Individual healthcare providers' behaviors are often constrained by factors that are beyond their own control (1). The assumption that all behaviors are largely under conscious control has taken a “theoretical battering” due to research showing the importance of non-conscious processes that operate in organizations (2). Research suggests that many healthcare provider behaviors that are repeatedly performed become non-reflective and more or less automatic (3). Individual behavior is also constrained by factors at collective levels (1). Collective levels include interpersonal (e.g., relations between healthcare providers), group (e.g., healthcare professionals providing care in a breast medical oncology practice), intraorganizational (e.g., hospital culture), and interorganizational (e.g., accreditation standards). Collective-level influences may also be largely non-conscious, having become internalized and taken for granted (e.g., norms and values of a professional culture) (1).

Various implementation determinant frameworks include factors at the organizational level. For example, the Consolidated Framework for Implementation Research (CFIR) and the Exploration, Preparation, Implementation, Sustainment (EPIS) Framework include inner setting (i.e., intraorganizational) and outer setting (i.e., interorganizational) domain (4, 5). Domains are comprised of constructs (i.e., explanatory concepts that cannot be directly observed but can be inferred from observed data) (6). Organization-level domains include constructs such as structural characteristics (“the social architecture, age, maturity, and size of an organization”), cosmopolitanism (“the degree to which an organization is networked with other external organizations”), and funding (“fiscal support provided by the system in which implementation occurs”) (7–9). The Theoretical Domains Framework (TDF) similarly includes the environmental context and resources domain, which includes constructs such as material resources and barriers and facilitators (10).

Commonly used implementation determinant frameworks encourage implementation scientists to consider organizational influences on implementation; however, the scope of organization-level constructs described in extant frameworks is limited. Furthermore, determinant frameworks often lack explanations of the mechanisms underlying the relationships between organization-level constructs and implementation. Extant frameworks' limited scope impedes progress in implementation science by obscuring the influence of organization-level constructs that may drive implementation outcomes. A substantial body of work in industries other than healthcare provides evidence of the significant influence of organizational influences on implementation, pointing to high-leverage strategies to promote organizational change. Organization theory has been applied to educational and budgetary reform, elucidating the critical importance of addressing power dynamics among leadership and fostering positive change culture to facilitate implementation (11, 12). In the non-profit industry, organization theory can be used to build capacity, assist with decision-making, narrow target populations, and clarify organizational needs (13).

Failing to account for the critical influence of organization-level constructs on implementation introduces omitted variable bias–i.e., the faulty attribution of the influence of the omitted variable(s) to variables that were included (14). In the case of implementation research, this may amount to, for example, attributing the influence of organizational inertia (i.e., resistance to change) to a construct that is related but distinct (e.g., readiness for implementation) or an unrelated construct (e.g., individual provider motivation). The misattribution of omitted organization-level constructs to the constructs that extant implementation frameworks include has important implications for subsequent stages of implementation research, such as selecting and identifying strategies to target the constructs influencing implementation.

Many extant implementation determinant frameworks are conceptual frameworks, in that they offer a menu of constructs thought to influence implementation, but they do not address how change takes place or any causal mechanisms, which is critical for falsifying hypothesized relationships through empirical study (15). The ability to falsify hypothesized relationships between constructs depends on explanations of the mechanisms underlying relationships between constructs that are derived from theory (16). The constructs in conceptual frameworks such as the CFIR derive from a combination of theory and empirical studies. For example, the CFIR peer pressure construct derives from Institutional Theory, but patient needs and resources derives from a combination of empirical evidence and other conceptual frameworks rather than theory.

In contrast to conceptual frameworks, theoretical frameworks are based on theories, which propose mechanisms underlying the relationship between constructs and implementation. One commonly used theoretical framework in implementation science is the TDF. As a theoretical framework, the TDF can be used to identify mechanisms proposed in included theories; however, the TDF does not offer nuanced insight into organization-level influences on implementation. The TDF's environmental context and resources domains contains constructs that derive from several theories that are identified as organization theories; however, many of the included theories are not in fact organization theories (e.g., decision-making theory). As such, the TDF is limited in its contributions to understanding organization-level influences on implementation.

