Table 2, Summary of framework approaches - Systematic Reviews of Complex Multicomponent Health Care Interventions

Purpose of the Systematic Review	Approach and Description	Example
Effectiveness of a class or bundle of interventions	Holistic: The intervention is considered a unit and is not disaggregated by component.	In a review of methods used for quality improvement effectiveness, authors categorized interventions into one of nine categories, including “multiple intervention” category, for those studies where more than one component was studied, but the effect of a component could not be discerned from the article. One multiple intervention consisted of an initial educational session to inform physicians of appropriate protocols for antibiotic prescribing for acute respiratory infections, complemented by exam room posters to encourage sustained use of guidelines. These intervention components were considered interdependent and disaggregation into component parts may have resulted in a loss of their cumulative effect.⁵
Effectiveness of interventions based on intervention features or other factors	Component: Group by intervention components or activities.	In a systematic review of fall prevention strategies, within “exercise and PT” authors included all trials that had exercise/PT as the sole intervention or part of a multicomponent intervention. Sensitivity analysis was done without the multicomponent intervention studies, and results were similar.⁸
Active component: Group by the intervention components considered key to intervention effectiveness.	Authors of a systematic review of nonpharmacologic interventions for post-discharge heart failure care categorized interventions by the primary component reported in each study, and analyzed studies by component. Primary intervention components included telephone follow-up only, increased visits to cardiology clinics, home visits, home visits and increased clinic visits, home telemonitoring, multidisciplinary care, and self-care instruction.⁹
Function: Group by the intended purpose of the intervention or components, rather than its structure or form.	A review of collaborative care interventions for depression defined three components: the introduction of a new role of the case manager into primary care, the introduction of mechanisms to foster closer liaison between primary care clinicians and mental health specialists; and introduction of mechanisms to collect and share information on the progress of individual patients.¹⁰
Other intervention characteristic: Intervention features such as breadth of coverage,¹¹^,¹² or by intensity, frequency, duration, or target.	Based on results of previous systematic reviews, authors of a review of interventions to prevent falls in older adults categorized the interventions as comprehensive if they included treatment and education to comprehensively address risks, conditions, or functional limitations identified through the assessment; or noncomprehensive if they provided only referral or limited treatment of selected risks, conditions, or functional limitations.⁸^,¹¹
Theory: Group conceptually similar interventions using theory underlying interventions.	An evidence review on models of health-related lifestyle advisors grouped interventions by underlying theory to assess the influence on intervention effectiveness. Interventions involving theoretical underpinning seemed to have no bearing on intervention success.⁷
Topic-specific typology: Classifies interventions into mutually exclusive categories based on characteristics specific to the topic.	For example, a review of heart failure organization of care after hospitalization grouped multicomponent interventions as a case management model which monitored patients by telephone and home visits; a clinical model which followed patients in specialist clinics; or a multidisciplinary model, which bridged the gap between hospital admission and discharge. Within these different models of care, they noted similarities of components, such as phone followup, education, self-management, and dietary advice.¹³
Context: Group by relevant social, political, environment, seasonal or other contexts that can act as an effect modifier.	A review of qualitative and quantitative literature focused on interventions to promote healthy eating in children. Authors identified children's attitudes about healthy eating, and these were used in sensitivity analysis to determine factors which affect outcomes. The authors then categorized studies by whether they addressed these factors. For example, authors found that interventions that reduced or removed any emphasis on health messages had larger increases in fruit and/or vegetables consumed by children.¹⁴
Implementation factors: Group by implementation features of the intervention such as adaptation and facilitation.	In a review of the effectiveness of home visiting programs, the authors found intentional changes and refinements to models over time, as well as other adaptations due to time and resource constraints, community norms, and characteristics of the target population. For this reason the authors reported adapted models separately from nonadapted models. They did not consider the “original” model and adapted model the same.¹⁵
Structure, process, outcome (Donabedian): The intervention and its environment as a system are organized into structure, process and outcome.	This model was used to frame a review comparing specialist to generalist care. The analytic framework was organized by structure (provider characteristics and potential confounders related to practice and patient characteristics), process (resource utilization, appropriateness of care, and process outcomes), and outcome (clinical outcomes, disability, satisfaction); and also indicated other mediating factors such as process mediators (attributes of primary care).¹⁶
Factors influencing the success or failure of an intervention	Realist approach: Focuses on the interaction between intervention mechanism, context and outcome to make statements about situations in which programs may be more likely to be effective.	A realist review was conducted on knowledge translation interventions to enable evidence –informed health care. This review resulted in the formation of a hypothesis: that change agents who are adequately supported and resourced (context), who role model the practices they espouse (mechanism), may impact more positively on achieving evidence-informed health care (outcome).¹⁷
Mechanisms of action: Makes explicit nonlinear relationships between the intervention, its components, context, and intermediate and final outcomes.	This systematic review focused on home-based environmental intervention for individuals with asthma. The analytic framework included the proposed mechanism of action for intervention components and illustrated how components of the multicomponent interventions were thought to reduce asthma morbidity through intersecting pathways: one with environmental assessment and remediation to change the physical environment, and a second via education to change behavior of a patient and household.¹⁸
Configurational: May identify necessary, sufficient, or both necessary and sufficient conditions for intervention success.	Using qualitative comparative analysis, authors explored the relationship between the patient-centered medical home and quality in 21 cases.¹⁹ The analysis found that provider performance reporting systems and diabetes team care were necessary but not sufficient conditions for good “optimal diabetes care” scores.

