System-Level Factors Associated With Use of Outpatient Specialty Palliative Care Among Patients With Advanced Cancer

PURPOSE

The proportion of patients with advanced cancer who receive outpatient specialty palliative care (OSPC) is as low as 2.0%. Improved understanding of the system-level factors influencing use of OSPC could inform adaptations to the delivery of palliative care to maximize access. We examined associations between OSPC use among patients with advanced solid tumors and oncology-OSPC clinic colocation and patient travel time to an OSPC clinic.

PATIENTS AND METHODS

We conducted a retrospective cohort study of patients with advanced solid tumors receiving oncologic treatment between January 1 and December 31, 2016, within a comprehensive cancer center network with well-established, oncology-specific OSPC clinics. Multivariable logistic regression analysis was used to evaluate the associations of clinic colocation and geographic access with OSPC use.

RESULTS

Of 9,485 patients with advanced solid tumors, 478 (5.0%) received OSPC services in 2016. After controlling for age, sex, marital status, cancer type, insurance, treatment intent, and illness severity, patients whose oncologist practices were colocated with OSPC clinics were more likely to use OSPC (odds ratio [OR], 19.2; 95% CI, 14.1 to 26.2). Compared with patients who lived > 90 minutes from an OSPC clinic, patients with travel times of < 30 minutes (OR, 3.2; 95% CI, 2.2 to 4.6) and 31 to 60 minutes (OR, 2.4; 95% CI, 1.6 to 3.6) were also more likely to use OSPC.

CONCLUSION

Among patients with advanced solid tumors, colocation of oncology and OSPC clinics and shorter patient travel time were associated with greater odds of using OSPC. Future efforts to increase OSPC use in this population should consider clinic colocation and travel burden.

Patterns of Biomarker Testing Rates and Appropriate Use of Targeted Therapy in the First-Line, Metastatic Non-Small Cell Lung Cancer Treatment Setting

Despite clear clinical benefit and guideline recommendations for predictive biomarker testing and subsequent first-line targeted therapy treatment in patients with non-small cell lung cancer (NSCLC), there is evidence that testing has not been widely embraced in the clinical setting. This study uses clinical pathways to understand biomarker testing patterns and ensuing first-line treatment decisions. Data of patients with metastatic NSCLC were analyzed for testing rates and treatment selection at 7 cancer programs using data input by providers into the pathways software. Findings were analyzed by type of provider (community or academic). Among providers using clinical pathways, biomarker testing rates were high and appropriate selection of targeted therapy was observed. Clinical pathways can act as a tool to assist oncology practices to promote testing of key biomarkers and subsequent selection of appropriate therapy.

Using clinical pathways to understand biomarker testing patterns

e18189

Background: Via Pathways’ Disease Committees incorporate biomarker testing recommendations into the pathway decision support when the literature supports improved outcome. For several years, the non-small cell lung pathway (NSCLP) has recommended testing for ALK, EGFR, and, more recently ROS1 targeted therapies. The decision support software (VP) requires the practitioner to register testing and results, including the decision not to test. Methods: For the twenty months ending 10/31/16, we analyzed all NSCLP patients treated at UPMC CancerCenter locations for first line, metastatic, non-squamous lung cancer per the VP. Results were analyzed in total and by type of provider (83 community providers vs. 3 academic providers who specialize in lung cancer). Provider’s answers to the biomarker testing question included “positive” results, “negative” results, “test pending”, and “not tested”. Results: Of the 684 treatment starts in the 1stline non-squamous NSCLP, 100 were patients of academic providers (AP) and 584 were patients of community providers (CP). AP indicated testing for ALK and EGFR in 100% of patients (n=100). CP indicated testing for ALK in 91% (n=532) and for EGFR in 92% (n=538). AP indicated testing for ROS1 in 96% (n=93) and CP indicated testing for ROS1 in 81% (n=473). Results are summarized in the table. Conclusions: Data from clinical pathways software can be utilized to understand biomarker testing rates among providers. In this example, the three AP tested for actionable mutations more frequently than the 83 CP, although both rates of testing appeared to be high. This data can be used to generate hypothesis about patterns of care that may include patient parameters, size of network or other factors that can then further analyzed and tested. Pathways serves as an important tool in promoting and monitoring quality care. [Table: see text]

