The WHO views the safe management of medicines as a global challenge.1 In the UK, guidance stresses the importance of improving the way medicines are managed at care transitions, such as admission to and discharge from hospital.2 However, there is continued evidence that systems managing medicines are not optimally calibrated, particularly during and after hospital discharge,3 when the responsibility for patient care shifts across organisations and clinicians.4

Medicines management is a system that supports the therapeutic use of medicines by patients, involving multiple healthcare organisations and staff with different clinical specialties and professional roles.5 There is no shortage of evidence about the points at which healthcare systems fail to provide safe care.6–9 Patients are not always well prepared to leave hospital and self-manage their ongoing treatment.10 The effective transfer of sufficient and accurate information between healthcare organisations remains inadequate in many cases4 and is compounded by boundaries between care providers who may not always have access to the same information about patients’ health. It is then unsurprising that discrepancies arise between medicines lists held by different care providers and patients.3 11 12

Heart failure (HF) is a chronic condition affecting 900 000 individuals in the UK and is projected to rise significantly with an ageing population.13 HF is a costly condition for the National Health Service (NHS) and is characterised by high rates of readmissions.14 HF symptoms and disease progression can be controlled through well-managed medicines; however, guidelines for their use are not always applied and cardiology medicines can also cause harm, such as kidney injury, if they are not monitored.15 Hence, the optimal management of medicines when leaving hospital is crucial to enhance quality of life, manage symptoms, prevent deterioration leading to hospital readmission and reduce mortality.16

Current thinking in patient safety has shifted focus from the deconstruction of events leading up to safety incidents (Safety I) to a more positive and proactive view of healthcare systems that identifies and values what goes right as well as pinpointing what goes wrong (Safety II).17 18 Thus Safety II focuses on preventing error while accepting that there is variability in the delivery of healthcare, acknowledging that patients do not always experience harm as a consequence of their care. Instead of reacting when things go wrong, organisations proactively anticipate developments, negative as well as positive. It offers recognition of good performance in the face of uncertainty, valuing flexibility, adaptability, foresight and knowledge of how systems operate.17 19 This in turn promotes a more dynamic attitude to performance through resilience which is the ability for a system and the individuals therein to adjust prior to, during or following changes or disturbances or in the face of ongoing, sustained pressure.18 20–22

This concept of resilience in healthcare has looked at specific points in the patient pathway such as handover of care between staff in one location such as a ward or performing specific roles.19 23 24 Only one previous study has explored how patients can enhance resilience in medicines management at and after hospital discharge through anticipating discrepancies and taking remedial action.25 No studies to date have explored resilience in medicines management at this care transition from multiple perspectives, including staff and patients and across different healthcare economies.

This study aimed to address this evidence gap by systematically investigating resilience in the medicines management system, using HF as an example condition. More specifically, the study was designed to understand how the system compensates for weaknesses and maximises opportunities in order to deliver safe yet optimal treatment. Its objectives were to explore the system of medicines management in multiple healthcare economies to highlight where resilience exists and identify where improvements to the system can be made to enhance resilience.

We used a mixed-methods design in four healthcare economies and their local primary care organisations (one comprising two hospitals and three comprising one hospital) in the north of England. Sites were selected to include university teaching hospitals and non-university teaching hospitals in different areas. Data from site observations, staff and patient qualitative interviews and documentary analysis (discharge letters and organisational and national policies) were collected between June 2016 and March 2017. NHS research ethics committee approval was sought and granted (16/NS/0018).

A total of 55 interviews with 20 patients with HF were conducted: 19 at discharge or shortly afterwards (time point 1); 19 approximately 2 weeks after discharge (time point 2); and 17 approximately 6 weeks after discharge (time point 3). We were unable to contact one patient from site 1 at time point 1; at site 2 one patient withdrew from the study after the first interview. One patient from site 3 was not interviewed at the third time point due to hospital readmission, and at site 4 one patient was too ill to continue after the second interview. Table 1 presents the number of patients interviewed for each site at the different time points. Table 2 outlines the gender and age of interviewed patients.

Forty-five interviews (table 3) were conducted with healthcare professionals: 19 with primary care staff (15 in four general practitioner (GP) surgeries, two community pharmacists and two community HF nurses) and 26 with secondary care staff. Table 4 presents the number of healthcare staff interviewed by site.

