Evaluating applications of the CFIR in low- and middle-income - - PowerPoint PPT Presentation

Evaluating applications of the CFIR in low- and middle-income countries

Arianna Rubin Means, Christopher Kemp, Marie Claire Gwayi-Chore, Sarah Gimbel, Kenneth Sherr, Brad Wagenaar, Judith Wasserheit, Bryan J. Weiner December 3, 2018

Introduction to the CFIR

Characteristics of the intervention

• Intervention

source

• Evidence strength

and quality

• Relative

advantage

• Adaptability
• Trialability
• Complexity
• Design quality
• Cost

Inner Setting

• Structural

characteristics

• Networks and

communications

• Culture
• Implementation

climate Outer Setting

• Patient needs and

resources

• Cosmopolitanism
• Peer pressure
• External policies

and incentives Individuals involved

• Knowledge and

beliefs about the intervention

• Self-efficacy
• Individual stage of

change

• Individual

identification with

• rganization
• Other personal

attributes Implementation process

• Planning
• Engaging
• Execution
• Reflection and

evaluation

Source: Damschroder, Laura J., et al. "Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science." Implementation science 4.1 (2009): 50.

Rapid Growth in Use of CFIR

20 40 60 80 2009 2011 2013 2015 2017 Publications with CFIR in Title/Abstract

HICs

• More resources
• Individualist
• Multiple health systems
• Not donor-driven

LMICs

• Fewer resources
• Collectivist
• National health system
• Donor-driven

Objectives

• Characterize application of the CFIR in LMICs
• Assess compatibility and relevance of CFIR domains

and constructs in these settings

• Identify opportunities to refine the CFIR to optimize utility

in these settings

Methods

• Systematic review
• February 1, 2018
• SCOPUS
• PubMed, EMBASE, CINAHL, PsycINFO
• Included peer-reviewed studies using CFIR in LMICs
• Data abstraction informed by Kirk et al. (2015)1
• Survey to corresponding authors:

1) Why they chose the CFIR as a guiding framework 2) Which domains and constructs were compatible, incompatible, or irrelevant and why 3) Ways in which the authors believe that the CFIR could be optimized or updated for use in global health contexts, if any

1Kirk et al. 2015. A systematic review of the use of the consolidated framework for implementation research.

Results: PRISMA flow diagram

Databases 374 Non-duplicates 166 Screened 166 Excluded 68 Full-text review 98 Excluded 71 48 Did not use CFIR 10 Not LMIC 5 Not peer reviewed 4 Multiple reasons 4 Not primary research Studies included 27

Results: study descriptive statistics

• 20 LMICs and territories
• Diverse health topics

– Chronic disease, clinical practice guidelines generally, health policy, hepatitis C, HIV, immunizations, maternal health, obesity, pediatric inpatient care, pediatric mental health, primary healthcare, surgery, tobacco cessation, and tuberculosis

• Most applied CFIR post-implementation (14, 52%)
• Most used CFIR to:

– Guide data analysis (14, 52%) – Contextualize findings (14, 52%)

• Only 2 (7%) investigated outcomes linked to CFIR constructs
• Unit of analysis:

– Health providers (13, 48%) – Organizations (11, 41%)

Results: constructs used

2 4 6 8 10

Reflecting & Evaluating Planning Opinion Leaders Formally Appointed Internal… External Change Agents Executing Engaging Champions Self-efficacy Other Personal Attributes Knowledge & Beliefs about the… Individual Stage of Change Individual Identification with Organization Peer Pressure Patient Needs & Resources External Policy & Incentives Cosmopolitanism Tension for Change Structural Characteristics Relative Priority Organizational Incentives & Rewards Networks & Communications Learning Climate Leadership Engagement Goals and Feedback Culture Compatibility Available Resources Access to Knowledge & Information Trialability Relative Advantage Intervention Source Evidence Strength & Quality Design Quality & Packaging Cost Complexity Adaptability

Intervention characteristics Inner setting Outer setting Characteristics of individuals Process of implementation

