patients required pain medication, and about half of patients had urinary incontinence, which did not change pre- and post-diagnosis. Patients who remained limited had higher rates of ED visits (Risk ratio [RR] 1.05, p < .01) compared to those never limited. Those who worsened had higher rate of hospitalization (RR 1.44, p < .01) and ED visits (RR 1.63, p < .01). 12-month and 5-year survival was 46.7% and 6.1%, respectively. Factors associated with worse survival in a multivariable model included: remaining limited at both assessments (OR 1.52, p < .01), worsening from prior (OR 2.00, p = .01), as well as older age and higher cancer stage. Conclusions: Older adults with colon cancer residing in nursing homes have high prevalence of geriatric conditions and differential health care use and survival based on the presence of geriatric conditions, highlighting the need to consider geriatric conditions when providing cancer care to this population.

Depression among older adults with gastrointestinal (GI) malignancies: Results from the Cancer and Aging Resilience Evaluation (CARE) registry

92

Background: Depression among older adults with cancer is often under recognized and under treated. This study characterizes the burden of depression in older adults with GI malignancies and its relationship with geriatric assessment (GA) impairments, health-related quality of life (HRQOL), and healthcare utilization. Methods: Since September 2017, patients ≥60 years in GI oncology clinics at UAB were asked to complete a CARE GA questionnaire. We examined depression using the Patient-Reported Outcomes Measurement Information System (PROMIS) Depression 4 item short form; moderate/severe (mod/sev) depression was defined by a T-score ≥60. GA impairments, HRQOL (PROMIS Global-10), and healthcare utilization were compared between those with and without mod/sev depression using a Chi-squared or t-test. Results: A total of 462 enrolled; mean age was 69 (range 60-96), majority white (72%), males (52%), and with advanced stage (III/IV, 69.1%). Most common cancers were colorectal (36%), pancreatic (23%), or hepatobiliary (15%). Overall, 55 patients reported mild depressive symptoms (12%) and 60 (13%) were found to have mod/sev depression. Depressed participants did not differ in demographics or GI cancer type/stage. Those reporting mod/sev depression were more likely to report falls (44 v 19%, p<.001), impaired performance status (63 v 27%, p<.001), dependence in activities of daily living (ADL; 45 v 14%, p<.001), dependence in instrumental ADL (84 v 45%, p<.001), cognitive dysfunction (40 v 5%, p<.001), financial distress (37 v 23%, p=.03), anxiety (76 v 10%, p<.001), fatigue (88 v 54%, p<.001), and pain (70 v 32%, p<.001). Depressed patients reported lower physical (32 v 45%, p<.001) and mental (36 v 50%, p<.001) HRQOL sub-scores, as well as more emergency room visits (67 v 49%, p=.009), but no difference in hospitalizations. Conclusions: More than one out of eight older adults with a GI malignancy reported mod/sev depression, which was associated with impairment in several geriatric domains and overall quality of life. As depression is a pleiotropic yet treatable comorbidity, oncologists should prioritize its screening and treatment.

Utilizing visualization to qualitatively evaluate electronic health record-derived database limitations