We identified four primary inter-related themes concerning organisational resilience and termed these: gaps, traps, bridges and props. Examples representing each theme are shown in tables 5–8.

‘Gaps’ were defined as a discontinuity in key processes that form the medicines management system and had the potential to result in poorly optimised medicines. Approaches to preparing discharge information varied across sites, with information sometimes being missed due to a lack of preparation time. Gaps were also evident in the information shared with primary care and in the preparation of patients to use their medicines. For the latter, we identified no standardised processes for informing patients about their medicines and, while hospital policies stipulated that patients should be informed, and gave details of the types of information patients should have, there was no guidance on optimal methods for informing patients about their medicines or training, so patients’ experiences of receiving medicines were inconsistent and information was deficient for some.

After discharge we found gaps in the continuity of care, for example, not all patients had a community pharmacy Medicines Use Review because pharmacies did not routinely receive information about the patients’ medicines at discharge. Waiting times for specialist staff follow-up varied considerably and were sometimes lengthy, for example, waiting times for an appointment with community HF specialist nurses who would manage medicines titration was sometimes as long as 3 months after discharge.

We defined ‘traps’ as features of the way the medicines management system was designed or managed that might produce medication errors defined as a ‘failure in the treatment process that leads to, or has the potential to lead to, harm to the patient’29 30 or unintended adverse medication events. These were evident in the coordination of discharges, for example, the pressure on ward staff to expedite discharges as quickly as possible appeared to impact on the effective preparation of discharge information and on educating patients about their discharge medicines.

Staff preparing discharge information were often interrupted, could not always locate patients’ notes and none reported receiving training about safe practices with medicines at discharge to primary care. We also found error traps after discharge, such as a lack of time and resources in GP surgeries to process discharge information. Finally, there was evidence that patients’ lack of knowledge about the purpose of their medicines could potentially cause confusion particularly when the changes made in hospital led patients to have different supplies or multiple multicompartment ‘compliance aid’ tablet boxes. Like hospital staff, none of the primary care staff had received formal training about safe practices with medicines at discharge to primary care.

‘Bridges’ were identified as formalised features of the medicines management system that had been made permanent and promoted the safety and continuity of medicines management. They ensured that, despite varying conditions, care could be delivered successfully to patients with HF.

Bridges also included methods of communicating with primary care about treatment, for example, when hospitals sent an electronic copy of the patient’s discharge summary to their GP. In this case, summaries were put together by multidisciplinary teams including junior doctors, nurses and pharmacists who would check and add information about medicines that would be useful to the primary care team. After discharge, two participating hospital trusts ran pharmacist-led titration clinics to ensure that medicine doses were optimised. Titration clinics also meant patients would be seen more quickly than if they had to see a consultant cardiologist. One cardiology pharmacist explained that the titration clinic ensured patients’ medicines were adjusted as and when appropriate, in light of some GPs not feeling confident about changing them.

GP practices differed in how they processed discharge information. In one practice, administrative staff would review the discharge summary and forward actions to practice staff if medicines information needed to be changed. In another practice, this task was the responsibility of the GP, who would forward actions to practice staff and book any tests needed as a consequence of any changes in medicines occurring during hospital stay (ie, blood tests). One GP reported that his practice had re-engineered their processing of discharge summaries to include a multidisciplinary team comprising administrative staff, a practice-based pharmacist (whose post was created in response to recognised safety risks) and GPs, with the pharmacist taking responsibility for coordinating the process.

Practice pharmacists perceived that their specialist knowledge improved as a consequence of being involved in the discharge process, while further expediting the safe management of medicines for patients after discharge. Some practice staff described having targets in place linked to time taken to process discharge summaries, with some practices prioritising processing driven by the risk of readmission. One data quality manager explained that they tried to process discharges within 24 hours of receiving information from the hospital, including reconciling medicines, but also explained that they had a maximum of a week to complete it.

‘Props’ were informal, temporary or impromptu actions taken by patients or healthcare staff to avoid potential adverse events, such as medication errors. Props were sometimes developed in response to risks in the working environment, such as interruptions during medicines rounds.