Results: construct compatibility

Domain Construct Compatible (n) Incompatible (n) Irrelevant (n) Intervention Adaptability 16 1 Complexity 16 1 Cost 12 2 3 Design Quality & Packaging 12 2 3 Evidence Strength & Quality 13 3 1 Intervention Source 16 1 Relative Advantage 10 2 5 Trialability 9 3 5 Inner setting Culture 17 Implementation Climate 16 1 Networks & Communications 12 2 3 Readiness for Implementation 15 1 1 Structural Characteristics 14 3 Outer setting Cosmopolitanism 12 1 4 External Policy & Incentives 16 1 Patient Needs & Resources 9 6 2 Peer Pressure 14 1 2 Characteristics of individuals Individual Identification with Org 10 4 2 Individual Stage of Change 9 5 2 Knowledge & Beliefs about Interv. 16 Other Personal Attributes 10 4 2 Self-efficacy 13 2 1 Process Engaging 17 Executing 16 1 Planning 16 1 Reflecting & Evaluating 16 1

Results: incompatible constructs

Domain Construct Author feedback examples Outer setting Patient Needs & Resources Health care settings in these contexts are not patient-centered or do not have that focus. Therefore, it's difficult to apply this construct. No patients are involved in the intervention at the level of the health system in LMICs Characteristics

• f individuals

Individual Stage of Change The concept of individuality within the health care team was not compatible with the countries and settings where this study was conducted These interventions look not at individuals but at teams and organizations involved in the implementation

Results: CFIR relevance in LMICs

Circumstances in which CFIR is not relevant in LMICs In some settings health policy decisions are made from top down, and recipient will not have much

• ption nor alternatives. In such conditions, CFIR

individual and process domains might reflect skewed and over optimistic results Contexts vary largely such as the health systems and not only internally but the social norms, culture of the people and the political environment/economy...Therefore, it might be good to consider the macro-level factors as well

Results: adaptations or improvements

Possible adaptations or improvements to the CFIR in LMICs It will be good if the CFIR can communicate more

• n how it can be used or applied in larger scale of

actions such as implementation of national policy and strategy, not only at an intervention level I think there should be an emphasis on resource constraints as they exist in both the inner and

• uter setting. This was a common theme that is

not very explicitly presented in the CFIR

Results: additions to domains & constructs

Domains or constructs that should be added to increase relevancy of CFIR in LMICs More systems-based domains and constructs could be added in response to national and global actions such as accountability, governance & politics (both national and international) and legal and regulatory process. These factors play an important role in influencing the implementation of national policy It would be excellent if it could be adapted for use in researching health systems. In addition, if rather than, individuals there could be a domain for teams…I believe adding the domain of collective efficacy to characteristics of individuals would be useful

Results: additions to domains & constructs

Domains or constructs that should be added to increase relevancy of CFIR in LMICs More systems-based domains and constructs could be added in response to national and global actions such as accountability, governance & politics (both national and international) and legal and regulatory process. These factors play an important role in influencing the implementation of national policy It would be excellent if it could be adapted for use in researching health systems. In addition, if rather than, individuals there could be a domain for teams…I believe adding the domain of collective efficacy to characteristics of individuals would be useful. But also: I don't think any more should be added, it is already complex enough

Main messages from author responses

Implementation research in LMICs often requires study of health systems, at multiple operational and administrative levels that do not necessarily fit within existing Inner and Outer setting domains Many LMICs have hierarchical structures that are not compatible with Western conceptualizations of the role of the individual in affecting implementation Due to the resource constraints facing LMICs, sustainability

• f interventions as well as resource source and flow must

also be accounted for

1 2 3

Limitations

• Missing perspectives of those who chose not to use CFIR
• No comparison against other determinants frameworks

– e.g. COACH, developed for use in LMICs2

2Bergstrom et al. 2015. Health system context and implementation of evidence-based practices…