Organization theories provide explanations for the complex interactions within and between organizations and their context (environment, surrounding policies, cultural norms). These theories not only describe and explain these interactions, but can also be used to predict implementation outcomes based on contextual factors. Organization theories have the potential to explain how policies, institutions, funding, and workforce dynamics affect implementation outcomes (17). Organization theories have been historically used to an explanatory tool in fields of education, nonprofit organizations, management, and health services research, dating back to the 1950 s (11–13, 18, 19). These theories, while widely used and published, remain largely inaccessible outside of organization science. Organization theories provide their own inventory of constructs, which often require significant training to apply with fidelity.

To equip implementation scientists with understanding of a broader scope of organization-level constructs and their hypothesized influence on implementation, a comprehensive yet accessible framework of organizational influences on implementation is needed. In this paper, we describe the development of the Organization Theory for Implementation Science (OTIS) framework, which summarizes constructs from nine organization theories identified as relevant to implementation in preliminary studies (20). Our overarching goal is to increase implementation scientists' familiarity with and conceptualization of the myriad organizational factors that influence implementation through mechanisms clearly articulated by organization theories.

We developed the OTIS framework using a combination of concept mapping and iterative consensus-building, with support for interpretation from members of the Cancer Prevention and Control Research Network (CPCRN) OTIS workgroup (21). CPCRN is a national network of academic, public health, and community partners whose work focuses on reducing the burden of cancer within specific workgroup and interest group projects. CPCRN OTIS workgroup members include investigators conducting research at the intersection of implementation science and cancer prevention and control. This study was approved by the Wake Forest University School of Medicine IRB (IRB00072134) on 6/2/21.

This study describes how we created the OTIS framework to increase implementation scientists' familiarity with and conceptualization of the diverse set of organizational influences on implementation. Increasing implementation scientists' conceptualization of organizational influences may contribute to more comprehensive understanding of the key drivers of implementation and, in turn, our ability to identify and select strategies to accelerate the translation of evidence into practice, as found in other industries, such as business and education (11, 12, 17, 18). Our efforts yielded six conceptually distinct domains, encompassing 70 constructs from nine organization theories with relevance to implementation. Distilling many constructs from several theories into a limited number of domains limits the burden on implementation scientists to account for the vast array of potentially important organizational influences on implementation. The six domains that we identified in this study reflect concepts that are central to organization theory, including power, structure, autonomy, control (20), but which are less commonly addressed in implementation science. The concepts reflected in the OTIS framework offer perspective on key questions in implementation science, such as how and why organizations adopt, implement, and sustain evidence-based practices—or resist doing so.

The OTIS framework considerably expands upon existing implementation determinant frameworks' conceptualization of organizational influences on implementation. OTIS includes constructs such as specialization, which is not explicitly captured in the CFIR or EPIS frameworks, but which may influence the decision to adopt an evidence-based practice. For example, a study of determinants of low-value use of computed tomography to evaluate microscopic hematuria found that, while urologists' evaluation practices changed following the American Urological Association's revised guidelines, primary care providers' evaluation practices often went unchanged, highlighting the need to tailor strategies for the various specialties involved in implementation (27, 28). OTIS greatly expands upon the TDF's organization-level environmental context and resources domain with more nuanced domains, such as interorganizational relationships. Future efforts should systematically map OTIS onto extant determinant frameworks to clearly articulate OTIS's unique contribution. For example, OTIS's Governance and Operations and Tasks and Processes domains include several constructs that may add critical nuance to EPIS's Funding/Contracting construct. Before systematic mapping of OTIS onto extant frameworks, OTIS may be used in its current form in conjunction with other frameworks, such as the CFIR and TDF, which are increasingly used in combination and already capture intra-organizational constructs, such as climate and leadership (28). For example, some OTIS domains or constructs that appear not to be captured in CFIR (e.g., tasks and processes; stability of the population of organizations; normative pressures) could be included in implementation determinant studies.

OTIS also expands upon commonly used implementation frameworks by allowing users to access organization theories articulate the mechanisms underlying relationships between included constructs and implementation. For example, the EPIS framework identifies sociopolitical influences on implementation (e.g., legislation; monitoring and review); however, EPIS does not articulate how or why these constructs influence implementation. In contrast, OTIS's basis in theory allows users to identify hypothesized relationships between included constructs and implementation, as clearly articulated in publicly available OTIS abstraction forms (28). Specifically, users may consult the propositions section of OTIS abstraction forms to identify mechanisms underlying included constructs. For example, OTIS describes how coercive influences of governments and accrediting bodies exert normative pressure (Interorganizational Relationships domain) on healthcare organizations to comply with legislation and monitoring by virtue of organizations' dependence on these governing bodies for permission to operate. Therefore, OTIS could be used in conjunction with extant frameworks to explain the mechanisms underlying constructs' influence on implementation (29). Clearly articulated mechanisms are critical for identifying strategies that are best-suited to influence the construct identified as influencing implementation.

