24.12.2012 Views

Annual Meeting Proceedings Part 1 - American Society of Clinical ...

Annual Meeting Proceedings Part 1 - American Society of Clinical ...

Annual Meeting Proceedings Part 1 - American Society of Clinical ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

6093 General Poster Session (Board #8B), Mon, 1:15 PM-5:15 PM<br />

Unanticipated hospital admissions in patients undergoing radiotherapy<br />

with or without concurrent chemotherapy: Incidence and predictive factors.<br />

Presenting Author: Nabeel H. Arastu, University <strong>of</strong> North Carolina at Chapel<br />

Hill and Brody School <strong>of</strong> Medicine at East Carolina University, Greenville,<br />

NC<br />

Background: Unanticipated admissions are burdensome for patients and<br />

the healthcare system. An improved understanding <strong>of</strong> their frequency and<br />

predictive factors can inform initiatives to prevent such admissions and<br />

mitigate their associated human and financial costs. Methods: Electronic<br />

medical records <strong>of</strong> 500 patients undergoing external beam radiotherapy<br />

(RT) at our center in 2010 were reviewed. Unanticipated admission within<br />

90 days <strong>of</strong> initiating RT, and associated clinical factors, were recorded.<br />

Chi-squared and uni- and multivariate logistic regression was used to<br />

examine factors associated with admission. Results: Unanticipated admissions<br />

occurred in 20% (101/500) <strong>of</strong> patients, mean length <strong>of</strong> stay was 4<br />

days (range 1-16), and the mean interval between the start <strong>of</strong> RT and<br />

admission was 32 days (0-86 days). The most common indications for<br />

admissions were pain (19% <strong>of</strong> admissions), respiratory distress (15%), and<br />

neurologic symptoms (13%). On univariable analysis, 33% <strong>of</strong> patients<br />

treated for palliative intent were admitted (vs. 16% <strong>of</strong> curative intent<br />

patients, p�0.001), as were 26% <strong>of</strong> patients receiving concurrent chemotherapy<br />

(vs. 17% receiving RT alone, p�0.02). Multivariable analysis<br />

showed treatment intent, chemotherapy, and marital status to be associated<br />

with unplanned admissions (Table). A highly variable rate <strong>of</strong> unanticipated<br />

admission per diagnosis was observed (e.g. 4% for breast, 19% for<br />

GI/GU/GYN/ENT, and 37% for metastatic sites). Conclusions: Rates <strong>of</strong><br />

unanticipated admissions are �20% in patients undergoing RT. Approximately<br />

1/3 <strong>of</strong> patients receiving palliative RT, and more than 1/4 receiving<br />

concurrent chemoradiation, experienced an unplanned admission. Prophylactic<br />

measures should be studied in these high-risk patients to reduce<br />

admission rates, as unplanned admission may be an important quality <strong>of</strong><br />

care indicator in oncology.<br />

Odds <strong>of</strong> an unanticipated admission: Multivariate analysis.<br />

Covariate OR P value<br />

Concurrent chemotherapy (vs. RT alone) 3.4 �0.001<br />

Palliative intent (vs. curative) 2.7 �0.001<br />

Not married (vs. married) 2.3 �0.001<br />

6096 General Poster Session (Board #8E), Mon, 1:15 PM-5:15 PM<br />

Cost-effectiveness <strong>of</strong> adjuvant radiotherapy for older women with early<br />

hormone-receptor positive breast cancer. Presenting Author: Katherine<br />

Elizabeth Reeder-Hayes, University <strong>of</strong> North Carolina Hospital, Chapel Hill,<br />

