Improving patient discharge and reducing hospital readmissions by using Intervention Mapping

We structured the problem analysis by using the PRECEDE PROCEED model [24] (see Additional file 1), to analyse and describe the scale, causes, and consequences of the health problem and to identify the target population.

In step 2, we identified the desired outcomes of the intervention and formulated specific performance objectives for the target population, such as writing a complete, accurate and timely discharge letter by the hospital physician. This resulted in a step-by-step checklist of what needs to be accomplished in order to obtain the desired outcomes [28].

It is important to identify what steps need to be tweaked in order to affect the performance objective, and ultimately the intervention outcome [28]. We identified the most important determinants (e.g., lack of knowledge and understanding between hospital and primary care providers) that need to be changed and combined these with performance objectives to formulate our change objectives. These change objectives specified who and what will change as a result of the intervention.

We selected theory-based methods that relate to the change objectives in step 2. These methods were required to change the behavioural and environmental determinants of ineffective hospital discharge. Subsequently, these methods were translated into practical strategies.

In this step, we provide suggestions for the design of the intervention by considering the target group and local setting [18]. The intervention studies identified in step 3 were classified independently by two researchers (GH and MZ) according to the Oxford Centre for Evidence-Based Medicine - Levels of Evidence from 2009 onward [33].

We made suggestions for developing an implementation plan for accomplishing program adoption, and for evaluating the effects and feasibility of the intervention program. The suggestions were based on literature regarding effective implementation strategies [17, 30, 34‐36], existing implementation toolboxes [37, 38], and a literature review on methods to evaluate complex interventions in health care [35, 39, 40].

The health problem and the underlying causes are presented in Figure 1. The published studies demonstrate that one in five patients experience an adverse event within 3 weeks after hospital discharge, of which one in three was considered preventable [41]. Three per cent of the adverse events led to permanent disability, including death. The one month unplanned readmission rates varied between 13% [42], and 20% [43]. Unnecessary hospital readmissions lead to considerable suffering, harm and extra costs. Friedman and Basu estimated hospital costs for preventable readmissions during 6 months at about $730 million [7]. Jencks estimated total hospital costs at $44 billion per year for rehospitalisations among Medicare patients within 30 days of hospital discharge [44].We found that ineffective handovers that lead to patient readmissions are caused by poor information exchange, poor coordination of care and poor communication between hospital and primary care providers, and between care providers and patients. The underlying causes include attitudinal and behavioural factors (e.g., lack of understanding of the needs of the counterpart, a distant relationship and a lack of collaborative attitude between hospital and primary care providers), organisational factors (e.g., lack of guidelines), technical factors (lack of a shared electronic information system) or patient factors (e.g., patients are less skilled or don’t dare to speak up). All the identified causes and their underlying factors are summarised in Figure 1.

Our literature review identified a raft of change methods, such as knowledge transfer, active listening and guided practice from the Social Cognitive Theory (SCT) [45], consciousness raising from the Transtheoretical Model [46, 47], shifting perspectives and interpersonal contact from the Intergroup Contact Theory [48] and standardised working processes from the SCT and Rational Decision-making theories [30] as influencers of the behavioural and environmental determinants of ineffective hospital discharge. Goal-setting and implementation intentions were derived from theories of Goal Directed Behaviour [49, 50], and multi-disciplinary collaboration and case management from theories of Integrated Care [51]. These theory-based methods were subsequently operationalised into practical strategies and corresponding activities and materials for the targeted population [52‐76] as shown in Table 3.

We formulated a wide variety of change objectives at the individual clinician and patient levels, the interpersonal level, organisational and technical levels that need to be considered in order to tackle ineffective handovers at discharge more reliably (Additional file 3). Given these change objectives the intervention likely needs to be multi-faceted and needs to be tailored to the needs encountered in the local setting. Table 3 shows a framework with examples of strategies and related materials and activities guiding healthcare providers and policy makers in the development of their intervention. The list of all identified strategies and related materials, level of evidence and references are available upon request.

Many interventions were evaluated in well-designed studies. For example, the use of standardised discharge practices such as the use of discharge letter templates, discharge planning guidelines and medication reconciliation checklists are effective strategies [65‐67, 70]. The use of a shared electronic patient information platform facilitates discharge communication between hospital and primary care providers [66, 71‐73]. There is evidence demonstrating that the patient’s role in the discharge process is enhanced by the provision of written and verbal discharge information and by assistance and guidance in self-management (e.g., discharge counselling, follow-up calls or home-based visits and a patient discharge record or question form) [74, 76]. However, many promising interventions have not been evaluated properly or were tested using weak study designs. For example, the effects of lectures and exercises on discharge practice in the medical curriculum, and regular group discussions involving hospital and primary care providers are largely unknown [52].

Moreover, there is limited evidence on the effects of reinforcement by using discharge planning reminders, mandatory administrative tasks or financial incentives and penalties [77].

Insight also lacks into the effects of strategies to increase care provider reflections on discharge practices (e.g., use of a standardised feedback form, video reflection, role play or simulation of discharge consultations) [52] and regarding the use of teach-back to check the patient’s understanding of their medical and medication history [63, 78].

Commitment from and ownership by the target group is essential to successful implementation [79, 80]. The awareness among end users is enhanced when they are directly involved in the development or modification of the innovation, in mounting the implementation plan, and in selecting the implementation strategies to be used [35].

Moreover, uptake of policies and protocols, reimbursement and the consideration of patients’ preferences are necessary for a sustainable implementation [81].

Strategies that address the barriers to change are required to implement interventions in daily practice [36]. Most theories on implementing interventions in health care emphasise that an analysis of the barriers to change practice is a prerequisite to selecting or developing an effective implementation strategy [17]. An implementation plan should be developed specifically after selecting the implementation strategies to tackle the barriers. This plan should be compatible with the target group and settings in which the implementation will take place. Good management and planning of implementation activities (i.e., what, when, where, how and by whom) also appears to be a requisite for successful implementation of innovations in patient care [35].

Formative and summative effect evaluation should be carried out using hospital readmission and adverse events as defined patient outcome effects to evaluate whether the intervention led to the desired degree of change. The formulated performance objectives in step 2 can be operationalised in measurable process indicators, for example by assessing the proportion of patients discharged with a complete discharge letter and assessing the proportion of patients discharged after medication reconciliation.

A process evaluation should be performed to understand the effect, success or failure of the intervention and to get an impression of its feasibility, generalizability and its acceptability in the target population. The process evaluation gives insight into the black box of the implementation process and can explain the variation in results in evaluating interventions. The activities carried out as part of the intervention, the actual exposure of participants to these activities, and their experience of these activities should be studied [40].