Expanding indications of existing drugs and associated costs under risk reimbursement models

e18306

Background: New payment models in oncology will likely include taking risk on drug costs. While new drugs could be excluded, existing drugs with expanded indications are problematic as the payer has limited data needed to exclude these costs. Recent label expansion for pembrolizumab (PB) in 1st line NSCLC for > 50% PD-L1 expression is an example of this risk. A broader indication in the future will likely result in use of PB for almost all 1stline patients. Methods: We reviewed treatment starts documented by providers in the Via Portal (VP) for 1st, maintenance (MT) and 2nd lines of therapy in non-squamous mutation-negative NSCLC for the 12 months ended 11/30/16. We assumed 6 month treatment lengths and calculated cost to Medicare (MC). We calculated a hypothetical alternative where PB received full 1st line and maintenance indication given with pemetrexed (PM) and carboplatin (per recent FDA filing and results of KEYNOTE-021 trial1). We assumed 2nd line patients received either docetaxel or ramucirumab/docetaxel (using 3rd line treatments in VP). Results: Total MC cost in the actual arm (n = 208) was $13,943,412, $12,190,264 and $11,614,488 for 1st, MT and 2nd line, respectively. The total MC cost in the hypothetical arm (n = 208) was $25,477,452, $25,687,168 and $6,415,575 for 1st, MT and 2ndline, respectively. This represents a total increase of $19,832,031 (52.5%) or $95,346 per patient. Conclusions: Bundled rates and shared savings that include oncology drugs are subject to significant financial risk for providers when existing drugs receive expanded indications to a new, larger population of patients. Langer CJ, et al. Carboplatin and pemetrexed with or without pembrolizumab for advanced, non-squamous non-small-cell lung cancer: a randomised, phase 2 cohort of the open-label KEYNOTE-021 study. Lancet Oncol. 2016;17(11):1497-1508.

Developing and piloting an electronic telephone triage application

e21680

Background: With the shift of cancer care to the outpatient setting, telephone triage is a vital part of today’s oncology practice. UPMC CancerCenter (UPMC) desires to streamline the current telephone triage process across its 31 outpatient clinics. UPMC and Via Oncology collaborated to develop and pilot a prototype of an electronic telephone triage application. The key components of the prototype were defined as decision support, EHR integration, reporting, and workflow management. Methods: A prototype application, Symptom Manager (SM), was developed and piloted with 4 nurses at 3 outpatient clinics. Front office staff registered inbound calls in SM, which populated a phone triage queue. The nurse was prompted to enter information about the call. Decision support materials were displayed for reference, including algorithms for symptom assessment, adjudication, and treatment. The output was a structured summary of the encounter that was copied to the patient’s chart. From the queue, the nurse was also prompted to make outbound calls for symptom reassessment. A feedback session with the nurse users and institution leadership was held following the pilot. Results: A total of 235 inbound calls were captured in the application, of which 74 (31.5%) were symptom-related. Additional data fields collected during the pilot included symptom, adjudication, time elapsed for callback, and symptom status at follow up. The ability to extract and report on data of this type was viewed as valuable for quality and process improvement. The nurse users praised SM’s ease of use, but prototype performance issues shortened the pilot. They indicated that the standardized documentation increased efficiency. Requests for future enhancements included deeper EHR integration and additional fields for data capture. Conclusions: This pilot provided the feedback necessary to develop the latest version of SM which begins roll out across UPMC in November 2016 and includes EMR integration. An update on the software and additional data captured during the pilot will be provided at presentation.

Clinical Pathways: Management of Quality and Cost in Oncology Networks in the Metastatic Colorectal Cancer Setting

PURPOSE

Via Pathways (clinical pathways for cancer) provide evidence-based guidance for specific patient presentations based on the merit of efficacy, then toxicity, and finally cost (if efficacy and toxicity are comparable). We evaluated the impact of a change to the guidance in the metastatic colorectal cancer (mCRC) setting across two large, integrated health networks.

METHODS

Cetuximab and panitumumab were determined to have equal efficacy in the treatment of mCRC with no significant difference in toxicity based on recent data from key clinical studies. A cost analysis using Centers for Medicare and Medicaid Services average sales data determined a cost advantage for panitumumab. A substitution of panitumumab for cetuximab in the clinical pathway for all mCRC lines of therapy was initiated as of August 2014.

RESULTS

In the preimplementation period, 86 (93.5%) and six (6.5%) treatment selections were for cetuximab and panitumumab, respectively. After the pathway change was implemented, 13 (18.1%) and 59 (81.9%) treatment selections were for cetuximab and panitumumab, respectively. The change in prescribing habits was rapidly altered by the pathway change. The estimated annualized cost savings for the two health networks resulting from the response to the pathway change was $711,021.