317

Background: Electronic health record (EHR) databases are a promising platform for clinical research using real-world data. However, information on potential limitations of these data sources is lacking. We sought to understand how data visualization might be used to identify data inconsistencies and the applicability of previously validated claims-based algorithms used to identify patients with metastatic breast cancer (MBC). Methods: This retrospective study utilized ASCO’s CancerLinQ Discovery database derived from EHR data. Subjects included women ≥18 years treated for MBC diagnosed ≥1980. Subjects with MBC were identified using two billing codes for metastasis on separate dates following primary breast cancer diagnosis. Treatment course sequences were visualized. Patients were represented by a horizontal bar on the Y-axis. Treatments were displayed using colored bars (blue: chemotherapy, red: endocrine therapy, green: HER2 targeted, orange: novel therapy) with time of treatment on the X-axis. Visualizations were qualitatively evaluated, and treatment patterns inconsistent with clinical practice were identified. Results: We identified 4,760 women treated for MBC using billing codes for primary breast cancer diagnosis and distant metastasis. Most patients (96%) had a primary breast cancer diagnosed in 2000 later. Treatment patterns inconsistent with clinical practice identified using the visualization technique included: 1% of patients received adjuvant chemotherapy continuously for ≥1.5 years, suggesting missed coding for metastatic disease; 5% of patients did not receive any treatment in the year following metastasis, suggesting the billing code may have been used in workup and not for confirmed metastatic disease. Among patients with MBC, 50% identified as HR+ across all records had not received hormone therapy, while 39% identified as HR- across all records received hormone therapy. Conclusions: Because previously validated algorithms may not translate well to EHR databases, quality auditing should always be performed. The proposed data visualization can be used for improving algorithms, qualitatively identifying errors, and avoiding biased or inaccurate results.

Health insurance literacy, status, and financial toxicity in women receiving treatment for metastatic breast cancer

96

Background: Though uninsurance rates declined after the Affordable Care Act, the number of underinsured, or individuals who spend > 10% of their income on out-of-pocket (OOP) medical costs, continues to rise. In patients with metastatic breast cancer (MBC), underinsurance may lead to financial toxicity (FT), or patient-level financial burden and distress, since diagnosis and treatment is extremely costly. This study explores health insurance literacy and the association between FT and health insurance status in women receiving treatment for MBC. Methods: This cross-sectional study utilized survey data collected from 2017-2019 in women age ≥18 receiving treatment for MBC at two academic medical centers in Alabama. FT was measured by the Comprehensive Score for Financial Toxicity (COST) tool (11-item scale from 0-44, with lower scores indicating worse FT). Health insurance status and OOP costs were self-reported. Effect sizes were calculated using Cohen’s d or Cramer’s V. Mixed and generalized linear models clustered by site and treating medical oncologist estimated the association between FT and health insurance status. Results: In 81 women with MBC, median COST score was 24 (interquartile range [IQR] 17-30), 44% had private insurance, 40% Medicare, and 16% Medicaid. Though 25% and 33% of surveyed patients did not know their health insurance premium or deductible cost, respectively, privately insured patients more often knew the cost of their premiums (97%; V = 0.58) and deductibles (81%; V = 0.33) compared to publicly insured patients. In adjusted models, FT levels did not differ significantly based on health insurance type (private insurance COST 21, 95% confidence interval [CI] 18-25; Medicaid COST 23, 95% CI 17-29; Medicare COST 24, 95% CI 20-27). However, risk of severe FT (COST ≤13) was 147% higher for privately insured patients versus Medicare beneficiaries (risk ratio 2.47, 95% CI 1.44-4.21). Conclusions: Despite higher levels of health insurance literacy, privately insured patients receiving treatment for MBC may be at increased risk of severe FT. Further research is needed to understand causes of underinsurance in patients with MBC, which could lead to cancer-related FT.