Adaptable periphery Core components

INTERVENTION (unadapted) INDIVIDUALS INVOLVED CHARACTERISTICS OF SYSTEMS

Adaptable periphery Core components

INTERVENTION (adapted) OUTER SETTING PROCESS INNER SETTING

CFIR (now)

CFIR (adapted for LMICs)

Adaptable periphery Core components

INTERVENTION (unadapted) INDIVIDUALS INVOLVED CHARACTERISTICS OF SYSTEMS

Adaptable periphery Core components

INTERVENTION (adapted) OUTER SETTING PROCESS INNER SETTING

Additional constructs for use in LMICs

Construct Definition Authority norms Perceived political, social, or administrative norms or hierarchies Implementation level The specific administrative level(s) that do or do not interact to influence implementation Donor priorities Stakeholders’ perception regarding the degree to which donor preferences and priorities influence implementation Community characteristics The extent to which community characteristics affect the willingness or ability for

• rganizations to engage in implementation

Sustainability The observed or expected likelihood for implementation to take place with consistent or improved outcomes over relevant durations of time Scalability The observed or expected likelihood for implementation to be replicated across heterogenous geographic or practice settings Strategic policy alignment The degree to which health system policies are aligned with perceived needs and priorities of relevant stakeholders Resource continuity The presence of sufficient resources (financial, human, or material) over durations of time necessary for ongoing implementation at scale Resource source The origin of available resources used to test, launch, and sustain implementation

NEW Characteristics of Systems Domain

Additional constructs for use in LMICs

Domain Construct Definition Inner Setting Team dynamics Socio-cultural norms for team work and task allocation within an organization Characteristics of Intervention Feasibility* Stakeholders’ perceptions of the extent to which implementation can effectively take place within a given setting due to both inherent and external facilitators and barriers Characteristics of Intervention Workload capacity The degree to which an intervention is or is not compatible with current work or patient loads of organizations or care teams Process of implementation Decision making* The type, duration and timing of the activities involved in making decisions about the intervention.

NEW constructs to add to existing domains

* Constructs also appear in Smith 2014, a methods report prepared by RTI for AHRQ that adapted the CFIR for three complex system interventions involving (1) process redesign for improved efficiency and reduced costs, (2) patient-centered medical homes, and (3) care transitions

Conclusion

• The CFIR is a useful framework for global IS practitioners as it allows for

identification of implementation facilitators and barriers across settings

• 27 studies have used the CFIR in LMICs, the majority of which only

applied the framework post data collection for data analysis/interpretation. This application is similar to findings from studies in HICs (Kirk 2015).

• Constructs identified as more or less useful by authors often align with

unique attributes of LMICs versus HICs (e.g. collectivist versus individualistic societies, low resource versus higher resourced settings, etc)

• Our proposed Characteristics of Systems domain would provide

global IS practitioners opportunities to account for health systems-level facilitators and barriers independent of the implementing organization

Thank you!

Questions? aerubin@uw.edu

Adoption Implementation Scale-up

Implementation research in LMICs

Influenced by available evidence base, intervention characteristics, etc. Influenced by

• rganizational

willingness, presence

• f champions, etc.

Influenced by resource availability, national, regional, and global policies, etc.

Comparison with Kirk et al. (2015)

• 20%
• 10%

0% 10% 20%

Reflecting & Evaluating Planning Opinion Leaders Formally Appointed Internal Implementation Leaders External Change Agents Executing Engaging Champions Self-efficacy Other Personal Attributes Knowledge & Beliefs about the Intervention Individual Stage of Change Individual Identification with Organization Peer Pressure Patient Needs & Resources External Policy & Incentives Cosmopolitanism Tension for Change Structural Characteristics Relative Priority Organizational Incentives & Rewards Networks & Communications Learning Climate Leadership Engagement Goals and Feedback Culture Compatibility Available Resources Access to Knowledge & Information Trialability Relative Advantage Intervention Source Evidence Strength & Quality Design Quality & Packaging Cost Complexity Adaptability

Intervention characteristics Inner setting Outer setting Characteristics of individuals Process of implementation