Translating evidence to patient care through caregivers: a systematic review of caregiver-mediated interventions

Patient- and family-centered care (PFCC) research, where patients and families partner with providers in the planning and delivery of healthcare, represents a shift from solely studying provider-identified priorities [1, 2]. PFCC represents a cultural change in healthcare that acknowledges that patients and their caregivers are central figures in decision-making and the delivery of care [1, 2]. Perceptions of PFCC and its benefits are high among patients and family members, though there are barriers to its implementation, including a purported lack of provider knowledge of PFCC, system barriers (e.g., lack of personnel, visiting hours), and a perceived increase in provider burden [3‐5]. The uptake of PFCC interventions is limited by these barriers, and the science of PFCC is lagging behind scientific, clinical, operational, and public interest and desire.

PFCC may provide a mechanism for quality improvement initiatives designed to implement evidence-informed practices into patient care. Patients and caregivers can act as agents of change, making evidence more accessible and providing a method for implementing science into patient care, regardless of the clinical setting [2]. Due to the nature of their illness and the use of invasive therapies, some patients may not be able to act as active agents of change [6]; children, and a growing number of adults with complex care needs, represent populations who require a caregiver. This presents a unique role for caregivers to be engaged in PFCC, though the evidence supporting the delivery and effectiveness of caregiver-mediated interventions is limited. Because of the growing role of patients and caregivers in care, there is an opportunity to engage and empower them to promote the uptake of science into patient care and the practice of evidence-informed medicine.

In order to effectively carry out PFCC, we must create, implement, and measure the effect of PFCC interventions. A recent systematic review of patient-mediated interventions suggested benefit [7]. There is a large and diverse literature on caregiver-mediated interventions, though no systematic review has been conducted. We therefore asked the question: Are caregiver-mediated interventions (interventions delivered in some part by the caregiver) effective in improving patient, caregiver, provider, or health system outcomes?

This systematic review was conducted and reported per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) criteria [8] (see Additional file 1 for the checklist). An a priori protocol was published online through PROSPERO (registration number CRD42016052509; Additional file 2). Ethical approval was not required, as previously published data were used.

We utilized systematic review methodology to synthesize the literature on the effectiveness of caregiver-mediated interventions. There is a large and recent body of literature (including many randomized controlled trials [RCTs]) suggesting that complex interventions comprised of inform-activate and inform-activate-collaborate engagement strategies are associated with improved patient and caregiver outcomes. In contrast, there is very little literature regarding provider and health system outcomes, though all interventions involved care providers and/or the health system in some way. The potential for unintended consequences or harm from caregiver-mediated interventions was not explored in the included studies, and much of the literature focused on the parent-child relationship.

That interventions with an inform engagement strategy always included parts of activate and/or collaborate engagement strategies is noteworthy, as research suggests that education-only interventions (inform in this language) in various settings and populations are some of the weakest [68‐71]. Education with prompts for action or education with active collaboration may circumvent the limitations of education alone [68‐71]. Most interventions were mediated by parents or family caregivers (any family member, though not defined specifically), though there was heterogeneity in both the type of caregiver included and in how caregiver was defined. It is important to recognize that one caregiver is not the same as another; they may differ in their commitment, readiness, and capacity to fulfill the role. Validated tools to assess caregiver readiness (Willingness to Care Scale [72], Caregiver Self Expectations Inventory [73]) should be employed alongside caregiver-mediated interventions in this context [74]. Interventions to support caregivers once they are already in a caregiving role dominate the literature [75‐77], with little or no information supporting the use of interventions to prepare the caregiver prior to entering the caregiving role. Nearly all studies of adults found improvement in both patient and caregiver outcomes, though in pediatric studies only half found improvement in patient and caregiver outcomes. Improved outcomes may be dependent on whether the caregiver-patient relationship is parent-child (i.e., parents providing support to children) versus child-parent (i.e., children providing support to parents).