Visualizing treatment patterns and survival in metastatic breast cancer

316

Background: Optimal treatment sequencing (i.e., the order in which drugs are given) for metastatic breast cancer (MBC) is unknown. We aimed to develop an approach to visualize treatment patterns and survival in MBC. Methods: This retrospective study utilized ASCO’s CancerLinQ Discovery® database generated from electronic health records. Subjects included 3,312 women aged ≥18 years who were diagnosed with and received treatment for MBC after 1980. Hormone receptor (HR) status was determined by concordant diagnosis and treatment records. Human epidermal growth factor (HER2) status was determined by delivery of HER2-targeted therapy. Ordered and administered treatments were included. We created spatiotemporal plots of treatment patterns for HR+/HER2-, HER2+, and triple negative (TN) MBC. Individuals were represented on the Y-axis, and time on the X-axis with development of MBC aligned at time 0. Treatment classes were identified by colors: hormone therapies in shades of red, chemotherapies in shades of blue, HER2-targeted therapies in shades of green, and novel therapies in shades of orange. Concurrent treatments were represented by split bars. An overlaid Kaplan-Meier curve allowed for observations about the relationship between survival and treatment. Results: We developed a novel visualization approach to simultaneously display heterogeneous, longitudinal treatments and survival. Median survival after first documentation of MBC was 3.1 (IQR 1.4-7.2), 1.3 (IQR 0.6-2.8), and 2.6 (IQR 1.0-5.2) years for HR+/HER2-, TN, and HER2+ MBC, respectively. Patients with longer survival often had long duration of initial therapy, suggesting a more indolent or responsive disease. Substantial heterogeneity in treatment sequencing was observed for HR+HER2- and TN cohorts. In the HER2+ cohort, HER2-targeted therapy was commonly administered for the duration of treatment with more homogeneous sequencing. Conclusions: This novel visualization approach allows for observing the relationship between treatment patterns and survival, which is challenging to demonstrate with traditional quantitative methods. This approach can generate hypotheses regarding impact of treatment patterns on survival.

Impact of travel time on healthcare costs and resource utilization by phase of care for older cancer patients

6560

Background: Closures of hospitals and clinics may have unintended consequences, including increasing patient travel time. Increased patient travel time to healthcare facilities has the potential to adversely impact patient outcomes. Limited data exist on the impact of travel time on healthcare costs and resource utilization. Methods: This retrospective cohort study from 2012-2015 evaluated drive time to cancer care site for Medicare beneficiaries age ≥ 65 in the Southeastern US. The primary outcome was Medicare spending by phase of care (initial, survivorship, end of life [EOL]). Secondary outcomes included resource utilization measured by hospitalization rates, hospitalization sites, intensive care unit (ICU) admissions, and chemotherapy-related hospitalization rates. Hierarchical linear models with patients clustered within cancer care site and adjusted for pertinent covariates were used to determine the effects of drive time on average monthly phase-specific Medicare spending. Results: Median drive time was 32 minutes (IQR 18-59) for the 23,382 included Medicare beneficiaries, with 24% of patients driving > 1 hour to their cancer care site. During the initial phase of care, Medicare spending was 14% higher for patients traveling > 1 hour than those traveling ≤ 30 minutes.Hospitalizationrates were 4-13% higher for patients traveling > 1 hour vs. ≤ 30 minutes in the initial (61 vs. 54), survivorship (27 vs. 26), and EOL (310 vs. 86) phases of care (all p < .05). The majority of patients traveling > 1 hour were hospitalized at a local hospital rather than at their cancer care site, whereas the converse was true for patients traveling ≤ 30 minutes. Conclusions: As healthcare locations close, patients living farther from treatment sites may experience more limited access to care, and healthcare spending could increase for Medicare.

Cumulative burden of new-onset chronic health conditions (CHCs) among older cancer survivors