Members of the CPCRN OTIS workgroup are currently applying the OTIS framework in the following projects: Project 1 a) tests the conceptual validity and applicability of the OTIS framework in community oncology practices and b) develops, tests, and disseminates tools using OTIS in implementation research, including a qualitative interview guide and codebook. Project 2 is an American Society of Clinical Oncology collaborative study on Sexual Orientation and Gender Identity data collection. OTIS will be used in this project to a) reanalyze data that have been previously analyzed using the CFIR, and b) compare results between CFIR and OTIS findings. Project 3 applies OTIS to a CDC-funded U01 cooperative agreement to reduce health inequities for cancer survivors in the District of Columbia. OTIS will be used to a) to build community coalitions of approximately 10 organizations to improve infrastructure and communication and b) to think consider power dynamics and the elimination of disparities and health inequalities. Some limitations of our study should be noted. Concept mapping requires participants to have pre-existing knowledge and experience with the topic they are mapping, limiting the pool of potential participants. There is a limited population of researchers with the required familiarity of organization theories and implementation science to participate in concept mapping. As a result, our purposive sampling approach was necessary to increase the likelihood that participants would understand included constructs enough to sort and rate them. However, it is possible that participants lacking refined expertise in implementation or organization science would have valuable perspective on included constructs. For example, hospital administrators may lack fluency in the terminology included in organization theories, but they may have unique insight into how, for example, normative pressure from professional organizations influences implementation. Future work should refine the language used in OTIS to increase its accessibility to an audience without expertise in organization science. Additionally, the clusters that Global Max generated, in many cases, lacked coherence as indicated by the stress value of 0.32, suggesting variation in concept mapping participants' interpretation of the constructs and their relationships. To address this concern, OTIS workgroup members used their expertise to reorganize many clusters in our consensus-building process, potentially suggesting the limited utility of concept mapping for developing the framework. We view the OTIS framework as a living document to be revised through application. For example, implementation scientists may find through qualitative interview data collection that study participants describe OTIS constructs in combinations not reflected in the domains identified in this study. Future iterations of the OTIS framework will be revised to reflect empirical evidence.

Despite these limitations and the need for continued development, OTIS may be used in its current form in implementation research. OTIS could be used to inform data collection or analysis. For example, OTIS could be used to develop guides for interviews with cancer program leadership to understand the potential influence of participation in quality improvement networks, professional norms, and the ability to recruit providers influence compliance with cancer program accreditation standards (30). We plan to use OTIS to analyze data that were previously collected regarding factors influencing cancer programs' implementation of exercise interventions. In each of these cases, OTIS offers researchers the tools necessary to understand the mechanisms underlying factors that influence implementation, pointing toward strategies to facilitate implementation (e.g., strengthening or reorganizing quality improvement networks to support compliance with accreditation standards).

We distilled 70 constructs from nine organization theories into six domains in the OTIS framework. The OTIS framework has several potential benefits. First, OTIS may enhance implementation scientists' consideration of organization-level constructs, which to date has been insufficient (17). Second, OTIS adds nuance to relatively limited conceptualizations of organizational influences in extant implementation determinant frameworks, such as the CFIR, EPIS, and TDF. Third, OTIS may increase the use of theories in implementation science. Evidence suggests that the use of theories, models, and frameworks in implementation science is inconsistent and often inappropriate (31). Unlike conceptual frameworks, which offer a menu of constructs thought to influence implementation, theoretical frameworks including OTIS are based on theories, which propose mechanisms underlying the relationship between constructs and implementation. OTIS links implementation scientists to theories that may explain the phenomena underlying complex implementation problems, such as slow uptake or poor sustainment. Future efforts should include expanding extant frameworks with OTIS's unique domains and constructs; refining OTIS's language to increase its accessibility to an audience without expertise in organization science; and revising OTIS to reflect empirical evidence.