CONCLUSION

This study demonstrates that clinical pathways can act as a tool to assist oncology practices in decreasing costs and quickly responding to changing treatment paradigms by providing clinicians with consensus-driven treatment recommendations that incorporate the most up-to-date clinical trial results, toxicity considerations, and regimen cost information.

Developing and piloting an electronic telephone triage application

137

Background: With the shift of cancer care to the outpatient setting, telephone triage is a vital part of today’s oncology practice. UPMC CancerCenter (UPMC) desires to streamline the current telephone triage process across its 31 outpatient clinics. UPMC and Via Oncology collaborated to develop and pilot a prototype of an electronic telephone triage application. The key components of the prototype were defined as decision support, EHR integration, reporting, and workflow management. Methods: A prototype application, Symptom Manager (SM), was developed and piloted with 4 nurses at 3 outpatient clinics. Front office staff registered inbound calls in SM, which populated a phone triage queue. The nurse was prompted to enter information about the call. Decision support materials were displayed for reference, including algorithms for symptom assessment, adjudication, and treatment. The output was a structured summary of the encounter that was copied to the patient’s chart. From the queue, the nurse was also prompted to make outbound calls for symptom reassessment. A feedback session with the nurse users and institution leadership was held following the pilot. Results: A total of 235 inbound calls were captured in the application, of which 74 (31.5%) were symptom-related. Additional data fields collected during the pilot included symptom, adjudication, time elapsed for callback, and symptom status at follow up. The ability to extract and report on data of this type was viewed as valuable for quality and process improvement. The nurse users praised SM’s ease of use, but prototype performance issues shortened the pilot. They indicated that the standardized documentation increased efficiency. Requests for future enhancements included deeper EHR integration and additional fields for data capture. Conclusions: This pilot provided the feedback necessary to develop the latest version of SM which begins roll out across UPMC in November 2016 and includes EMR integration. An update on the software and additional data captured during the pilot will be provided at presentation.

Developing and piloting an electronic telephone triage application

68

Background: Telephone triage is a vital part of today’s oncology practice and is often not standardized and measured to allow for quality improvement. UPMC CancerCenter (UPMC) has 31 outpatient clinics and is in need of a tool to streamline the current telephone triage process. UPMC and Via Oncology collaborated to develop and pilot a prototype of an electronic telephone triage application. The key components of the prototype were defined as decision support, EHR integration, reporting and workflow management. Methods: A prototype application, Symptom Manager (SM), was developed and piloted with 4 nurses at 3 outpatient clinics. Front office staff registered inbound calls in SM, which populated a phone triage queue. The nurse was prompted to enter information about the call. Decision support materials were displayed for reference, including algorithms for symptom assessment, adjudication and treatment. The output was a structured summary of the encounter that was copied to the patient’s chart. From the queue, the nurse was prompted to make outbound calls for symptom reassessment. A feedback session with nurse users and institution leadership was held following the pilot. Results: A total of 235 inbound calls were captured in the application, of which 74 (31.5%) were symptom-related. Additional data fields collected during the pilot included symptom, adjudication, time elapsed for callback, and symptom status at follow up. The ability to extract and report on data of this type was viewed as valuable for quality and process improvement. The nurse users praised SM’s ease of use, but prototype performance issues shortened the pilot. When referencing the decision support materials, nurses felt more confident consulting providers and during independent decision-making. They indicated that the standardized documentation increased efficiency. Requests for future enhancements included deeper EHR integration and additional fields for data capture. Conclusions: This prototype fulfilled the defined key components of a tool of this type. SM is currently being refined and incorporated into the Via Portal, Via Oncology’s decision support software. An update on the software and additional data captured during the pilot will be provided at presentation.

Abstract 2594: Clinical pathways as a platform to support clinical research

Introduction

Clinical trials are essential to advancing cancer care; however, actual patient accrual rates generally fall well below the Institute of Medicine's stated goal of 10% of all newly diagnosed patients. To support the goal of increasing clinical trial accruals, Via Oncology (VO) uses its clinical pathways decision support tool (Via Portal or VP) to promote awareness of locally available clinical trials and provide associated analytics for its cancer center (CC) customers.