11528

Background: In the US there are an estimated 11 million survivors of cancer diagnosed at ≥65y of age. Description of the morbidity in these survivors has been limited to single complications or to prevalence of comorbidities. The cumulative burden of CHCs remains unstudied, and is critically needed to inform healthcare delivery in this burgeoning population. Methods: Using SEER-Medicare, we identified 300,082 patients with breast (34%), prostate (33%), colorectal (16%), non-small cell lung (NSCLC 10%) or non-Hodgkin lymphoma (NHL 7%) diagnosed between 2000 and 2011 at age > 65y (mean age at diagnosis: 75y; 47% males, 88% non-Hispanic whites). An age-, race-, and sex-similar non-cancer cohort (n = 97,842) was assembled. New-onset non-malignant health conditions (n = 109) were consolidated into 10 organ-specific CHCs. Inpatient CHC visits were used to describe severe CHCs. The cumulative incidence (CI) and cumulative burden (CB) of CHCs was described up to 10y from cancer diagnosis and by attained age – up to six months prior to death or until 12/31/2013. Subsequent malignant neoplasms (SMNs) were described 10y from primary cancer diagnosis. Results: The 10y CI of any CHC and severe CHC was 98% (95%CI 98-99%) and 73% (72-73%) in cancer patients and 92% (91-92%) and 55% (54-55%) in non-cancer controls (hazard ratio [HR10y]: 1.65, 95%CI 1.64-1.66). Cardiovascular conditions were the largest contributor to severe non-malignant CHCs (10y CI: 49%-69%). Prostate cancer survivors had the highest 10y CI for SMNs (19.4%). The CI for severe CHCs was 44% by age 80y and 85% by age 90y, compared to 34% and 54% in controls (p < 0.001). The 10y CB of CHCs was highest among NSCLC (42 CHCs/survivor) and NHL (41 CHCs/survivor) survivors; in comparison, the 10y CB was 31 CHCs/individual in controls. Colorectal cancer survivors had greatest overall burden at age 80y (27 CHCs/survivor) and 90y (36 CHCs/survivor), compared to 13 and 16 CHCs/individual in controls. Conclusions: The cumulative burden of new-onset multimorbidity among older cancer survivors is substantially greater than that of their non-cancer counterparts, providing quantifiable evidence that survivor-adapted healthcare management policies and risk-based interventions are needed.

Accelerated sarcopenia and outcomes in older adults with cancer: The Health ABC Study

11526

Background: Progressive loss of muscle mass and strength (sarcopenia) is a well-known phenomenon of aging; however, little is known about the contribution of a cancer diagnosis to sarcopenia and its subsequent impact on disability. Using a prospective cohort of older adults from pre- to post-cancer diagnosis and a similarly-followed non-cancer cohort, we examined the trajectory of sarcopenia measures and their association with overall survival (OS) and major disability among those with cancer. Methods: The Health, Aging, and Body Composition (Health ABC) Study is a prospective longitudinal study where 3,075 community-dwelling older adults (70-79y) underwent 6 annual assessments of body composition and were followed for development of sentinel events (cancer, disability, death). Appendicular lean mass (ALM [kg]) was a sum of DXA-based lean tissue of all extremities. Hand grip strength (HGS [kg]) was averaged from 2 trials per hand. Gait speed (GS) was evaluated over a 20m course. We used linear mixed effect models to compare the change in ALM, HGS, and GS between individuals who subsequently developed cancer and those who did not, adjusting for age, race, gender, enrollment site. Among patients with cancer, we used multivariable cox regression for time from cancer diagnosis to mortality and major disability (cane/walker, inability to walk 0.25 mile/climb 10 steps, assistance with activities of daily living) treating sarcopenia measures as time-varying covariates. Results: Mean age at enrollment was 75y; 52% female; 42% black; 515 new cancers (prostate: 23%, colorectal: 15%, lung: 13%, breast: 11%). Compared with non-cancer controls, we found significantly steeper declines in HGS ( p= 0.03) and GS ( p< 0.001), and a trend in ALM ( p= 0.07) prior to cancer diagnosis; and a significantly steeper decline in ALM ( p< 0.001), but no difference in HGS ( p= 0.6) or GS ( p= 0.4) after cancer diagnosis. Slow GS was associated with a 44% increase in mortality ( p= 0.02) and a 70% increase in disability ( p= 0.02), but not ALM or HGS. Conclusions: Accelerated loss in sarcopenia measures both prior to and after a cancer diagnosis, and association with disability and mortality in older adults with cancer, present opportunities for targeted interventions.

Geriatric Assessment Predicts Hospitalization Frequency and Long-Term Care Use in Older Adult Cancer Survivors

PURPOSE

The association between geriatric assessment (GA)-identified impairments and long-term health care use in older cancer survivors remains unknown. Our objective was to evaluate whether a GA performed at cancer diagnosis was predictive of hospitalizations and long-term care (LTC) use in older adult cancer survivors.

