4 results
3510 Academic influence in gynecologic oncology is associated with industry funding: an analysis of the Open Payments database
- David Samuel, Shelby Adler, Nicole Vilardo, Gregory Gressel
-
- Journal:
- Journal of Clinical and Translational Science / Volume 3 / Issue s1 / March 2019
- Published online by Cambridge University Press:
- 26 March 2019, p. 32
-
- Article
-
- You have access Access
- Open access
- Export citation
-
OBJECTIVES/SPECIFIC AIMS: Industry payments to physicians can present a conflict of interest. The Physician Payments Sunshine Act mandates the disclosure of these financial relationships to increase transparency. Recent studies in other surgical specialties have shown that research productivity is associated with greater industry funding. In this study, we characterize the relationship between academic influence and industry funding among academic gynecologic oncologists. METHODS/STUDY POPULATION: Departmental websites were used to identify academic gynecologist oncologists and their demographic information. The Hirsch index (h-index) relates an author’s number of publications to number of times referenced by other publications, a validated measure of an author’s academic influence. This was obtained from the Scopus database. The Center for Medicaid and Medicare Services Open Payments online database was searched for all industry payments in 2017. The NIH Reporter online database was searched for active grants. Goodness of fit testing showed that all variables followed nonparametric distributions. Medians were compared using Mann-Whitney U tests and Kruskal-Wallis analysis of variance with post-hoc Dunn’s test. RESULTS/ANTICIPATED RESULTS: Four hundred and sixty-six academic gynecologic oncologists were included in the analysis. In 2017, 89.7% of this group received industry funding totaling $41.4 million. Median industry funding was $453 [IQR $67-19684] and median h-index was 14 [IQR 8-26]. Only 8.1% of gynecologic oncologists were NIH grant recipients and they received significantly higher industry payments ($357 vs. 11,168, P<0.01). Gender and academic rank were not associated with industry funding. Gynecologic oncologists in the highest decile of industry funding received a median payment of $447,651[N=46, IQR $285,770 – 896,310] totaling $36.5 million. The median h-index for this top-earning decile was 23 [N=46, IQR 16.5-30.3]. When stratified by payment amount, median h index increased but only reached statistical significance in the highest cohort receiving >$100,000 (N = 63, P<0.05). DISCUSSION/SIGNIFICANCE OF IMPACT: The majority of academic gynecologic oncologists receive industry funding although there are large variations in payments. Those receiving the largest payments are more likely to hold NIH grants and have greater academic influence.
3515 Readability of online patient education materials on gynecologic malignancies from major medical associations
- David Samuel, Nicole Vilardo, Sara Isani, Gregory Gressel
-
- Journal:
- Journal of Clinical and Translational Science / Volume 3 / Issue s1 / March 2019
- Published online by Cambridge University Press:
- 26 March 2019, pp. 28-29
-
- Article
-
- You have access Access
- Open access
- Export citation
-
OBJECTIVES/SPECIFIC AIMS: Patients are increasingly using online materials to learn about gynecologic cancer. Recent studies demonstrate that 85-96% of patients with a gynecologic malignancy utilize the Internet as a health resource. Providers can refer patients to educational materials produced by major medical associations available on their websites. However, patient educational materials (PEMs) published by professional organizations from other surgical specialties have been shown to be difficult to read for the average American. The NIH and AMA recommend that PEMs be written between a sixth and eighth grade reading level. In this study, we assess the readability of online PEMs on gynecologic cancer published by major medical associations. METHODS/STUDY POPULATION: Seven national medical association websites with PEMs on gynecologic malignancy were surveyed: American College of Obstetricians and Gynecologists, Center for Disease Control, Foundation for Women’s Cancer, National Cancer Institute, National Cervical Cancer Coalition, National Ovarian Cancer Coalition, and Society of Gynecologic Oncology. Online PEMs were identified and analyzed using five validated readability indices. One-way ANOVA and Tukey’s test were performed to detect differences in readability between publishers. RESULTS/ANTICIPATED RESULTS: Two hundred and thirty PEMs were included in this analysis. Mean readability grade levels with standard deviation were: 11.3 (2.8) for Coleman-Liau index; 11.8 (3.2) for Flesch-Kincaid; 11.1 (1.2) for FORCAST formula; 12.5 (2.7) for Gunning FOG formula; 12.1 (2.6) for New Dale-Chall formula; and 13.5 (2.5) for SMOG formula. Overall, PEMs were written at a mean 12th grade reading level. Only 4.3% of articles were written at an 8th grade reading level or below. ANOVA demonstrated a significant difference in readability between publishing associations (p<0.01). PEMs from the Center for Disease Control had a mean 10th grade reading level and were significantly lower than all other organizations. PEMs from The Foundation for Women’s Cancer had a mean 13th grade reading level and were significantly higher than most other organizations. DISCUSSION/SIGNIFICANCE OF IMPACT: Gynecologic oncology PEMs available from major medical association are written well above the recommended sixth to eight grade reading level. Simplifying PEMs may improve patient understanding of their disease and facilitate physician-patient communication.
