Lee Cheng

Usage of erythropoiesis-stimulating agents in cancer patients at an academic cancer center and experience with specific care management tools for anemia†

In 2007, the US Food and Drug Administration (FDA) issued regulatory alerts for use of erythropoiesis‐stimulating agents (ESAs) in cancer patients with anemia after clinical trials and meta‐analysis data found that high ESA doses were associated with adverse outcomes in patients. In response to these findings, specific patient management tools for anemia (consisting in an algorithm and prescribing order set) were developed by a multidisciplinary team at The University of Texas MD Anderson Cancer Center.

Provider and Patient Gender Influence on Timing of Do-Not-Resuscitate Orders in Hospitalized Patients with Cancer

BACKGROUND End-of-life decisions and advance directives require timely physician-patient discussions but barriers exist to these discussions. OBJECTIVE To evaluate the influence of physician and patient gender on the timing of inpatient do-not-resuscitate (DNR) orders. DESIGN Retrospective cohort study. SETTING/SUBJECTS All adult patients (≥18 years) with cancer who received inpatient DNR orders at The University of Texas MD Anderson Cancer Center between January 2011 and December 2013. MEASUREMENTS Gender interaction between physicians and patients towards timing of the DNR order. RESULTS We identified 4,157 unique patients with a cancer diagnosis. These patients were treated by 353 physicians, of whom 123 (34.8%) were females and 230 (65.2%) were males. Multivariate analysis showed female patients were 1.3 times more likely to have early DNR orders written during hospital admission than were male patients (odds ratio [OR] 1.27; 95% confidence interval [CI] 1.07-1.50). When comparing gender interaction between physicians and patients, our results showed that female physicians were 1.5 times more likely to write early DNR orders with their female patients than for their male patients (OR, 1.48; 95% CI, 1.13-1.94). Same gender physician-patient dyads were not found between male physician and their patients (OR, 1.09; 95% CI, 0.91-1.31). Higher age, more comorbid conditions, and progression of diseases were also associated with early DNR orders (all p < 0.01). CONCLUSION Female patients are more likely to receive early DNR orders from their female physicians. Gender and gender interaction between physician and patients may potentially influence the timing of receiving DNR order.

Factors related to end-of-life (EOL) chemotherapy in solid tumor (ST) patients.

92 Background: Quality performance measures for cancer care, including use of chemotherapy in the last two-weeks of life, will be required for reporting. In this study, we evaluated the use of chemotherapy in ST patients (pts) within 14 days of EOL relative to several factors that may influence this practice. METHODS Adult pts (≥18 years) treated for ST at our institution, deceased December 01, 2010 through May 31, 2012, were retrospectively studied. Data on demographics, chemotherapy (excluded: hormones) within 14 days EOL, comorbidities, and cancer diagnoses were from administrative databases. Logistic regression analysis was performed for association of EOL chemotherapy with age, gender, ethnicity, comorbidities, cancer types, and metastatic status. RESULTS 5,607 pts met study criteria: median age 64 years; 48% female; 76% metastatic disease. EOL chemotherapy frequency was 3.9% overall, 4.6% in metastatic disease versus 1.7% in non-metastatic disease (p<0.01). Of 23 patients who received chemotherapy and had non-metastatic disease, the major tumor sites were: brain/other nerve system (34.8%) and lung/bronchus (21.7%).The top 10 frequencies in chemotherapy use by tumour sites were: melanoma (6.7%), female breast (5.5%), lung and bronchus (4.9%), pancreas (4.2%), brain/other never system (3.9%), head and neck (3.8%), female genital system, excluding ovary (3.4%), ovary (3.2%), liver/intrahepatic bile duct (3.0%), colon and rectum (2.6%). By regression analysis, the factor statistically significantly associated with receiving chemotherapy was metastatic disease (odds ratio [OR], 3.29; 95% CI,1.96-5.54), while factors associated with significantly less treatment were age ≥ 65 (OR, 0.66; 95% CI, 0.49-0.90); and any comorbid conditions (≥1 versus 0) (OR, 0.58; 95% CI, 0.38-0.89). CONCLUSIONS A small portion of ST patients who died received EOL chemotherapy (3.9%). However, metastatic disease and diagnosis category influenced treatment. Melanoma and breast cancer patients had higher frequency of EOL treatment. Older age and comorbidities were associated with less treatment. Variation in EOL treatment of ST patients is thus influenced by several clinical factors, but did not seem influenced by gender or ethnicity factors.

Analysis of the results of implementation of a pregnancy screening policy at a comprehensive cancer center.