Purpose of the Systematic Review

Approach and Description

Example

Effectiveness of a class or bundle of interventions

Holistic: The intervention is considered a unit and is not disaggregated by component.

In a review of methods used for quality improvement effectiveness, authors categorized interventions into one of nine categories, including “multiple intervention” category, for those studies where more than one component was studied, but the effect of a component could not be discerned from the article. One multiple intervention consisted of an initial educational session to inform physicians of appropriate protocols for antibiotic prescribing for acute respiratory infections, complemented by exam room posters to encourage sustained use of guidelines. These intervention components were considered interdependent and disaggregation into component parts may have resulted in a loss of their cumulative effect.⁵

Effectiveness of interventions based on intervention features or other factors

Component: Group by intervention components or activities.

In a systematic review of fall prevention strategies, within “exercise and PT” authors included all trials that had exercise/PT as the sole intervention or part of a multicomponent intervention. Sensitivity analysis was done without the multicomponent intervention studies, and results were similar.⁸

Active component: Group by the intervention components considered key to intervention effectiveness.

Authors of a systematic review of nonpharmacologic interventions for post-discharge heart failure care categorized interventions by the primary component reported in each study, and analyzed studies by component. Primary intervention components included telephone follow-up only, increased visits to cardiology clinics, home visits, home visits and increased clinic visits, home telemonitoring, multidisciplinary care, and self-care instruction.⁹

Function: Group by the intended purpose of the intervention or components, rather than its structure or form.

A review of collaborative care interventions for depression defined three components: the introduction of a new role of the case manager into primary care, the introduction of mechanisms to foster closer liaison between primary care clinicians and mental health specialists; and introduction of mechanisms to collect and share information on the progress of individual patients.¹⁰

Other intervention characteristic: Intervention features such as breadth of coverage,¹¹^,¹² or by intensity, frequency, duration, or target.

Based on results of previous systematic reviews, authors of a review of interventions to prevent falls in older adults categorized the interventions as comprehensive if they included treatment and education to comprehensively address risks, conditions, or functional limitations identified through the assessment; or noncomprehensive if they provided only referral or limited treatment of selected risks, conditions, or functional limitations.⁸^,¹¹

Theory: Group conceptually similar interventions using theory underlying interventions.

An evidence review on models of health-related lifestyle advisors grouped interventions by underlying theory to assess the influence on intervention effectiveness. Interventions involving theoretical underpinning seemed to have no bearing on intervention success.⁷

Topic-specific typology: Classifies interventions into mutually exclusive categories based on characteristics specific to the topic.

For example, a review of heart failure organization of care after hospitalization grouped multicomponent interventions as a case management model which monitored patients by telephone and home visits; a clinical model which followed patients in specialist clinics; or a multidisciplinary model, which bridged the gap between hospital admission and discharge. Within these different models of care, they noted similarities of components, such as phone followup, education, self-management, and dietary advice.¹³

Context: Group by relevant social, political, environment, seasonal or other contexts that can act as an effect modifier.

A review of qualitative and quantitative literature focused on interventions to promote healthy eating in children. Authors identified children's attitudes about healthy eating, and these were used in sensitivity analysis to determine factors which affect outcomes. The authors then categorized studies by whether they addressed these factors. For example, authors found that interventions that reduced or removed any emphasis on health messages had larger increases in fruit and/or vegetables consumed by children.¹⁴

Implementation factors: Group by implementation features of the intervention such as adaptation and facilitation.

In a review of the effectiveness of home visiting programs, the authors found intentional changes and refinements to models over time, as well as other adaptations due to time and resource constraints, community norms, and characteristics of the target population. For this reason the authors reported adapted models separately from nonadapted models. They did not consider the “original” model and adapted model the same.¹⁵

Structure, process, outcome (Donabedian): The intervention and its environment as a system are organized into structure, process and outcome.

This model was used to frame a review comparing specialist to generalist care. The analytic framework was organized by structure (provider characteristics and potential confounders related to practice and patient characteristics), process (resource utilization, appropriateness of care, and process outcomes), and outcome (clinical outcomes, disability, satisfaction); and also indicated other mediating factors such as process mediators (attributes of primary care).¹⁶

Factors influencing the success or failure of an intervention

Realist approach: Focuses on the interaction between intervention mechanism, context and outcome to make statements about situations in which programs may be more likely to be effective.

A realist review was conducted on knowledge translation interventions to enable evidence –informed health care. This review resulted in the formation of a hypothesis: that change agents who are adequately supported and resourced (context), who role model the practices they espouse (mechanism), may impact more positively on achieving evidence-informed health care (outcome).¹⁷

Mechanisms of action: Makes explicit nonlinear relationships between the intervention, its components, context, and intermediate and final outcomes.

This systematic review focused on home-based environmental intervention for individuals with asthma. The analytic framework included the proposed mechanism of action for intervention components and illustrated how components of the multicomponent interventions were thought to reduce asthma morbidity through intersecting pathways: one with environmental assessment and remediation to change the physical environment, and a second via education to change behavior of a patient and household.¹⁸

Configurational: May identify necessary, sufficient, or both necessary and sufficient conditions for intervention success.

Using qualitative comparative analysis, authors explored the relationship between the patient-centered medical home and quality in 21 cases.¹⁹ The analysis found that provider performance reporting systems and diabetes team care were necessary but not sufficient conditions for good “optimal diabetes care” scores.

From: Results

Contents

Bookshelf

Systematic Reviews of Complex Multicomponent Health Care Interventions [Internet].

Table 2Summary of framework approaches

Views

Other titles in these collections

Recent Activity