EMMA McGINNIS
Graduate Student, Colorado State University

Evidence from randomized controlled trials indicates that supervised exercise and behavior change counseling improves long-term physical activity (PA) adherence in cancer survivors. However, translating this work to real-world settings remains a challenge. This study examined the feasibility and acceptability of implementing PA behavior change counseling (PABCC) sessions into an existing cancer-exercise program.

Cancer survivors enrolled in BfitBwell, a 12-week exercise program at the Anschutz Health and Wellness Center.

Participants were randomized to receive (1) the standard BfitBwell program, or (2) BfitBwell plus six evidence-based PABCC sessions. Feasibility was assessed by participant representativeness, process fidelity, and time and cost to deliver the PABCC sessions. Acceptability was based on reasons for declining participation, adherence, and participant satisfaction with the PABCC sessions.

From July 2019- February 2020, N=93 enrolled in the BfitBwell program and N=33 (35.5%) enrolled in the study. Study participants were mostly female (63.6%), diagnosed with breast cancer (39.3%), and an average age of 54.3 &plusmn 12.37 years. There were no differences in age, sex, or cancer diagnosis between those who consented to the study, and participants in the BfitBwell database In the 26 applicable fidelity measures, no items were missed. To date, 42 hours have been spent delivering the PABCC sessions, and costs were printed workbooks ($21.75 per workbook). Reasons for declining to participate in the study were not interested (n=8), not able to guarantee class attendance (n=15), unable to make scheduled class times (n=12), or other (n=19). PABCC participants (n=6) attended M= 5.33&plusmn0.52 PABCC sessions (89% adherence rate). The satisfaction questionnaire has been completed by n=5, and 100% reported they enjoyed the PABCC sessions, most (80%) thought attending PABCC sessions will improve their ability to continue to exercise after BfitBwell, and 100% reported PABCC sessions were an added time burden.

Study participants were representative of BfitBwell participants, and found the PABCC sessions useful, but time intensive. Reasons for declining to participate in the study suggest that alterations in time commitment or intervention modality should be considered for continued implementation of PABCC in BfitBwell. To date n=13 have completed the study, and n=6 are currently enrolled. Data collection will continue through May, 2020.

CHERYL KELLY
Investigator, Kaiser Permanente Colorado

Kaiser Permanente (KP) is investing in a close-loop social needs referral platform called Thrive Local. Thrive Local links clinical care to social care through data integration with community-based partner organizations. Thrive Local consists of three key functions: a resource directory that provides up-to-date and searchable information on community resources; geographically-based networks of social service organizations (i.e., Community Network); and a technology platform that engages clinicians and staff, members, and employees of community organizations to make rapid, secure referrals between health care providers and social care providers and track the outcomes of those referrals. The purpose of this poster is to describe the methods and approach being used to evaluate this initiative designed for real-world impact.

There are several study populations included in the evaluation. KP providers and staff who are trained to deliver the intervention and must modify existing workflows to integrate screening and referral of basic needs. Community-based organizations (CBOs) that are part of the community partner network in each region and are likely to experience an increase in referred clients seeking services. KP members who are screened for basic social needs and referred for services.

An evaluation of Thrive Local across eight KP regions is being led by Dr. Allen Cheadle from the Center for Community Health and Evaluation and Dr. Cheryl Kelly from the Partners in Evaluation & Research Center using the RE-AIM framework. The implementation evaluation methods include interviews with KP providers, CBOs, and patients as well review of secondary sources (e.g., progress reports and other documents generated). Reach is being measured by data generated by the referral platform on the number and type of referrals and their resolution. Effectiveness is being measured by a patient-reported outcomes and secondary use of electronic health records. The patient survey will assess whether social needs are being met and whether patients they are able to address needs that may arise in the future. Additionally, data on referrals and progress towards closing social needs will be matched to patient medical record number to assess longer-term changes in health outcomes and healthcare utilization.

This study is beginning implementation despite several challenges, including variability of implementation across regions, a diverse set of stakeholders, and a staggered roll-out of a common tool to screen for social needs. This poster presentation will share the evaluation methods and measures and details about how we are incorporating patient-reported outcomes and objective measures of health and healthcare utilization to evaluate the impact of a closed-loop referral platform to address patient’s social needs.

TYLER ANSTETT
Assistant Professor, University of Colorado School of Medicine

Optimal utilization of blood products requires a balance between clinical benefit and unnecessary costs and risks associated with transfusions. Excess packed red blood cell (pRBC) transfusions are associated with harm to patients and additional costs to patients and health systems. The American Association of Blood Banks recommends a restrictive threshold of a pre-transfusion hemoglobin (Hgb) level of less than 7.0g/dL and single unit pRBC transfusions for the vast majority of hospitalized patients. Other institutions have significantly improved pRBC utilization through Electronic Health Record (EHR) focused interventions. In this project, we sought to conduct a needs assessment across the UCHealth system and evaluate the optimal strategy for encouraging prudent transfusions.

We examined the inpatient transfusion practices at five UCHealth hospitals across three regions: North, South, and Denver Metro between 2/1/2019 – 1/31/2020.

Data was acquired from the Clarity tabulation of the EPIC EHR. We included adult patients (> 17 years), who received a blood transfusion during an inpatient encounter. We excluded both perioperative hospital units and outpatient transfusions. We analyzed the data by the pretransfusion hemoglobin value as well as the number of pRBCs transfused for each Hgb value. We also examined the blood transfusion ordering interfaces at each hospital to evaluate for trends.

18,055 units of pRBCs were transfused during 13,804 transfusion administrations across all regions during the study period. 7,015 (51%) of transfusions were for pre-Hgb ≥ 7.0, and of all transfused units 3,471 (25%) of transfusions included two or more units of pRBC for pre-Hgb greater than 6.0. There was some heterogeneity across the institutions with transfusions for ≥ 7.0 ranging from 44% to 68%. Analysis of the ordering interface revealed regional differences that did not impact ordering practices, with University Hospital having the least directive interface but the lowest percentage of transfusions for Hgb ≥7.0g/dL at 41%. None of the blood transfusion ordering interfaces include proven strategies for reducing unnecessary transfusions.

As demonstrated by the pre-Hgb level of the transfusions and number of units transfused each time, these data support the need for improvement across the UCHealth system. We are embarking on a clinician level randomized user-centered, design-focused pragmatic factorial trial manipulating the blood transfusion ordering interface to evaluate the most effective method for reducing unnecessary blood transfusions. The trial design will account for the two main interventions (Overutilization of pRBC for pre-Hgb ≥ 7.0 and multiple units per transfusion order with pre-Hgb ≥ 6.0). We currently working with UCHealth leadership for approval of the design.