Methods

VO works continuously with each cancer center to identify and embed their CC specific clinical trials in the applicable disease within the VP. In their daily practice, providers navigate the VP by entering information about each patient's specific state and stage of disease. If a clinical trial is locally available anywhere within the CC's network, that trial is presented as the first treatment recommendation before any standard of care pathway recommendations. Links to trial specific documents are also presented. If the provider feels the patient might be appropriate for that trial, a secure email notification is generated for the trial coordinator. If the patient is not accruing to the trial, the provider must select the reason for non-accrual from a structured list of reasons. Additionally, the VP can be interfaced with trial management systems for timely opening and closing of trials within the VP.

Data captured in the VP for the twelve months ended October 31, 2015 was retrospectively analyzed to calculate accrual rates for patients potentially eligible for a clinical trial. The distribution of reasons for not accruing patients to clinical trials was also calculated and analyzed.

Results

During this time period, 10,000 patients had a high level

clinical presentation that matched an open clinical trial embedded in the VP

(the patients were not specifically matched for trial inclusion/exclusion

criteria). Of the total, 1,391 (13.9%) of patients were documented as accrued to a clinical trial. A total of 9,661 reasons for not accruing the patient to a trial were captured and distributed as follows: patient not eligible (34.4%), patient not interested (14.7%), patient accrued to other trial (0.8%), financial burden or insurance does not cover (0.5%), trial not available at site (25.6%), trial already selected (0.8%), other reason(23.2%).

Conclusions

VP also acts as a tool to promote clinical trial accruals and measure reasons for non-accruals. The reasons for not accruing provide insight to the barriers to accrual that patients and practices face. Additional reasons could be defined in the future to reduce the incidence of other non-defined reasons.

Citation Format: Peter G. Ellis, Aurora Health Care, Baystate Health, Broward Health, Carroll Hospital, IU Health, TCCBD, UPMC. Clinical pathways as a platform to support clinical research. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2594.

Pathways Clinical Decision Support for Appropriate Use of Key Biomarkers

Purpose:

Breast cancer diagnostics have the ability to predict disease recurrence and the benefit of chemotherapy. This study measures the use of a diagnostic assay, Oncotype DX, when embedded in a breast cancer decision support algorithm and, on the basis of the assay results, the use of chemotherapy in the adjuvant setting.

Methods:

UPMC CancerCenter retrospectively reviewed patients with estrogen receptor–positive, human epidermal growth factor receptor 2 (HER2)Neu–negative disease with zero to three positive nodes navigated in the Via Pathways decision support portal during a 12-month period. The breast algorithm prompted input of the assay recurrence score (RS) and then recommended hormonal therapy alone (HT) for low RS, or chemotherapy followed by HT for high RS. The patient’s RS was correlated with the treatment decision.

Results:

During this time period, 643 patients had ER-positive, HER2Neu-negative disease with zero to three positive nodes. Of those, 596 (92.7%) had diagnostic testing to determine chemotherapy plus HT versus HT alone, and 47 had chemotherapy followed by HT without an RS. For node-negative patients classified with low or high RS, pathway treatment adherence rates were 99.7% and 96.6%, respectively; node-positive patients had 95.7% and 87.5% adherence rates, respectively.

Conclusion:

This analysis demonstrates the use of a clinical pathway to measure the adoption of a diagnostic test, the Oncotype DX breast assay, and the use of the appropriate therapy on the basis of the RS. As more diagnostics are established to aid in the personalized treatment of diseases, pathways may be important in maintaining clinician awareness of the appropriate disease presentations where these tests should be used, measuring usage of these tests, and tracking the treatment decisions on the basis of test results.

Actionable biomarkers in a non-small cell lung cancer (NSCLC) clinical pathway (CP)

155

Background: Oncologists use Via Pathways (VP) to drive standardization to best evidence-based cancer care. VP include guidance for biomarker testing and associated treatments. Practices collaborated to analyze the first line non squamous carcinoma NSCLC VP. An unexpectedly large percentage of treatment decisions captured by oncologists in the VP portal had unknown ALK translocation or EGFR mutation test results. The VP was modified to require input of whether or not the test was ordered. This analysis examines the impact of this change. Methods: ALK and EGFR testing responses from the VP database were analyzed pre and post implementation of the change. Treatment decisions were reviewed to analyze prescribing patterns. Results: From 1/1/2014 to 1/29/2015, ~40% of EGFR or ALK testing responses were unknown. Post-implementation (1/30/2015 to 5/31/2015), 29.8% were charted as awaiting ALK results and 9.4% did not order the test; 29.5% were charted as awaiting EGFR results and 7.6% did not order the test. For the combined period, appropriate targeted agents were selected for 91.9% (n = 37) and 86.5% (n = 133) of patients who were ALK translocation positive and EGFR sensitizing mutation positive, respectively. Conclusions: Results demonstrate the ability of CP to promote testing of key biomarkers and use of appropriate targeted agents. Frequent evaluation and modification of pathway content is needed to ensure practice patterns are accurately captured. [Table: see text]