3377 A cross-sectional analysis of opioid prescribing patterns among gynecologic oncologists using Medicare fee-for service provider utilization & payment data
- David Samuel, Devin Miller, Sara Isani, Dennis Kuo, Gregory Gressel
-
- Journal:
- Journal of Clinical and Translational Science / Volume 3 / Issue s1 / March 2019
- Published online by Cambridge University Press:
- 26 March 2019, p. 31
-
- Article
-
- You have access Access
- Open access
- Export citation
-
OBJECTIVES/SPECIFIC AIMS: Opioids are the first-line treatment for moderate to severe cancer-related pain. Increased awareness of opioid prescription misuse and adverse outcomes has prompted statements on their use from multiple national medical groups. In this study we characterize national-level opioid prescription patterns among gynecologic oncologists treating Medicare beneficiaries. METHODS/STUDY POPULATION: The Centers for Medicare and Medicaid Services (CMS) database was used to access Medicare Part D beneficiary data (2016). All available opioid claims prescribed by gynecologic oncologists were identified. Medication type, prescription length and other prescribing factors were recorded. Physician demographics were obtained from departmental websites and accrediting bodies. Physicians with <10 opioid claims are not included in the CMS database. Bivariate statistical analysis including chi-squared, Fisher’s exact test and Wilcoxon rank-sum test were performed to compare variables with threshold for significance set at p<0.05. Linear regression modeling was also performed to examine association of gender with number of opioids prescribed. RESULTS/ANTICIPATED RESULTS: A total of 494 board-certified gynecologic oncologists were included in this analysis. In 2016, gynecologic oncologists wrote 23,584 opioid prescriptions for 267,824 days of treatment (average of 9.24 prescribed days per claim). The most commonly prescribed opioid was oxycodone/acetaminophen (41%). Male physicians had significantly more opioid prescription claims than females (p<0.01) including after adjusting for differences in years of experience. The majority of physicians had 11-50 opioid prescription claims (68%). A minority were high prescribing physicians with >100 opioid claims (11%). Of these, the overwhelming majority were male (82%) and late career (46%, >15 years since board certification). Physicians in the South had the greatest number of opioid prescription claims and significantly more than physicians in the Northeast, who had the fewest (p<0.01). Mean number of opioid claims increased with increasing years of experience (p<0.05). DISCUSSION/SIGNIFICANCE OF IMPACT: Among gynecologic oncologists, there were gender-based, regional and experience-related variations in opioid prescribing in the Medicare population in 2016. Further longitudinal studies are required to elucidate secular trends in opioid prescription practice.
2317 Uterine serous carcinoma is associated with a high risk of venous thromboembolism regardless of latency from surgical staging
- Gregory M. Gressel, Lauren Turker, Shayan Dioun, Nicole S. Nevadunsky
-
- Journal:
- Journal of Clinical and Translational Science / Volume 2 / Issue S1 / June 2018
- Published online by Cambridge University Press:
- 21 November 2018, pp. 50-51
-
- Article
-
- You have access Access
- Open access
- Export citation
-
OBJECTIVES/SPECIFIC AIMS: Patients with gynecologic cancer are known to have an increased risk of venous thromboembolism (VTE) in the postoperative period secondary to hypercoagulability from both malignancy and pelvic surgery. Recent literature suggests that chemotherapy itself may be thrombogenic and prophylaxis may be beneficial in ambulatory patients receiving chemotherapy. Although extended VTE prophylaxis is commonly given after surgical staging, administration of prophylactic anticoagulation during chemotherapy or radiation treatment is not routinely performed. This study seeks to characterize risk factors and timing of VTE in a cohort of women diagnosed with uterine serous carcinoma (USC). METHODS/STUDY POPULATION: After institutional review board approval, a cross-sectional study was performed of all women diagnosed with USC between January 1999 and January 2016 at Albert Einstein College of Medicine. Data analysis was performed using Stata version 14.2 (Stata Statistical Software: Release 14, 2015. College Station, TX: StataCorp LP). Baseline clinical data was analyzed to calculate descriptive statistics. Normality of continuous variables was visually assessed and if no substantial violations were noted, data was reported as means±standard deviations. Otherwise, they were reported as medians with interquartile ranges. Categorical data was presented as number of patients with percentages. Bivariate analysis was performed to assess the association between clinical variables and diagnosis of VTE. Continuous variables (age, body mass index, number of risk factors for VTE) were visually assessed for normality. Levene’s test was used to assess for equal variance among groups. If no substantial violations were noted, means and standard deviations were calculated using 2 sample t-test for equal variance. Variables violating normality assumptions were analyzed using the Mann-Whitney U-test, calculating medians and interquartile ranges. Categorical and dichotomous