Hospital staff told us that they suspected recommendations made by the hospital (eg, the uptitration of doses which is critical in HF) may not be acted on in primary care. Hence, staff created solutions to prevent a break in the ongoing treatment, giving patients an extra copy of their discharge letter to take to the GP. Some staff members described being cognisant of how discharge information can be difficult for patients to understand and would take extra time to explain the discharge summary and any abbreviations contained within it. One staff nurse at site 1 described having to make protected time to hold these discussions with patients, drawing curtains around the patients’ beds to prevent any disruption. Some patients reported being discharged with an insufficient amount of medicines, leading them to seek community pharmacists help to provide them with emergency supplies until they could see a GP. Some patients also proactively provided the necessary links between community pharmacy, GPs and the hospital after discharge. For example, one patient called the community pharmacy to ask what information, if any, they had been provided with about his medicines. Another patient provided their GP practice with information about dose changes.

Finally, community pharmacists stepped in to organise supplies for patients when something had gone wrong and the patient was unable to get the correct medicines.

Notwithstanding a considerable body of research that has illuminated the sometimes alarming levels of preventable harm in healthcare systems and how this could be reduced,31 32 this study suggests that there are opportunities to enhance the system that manages medicines across multiple organisations. The study also provides a positive perspective on the strategies developed and actions taken by healthcare organisations, their staff and patients to provide care successfully in the face of continued pressure and gaps that appear between and between organisations in this complex system.33 The multiple perspectives of patients and multidisciplinary staff, independent observation of practice and documentary analysis collected through mixed methods allowed the possibility of triangulating data from multiple sources to offer a thorough description of a complex system. In doing so we also present a framework within which to explore and understand resilience in healthcare systems: that of ‘bridges’ and ‘props’ set against a backdrop of ‘gaps’ and ‘traps’. Moreover, this study explores a whole healthcare system inclusive of its transitions, to reveal the context of that system. In contrast, previous studies, although adding to an understanding of system resilience, have examined these problems from solely a health professional perspective23 or a patient perspective.25

Bridges and props either provided permanent system adaptations to potential gaps in care, or temporary fixes, usually implemented by individuals or small teams. Sometimes the props were put in place despite organisational pressure, for example, to discharge patients and free beds. We also draw out the dissonance between what healthcare professionals believe should happen and the reality of contemporary practice. This was clear from the differences between the recommendations for hospital discharge from national guidance and local policy, where, for example, patients must be fully informed about their medicines and any changes, and the overall discharge process—which in the settings observed, may lack depth and the necessary detail, or appear rushed. This was sometimes due to different local conditions, such as the number of discharges that needed to be completed in a day, but also to local policies that lacked sufficient detail and were not supported by staff training. Healthcare systems are complex and non-linear and the Safety II paradigm asserts that success and failures are products of the same variable system performance and that linear models of events such as medication errors cannot reflect the complexity of modern healthcare systems.34 An enhanced view of the system using a Safety II lens allows healthcare organisations and policymakers to understand and close the gap between work as imagined versus work as done.35 This view also provides a better understanding of how policies and guidelines are actually interpreted and whether they are implemented in healthcare organisations by staff who adjust their performance to deliver care in a complex system.

Resilient systems are able to learn from their clinical experience (both positive and negative), adapt to it and respond to provide successful outcomes.36 37 It was evident that staff were able to anticipate system vulnerabilities, for example, in the transfer of discharge information, and take compensatory adaptive action in the form of ‘props’. As found by a previous study, patients also took remedial action, such as providing missing information about medicines changes to staff.25 Resilient systems can monitor, learn and anticipate opportunities to improve. A better understanding and acceptance of the error traps in the system present healthcare organisations with the opportunity to learn about how the system operates, particularly when it is under pressure and presents a basis to improve. A better knowledge of gaps allows staff to anticipate where problems may occur and take action to avoid them. Props in the system are indicators of how flexible staff and teams are and healthcare systems can learn from the temporary fixes put in place and knowing where bridges have successfully joined up care can help systems learn and be better placed to innovate elsewhere. There are opportunities to learn from the ‘ordinary performance adjustments’ that staff undertake to better understand how to keep patients safe,37 thereby formalising system props into bridges.

There are numerous opportunities for patient safety to be compromised and medicines to be suboptimally managed during this care transition. However, there are also cross-organisational bridges and temporary fixes in the form of props, put in place by individuals, including patients and carers, and teams to maximise the opportunity for safe and optimal care to be delivered. Investigating gaps and traps in the healthcare system and identifying existing compensatory props and bridges allow the illustration of areas where healthcare can be improved and fragmented communication minimised during care transitions.