The usefulness of clinical pathways (CP) in managing quality and cost in oncology networks

715

Background: UPMC CancerCenter (UPMC) and Indiana University Health (IUH) utilize Via Pathways (VP) for their CP initiative. VP are developed and maintained by disease committees that evaluate therapies on merit of efficacy, then toxicity, and finally cost (if efficacy and toxicity are comparable) to provide a recommendation for specific patient presentations. Recent data from key studies regarding the use of panitumumab (PAN) or cetuximab (CET) in the setting of metastatic colo-rectal cancer by the colo-rectal committee led to the determination that both treatments were equally effective with no significant difference in toxicity (Price et al., 2014; Peeters et al., 2014). A subsequent cost comparison utilizing CMS average sales prices demonstrated an approximate 14% monthly cost advantage for PAN. A substitution of PAN for CET across all metastatic lines of therapy in the pathway was initiated as of August 2014. This analysis was undertaken to understand the impact of this change. Methods: We reviewed the UPMC and IUH data within the VP database to compare new treatment starts of CET and PAN before and after this change was put in place. Results: There was rapid change in the selection behavior between the two drugs as demonstrated in the table. Conclusions: The results in this simple example of substituting equivalent drugs based on cost exemplify the power of CP to rapidly change prescribing habits across two large cancer networks. CP serve as an invaluable tool to allow oncology practices to quickly respond to the changing treatment norms of oncology care. [Table: see text]

Incorporation of a patient question prompt list into a pancreatic cancer pathway

20

Background: Metastatic pancreatic cancer is an aggressive disease affording a life expectancy of less than 6 months (Worni et al., 2013). Timely end-of-life discussions are critical. Patient question prompt lists (QPL) have been shown to facilitate discussions around distressing topics and increase patient satisfaction (Brandes et al., 2014; Clayton et al., 2007). We evaluated the use of a QPL embedded in a pancreatic pathway provided by Via Pathways (VP) to enhance communication between oncologists and patients with metastatic disease. Methods: VP are disease-specific decision support algorithms delivered in a web-based portal and used by oncologists at the point of care. Through a committee of oncologists and palliative care specialists, a QPL was developed and incorporated into the VP pancreatic pathway. The goal of the QPL was to empower patients to have informed discussions of available cancer therapies, treatment goals, and personal preferences. Users of the pathway were required to indicate if the QPL was used, then had to provide it to the patient. The QPL prompt was placed in the first line metastatic branch of the pancreatic pathway and piloted from 12/1/14 to 2/28/15. Following the pilot, users were surveyed through email. Results: Results are shown in the table. Conclusions: Based on the results, the pilot had limited success. Although there was an even distribution of those indicating the QPL was and was not provided to the patient, survey results suggested the majority of providers were unaware of the QPL. Our conclusion is that our approach must be less of a burden to the oncologist. Current focus is on automated delivery of advance care planning materials directly to the patient, removing that task from the oncologist workflow. [Table: see text]

Clinical pathways and quality metrics

273

Background: UPMC CancerCenter (UPMC) participates in a number of nationally recognized quality programs and strives to maximize quality through continuous improvement of the quality measures endorsed by these programs. Clinical pathways are potential aid in driving quality improvement through decision support and measurement. Methods: When deficiencies are identified, UPMC’s quality team develops improvement strategies including working with its pathways vendor (Via Oncology) on the placement of key decision support prompts and required data fields within its physician-facing pathways portal. Gaps in data completeness have also been identified as an opportunity to use pathways for more complete data collection during chart abstraction as well as interim reporting. Results: Due to the success above, UPMC continues to actively engage with pathways as a tool to drive adherence to and capture complete data for quality measures. If release of the UPMC data is approved by the quality measurement organization(s), UPMC will share the results of its quality measures in the full poster. Conclusions: Pathways are a tool for promoting adherence to quality measures by oncologists through the use of a point of care decision support system. Pathways are also a source of reporting quality measures across all patients in time period which allows for interim measurements and proactive identification of improvement areas. [Table: see text]

