An innovative, systems-based conceptual framework of physician cancer screening decision-making

Background: Although evidence-based screening guidelines for different types of cancer have been developed, screening rates continue to be suboptimal, particularly among vulnerable populations. While a host of individual and community factors impact cancer screening, a physician's recommendation and follow-up are key drivers of individuals' receipt of guideline-concordant screening services. Yet, a comprehensive, systems-based model of physician cancer screening decision-making is lacking. Objectives: 1) To systematically review the literature on physician cancer screening decision-making; and 2) to propose an innovative, systems-based framework of physician cancer screening decisions. Methods: We conducted a systematic literature review of physician practice patterns using PubMed and other databases; reviewed behavioral prediction theoretical models; and developed a comprehensive conceptual framework of physician behaviors. Results: Studies of physicians' screening practices have documented that large proportions of physicians do not deliver screening services consistent with national guidelines and that variations exist in the quality of care contingent on physician and practice characteristics. Our proposed framework of physician cancer screening decision-making is based on the reasoned action approach, a family of behavioral prediction models including: Theory of Reasoned Action, Theory of Planned Behavior, and Integrative Model. We propose that physician screening intentions (i.e. decisions) would be predicted proximally by their attitudes toward screening, beliefs about norms related to screening (e.g., from other physicians in their practice), and feelings of control over making successful screening recommendations, which in turn would be impacted more distally by such factors as physicians' perceived barriers to and benefits of screening, as well as patients' beliefs and attitudes about screening. Moderators impacting those proximal and distal mediators include physician characteristics (training, specialty, demographics), practice characteristics and support systems (e.g. electronic health records, reminders), and the community context (community attitudes, economic factors, health care resources). Conclusion: While recent reviews have emphasized the importance of distal mediators, practice-level, and community-level factors in determining physician behaviors, a comprehensive framework for articulating those associations has not been put forward. We developed a framework of physician screening behavior that is systems-based, grounded in the reasoned action approach, and supported by the empirical evidence. The reasoned action approach has broad support in the literature, with meta-analyses demonstrating that it is able to account for 50-60% of behavioral decision-making variance on average. Our framework can guide health services research and be used to identify, develop, and evaluate physician and patient interventions to promote cancer screening and reduce cancer morbidity and mortality.