# 46 Research Abstract Format PROJECT NAME: Mixed Methods Qualitative Research as a Means for Effective Quality Improvement Research Institution: UT Southwestern Primary Author: Stephen Inrig Secondary Author: Jasmin Tiro Other Authors: Simon Craddock Lee; Bijal Balasubramanian Project Category: Research Please complete the following sections. Submission is limited to a maximum word count of 500 (not including text in graphs). Background: As the second highest cause of cancer death in the US, colorectal cancer (CRC) is an important health problem. CRC screening and appropriate diagnostic follow-up is essential to reducing CRC incidence and mortality, but screening rates are low in the US, particularly among uninsured, low-income minorities in safety-net systems. The complex CRC screening process occurs at multiple levels within a system and involves numerous steps and transitions within and between multiple services. No single data collection method adequately captures this complexity, so improving cancer care delivery requires new ways of observing and measuring these complex care processes and multilevel influences. Our safety-net based research team (The Parkland-UT Southwestern PROSPR Center) is part of an NCI-funded network (PROSPR, or Population-based Research Optimizing Screening through Personalized Regimens) that is assessing factors that influence cancer screening completion across numerous health systems. One of our projects’ aims is to apply a mixed methods research design to observe these multi-level, multi-step processes in order to determine optimal intervention points that will improve this complex care delivery process. Methods: Our multi-phase mixed-methods study triangulated methods, data sources, and investigators to comprehensively characterize the screening process and evaluate it through quantitative analyses [Figure 1; Table 1]. Organizational variables (like policies, protocols, or culture) vary between primary care clinics or vary longitudinally either at the clinic- or system-level. We will use hierarchical random intercept logistic regression modeling to identify which organizational and patient-level characteristics predict CRC screening completion and follow-up. Our multi-level modeling strategy accounts for the nesting of patients within providers and clinics as well as the random variation across clusters of patients, physicians, and clinics. We will evaluate two main outcomes: a) completion of CRC screening among primary care patients at average risk for CRC, and b) completion of CRC diagnostic evaluation among patients with abnormal screening results. Through a series of cross-tabulations, we will explore which intermediate steps are associated with the highest number of failures to complete screening. Phase 1 Phase 2 EMR Abstraction to Rank Order Clinics Phase 3 Organizational Survey Document Analysis Semi-structured Interviews Participant Observation Participant Observation Hierarchical Models (Qualitative and Quantitative-EMR Abstraction) Figure 1: Multi-phase Mixed Methods Design Table 1: Mixed Methodology Listing Method Purpose Examples of Products EMR abstraction Identify primary care clinics with highest and lowest completion rate for entire CRC screening process Identify screening process steps and interfaces with largest number of failures Understand origins, development, implementation, and prioritization of CRC screening Characterize organizational culture, structure, and formal protocols of the CRC screening process, including guideline dissemination and training of care teams Describe organizational structure, a broad range of clinical and non-clinical care behaviors as they relate to organizational protocols for CRC screening processes Evaluate functionality of the system for referring patients with abnormal screening tests Clarify observations and understand organizational culture (values, beliefs, and norms) Elucidate decision-making pathways for CRC screening processes at the network- and clinic levels Explore experiences and perceptions of whether organizational protocols are compatible with the situation of socially disadvantaged, safety-net patients Supplement the qualitative measurement of organizational culture and processes in the primary care clinics with highest and lowest completion rate Measure organizational culture, CRC screening processes, and protocols in all 11 primary care clinics Document Analysis Participant Observation Semi-structured interviews with leaders and care teams Organizational survey Rank order and select clinics for qualitative analysis Selection of problematic steps and interfaces for qualitative analysis Photocopies of documents scanned into database using Optical Character Recognition Chronology of CRC screening policy implementation Comprehensive report outlining CRC screening-related policies Detailed descriptive field notes, transcribed as text and entered into database Flowcharts depicting team members roles, responsibilities, relationships, and behaviors across range of CRC steps and interfaces Audio-recordings and transcripts Understanding of practice member experiences with CRC screening process and their beliefs about the value of EMR to improve delivery of screening processes Clarify processes not easily observed, or confusing during participant-observation (e.g., values, beliefs, and attitudes regarding CRC screening process) Identify variations in use of CRC screening protocols across 11 clinics & between staff within the clinics Assess degree of agreement between participantobservation with clinic staff perspectives Evaluate how these variables modify the relationship between patients’ degree of social disadvantage and completion of guideline-based CRC screening and follow-up Results: Thus far we have chronicled the CRC care continuum as it exists in Parkland’s safety-net setting and begun determining clinic-level variation in CRC screening completion [Figure 2]. Our mixed methods design has identified several likely nodes where screen failure and patient dropoff can occur. Figure 2: CRC Care Continuum Model Conclusions: Using multi-phase mixed methods, we are capturing an otherwise opaque system that has evolved in response to changing healthcare systems, technologies, staff, resources, and care delivery methods. This will allow us to identify small but significant nuances of care delivery and systems functioning that influence CRC screening delivery. The need for in-depth understanding of such processes and systems is clear, particularly for Quality Improvement interventions. In the era of Health Reform, researchers and QI professionals must find more effective ways to promote health and organize care. Mixed Methods Designs will play an important role describing, comparing, intervening in, and evaluating outcomes of complex healthcare systems and we urge researchers, policy makers, and others to employ them as a means of maximizing efficiency, reducing costs, and improving patient-centered outcomes.