Pathways clinical decision support for appropriate use of key biomarker

172

Background: UPMC CancerCenter (UPMC) has utilized clinical pathways for almost ten years in an effort to ensure standardization to the evidence based care for its patients. UPMC oncologists participate in the various pathways disease committees that develop and maintain the pathways content and utilize the pathways through a web-based portal in their daily decision making and documentation. The pathways cover not only treatment recommendations but also guidance for work up such as recurrence risk tools (OncoType Dx) for node negative, HER2 negative, ER positive breast cancer who are candidates for chemotherapy. For patients with a low recurrence risk score, the pathways recommends hormonal therapy only, saving the patient both toxicities and costs of chemotherapy where appropriate. Methods: UPMC analyzed its use of chemotherapy in patients with a recurrence risk score of less than 19 through a retrospective review of physician-input data in its Via Pathways Portal for the twelve months ended May 31, 2014. During this time period, the Via Pathways recommended the recurrence risk test for node negative, HER2 negative, ER positive patients. For patients with a low risk score (less than 19), the Via Pathways recommended hormonal therapy only. For those low risk score patients receiving chemotherapy, the physician would document that an Off Pathway decision was being made, indicate the reason for going Off Pathway, and document the actual therapy delivered. Results: For the twelve months ended May 31, 2014, UPMC physicians documented 288 decisions for patient presentation of node negative, HER2 negative, ER positive patients with a recurrence risk score of less than 19. Of these decisions, 99% (n=284) were On Pathway for hormonal therapy. Of the remaining 1%, three (3) were for accrual to a clinical trial and one (1) was Off Pathway for chemotherapy. Conclusions: Pathways are a tool for promoting adherence to evidence based care by oncologists through the use of a point of care decision support system. Pathways have the potential to reduce costs and toxicities of treatment through the evidence based guidance developed by the pathways disease committees and the adherence to such guidance by oncologists utilizing the pathways decision support tool.

Development and implementation of oncology care pathways in an integrated care network: the Via Oncology Pathways experience

Physician involvement is integral to a successful clinical pathways program. The Via Pathways model used at the UPMC CancerCenter, in which the oncologists developing the clinical content are ultimately the oncologists who use the Pathways Portal, has proven to be successful in ensuring physician participation. In addition, an agile user interface and pragmatic layout of the tool are essential for incorporating pathways into the daily workflow of physicians. When appropriately developed and implemented, clinical pathways are an effective tool for standardizing care and ensuring quality.

Impact of a decision support tool on quality measures

6643

Background: UPMC CancerCenter received their ASCO QOPI certification after the Spring 2011 measurement period with a QOPI Certification Overall Quality Score of 78.2%. The results, however, demonstrated a number of metrics performing below QOPI mean scores. Physicians and staff performed an analysis of root causes for these underperforming metrics and identified a combination of steps to improve these metrics and the overall score. Methods: Interventions to ensure high capture of Core Metric #9 (staging within 30 days of diagnosis), #10 and #11(treatment intent and discussion with patient) were imbedded into the clinical pathways program (Via Oncology Pathways) used by the UPMC CancerCenter physicians beginning in February 2011. These data points were charted by physicians when utilizing the Via Pathways Portal and automatically displayed on the regimen order sets that are generated by that Portal. These order sets were used as official orders and therefore placed into the patient charts. Results: UPMC CancerCenter’s QOPI Certification Overall Quality Score improved by 7 percentage points to 85.39% for the Fall 2012 measurement period (as compared to Spring 2011). All of the 24 certification metrics for Fall 2012 were similar or improved as compared to Spring 2011. Core Metrics #9 and #10 improved by 18 and 22 percentage points, respectively Conclusions: Improvements in quality measures such as ASCO QOPI can be gained through targeted analysis of root cause and application of new interventions such as clinical pathways decision support. We feel that decision support tools such as pathways programs may be an important source of data for e-QOPI type programs and should be considered.

Using a pathways process to ensure measurable evidence-based standardized care in a large cancer network