METHODS

Older adults with GA performed between 3 months before through 6 months after diagnosis were included (N = 125). Patients with Medicare Parts A and B coverage and no managed care were identified. Hospitalizations and LTC use (skilled nursing or assisted living) were assessed up to 5 years postdiagnosis. GA risk measures were evaluated in separate Poisson models estimating the relative risk (RR) for hospital and LTC visits, adjusting for age and Charlson comorbidity score.

RESULTS

The mean age of patients was 74 years, and the majority were female (80%) and white (90%). Breast cancer (64%) and early-stage disease (stages 0 to III, 77%) were common. Prefrail/frail status (RR, 2.5; P < .001), instrumental activities of daily living impairment (RR, 5.47; P < .001), and limitations in climbing stairs (RR, 2.94; P < .001) were associated with increased hospitalizations. Prefrail/frail status (RR, 1.86; P < .007), instrumental activities of daily living impairment (RR, 4.58; P < .001), presence of falls (RR, 6.73; P < .001), prolonged Timed Up and Go (RR, 5.45; P < .001), and limitations in climbing stairs (RR, 1.89; P < .005) were associated with LTC use.

CONCLUSION

GA-identified impairments were associated with increased hospitalizations and LTC use among older adults with cancer. GA-focused interventions should be targeted toward high-risk patients to reduce long-term adverse health care use in this vulnerable population.

Integrating geriatric assessment into routine gastrointestinal (GI) consultation: The Cancer and Aging Resilience Evaluation (CARE)

667

Background: Integrating Geriatric Assessment (GA) in the management of older adults with cancer is recommended, yet rarely practiced in routine oncologic care. In this report, we describe the feasibility of integrating the routine incorporation of GA in the management of older adults with GI malignancies and characterize GA impairments. Methods: CARE was adapted from the Cancer and Aging Research Group GA with modifications to create a completely patient-reported version. The CARE assesses self-reported functional status, physical function, nutrition, social support, anxiety/depression, cognitive function, comorbidities, and social activities. Patients ≥ 60yo referred for consultation to the GI Oncology clinic were asked to complete the CARE (paper/pencil) on their first visit. The completed CARE was collected during nurse triage and submitted to the clinical team prior to the physician encounter. Feasibility was defined as completion of the CARE by ≥ 80% of eligible patients during the initial consultation. Results: Between September 2017 and August 2018, 199 eligible new patients attended the GI Oncology Clinic, 192 (96.5%) were approached, and 181 (90.4%) completed the CARE. Most patients (79.6%) felt the length of time to complete was appropriate (median time of 10 minutes [IQR 10-15 minutes]). The mean age was 70y (range 60-96), 54.3% were male, and 75.1% were non-Hispanic white. Common tumor types included colon (27.8%), pancreatic (21.2%), and rectal (10.2%) cancer; predominately advanced stage diseases (stage III: 26.9%; stage IV: 40.0%). GA impairments were prevalent: 48.6% reported dependence in Instrumental Activities of Daily Living, 18.0% reported dependence in Activities of Daily Living, 22.5% reported ≥ 1 fall, 29.4% reported a performance status ≥ 2, 51.3% were limited in walking one block, 75.7% reported polypharmacy (≥ 4 medications), and 84.3% had ≥ 1 comorbidity. Conclusions: Performing a GA in the routine care of older adults with GI malignancies is feasible, and GA impairments are common among older adults with GI malignancies. A fully patient-reported GA such as the CARE may facilitate broader incorporation of GA in the routine clinic work flow.