NC<br />

Background: Radiation therapy (XRT) following breast conserving surgery<br />

decreases local recurrence at the expense <strong>of</strong> additional morbidity and<br />

treatment costs. However, its utility in elderly women at low risk <strong>of</strong><br />

recurrence has been questioned. This study assessed the cost-effectiveness<br />

<strong>of</strong> adding XRT to hormonal therapy (HT) in women over 70 with stage I,<br />

hormone-receptor positive (HR�) breast cancer after breast conserving<br />

surgery. Methods: A decision tree model was used to assess the costs and<br />

benefits <strong>of</strong> XRT � HT versus HT alone in 10,000 women age 70� with<br />

stage I HR� breast cancer. Using a societal perspective, we considered<br />

medical costs and quality <strong>of</strong> life effects <strong>of</strong> initial treatment, recurrences,<br />

and metastatic disease as well as long-term XRT-associated complications<br />

including breast fibrosis, chronic pneumonitis and cardiac disease. Probabilities<br />

<strong>of</strong> recurrence and death were modeled on recent clinical trial<br />

results, while toxicity probabilities were taken from literature review. The<br />

primary health outcome was incremental quality-adjusted life years (QALYs)<br />

gained. One-way and probabilistic sensitivity analyses (PSA) were performed<br />

to assess the sensitivity <strong>of</strong> model results and conclusions to various<br />

parameter estimates. Results: In the base-case scenario, the incremental<br />

cost-effectiveness ratio (ICER) for the addition <strong>of</strong> XRT was $923,017/<br />

QALY. The ICER was highly sensitive to variations in utility weights,<br />

particularly those reflecting patient preferences for initial treatment with or<br />

without XRT and those reflecting the decrement in quality <strong>of</strong> life resulting<br />

from breast fibrosis. In PSA, XRT was associated with lower qualityadjusted<br />

life expectancy at higher cost in 58% <strong>of</strong> simulations. Conclusions:<br />

In women over 70 with stage I HR� breast cancer, the addition <strong>of</strong> XRT to<br />

HT is not cost-effective at a willingness-to-pay threshold <strong>of</strong> $100,000/<br />

QALY, and is associated with little or no improvement in quality-adjusted<br />

life expectancy. Providers should be aware that the cost-effectiveness <strong>of</strong><br />

XRT in this population is strongly influenced by patient preferences<br />

surrounding recurrence and toxicity risks, and should weigh these factors<br />

when making shared decisions with patients.<br />

Health Services Research<br />

405s<br />

6095 General Poster Session (Board #8D), Mon, 1:15 PM-5:15 PM<br />

Association between financial relationships with commercial interests and<br />

research merit at the ASCO <strong>Annual</strong> <strong>Meeting</strong> (AM). Presenting Author:<br />

Beverly Moy, Massachusetts General Hospital, Boston, MA<br />

Background: Financial relationships with commercial interests (COI) are<br />

common in cancer research. There are few data examining the correlation<br />

between COI and research merit. <strong>Meeting</strong> placement prominence (MP) and<br />

peer review score (PRS) are indicators <strong>of</strong> research merit. We examined the<br />

association between ASCO AM abstracts whose authors disclose COI and<br />

both MP and PRS. Methods: We reviewed abstracts presented at the ASCO<br />

AM in 2006 and 2008-2011. We evaluated associations between COI<br />

disclosed by any author and PRS and MP (order <strong>of</strong> prominence: plenary<br />

session (PS), clinical science symposium (CSS), oral presentation (OP),<br />

poster discussion session (PDS), vs. general poster session (GPS)). Chisquare<br />

tests, T-tests, and logistic regressions <strong>of</strong> COI were used to assess<br />

associations with MP, PRS, and year. Results: Of 12,446 total abstracts<br />

accepted for presentation, 78% <strong>of</strong> PS, 59% <strong>of</strong> CSS, 54% <strong>of</strong> OP, 52% <strong>of</strong><br />

PDS, and 39% <strong>of</strong> GPS report at least one COI. Abstracts selected for PS,<br />

