Article Text

Download PDFPDF

Transition of care from adult intensive care settings – implementing interventions to improve medication safety and patient outcomes
  1. Suzanne McCarthy1,
  2. Raisa Laaksonen2,
  3. Virginia Silvari1,3
  1. 1School of Pharmacy, University College Cork, Cork, Ireland
  2. 2Faculty of Pharmacy, University of Helsinki, Helsinki, Finland
  3. 3Pharmacy Department, Cork University Hospital, Cork, Ireland
  1. Correspondence to Dr Raisa Laaksonen, Faculty of Pharmacy, University of Helsinki, 00014 Helsinki, Finland; raisa.laaksonen{at}

Statistics from

On admission to an intensive care unit (ICU), patients’ regular long-term medications may be withheld while they are being stabilised. Such medications are sometimes not restarted during the rest of their hospital stay, even when transferred to a lower acuity ward or discharged from hospital.1 This puts patients discharged from an ICU at higher risk of unintentional medication discontinuation, which could lead to future exacerbation of chronic conditions. Additionally, ICU patients may have medications commenced in the acute stage of their ICU admission (eg, gastric acid secretion inhibitors) that might inadvertently be continued following transfer from the ICU.2

There is a growing body of evidence that care transitions, whether from inpatient to outpatient settings, or within a hospital stay between different specialties or departments, pose an elevated risk of patients experiencing negative outcomes such as medication errors or adverse events.3 4 A systematic review suggests that across five studies, the median rate of medication errors following hospital discharge is 53% per adult discharged patient.5 However, less is known about medication errors in adults transferred from ICU to general hospital wards; the limited research available suggests high levels of medication errors associated with this transition point with 46%–74% of patients experiencing a medication error.6 7 Commonly occurring errors include continuation of medication indicated only in the ICU, untreated indications and medications without an indication.6 There is a need to understand what interventions can be used to reduce medication errors, and the effectiveness of these interventions, when transitioning patients from the ICU setting.

Medication-related interventions to improve outcomes of care transition from intensive care

In this issue of BMJ Quality & Safety, Bourne and colleagues’ systematic review and meta-analysis is a welcome contribution to examine the impact of medication-related interventions on medication and patient outcomes for adult ICU patients transitioning to a hospital ward, as well as barriers and facilitators to their implementation.8 Seventeen studies evaluating such interventions were retrieved in October 2020, of which nine were single component interventions (education of staff; medication review; guidelines; electronic transfer/hand-over checklist or letter; or medicines reconciliation) and eight were multicomponent interventions (mainly education of staff; guidelines; and medication review). Bourne and colleagues used ROB 2.0 (Risk of Bias) and ROBINS-I (Risk of Bias in Non-Randomized Studies - of Interventions) Cochrane tools to assess risk of bias for randomised controlled trials and non-randomised studies, respectively; the risk of bias was serious for most studies. The overall quality of the body of evidence in the systematic review was low,8 as assessed by GRADE (Grading of Recommendations, Assessment, Development and Evaluations).9

Meta-analysis was undertaken for deprescribing interventions at ICU and hospital discharge points, revealing reductions in both the risk of inappropriate medication continuation at ICU discharge (odds ratio (OR)=0.45 (95%CI 0.31-0.63), n=9 studies) and hospital discharge (OR=0.39 (0.2-0.76), n=9 studies.8 Multicomponent deprescribing interventions (education of staff combined with guideline implementation, n=6 studies) were most effective to reduce the risk of inappropriate medication continuation at hospital discharge compared with usual care (OR=0.26 (0.13 to 0.55)). The GRADE assessment for these interventions was moderate-quality evidence. This effect was not observed in studies assessing the impact at ICU discharge (OR=0.5 (0.22 to 1.11), n=4 studies, low-quality evidence using GRADE assessment). Single component deprescribing interventions (education of staff, medication reconciliation or medication review) were not found to have an impact at hospital discharge (OR=0.77 (0.16 to 3.74), n=3 studies, very low-quality evidence using GRADE assessment). However, single component interventions did reduce inappropriate medication continuation at ICU discharge (OR=0.42 (0.24 to 0.74), n=5 studies, low-quality evidence using GRADE assessment).

Notwithstanding the inherent risk of bias with the vast majority of these studies,8 these findings highlight the complexity of implementing and assessing the impact of both single and multicomponent interventions that take place at one point during the hospital stay (for example, during an ICU admission or just prior to ICU discharge), on an outcome later in the patient journey (such as at hospital discharge). Following ICU discharge, other changes can occur to the patient’s care journey (eg, input from other healthcare professionals, changes in the patient’s status, changes to medications); and the further away the time point evaluating the effect of the intervention (eg, hospital discharge), the more likely it is that other factors may interfere with the effect of earlier interventions made in the ICU and induce bias in the intervention effect.

Bourne and colleagues report that eight studies reported comparisons of patient outcomes:8 one study demonstrated a significant reduction of potential adverse drug events, with risk of bias assessed as moderate. No impact on mortality, ICU readmission or hospital length of stay was observed in the other studies, with moderate (n=5 studies) and serious (n=2 studies) risk of bias.8 Does this mean that such interventions are of little value? Not necessarily, as the studies included in the systematic review did demonstrate positive effects on other outcomes such as reductions in drug-related problems, which in turn may reduce pill burden for the patient, reduce waste and reduce costs, all of which are important outcomes from such interventions.8 Evaluating the impact of medication-related interventions on patient outcomes can be difficult and requires careful consideration. While strategies such as medication reconciliation, guidelines and medication review are widely endorsed by health systems, the data presented by Bourne and colleagues8 suggest that we do not currently have sufficient high-quality evidence to confirm their impact on patient outcomes at ICU discharge.