252

Background: UPMC CancerCenter includes 40 sites of services in a 100 mile radius in Western Pennsylvania. Consistency and quality of care are critical to such a diverse network. In addition, the UPMC mission includes accrual to clinical trials. To meet these challenges, UPMC developed the Via Oncology Pathways. The program has served UPMC well for over seven years to date and is now a key foundation for UPMC’s overall healthcare reform strategy for quality and accountable care. Methods: Treatment algorithms were developed for 90% of cancer types by establishing committees of academic and community specialists. These committees interpret the literature and define the most efficacious, least toxic, and economically efficient treatment regimens appropriate for highly specific disease presentations (e.g., node +, er-, her2 +, PS 0-1). Clinical trials are also imbedded into the algorithms. Quarterly, these algorithms are reviewed by the committees to assess relevance, review network feedback, add newly available trials and address emerging data. Equally important to clinical content is its presentation to the practicing physician in a manner that allows real time usage and adds value to physician workflow. This is accomplished with a web portal that presents the individual pathways status through the physician’s daily schedule. Results: Over 120 oncologists at UPMC use Via Oncology Pathways in their daily practice. In 2011, UPMC physicians confirmed a pathways status for 94% of their patient visits (195,000) and achieved an On Pathway rate of 82% for their 18,000 treatment decisions. The database also includes patient presentations, reasons for going off pathway and reasons for not accruing to clinical trial. Lower hospitalization rates and mandated adoption of personalized medicine were also observed. Conclusions: When appropriately developed and implemented, clinical pathways are a solution to improving the quality and cost effectiveness of cancer care by enhancing physician decision-making, standardizing care and ensuring access to evidence-based personalized medicine. We continue to expand the scope of our pathways to include diagnostic studies, surveillance protocols and end of life prompts.

A phase II trial of weekly paclitaxel, 5-fluorouracil, and leucovorin as first-line treatment for metastatic breast cancer

PURPOSE

This phase II multicenter trial evaluated the efficacy and toxicity of weekly paclitaxel, 5-fluorouracil, and leucovorin administered as first-line therapy for metastatic breast cancer.

PATIENTS AND METHODS

The study enrolled 155 women with pathologically confirmed and measurable metastatic adenocarcinoma of the breast. Treatment consisted of paclitaxel 80 mg/m2, 5-fluorouracil 425 mg/m2, and leucovorin 20 mg/m2 administered weekly 4 x per 4-week cycle in the first 40 patients enrolled (group 1), and weekly 3 x per 4-week cycle in the subsequent 115 patients (group 2) enrolled. Hematologic growth factor support was not routinely used. Twenty patients with hepatic dysfunction were enrolled to assess the tolerability of the regimen in this population. All therapies were delivered in an outpatient setting.

RESULTS

The overall response rate was 48%, with 12-month estimated survival rates of 53% and 65% for treatment groups 1 and 2, respectively. Response rates were not statistically different between the two treatment schedules. Therapy was well tolerated when delivered on the every 3 of 4-week schedule, including patients with hepatic involvement and those age > or = 65.

CONCLUSION

Weekly therapy with paclitaxel and 5-fluorouracil with leucovorin is active as first-line therapy for metastatic breast cancer. Use of this regimen should be given consideration, particularly in patients who are not candidates for anthracycline-based therapy.

Phase II trial and pharmacokinetic assessment of intravenous melphalan in patients with advanced prostate cancer

Alkylating agents have been reported to yield response rates of up to 20% in hormone-refractory prostate cancer. Melphalan was studied in four small trials in which the drug was given orally. In this phase II trial, melphalan (30 mg/m2) was given intravenously every 28 days to 27 patients with hormone-refractory prostate cancer. Pharmacokinetic sampling was performed so as to describe the clearance of melphalan in this population and in an attempt to carry out pharmacodynamic modeling for toxicity and response. Prostate-specific antigen (PSA) was also assessed prospectively. No objective responses to this regimen were documented. The median survival for patients on this trial was 11.5 months. There was no correlation between drug clearance and measured creatinine clearance and no relationship between systemic exposure and toxicity. A decrease of > 50% in serum PSA that was sustained for > 6 weeks was documented in two patients, most notably in one patient who has survived for more than 29 months. Intravenous melphalan is not an active agent in hormone-refractory prostate cancer.

High-Dose Oral Tamoxifen, a Potential Multidrug-Resistance-Reversal Agent: Phase I Trial in Combination With Vinblastine

BACKGROUND

P-glycoprotein mediates resistance to natural-product anti-neoplastic agents like vinblastine through an active transport process resulting in reduced intracellular concentration of these agents. The triphenylethylene antiestrogen tamoxifen and its major metabolite N-desmethyltamoxifen at concentrations of 4-6 microM enhance the intracellular concentration of natural-product antineoplastics and augment the cytotoxicity of such drugs three-fold to 10-fold in a variety of human and murine cell lines.

PURPOSE

On the basis of these preclinical findings, we conducted a phase I clinical trial of high-dose, oral tamoxifen administered in conjunction with a 5-day continuous infusion of vinblastine.