CSS, OP, and PDS had more COI compared to those selected for GPS (p �<br />

0.05). Stock ownership COI were more frequently disclosed in PS (30%),<br />

CSS (30%), OP (22%), and PDS (22%) compared to GPS (16%) (p �<br />

0.05). Employment COI were more frequently reported among abstracts<br />

presented at PS (39%), CSS (37%), OP (27%), and PDS (27%) compared<br />

to GPS (21%) (p � 0.05). Consultant COI were more likely to have higher<br />

MP than GPS placement (OR for PS�5.5; CSS�2.4; OP�2.2; PDS�1.9).<br />

Similarly, honoraria COI were more likely to have higher MP than GPS<br />

placement (OR for PS�3.9; CSS�1.8; OP�2.1; PDS�1.7). Better PRS<br />

was associated with COI (OR 0.17; p � 0.05). The relationship <strong>of</strong> better<br />

PRS with any COI strengthened over time from 2006, 2008-2011 (PRS<br />

times year interaction OR�0.65, p�0.001). Conclusions: ASCO abstracts<br />

whose authors report COI have higher merit as measured by MP and PRS.<br />

This suggests a dependence on industry relationships for access to<br />

important data that will lead to prominent cancer research. These relationships<br />

will require further investigation and ongoing management. To our<br />

knowledge, this is the first study examining the scientific merit <strong>of</strong> research<br />

with relation to COI.<br />

6097 General Poster Session (Board #8F), Mon, 1:15 PM-5:15 PM<br />

The impact <strong>of</strong> race/ethnicity concordance between patients and their<br />

navigators in time to diagnostic resolution <strong>of</strong> breast and cervical cancer<br />

screening abnormalities. Presenting Author: Marjory Charlot, Boston University<br />

Medical Center, Boston, MA<br />

Background: Patient navigators have been shown to reduce cancer disparities<br />

among racial/ethnic minorities by improving timely diagnosis and<br />

treatment <strong>of</strong> cancer. For a group <strong>of</strong> navigators who received cultural<br />

competency training, we sought to determine if racial/ ethnic concordance<br />

<strong>of</strong> the navigator and patient improved time to diagnostic resolution <strong>of</strong><br />

cancer screening abnormalities. Methods: Demographic data on 1466<br />

patients and their 23 navigators from the Boston Patient Navigation<br />

Research Program were used to assess concordance by race and ethnicity.<br />

All participants with either breast (n�751) or cervical (n�715) screening<br />

abnormalities were followed up to one year. Kaplan-Meier survival curves<br />

and proportional hazards regression models examined the effect <strong>of</strong> race/<br />

ethnicity concordance on time to definitive diagnosis, adjusting for age,<br />

race, language, economic status, insurance status, and severity <strong>of</strong> screening<br />

abnormality. Analyses were performed separately for the breast and<br />

cervical groups. Results: Of the 1466 patients, 32% were White, 27%<br />

Black, 31% Hispanic, and 10% Asian. Navigators were 61% White, 17%<br />

Black, 13% Hispanic and 9% Asian. Fifty eight percent <strong>of</strong> patientnavigator<br />

pairs were concordant by race/ethnicity. Overall rate <strong>of</strong> diagnostic<br />

resolution <strong>of</strong> cancer screening abnormalities within 365 days was high at<br />

90%. In both the breast and cervical cancer screening groups, racial/ethnic<br />

concordance was not associated with time to diagnostic resolution, with an<br />

aHR 1.03 (95% CI: 0.85, 1.25) for breast patients and HR 1.02 (95% CI:<br />

0.85, 1.21) for cervical patients. Conclusions: Patient-navigator racial/<br />

ethnic concordance is not associated with time to diagnostic resolution <strong>of</strong><br />

cancer screening abnormalities, in part because overall rates <strong>of</strong> diagnostic<br />

resolution were high. Our data suggest that with cultural competency<br />

training, navigators can be equally effective with patients from different<br />

ethnic and cultural backgrounds.<br />

Visit abstract.asco.org and search by abstract for the full list <strong>of</strong> abstract authors and their disclosure information.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!