256 Background: Institutional implementation of a standard pregnancy screening policy for women will enhance the safe and timely delivery of treatment throughout the care continuum. In 2012, we developed and implemented an initial policy on pregnancy screening test (PST). The policy indicated that all providers should fully discuss with their female patients of child bearing potential regarding the potential risks of cancer treatment on fertility and the adverse effects on the fetus. The providers should also discuss the contraceptive options with female patients and their partners prior to diagnostic studies and/or treatment. In 2015, we revised the initial policy to further clarify timeframe for PST and clinical indicators and also developed an algorithm to support this best practice. We assessed practice changes as a result of the policy changes regarding PST at our institution. METHODS A retrospective data review was conducted to assess the number of PST over a 5 year period at The University of Texas MD Anderson Cancer Center. The timeframe includes both baseline and post-implementation policy and reflects annual changes from fiscal year (FY) 2010 to 2015. The pregnancy test was defined as measuring human chorionic gonadotropin (hCG) level in a urine sample. RESULTS We observed a significant increase in PSTs over time among female patients between ages 10 to 55 years. Over 5 years, there were 23,793 urine pregnancy tests performed. The number of urine pregnancy tests performed per FY were as shown on table below (trend test, p< 0.01). The percentage of PST performed among age groups (years): 1.8% of 10 to 15, 11.6% of 16 to 25, 22.3% of 26 to 35, 34.7% of 36 to 45, and 29.5% of 46-55 years old. CONCLUSIONS Wedemonstrated that developing a standard institutional policy on pregnancy screening tests with further enhancements contributed to a significant increase in the number of PSTs performed over 5 years at our comprehensive cancer center. The majority of PSTs were conducted among female patient between the ages of 26 and 55. [Table: see text].

Practice of venous thromboembolism (VTE) prophylaxis in hospitalized cancer patients at a comprehensive cancer center.

149 Background: VTE prophylaxis measures are endorsed by the National Quality Forum in alignment with quality indicators from the Centers for Medicare & Medicaid Services. Accordingly, documentation of VTE risk, prophylaxis measures, and contraindications are recommended for hospitalized patients. To standardize practice we embedded a VTE risk assessment and prophylaxis module into admission and post-surgical order sets (OS), starting August 15, 2011. METHODS A retrospective study of 9,065 cancer patients (≥18 years) admitted to The University of Texas MD Anderson Cancer Center between June 01, 2013 through September 30, 2013. Pharmacological prophylaxis was executed with low-molecular-weight heparin or unfractionated heparin. Mechanical prophylaxis was executed with graduated compression stockings and/or sequential compression devices. Chi-square testing was used to determine the association between categorical variables. All statistically significant levels were determined with P values < 0.05. RESULTS 7,366 (81%) of all hospital admissions had documented VTE risk assessment and prophylaxis through the standardized VTE module. Before implementation of the new OS, only 40% of eligible patients received an order for VTE prophylaxis. The majority of patients were designated high or moderate risk (91.1%). Patients with high risk were more likely to receive pharmacological prophylaxis than those with moderate risk (74.1% vs. 38.2%, P<0.01). The most frequent contraindications to pharmacological prophylaxis were major surgery with risk of bleeding and thrombocytopenia (Table). CONCLUSIONS Most patients received VTE prophylaxis based on VTE risk levels presented in a standardized OS. There is is limited information in the clinical literature about the impact of VTE prophylaxis on outcomes among cancer patients, we plan to assess anticoagulation-related outcomes in this cohort of patients. [Table: see text].

Chemotherapy use during the last 30 days of life in patients with cancer.

231 Background: Treatment options are increasingly available to patients with advanced cancer. Appropriately timed cession of chemotherapy is a quality-of-care measure in the Quality Oncology Practice Initiative of the American Society of Clinical Oncology. To gain insight into whether communication about end-of-life (EOL) issues might need to be improved, we analyzed chemotherapy use during the 30 days at the EOL (EOLCx) among patients at large cancer center. We also analyzed the presence of electronic DNR (eDNR) orders at the time of death in hospital as a surrogate indicator of EOL discussions. METHODS We reviewed the records of cancer patients (age≥18 years) who received their care at The University of Texas MD Anderson Cancer Center and died between December 2010 and May 2012. Logistic regression was used to measure associations of EOLCx and age, gender, ethnicity, comorbidities, cancer type, and cancer progression. RESULTS A total of 7,399 patients met the inclusion criteria. The median age was 64 years; 46% were female and 18% had hematologic malignancies (HM). 996 patients (14%) received EOLCx, and of these, 554 died in the hospital (7%, 554/7,399). Of those who died in the hospital, 93% had eDNR orders (no difference between solid tumors [ST] and HM). The EOLCx was higher for patients who died in the hospital than for those who died elsewhere (44% vs 7%) and higher for patients with HM than for those with ST (38% vs 8%). EOLCx was more common in patients with ST with metastases, HM without relapse, and HM with relapse than in those with ST without metastases [odds ratios (ORs), 3.3, 13.9, and 32.8, respectively; all p<0.05]. EOLCx was less common in older patients (≥65 years) than younger patients (OR, 0.6; p<0.05) and was more common in patients with comorbidities than in those without (OR, 1.3; p<0.05) and in patients who were not black or Hispanic than in white patients (OR, 1.6; p<0.05). No difference of EOLCx between genders was found. CONCLUSIONS EOLCx was more common in patients with HM with or without relapse than in patients with ST. Most patients had eDNR orders at the time of death in hospital. The results suggest that communication about prognosis and advanced planning about the goals of medical care is becoming critical part in overall oncologic care plan.