variables (VTE risk factors, race, stage) were examined using the χ2 test or Fisher’s exact test (if expected values for more than 20% of cells were less than 5). Odds ratios were reported with 95% confidence intervals. Using a backwards stepwise elimination approach, a multivariable logistic regression model was fit to accurately examine association of risk factors with VTE, adjusting for other covariates. The resulting model was assessed for calibration and discrimination using Hosmer-Lemeshow test for goodness of fit, classification table, and ROC curve. Regression diagnostics were run in order to identify potentially influential covariate patterns in the model. First-order interactions were assessed for using product interaction terms (interaction defined as p-value for the likelihood ratio test <0.05). The resulting model was assessed for calibration and discrimination using Hosmer-Lemeshow test for goodness of fit, classification table, and ROC curve. A Cox proportional hazards model was also fit in order to examine the association between individual covariates and time to clot development. Log-rank testing was performed to compare survivorship experience by groups and survivorship curves were generated using the Kaplan-Meier method. Assumptions of the proportional hazards model was confirmed visually using log-log plots and goodness of fit assessment. RESULTS/ANTICIPATED RESULTS: A total of 413 patients were identified for inclusion in the study. The majority of patients (83%) were of non-White race. Bivariate analysis revealed no significant associations between age, BMI, or race with diagnosis of VTE (p=0.75, 0.49, and 0.28, respectively). Patients who had more than 2 risk factors for VTE had a significantly increased likelihood of VTE diagnosis (p=0.02). There was a highly significant association between stage of USC and diagnosis of VTE (p=0.005). Patients with stage III and stage IV cancer were 2.4 and 3.5 times more likely to develop VTE than patients with stage I cancer (95% CI: 1.09–5.30, 1.74–6.83, respectively). Of the 70 patients who were diagnosed with VTE, most were not postoperative (64.3%) and a large proportion developed clots while receiving chemotherapy (35.7%). Patients who developed VTE while on chemotherapy had a median Khorana score of 1 (IQR: 1, 2). In logistic regression modeling examining association of VTE with potential risk factors, covariates selected as significant for inclusion at the p<0.25 level included cancer stage, composite number of risk factors, diabetes, hypertension, cardiovascular disease (CVD), and COPD. Composite risk score was identified to be a potential confounder of the relationship between individual risk factors and development of clot and was therefore left in the model for adjustment. After adjusting for other covariates, only stage 4 disease (OR: 2.66, 95% CI: 1.53, −4.64) and hypertension (OR: 2.90, 95% CI: 1.14–7.36) were associated with development of VTE and were included in the final model. No concerning violation of assumptions of logistic regression or interaction was identified. The Hosmer-Lemeshow goodness of fit test identified that the model was well-fit using 10 groupings (p=0.35) and receiver operator characteristic testing showed that the model had acceptable discrimination with a ROC value of 0.7. The final model was found to classify 83.1% of participants correctly. Regression diagnostics identified 4 potentially influential covariate patterns. These patterns were eliminated from the model and no meaningful differences were noted. Patients contributed a total of 16,414 person months of analysis time in study follow-up. A negative, linear association was noted between stage of cancer and time to clot development. Long-rank testing revealed a significant difference in failure by stage of disease (p<0.001) and presence of hypertension (p=0.03). Cox proportional hazard modeling revealed that after adjustment for other covariates, only cancer stage and the presence of cardiovascular disease were significantly associated with time to failure. Patients with cardiovascular disease had a 2.02-fold increased risk of CVD compared to those without CVD (95% CI: 1.16–3.47). Those with stage 3 and 4 cancer were 3.19 (95% CI: 1.53-6.64) and 8.05 (95% CI: 4.11–15.78) fold more likely to develop VTE compared to those with stage 1 disease, respectively. DISCUSSION/SIGNIFICANCE OF IMPACT: Our study demonstrated that patients with USC are at high risk of developing VTE at all time points after their disease diagnosis, not just those who have undergone recent surgery. This risk is highest for women with hypertension, CVD, and stages III and IV disease. The fact that patients who developed clots on chemotherapy had an average Khorana score of 1, suggesting that they would not have been successfully risk stratified using previously published tools. To the best of our knowledge, this is the first study to report a high hazard for VTE in patients with serious endometrial cancer even several months after surgical staging. Although this is a retrospective study and cannot make inferences about VTE incidence, it generates the hypothesis that extended VTE prophylaxis may be beneficial in this cohort of patients regardless of their latency from surgical staging. Large randomized studies are needed to test this hypothesis.