All included interventions were complex (including multiple components). Thus, we are uncertain of the ”active ingredients” of the interventions and how these components affect outcomes. The knowledge translation literature suggests that complex interventions may be no better than single-component interventions [68, 69]. If this holds true for caregiver-mediated interventions, then single-component interventions may represent a resource-efficient means of improving patient, caregiver, provider, and health system outcomes. We observed variation in reported outcomes according to the types of engagement pairings, suggesting that different combinations of interventions might have differing effects: the previously observed effect of simple-component versus complex interventions may not hold true for caregiver-mediated interventions. Inform-activate engagement strategies provide the strongest evidence for improving patient outcomes through caregiver-mediated interventions. Based on the results of the current study, we recommend interventions that provide both condition-specific information and prompts or tools to support action for managing a condition to the caregiver. For example, an intervention to encourage early mobilization of patients recovering from a serious illness might include both information about how the illness affects the patient’s conditioning (inform) paired with instructions and prompts for caregiver-directed bedside physical therapy (activate). It is unclear how these interventions will apply in other contexts, though they appear to work in children and adults with any family caregiver in an inpatient setting. There was a trend towards more positive than equivocal effects for improved caregiver outcomes in an outpatient setting. Inform-activate-collaborate engagement strategies are supported by the strongest evidence to improve caregiver outcomes. We recommend utilizing this type of strategy for infants or children in outpatient or community settings with any family caregiver. For example, there could be a three-component (inform-activate-collaborate) telephone coaching intervention for parents of children with asthma (activate) combined with lifestyle advice on how to avoid asthma triggers (inform) and the provision of contact information for extra information or clinical supports (collaborate). These interventions have the potential to improve caregiver outcomes. At this time, there have been few evaluations of interventions using inform-collaborate (e.g., providing information to support activities of daily living and a list of available resources) or activate-collaborate (e.g., creating action plans specific to a medical condition and the provision of contact information for extra support) engagement strategies. For interventions conducted in the community (either partially or fully), it appears that collaborate engagement supports are necessary to improve outcomes.

Despite the large body of literature, there are a number of key gaps that must be addressed: provider and health system outcomes, the role of caregiver, the potential harms of caregiver-mediated interventions; and the economic impact of these interventions. Only two studies explored provider outcomes of caregiver-mediated interventions. Providers represent an important mechanism for implementing PFCC research [4, 78, 79]; it is essential to measure the impact that caregiver-mediated interventions have on this population. Ten studies reported on health system outcomes, which included hospital and emergency room visits and associated costs. As all caregiver-mediated interventions engage with the healthcare system, it is important to measure the impact they have on these limited resources. Within our aging population, investigations should explore whether the inverse effect of the parent-child dyad applies to caregiver-mediated interventions—that is, when children become the caregiver, are the same outcomes observed? The evidence from parent-child relationships may not inform the child-parent caregiver relationship in this context. No included studies reported on potential harms of caregiver-mediated interventions (e.g., falls among patients being mobilized by caregivers, caregiver feelings of guilt related to intervention adherence, conflict with providers); unintended effects of these interventions are an important area that requires further study. Caregiver-mediated interventions represent an opportunity for healthcare cost savings by transferring select care responsibilities from the clinical team to the caregiver. Future studies should assess the financial impact of employing caregiver-mediated interventions.

This study has a number of strengths, including the use of rigorous methodology defined by an a priori published protocol and adherence to recommended reporting criteria. We utilized previously validated tools to assess both the quality and strength of the evidence presented. The search was executed in five bibliographic databases, it was not restricted by language, and it was intentionally broad to ensure that caregiver-mediated interventions across all areas of healthcare were included. By focusing on healthcare settings, home-based interventions were excluded, including many studies of caregiver-mediated interventions in persons with dementia. In spite of these strengths, there are limitations to note. Our findings are limited by the methodological quality of the included studies; the risk of bias was generally high, and the overall quality of the evidence was low-moderate. When restricted to studies of at least moderate quality, there was a trend towards observing more positive effects. Risk of bias tends to be high in studies where blinding does not occur, as is the case in many of these interventions, and does not necessarily mean they are methodologically poor [80]. The literature is rich with RCTs, the quality of which is limited by methodological factors inherent in this type of research (e.g., lack of blinding) that are unlikely to be overcome in the future. We did not search the gray literature and could have missed studies, though our comprehensive search strategy included databases indexing both North American and European journals and full-text hand-searching was completed. We noted a lack of scientific abstracts, limited use of applicable keywords, and variable terminology to describe caregiver-mediated interventions in the included studies. This highlights the importance of full-text hand-searching, as these papers may not be found through standard database searching. All included papers reported on complex interventions, and thus it is not possible to determine which parts of the interventions were effective. The lack of effect in some studies may be a failure of implementation (few process evaluations conducted to inform) or a failure of measurement (i.e., use of the wrong measure) and not the intervention itself.

There is a large body of research, including many RCTs, to support the use of caregiver-mediated interventions that inform and activate caregivers to improve patient and caregiver outcomes. Additional study is needed to evaluate provider and health system outcomes, interventions for relationships outside of the parent-child role, potential harms, and economic implications. We recommend consideration of inform-activate and inform-activate-collaborate interventions when implementing patient and family-oriented research.