Financial distress amongst older adults with gastrointestinal (GI) malignancies

517

Background: Many patients with cancer report financial distress (FD); however, the magnitude of FD in the growing number of older adults with cancer remains less clear, particularly in those with GI malignancies. The purpose of this study was to evaluate the proportion of older adults with GI malignancies reporting FD and to characterize geriatric assessment (GA) and cancer-related factors associated with FD. Methods: Older adults ( ≥ 60yrs) seen in the GI oncology clinic at the University of Alabama Birmingham (UAB) were asked to fill out a patient-reported GA, entitled the Cancer & Aging Resilience Evaluation (CARE), at their visit. The CARE includes questions pertaining to patient’s independence in Activities of Daily Living (ADLs), Instrumental Activities of Daily Living (IADLs), falls, physical function, polypharmacy, and comorbidity. A single item question regarding FD from the patient satisfaction questionnaire (PSQ-18) was included. FD was defined as agreement with the phrase “Do you have to pay for more medical care than you can afford.” Demographic and GA characteristics were compared between those with and without FD using Chi-square and t-tests. Results: 233 patients completed the CARE a median of 71 days after diagnosis. Median age 68y (60-96); 54.5% male and 76.0% non-Hispanic white. Most common cancer types included colorectal (39.1%) and pancreatic cancers (20.6%). A total of 62 patients (26.6%) had FD. Patients with FD were more likely to be younger (68.1 vs. 70.1y, p = 0.04), of black race (37.1% vs. 15.8%, p = 0.007), have low education ( ≤ high school: 74.2% vs. 59.6%, p = 0.02), have one or more falls (31.5% vs. 19.9%, p = 0.077), to be limited a lot in walking 1 block (54.4% vs. 27.4%, p = 0.0003), take more than 4 medications (88.3% vs. 70.8%, p = 0.007), to have more than one comorbid condition (93.1% vs. 82.6%, p = 0.052), to report impaired IADLs (61.3% vs. 43.9%, p = 0.055), and impaired ADL (27.4% vs. 14.6%, p = 0.069). No associations were found with GI cancer type or stage, marital status, time from diagnosis, or hearing/vision impairments. Conclusions: Over a quarter of the older adult population with GI malignancies report FD. Several GA and demographic factors were associated with FD that may help identify older patients at risk for FD.

Impact of travel time on hospitalizations and patient cost responsibility by phase of care for older patients with cancer

37

Background: Healthcare reimbursement changes are contributing to increased closures of community hospitals and oncology practices, which may lead cancer patients to travel greater distances for care. Limited data exists on the impact of travel time on hospitalization rates and patient cost responsibility by phase of care for older cancer patients. Methods: This was a secondary analysis of Medicare claims from 2012-2015 for cancer patients age ≥65 receiving care in the University of Alabama at Birmingham Cancer Community Network. Patient addresses were obtained from network data, hospitalizations from inpatient claims, and patient cost responsibility from inpatient, outpatient, skilled nursing facility, carrier, and durable medical equipment claims. Drive time was calculated from patient home to cancer care site (CCS). Phase of care-specific (initial, survivorship, end-of-life [EOL]) rates of hospitalizations overall and by CCS vs. other care site (OCS) were calculated per 100 person-years. Hierarchical linear models compared average monthly phase-specific costs by drive time to CCS. Results: Of 23,605 older cancer patients, median drive time to CCS was 32 minutes (IQR 18-59), with 24% driving ≥1 hour to CCS. Rates of hospitalizations by initial (n = 14,225), survivorship (n = 18,805), and EOL (n = 8,211) phases of care were 54, 26, and 301 per 100 person-years, respectively. Higher rates of hospitalizations at OCS vs. CCS were shown for patients traveling ≥1 hour to CCS (initial, survivorship, and EOL rate of 41 vs. 20, 21 vs. 6, and 220 vs. 95 per 100 person-years, respectively). Median monthly costs by phase were $401 (IQR $182-$814) for initial, $369 (IQR $123-$1046) for survivorship, and $2075 (IQR $1123-$3723) for EOL. Patients traveling ≥1 hour to their CCS had higher cost responsibility, with patients in initial, survivorship, and EOL phases having $303 (95% CI $130-$476), $75 (95% CI $46-$105), and $736 (95% CI $308-$1164) higher average costs per month than those traveling < 1 hour, respectively. Conclusions: Cancer patients traveling further to receive care are potentially vulnerable to higher cost responsibility and limited access to care if community hospitals close, especially at EOL.