METHODS

We studied 53 patients with advanced epithelial tumors. Tamoxifen was given orally as a loading dose on day 1, followed by two doses a day on days 2-13. Vinblastine was given as a 120-hour continuous infusion (1.5 mg/m2 per day) on days 9-13 of each tamoxifen course. The starting dose of tamoxifen was 40 mg/m2 administered twice a day following a loading dose of 150 mg/m2. The maximum dose was 260 mg/m2 twice a day following a loading dose of 680 mg/m2. Treatment cycles were repeated every 28 days.

RESULTS

The dose-limiting toxic effects of tamoxifen were neurologic and began within 3-5 days after the start of treatment. They consisted of tremor, hyperreflexia, dysmetria, unsteady gait, and dizziness. One patient experienced a grand mal seizure 24 hours after the last tamoxifen dose. Toxic effects were rapidly reversible. Asymptomatic prolongation of the QT interval on electrocardiogram occurred at doses of tamoxifen of 80 mg/m2 or higher given twice a day. No coagulation or ophthalmologic abnormalities occurred. Tamoxifen did not enhance the toxicity of vinblastine. Mean plasma concentrations of tamoxifen or N-desmethyltamoxifen at 260 mg/m2 tamoxifen given twice a day for 13 days were 6.04 and 6.56 microM, respectively. There was no relationship between plasma antiestrogen content and the development of neurotoxic effects.

CONCLUSIONS

Tamoxifen at 150 mg/m2 given twice a day following a loading dose of 400 mg/m2 results in plasma levels of tamoxifen and N-desmethyltamoxifen of 4 and 6 microM, respectively, without dose-limiting toxicity. We recommend this dose for phase II trials of tamoxifen to modulate P-glycoprotein-mediated drug resistance.

IMPLICATIONS

Our study demonstrates that high-dose tamoxifen can be safely administered and that plasma concentrations that may inhibit P-glycoprotein function can be achieved.

Assessment of the calcium and phosphorus nutrition in horses by analysis of urine

Studied were made to determine if a practical assessment of the calcium and phosphorus nutrition of horses could be obtained from an analysis of urine samples. The concentrations of Ca and P in urine samples changed markedly when groups of 4 mares were fed diets containing from 1.0 to 3.9 g Ca/kg and from 1.5 to 6.1 g P/kg, but serum concentrations of Ca and P remained relatively constant. The concentrations in single urine samples were considered unreliable indicators of excretion of the minerals because of variations in water excretion, and two methods to overcome this problem were examined. In one method, the creatinine clearance ratios (%Cr) of Ca and P were calculated. The other method was simpler, and concentrations of the minerals were expressed as the ratio of the total solute concentration. Both of these methods appeared to give reliable estimates of the excretion of Ca and P in urine, and there were significant relationships between the values obtained for horses fed diets containing a wide range of Ca and P. When the horses were fed adequate Ca, the Ca-%Cr was greater than 2.5% and P-%Cr was less than 4%. The corresponding ratios of mineral concentration to solute concentration were greater than 15 mumole Ca/mosmole and less than 15 mumole P/mosmole. When a low Ca (1.0 g/kg) and high P (4 g/kg) diet was fed, the excretion of Ca and P had changed markedly after 3 days and could be monitored in single samples of urine by either method. It was concluded that a practical assessment of the Ca and P nutrition of horses could be obtained from an analysis of urine samples to determine if the diet is either low, adequate or high in Ca, and high in P.

An examination of the calcium and phosphorus nutrition of thoroughbred racehorses

The calcium and phosphorus nutrition of thoroughbred racehorses was assessed by analysis of serum and urine samples collected from 90 horses in 1975 and 139 horses in 1980-81 at racetracks in Melbourne. Horses that were excreting greater than 15 mumole Ca/mosmole and which had a calcium to creatinine clearance ratio greater tha 2.5% were considered to have adequate Ca intake. Horses that were excreting greater than 15 mumole P/mosmole and which had a phosphorus to creatinine clearance ratio greater than 4% were considered to have excessive phosphorus intake. Sixty-percent of the horses sampled had adequate Ca intake, and 44% had excessive intakes of P. Twenty-five percent of the horses were excreting more P in urine than Ca. This would indicate these horses were subjected to nutritional secondary hyperparathyroidism, and horses entered in races by 53 of 99 trainers were in this category. It may be concluded that a high proportion (40%) of thoroughbred racehorses receive inadequate calcium nutrition while they are fed high-grain diets during racing.