Implementation issues for interventions on transition from intensive care

It is documented in the literature that implementation of evidence-based practices in the ICU setting is often suboptimal; the challenges include staffing, team structures, and workflow and time pressures associated with the care of a critically ill patient.10 Implementation science advocates that successful integration of evidence-based interventions must be accompanied by knowledge of implementation.11 To guide and support practitioners implementing interventions locally, Powell and colleagues have developed a compilation of 73 implementation strategies (the Expert Recommendations for Implementing Change strategies).12 When deciding on which intervention(s) reviewed by Bourne and colleagues8 should be implemented, the multidisciplinary team may consider how different intervention strategies can address the contextual needs of their intensive care setting. As part of this process, teams need to assess the factors affecting the implementation process and outcomes, including factors such as the characteristics of the innovation (eg, medicines reconciliation, deprescribing, education, etc), the characteristics of the setting, potential barriers and facilitators and the preferences of stakeholders.13 From the different medication safety interventions and their outcomes in the ICU setting reviewed by Bourne and colleagues,8 practitioners should select appropriate single or multiple interventions that map to their identified local deficits, for example, education and training to address knowledge and skill gaps, or recruiting trained staff, including pharmacists, if there is a workforce issue.14

The contextual barriers of the ICU setting need to be considered, how it differs from a general ward setting as well as within and between countries. Much of the research included in the systematic review was conducted in North America, mostly in the USA, with only three studies in Europe and one in Australia.8 Healthcare systems, resources, delivery and practice vary not only between these countries but also between countries not represented in the review. In addition, local medication safety research is required to identify local medication safety improvement needs and assess relevant barriers in that context before implementing the findings of any research. Thus, when considering the generalisability of the findings of this systematic review by Bourne and colleagues, it is important to consider the ICU context of the study sites in relation to one’s own ICU context.12 15

Bourne and colleagues reported that few studies explored the barriers to intervention implementation in the ICU context, with the most common being increased workload associated with the intervention itself.8 Multiprofessional collaboration was reported as a barrier to intervention delivery when there is limited collaboration,8 in which case increasing multiprofessional collaboration might be one of the interventions. Teamwork and communication are inherent to the safe and effective delivery of care to ICU patients and on their handover to other medical teams within the hospital setting.16 Future research should, therefore, consider how multiple professions can be effectively targeted by medication safety interventions.

Engaging stakeholders in the design and implementation of interventions is vital as there may be many barriers that, unless considered, would delay the translation of evidence into practice.10 17 Stakeholders include the many members of the multidisciplinary healthcare team who care for patients in the ICU and their transfer from the ICU, including clinicians, nursing staff and pharmacy staff. The European Association of Hospital Pharmacists has established a multiprofessional Special Interest Group for the Investigation of Medication Errors in Intensive Care Units; this group will be investigating the views and experiences of healthcare professionals working in the ICU regarding medication safety interventions. This work will support the development of recommendations for improving medication safety in the ICU setting across Europe.

Another important stakeholder group to consider is patients as well as their families and carers. The WHO Report on Medication Safety in Transitions of Care highlights the importance of engaging with, and empowering patients, families and caregivers so that they can be involved in ‘preventing, identifying and taking early action to minimise harm associated with medication’.18 It has been reported that patient involvement in care transitions is often lacking.19 A qualitative study reporting on the patient perspective of the discharge process concluded that early patient involvement and communication were needed to support the needs of patients at discharge.20 Bourne and colleagues also highlighted that patients and their family or carers were often not involved in the transition process from the ICU setting to other wards.8 As patients in the ICU are often critically ill and may not be in a position to participate in the transition process, it is important that medication safety initiatives also consider engaging with families and caregivers; future research should consider all these stakeholders in co-designing effective ICU practices.21 Action research has previously been used to address problems and challenges in ICU settings, involving healthcare professionals, mainly nurses, and patients and family.22 The spiral process of ‘planning, action, observation and reflection’ where ‘reflection’ leads back to ‘planning’, starting a new cycle could similarly be used to develop interventions to prevent medication errors.23


Bourne and colleagues have highlighted that better-designed studies are necessary to evaluate which medication-related interventions, and in which combination, should be put in place to decrease the risk of negative clinical outcomes during care transitions from the ICU.8 Future studies are also needed on how to implement such interventions in practice and overcome existing barriers.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.



  • SM and RL contributed equally.

  • Contributors SM and RL contributed equally to this paper and are to be regarded as first authors.

  • Funding The EAHP has received funding from BD (Becton, Dickinson and Company) for the running of the SIG. However, the SIG is independent of the pharmaceutical company providing research funding to the EAHP.

  • Disclaimer The funder of the SIG has had no involvement in the writing of the editorial or in the decision to submit it for publication.

  • Competing interests RL and VS are co-chairs and SM is a member of the EAHP SIG for the Investigation of Medication Errors in Intensive Care Units.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles