Transitions of care following acute kidney injury: a systems-based analysis Authors: Duncan McNab, Mark Jeffries, Kelly Howells, Paul Bowie, Robbie Foy, Caroline Sanders, Tom Blakeman Published: 2026 ABSTRACT BackgroundPeople affected by Acute Kidney Injury (AKI) experience poor short and long-term outcomes following discharge from hospital. Attempts to improve care have recognised the complexity of the healthcare system. We applied systems methods to explore everyday work and illuminate areas important for success, challenges faced and how people adapted to these challenges.MethodsTwenty-six patients who had recently been discharged from six hospitals after an admission complicated by AKI were interviewed. Focus groups were conducted with healthcare staff across six sites in England. Using the Functional Resonance Analysis Method (FRAM), the activities undertaken by patients and professionals across the primary–secondary care interface were modelled and interactions examined. The Systems Thinking for Everyday Work (STEW) principles were then applied to understand how people coped with system challenges to achieve successful outcomes.ResultsTwenty-two interacting activities were identified and modelled. These formed three interlinked subsystems that underpin effective management after an admission affected by AKI: (1) generating a post-AKI plan; (2) involving patients in their care; and (3) enacting the plan following discharge. Within these, seven themes captured the key challenges. Senior oversight of discharge planning was inconsistent, and staff preparing discharge documents felt under pressure and had to make trade-offs between competing priorities. Patients reported anxiety and vulnerability and yet were expected to be active participants in their care with little evidence of education and involvement. Finally, enacting the post-AKI discharge plan in the community was complicated by inadequate communication, imperfect systems with limited capacity and tensions between short-term stability and long-term condition management.ConclusionTo optimise follow-up care for people affected by AKI, development of improvement strategies should focus on co-design of interventions with patients and healthcare professionals to address the inherent trade-offs at discharge and strengthen coordination across care boundaries.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12882-026-04971-5. Background Acute Kidney Injury (AKI) is a sudden deterioration in kidney function, that impairs fluid, electrolyte and acid-base balance [1, 2]. AKI commonly occurs during acute illness such as an infection, with low fluid volume, low blood pressure and the effects of medication accounting for the majority of cases [2, 3]. In England, over two thirds of AKI episodes are community acquired,10–20% of hospital admission are complicated by AKI, and more than half a million people are affected each year [1, 2, 4]. It is associated with significant morbidity, mortality, and healthcare burden across hospital, primary care and community settings. After discharge, people affected by AKI experience higher rates of unplanned readmission, major cardiovascular events, and development or progression of kidney disease, some of which may be avoidable [5–10]. This is particularly true among critically ill patients, the elderly, and those with comorbid conditions such as diabetes or cardiovascular disease [1, 11]. Internationally, improving AKI follow-up care is a patient safety priority [3, 12–15]. The NHS England Think Kidneys Programme adopted a system-wide approach to improve AKI care, which included resources to aid identification and improve care in both hospitals and community settings [2]. The Royal College of General Practitioners led system level guidance to improve discharge care and primary care follow-up [16, 17]. Incorporated into National Institute for Health and Care Excellence (NICE) AKI Quality Standards, the guidelines promote timely clinical review, monitoring of kidney health status, optimisation of medication and support for self-management [18]. Delivering this requires effective collaborative working across the primary-secondary care interface [19]. A holistic approach to care is needed as people who have experienced an AKI report difficulty attending follow-up appointments and struggle with competing health priorities, especially those managing multiple long-term conditions [20]. Improving healthcare delivery in such a complex system is challenging [19, 21]. Complex systems involve multiple dynamic interactions between different people (patients, carers, and various healthcare professionals), the tasks performed (health monitoring, optimising medication, and taking medication), the clinical environment (ability to attend for monitoring and care delivered across the primary-secondary care interface), and external factors (national policies) [22]. These interactions are often unpredictable resulting in varying system conditions (e.g. mismatches in demand and capacity and variable resource allocations) and goal conflicts (e.g. prioritising short or long term goals, implementing effective care for different long-term conditions). These factors influence how people work and how they achieve success, or otherwise. Emerging perspectives highlight that, in complex systems such as healthcare, improvement efforts may be undermined if they are developed without a thorough appreciation of how the system truly works in both routine and challenging conditions. In response we applied complex systems methods to explore everyday working practices and illuminate key areas within the AKI systems that are important for success to inform the design of future improvement strategies. Methods Study design A multi-method approach entailed semi-structured interviews with patients and focus groups with health professionals from primary and secondary care to generate data to model and analyse the system. A Patient and Public Involvement and Engagement group, including people with lived experience of AKI and a carer, contributed at regular meetings throughout the study. Members reviewed transcripts from selected patient interviews and focus groups, met with the research team to discuss emerging themes, and offered reflections that shaped the analytic focus. They also engaged with the application of complex systems methods, commenting on early system models and providing feedback that informed subsequent iterations and the direction of data analysis. Participants and setting Using reported data on the organisation of AKI services in England, six NHS sites were purposively selected to reflect variation in how post-discharge AKI care was organised, including sites with a virtual ward for AKI follow-up, two with specialist AKI nurse teams, and one using a renal clinical mailbox through which GPs could seek consultant advice [23]. Clinical care teams in each secondary-care site identified patients whose admissions had been complicated by an episode of AKI of any stage (including dialysis) and subsequently been discharged to their home. Eligible patients were provided with an information sheet and if agreeable were contacted after discharge, consented and interviewed by two of the research team (MJ and KH). The same clinical care teams were asked to recruit a range of healthcare staff involved in the management of patients with AKI at and beyond discharge to participate in focus groups to explore actions and reflections on AKI management. Eligible participants received written information and written consent was obtained prior to focus groups. Data collection Semi-structured interviews with 26 patients were conducted remotely up to four weeks after discharge between January 2023 and May 2024 (Table 1). These lasted approximately one hour and were conducted in English. Patients represented all stages of AKI and represented a wide range of socioeconomic backgrounds, as reflected in their Index of Multiple Deprivation scores (Table 1). Interviews explored patient experiences of admissions complicated by AKI with a particular focus on navigation and coordination of care following discharge. All participants were invited to take part in a follow-up interview four to six months later to explore their experience of follow-up care over this time; this was completed for 11 patients. Interviews were considered the best option to capture in-depth reflections in a private context. Table 1Characteristics of participants who underwent interviewParticipants from each hospital siteParticipant IDAge (by decade)GenderEthnicityIndex of Multiple Deprivation (IMD) ScoreAKI severityA130–40FemaleWhite British5Stage 2A260–70Female Unavailable 8Stage 3A320–30Female Unavailable 5 Unavailable A420–30FemaleWhite British5 Unavailable A550–60FemaleWhite British5 Unavailable B150–60Female Unavailable 7Stage 3B270–80Male Unavailable 6Stage 3C150–60FemaleWhite British5Stage 1C260–70MaleWhite British1 Unavailable C360–70FemaleWhite British6 Unavailable C470–80MaleWhite British3Stage 1C560–70FemaleWhite British6 Unavailable C660–70MaleBritish Indian2Stage 3D160–70FemaleWhite British6Stage 3D270–80MaleWhite British10 Unavailable D360–70FemaleWhite British4Stage 3D450–60MaleWhite British1 Unavailable D570 − 60MaleWhite British8Stage 3D660–70MaleWhite British7Stage 3E150–60MaleWhite British Unavailable Stage 3E260–70MaleWhite British8 Unavailable E350–60Female Unavailable Unavailable Unavailable E470–80MaleWhite British2Stage 3F180–90MaleWhite British6Stage 3F250–60FemaleWhite British7Stage 3F370–80FemaleWhie British2Stage 3 Table 1Characteristics of participants who underwent interviewParticipants from each hospital siteParticipant IDAge (by decade)GenderEthnicityIndex of Multiple Deprivation (IMD) ScoreAKI severityA130–40FemaleWhite British5Stage 2A260–70Female Unavailable 8Stage 3A320–30Female Unavailable 5 Unavailable A420–30FemaleWhite British5 Unavailable A550–60FemaleWhite British5 Unavailable B150–60Female Unavailable 7Stage 3B270–80Male Unavailable 6Stage 3C150–60FemaleWhite British5Stage 1C260–70MaleWhite British1 Unavailable C360–70FemaleWhite British6 Unavailable C470–80MaleWhite British3Stage 1C560–70FemaleWhite British6 Unavailable C660–70MaleBritish Indian2Stage 3D160–70FemaleWhite British6Stage 3D270–80MaleWhite British10 Unavailable D360–70FemaleWhite British4Stage 3D450–60MaleWhite British1 Unavailable D570 − 60MaleWhite British8Stage 3D660–70MaleWhite British7Stage 3E150–60MaleWhite British Unavailable Stage 3E260–70MaleWhite British8 Unavailable E350–60Female Unavailable Unavailable Unavailable E470–80MaleWhite British2Stage 3F180–90MaleWhite British6Stage 3F250–60FemaleWhite British7Stage 3F370–80FemaleWhie British2Stage 3 Characteristics of participants who underwent interview Six focus groups lasting between 60 and 90 min and were conducted by experienced qualitative researchers (MJ and KH) between September 2023 and May 2024. Focus groups were chosen to facilitate recruitment and encourage discussion and interaction between staff on the organisation and quality of care through discharge and during community follow-up. Four of the six sites arranged and recruited healthcare staff for focus groups which included renal and other medical consultants, renal and AKI specialist nurses, resident doctors and pharmacists. A general practitioner (GP) attended one focus group remotely. At the remaining two sites, involvement of primary care staff (GPs and GP based pharmacists) was organised through hospital staff and the research team networks. A total of 35 professionals from primary and secondary care participated. Both interviews and focus groups were recorded and transcribed verbatim (Table 2). Table 2Characteristics of focus group participants by site and roleFocus Group/siteRenal ConsultantOther Consultant including Nurse ConsultantRegistrar or Research FellowRenal NurseACP/ANPAKI Specialist NursePharmacistJunior Doctor (FY1 or 2 /IMT)Research NurseGPTotal by siteFG12111 (S)27FG211111 5 FG311121 6 FG42111 (S)2 7 FG52 (P)13FG65 (2 S, 3 P)27Total by role 4 3 3 3 1 2 9 5 1 4 35 ACP/ANP = Advanced Clinical Practitioner/Advanced Nurse Practitioner; S = Secondary Care; P = Primary Care; FY = Foundation Year; IMT = Internal Medical Training Table 2Characteristics of focus group participants by site and roleFocus Group/siteRenal ConsultantOther Consultant including Nurse ConsultantRegistrar or Research FellowRenal NurseACP/ANPAKI Specialist NursePharmacistJunior Doctor (FY1 or 2 /IMT)Research NurseGPTotal by siteFG12111 (S)27FG211111 5 FG311121 6 FG42111 (S)2 7 FG52 (P)13FG65 (2 S, 3 P)27Total by role 4 3 3 3 1 2 9 5 1 4 35 ACP/ANP = Advanced Clinical Practitioner/Advanced Nurse Practitioner; S = Secondary Care; P = Primary Care; FY = Foundation Year; IMT = Internal Medical Training Characteristics of focus group participants by site and role ACP/ANP = Advanced Clinical Practitioner/Advanced Nurse Practitioner; S = Secondary Care; P = Primary Care; FY = Foundation Year; IMT = Internal Medical Training Data analysis The research team met online on a fortnightly basis over nine months. Prior to each meeting an anonymised transcript was reviewed independently by each researcher. Transcripts of patient interviews were analysed first followed by transcripts from secondary care health professionals and community healthcare teams. Two methods for exploring work within complex systems were used to analyse transcripts. The Functional Resonance Analysis Method (FRAM) was used to synthesise data collected from across the work system to model and explore the interactions between different activities required for effective management of patients following discharge from hospital [24]. The Systems Thinking for Everyday Work (STEW) principles were used to illuminate the challenges people faced and the adaptations they employed within the work system [25]. FRAM The FRAM is a method to model and explore complex, socio-technical systems by examining how everyday work is carried out and how one activity influences other. It involves identifying the activities (or ‘functions’) performed as part of normal everyday work and mapping how each function is influenced by others through six defined aspects (Box 1). The FRAM Model Visualiser software package was used to create the FRAM model, four iterative steps were applied to each transcript to develop and improve the model as more data was analysed (26, Box 1). Box 1Functional Resonance Analysis Method – aspects and steps involved in creating FRAM modelBox 1 – Functional Resonance Analysis MethodAspects of functions [24]Input - what starts the functionOutput - what emerges from the functionPrecondition - what condition that must be met before the function can startResources - anything (people, information, materials) needed to carry out the function or anything that is used up by the functionControl - anything that controls or monitors the functionTime - how time can influence the function – such as a time constraint or a need for the function to be complete by a certain time Iterative steps involved in the creating the FRAM model Step one: Each researcher identified activities (or sets of activities) described in the transcripts as proposed ‘functions’. An example is the function < Complete discharge document>. These were discussed by the researchers to determine if it described a distinct activity. The name given to the function was then agreed to make sure that it represented the activity being performed. Initially patient-led activities were identified. Then, functions undertaken in a hospital setting were added followed by those undertaken in community.Step two: Each function was entered into the FRAM Model Visualiser Software package to develop the FRAM model. In this software package, functions are represented by hexagons.[26]Step three: Links between functions were identified by researchers through further analysis of the transcripts to determine the aspects of each function. For example, for the function < Complete discharge document > one of the outputs is - discharge document is completed and sent to the GP. Interactions are then modelled by identifying the same aspect in a different function. For example, the function < Process discharge document in GP > has the precondition – discharge document is completed and sent to GP. In this way links between functions are built by the researchers through analysis of transcript data.Step four: Researchers identified reports of variation in the transcript data of each function output that influenced the care received by patients were identified. Box 1Functional Resonance Analysis Method – aspects and steps involved in creating FRAM modelBox 1 – Functional Resonance Analysis MethodAspects of functions [24]Input - what starts the functionOutput - what emerges from the functionPrecondition - what condition that must be met before the function can startResources - anything (people, information, materials) needed to carry out the function or anything that is used up by the functionControl - anything that controls or monitors the functionTime - how time can influence the function – such as a time constraint or a need for the function to be complete by a certain time Iterative steps involved in the creating the FRAM model Step one: Each researcher identified activities (or sets of activities) described in the transcripts as proposed ‘functions’. An example is the function < Complete discharge document>. These were discussed by the researchers to determine if it described a distinct activity. The name given to the function was then agreed to make sure that it represented the activity being performed. Initially patient-led activities were identified. Then, functions undertaken in a hospital setting were added followed by those undertaken in community.Step two: Each function was entered into the FRAM Model Visualiser Software package to develop the FRAM model. In this software package, functions are represented by hexagons.[26]Step three: Links between functions were identified by researchers through further analysis of the transcripts to determine the aspects of each function. For example, for the function < Complete discharge document > one of the outputs is - discharge document is completed and sent to the GP. Interactions are then modelled by identifying the same aspect in a different function. For example, the function < Process discharge document in GP > has the precondition – discharge document is completed and sent to GP. In this way links between functions are built by the researchers through analysis of transcript data.Step four: Researchers identified reports of variation in the transcript data of each function output that influenced the care received by patients were identified. Functional Resonance Analysis Method – aspects and steps involved in creating FRAM model Input - what starts the function Output - what emerges from the function Precondition - what condition that must be met before the function can start Resources - anything (people, information, materials) needed to carry out the function or anything that is used up by the function Control - anything that controls or monitors the function Time - how time can influence the function – such as a time constraint or a need for the function to be complete by a certain time Step one: Each researcher identified activities (or sets of activities) described in the transcripts as proposed ‘functions’. An example is the function < Complete discharge document>. These were discussed by the researchers to determine if it described a distinct activity. The name given to the function was then agreed to make sure that it represented the activity being performed. Initially patient-led activities were identified. Then, functions undertaken in a hospital setting were added followed by those undertaken in community. Step two: Each function was entered into the FRAM Model Visualiser Software package to develop the FRAM model. In this software package, functions are represented by hexagons.[26] Step three: Links between functions were identified by researchers through further analysis of the transcripts to determine the aspects of each function. For example, for the function < Complete discharge document > one of the outputs is - discharge document is completed and sent to the GP. Interactions are then modelled by identifying the same aspect in a different function. For example, the function < Process discharge document in GP > has the precondition – discharge document is completed and sent to GP. In this way links between functions are built by the researchers through analysis of transcript data. Step four: Researchers identified reports of variation in the transcript data of each function output that influenced the care received by patients were identified. STEW The Systems Thinking for Everyday Work (STEW) framework is a systems-based method to aid the exploration of how people adapt the way they work due to the system conditions they face (such as demand, capacity, resource availability and constraints) and how these influence interactions with other parts of the system. To perform the STEW analysis each researcher identified participant accounts that related to the principles (Box 1). Multiple perspectives were considered to develop a consensus of the foundation concept (the purpose of the system), evidence of how work conditions (demand, capacity, resources and constraints) and interactions influenced the work people did. In addition, the ways people adapted their work to achieve system goals was identified. This included workarounds (for example, if information was missing) and trade-offs (balancing different priorities such as efficiency vs. thoroughness or long-term vs. short-term goals). Box 2Description of system thinking for everyday work principlesBox 2 -System Thinking for Everyday Work (STEW) Principles [25]1. Foundation conceptAttempt to explore and understand overall system safety and success, rather than focussing on isolated parts, events or outcomes.2. Seek Multiple PerspectivesAppreciate that people, at all organisational levels and regardless of responsibilities and status, are the local experts in the work they do. Explore with them how they achieve success.3. Consider Work ConditionsAppreciate that the interacting combination of demand, capacity, resource availability and constraints influence the way people undertake work at any given time. Explore the influence of conditions of work on safety and success.4. Analyse Interactions and Work FlowAppreciate that interactions between people, tasks, equipment, environments (e.g. physical, social, organisational) and external influences (e.g. national policy, regulatory obligations) are complex and dynamic and affect care system performance and human wellbeing (e.g. patients and staff). Explore influence of interactions and flow on safety and success.5. Explore Performance VariabilityAppreciate that people continuously adapt and vary how they work to achieve a successful outcome based on their own goals and the system conditions they face. Explore how important performance variability is to success and how it can be developed and performed safely.6. Understand Why Decisions Make Sense at the TimeExplore why staff make decisions and how this influences success; people do what makes sense to them based on the system conditions experienced at the time. Box 2Description of system thinking for everyday work principlesBox 2 -System Thinking for Everyday Work (STEW) Principles [25]1. Foundation conceptAttempt to explore and understand overall system safety and success, rather than focussing on isolated parts, events or outcomes.2. Seek Multiple PerspectivesAppreciate that people, at all organisational levels and regardless of responsibilities and status, are the local experts in the work they do. Explore with them how they achieve success.3. Consider Work ConditionsAppreciate that the interacting combination of demand, capacity, resource availability and constraints influence the way people undertake work at any given time. Explore the influence of conditions of work on safety and success.4. Analyse Interactions and Work FlowAppreciate that interactions between people, tasks, equipment, environments (e.g. physical, social, organisational) and external influences (e.g. national policy, regulatory obligations) are complex and dynamic and affect care system performance and human wellbeing (e.g. patients and staff). Explore influence of interactions and flow on safety and success.5. Explore Performance VariabilityAppreciate that people continuously adapt and vary how they work to achieve a successful outcome based on their own goals and the system conditions they face. Explore how important performance variability is to success and how it can be developed and performed safely.6. Understand Why Decisions Make Sense at the TimeExplore why staff make decisions and how this influences success; people do what makes sense to them based on the system conditions experienced at the time. Description of system thinking for everyday work principles 1. Foundation concept Attempt to explore and understand overall system safety and success, rather than focussing on isolated parts, events or outcomes. 2. Seek Multiple Perspectives Appreciate that people, at all organisational levels and regardless of responsibilities and status, are the local experts in the work they do. Explore with them how they achieve success. 3. Consider Work Conditions Appreciate that the interacting combination of demand, capacity, resource availability and constraints influence the way people undertake work at any given time. Explore the influence of conditions of work on safety and success. 4. Analyse Interactions and Work Flow Appreciate that interactions between people, tasks, equipment, environments (e.g. physical, social, organisational) and external influences (e.g. national policy, regulatory obligations) are complex and dynamic and affect care system performance and human wellbeing (e.g. patients and staff). Explore influence of interactions and flow on safety and success. 5. Explore Performance Variability Appreciate that people continuously adapt and vary how they work to achieve a successful outcome based on their own goals and the system conditions they face. Explore how important performance variability is to success and how it can be developed and performed safely. 6. Understand Why Decisions Make Sense at the Time Explore why staff make decisions and how this influences success; people do what makes sense to them based on the system conditions experienced at the time. Application of FRAM and STEW The FRAM model was studied to explore how interactions between activities influenced post-discharge AKI care, including how variability in one activity affected others. STEW complimented this by illuminating the challenges faced by patients and staff and the adaptive strategies they used to complete these activities successfully. Results Analysis of transcripts identified 22 functions which were modelled using the FRAM (Fig. 1). Examination of this model and the interactions between its functions identified three interlinked subsystems central to successful post-discharge AKI management: (1) generating of a post-AKI plan; (2) involving patients as active participants in their care and (3) enacting the plan. Generating and enacting the plan were linked through the key activity of completing the discharge document and the use of outpatient services. Patient involvement was shaped by in-hospital factors and influenced behaviours following discharge. Within these subsystems, seven themes described the challenges faced by people within the system. Key quotations are included to illustrate each theme. Quotations representing each STEW principle mentioned below are provided in the supplementary file (Supplementary file). Fig. 1Functional Resonance Analysis Method model describing discharge following Acute Kidney Injury. Key - Each hexagon is an activity or function - Red = patient functions, Blue = secondary care health care professional functions, Green = community based health care professional functions. The small circles at each angle of the hexagons represents the aspects of that function – input, output, preconditions, resources, time and control. Red circles at angles of some hexagon represent important influences on that function that were identified from the data but were out with the boundary of the analysis – such as providing capacity to perform follow-up Functional Resonance Analysis Method model describing discharge following Acute Kidney Injury. Key - Each hexagon is an activity or function - Red = patient functions, Blue = secondary care health care professional functions, Green = community based health care professional functions. The small circles at each angle of the hexagons represents the aspects of that function – input, output, preconditions, resources, time and control. Red circles at angles of some hexagon represent important influences on that function that were identified from the data but were out with the boundary of the analysis – such as providing capacity to perform follow-up Subsystem A - Generating a post-AKI plan Generating a post-AKI management plan involved the FRAM functions < Plan made by senior clinician for discharge > and < Complete discharge document>. Variability in the output of these functions influenced decisions at discharge and in the community. The STEW analysis illuminated the pressures felt by people as part of everyday work. These were the availability of senior clinical oversight of the discharge plan; the sense of pressure felt when completing discharge documents; and making trade-offs at discharge. Availability of senior clinical oversight of discharge plan The availability of a discharge plan made by senior clinicians influenced several FRAM functions. The discharge plan set the parameters for assessment of medical and social readiness for discharge and the strategy for ongoing care. It was a precondition for involvement of the discharge team and the main resource to aid completion of the discharge document. Discharge plans made by senior clinicians were considered an essential resource but were often not available or were vague and lacked all the information needed (Supplementary file, quote A1 and A2).If you don’t know the information about what follow-up and what bloods, then you’re going to be less inclined to fill it in, because it’s not for you guys to be deciding on what bloods are going to be done and when. And if someone’s not given you that plan already… (FG1, R5 AKI specialist nurse) If you don’t know the information about what follow-up and what bloods, then you’re going to be less inclined to fill it in, because it’s not for you guys to be deciding on what bloods are going to be done and when. And if someone’s not given you that plan already… (FG1, R5 AKI specialist nurse) Feeling a sense of pressure and stress when completing discharge documents Resident doctors (and occasionally pharmacists) were largely responsible for preparing discharge documents. It was deemed a challenging task that took place in a highly pressurised environment because discharge documents had to be completed quickly to create bed capacity (quote A3).Beds are gold dust at the moment…So if somebody can go home, if the doctor said, yeah, they can go, the…I’m not saying the medical care isn’t as good, but it’s a case of get them out, anything that can be reviewed in community, review it in community. If we’ve treated the main cause of admission, then they can go out. And everything else can be followed up (FG1/1 -Pharmacist - Secondary Care) Beds are gold dust at the moment…So if somebody can go home, if the doctor said, yeah, they can go, the…I’m not saying the medical care isn’t as good, but it’s a case of get them out, anything that can be reviewed in community, review it in community. If we’ve treated the main cause of admission, then they can go out. And everything else can be followed up (FG1/1 -Pharmacist - Secondary Care) Clinical details of AKI could be difficult to confirm. AKI may not have been the main medical concern during admission and may have resolved (quote A4). This resulted in extra work to confirm or refute and, if necessary, include information on an AKI diagnosis for patients whom the clinician may not know (quote A5). It was challenging to manage patients whose serum creatinine levels had improved but had not returned to baseline. Stability of kidney function was a mediator rather than a precondition for discharge. Indeed, social dimensions of care (arranging discharge when social care packages are in place) may be prioritised over full recovery of kidney function to maintain bed flow (quote A6). Making trade-offs and workarounds at discharge To cope with these challenges, doctors and pharmacists had to make trade-offs to navigate a sense of immediate pressure for being responsible for maintaining discharge flow whilst also needing to provide care in keeping with professional guidance and standards. This resulted in variability in the output of the function < Complete discharge document>. If information on post discharge monitoring or medication optimisation was not available, those completing discharge documents were not keen to ask senior clinicians for advice (quote A7). Instead, they would often defer decisions ‘downstream’ by requesting early monitoring post discharge and asking primary care teams to consider whether to restart medication temporarily stopped during admissions. In this way they attempted to maintain safety and allay fears of litigation if something went wrong; however, it resulted in increased work in primary care with little thought to the consequence of this approach (quote A8 and A9).I think …possibly one of the other problems is that the person doing the…leading the ward round, whether it’s a consultant or registrar or whatever, might make the decision they go home, but then they’ve gone off to clinic and then these little questions arise about what should we do about this medication, what do we do with this. And you’re left with fairly junior staff on the ward… I would be like, I don’t know. GP can sort that out. (FG1/2 - Renal Registrar) I think …possibly one of the other problems is that the person doing the…leading the ward round, whether it’s a consultant or registrar or whatever, might make the decision they go home, but then they’ve gone off to clinic and then these little questions arise about what should we do about this medication, what do we do with this. And you’re left with fairly junior staff on the ward… I would be like, I don’t know. GP can sort that out. (FG1/2 - Renal Registrar) In some settings, a computer-based template had been designed consisting of structured questions to encourage inclusion of relevant details about the patient’s AKI in the discharge document. The template took longer to complete and increased pressure felt by resident doctors; therefore, a workaround was employed by writing ‘not applicable’ in the template to avoid answering several additional questions and instead providing shorter free text details (quote A10). Subsystem B - Involving patients in care Patients described working to recover while in hospital and maintain health after discharge. These FRAM functions were influenced by patients being able to make sense of their kidney health which was influenced by patient anxiety and the provision of information. Patient anxiety and vulnerability were major challenges, particularly given expectations placed on patients to play an active role in managing their health despite limited education and involvement in care planning. Patient anxiety and vulnerability Across patient participant accounts there was a sense of anxiety and emotional vulnerability that was evident from diagnoses through admission and discharge and into post-discharge care (quote B1). This vulnerability was dynamic in that it was shaped by and also shaped interactions with healthcare professionals (quote B2).I obviously was scared then, that I was going to be left with very, very poor functioning kidneys, and I didn’t know what that…I mean my mind was racing. I’d already got myself on dialysis in my mind, and I was sort of thinking that I was going to be taking steroids and I’d have a moon face, you know how your mind runs away with you? (Patient D3) I obviously was scared then, that I was going to be left with very, very poor functioning kidneys, and I didn’t know what that…I mean my mind was racing. I’d already got myself on dialysis in my mind, and I was sort of thinking that I was going to be taking steroids and I’d have a moon face, you know how your mind runs away with you? (Patient D3) Introducing concepts and framing discussions about AKI was challenging due to the health of patients, a fear of exacerbating existing anxiety and lack of time for discussions (quote B3). Perhaps as a result, there was a paucity of accounts describing patient education about AKI. Patient capacity to participate in discussions to enhance understanding of AKI and how to maintain health varied during admissions depending on the patient’s health and between patients (quote B4).I think, to be honest, I was obviously not out of it totally but I wasn’t with it for the first few days, so people were coming in and telling me things and telling me what was going to happen, but I’ve got to be honest I didn’t register it, I couldn’t take it in. (Patient E) I think, to be honest, I was obviously not out of it totally but I wasn’t with it for the first few days, so people were coming in and telling me things and telling me what was going to happen, but I’ve got to be honest I didn’t register it, I couldn’t take it in. (Patient E) Other medical priorities resulted in less attention being paid to AKI especially if the AKI had resolved (quote B5). Patients sought information and support from other sources including other patients and reading. Some suggested a ‘guardian angel’ was needed, often a nurse or carer who recognised their vulnerability and helped them obtain information and arrange reviews (quote B6). Renal specialist nurse-led services delivered education to some patients while in hospital and after discharge. It was felt that reviews a few months after discharge were a useful space for education and understanding (quote B7). Expectation on patients Although healthcare professionals recognised that vulnerable patients may have limited capacity, patients were still generally expected to play an active role in improving and maintaining their own health. This included taking medication as prescribed, adhering to lifestyle recommendations and attending follow-up appointments with healthcare professionals in primary care (quote B8). Indeed, they were often expected to organise follow-up necessary to maintain health (quote B9).So, you know, if you, sort of, like, say you need to go and see your GP in a week or…you know, they don’t always do it. And so, they end up, sort of, worse than they left sometimes. You know, so that’s quite a…sometimes quite a revolving door, as we call it. You get the same patients coming back in the system (FG3/4 - Nurse) So, you know, if you, sort of, like, say you need to go and see your GP in a week or…you know, they don’t always do it. And so, they end up, sort of, worse than they left sometimes. You know, so that’s quite a…sometimes quite a revolving door, as we call it. You get the same patients coming back in the system (FG3/4 - Nurse) Despite this expectation, there was an absence of accounts articulating patient involvement in discharge planning and limited evidence of patients receiving education on how to maintain their health. Subsystem C - Enacting a post-AKI plan The FRAM functions needed to enact an effective post-AKI management were < Perform monitoring test>, < Review results of monitoring tests > and < Optimise ongoing medication>. Adequate information in the discharge document influenced all of these functions. Patients’ ability to attend review appointments was a precondition for attendance which was in turn a precondition for performing monitoring tests. It was influenced by their understanding of their kidney health and how to maintain their health post-discharge. The challenges of enacting the post-discharge plan were effective information transfer and optimising monitoring and medication. Effective information transfer The discharge document was the main mechanism for communicating care requirements to primary care and renal outpatient teams. It directed the timing and intensity of follow-up and medication optimisation. The thoroughness of the information in relation to AKI diagnosis (and staging), why medication was stopped, when it should be restarted, optimal doses, and how patients should be monitored varied considerably (quote C1).The comms are really, I find, unclear, so…so the discharge letter will say, stop, you know, their ACE inhibitor, their ARB, GP to consider re-introducing medicines…stop the metformin, consider re-introducing, stop… So there’s these lists of stops and reconsider, but they don’t…it’s quite unclear at what point …when to restart them. How. And actually, if you’ve got someone there, think, oh, actually, should this be re-started now? (FG5/3 - Pharmacist – Primary Care) The comms are really, I find, unclear, so…so the discharge letter will say, stop, you know, their ACE inhibitor, their ARB, GP to consider re-introducing medicines…stop the metformin, consider re-introducing, stop… So there’s these lists of stops and reconsider, but they don’t…it’s quite unclear at what point …when to restart them. How. And actually, if you’ve got someone there, think, oh, actually, should this be re-started now? (FG5/3 - Pharmacist – Primary Care) Reported delays in receiving discharge documents and limited capacity in primary care teams to provide monitoring appointments contributed to delays in follow-up (quote C2). Vague statements in the discharge document were often included with little thought in terms of clarity in purpose and system capacity for this to be achieved. Often GPs were asked to arrange blood tests, review results and make decisions on medication optimisation. This shifted decision making and created further workload in primary care (quote C3).The problem is that there is no thought as to why they’re doing that, and where do we go from here. You know, me as a GP, I’ve no problem checking bloods….What I want to know is that I’m doing it for a reason, and that I’ve got a plan of where I go next with that. (FG5/2, General Practitioner) The problem is that there is no thought as to why they’re doing that, and where do we go from here. You know, me as a GP, I’ve no problem checking bloods….What I want to know is that I’m doing it for a reason, and that I’ve got a plan of where I go next with that. (FG5/2, General Practitioner) There were concerns that relevant medicines were not restarted due to the ambiguity of discharge documents (quote C4). There was considered a lack of understanding and assumptions by secondary care about systems within primary care. Those who had primary care experience may be more likely to provide effective information to support arranging tests and acting on blood results, as well as considering the challenges of dealing with patients with multimorbidity who require monitoring of a number of conditions, and those who use medication compliance aids (quote C5). Optimising post-AKI monitoring and medication Primary care teams or specialist AKI services were responsible for optimising post-AKI care and variable systems were reported for coding diagnosis, initiating follow-up, interpreting results and optimising medication and further monitoring. Discharge documents often suggested specific tasks (such as rechecking bloods in two weeks) which could be arranged effectively by administrative staff. However, there was concern that if the result of these blood tests was normal, it may be filed with no further action and an opportunity to optimise medication may be missed (quote C7 and C8).They can be processed by, you know, reception staff, who have really limited clinical knowledge. And that’s the concern, is if there’s not sufficient amount of detail in there, that there has been an AKI, and these meds have been stopped for that reason, and they need to get back on, it’s quite, it is possible that a non-clinical person might be making, taking those medications off that patient’s repeat, and it doesn’t get, then, ever seen by a clinician, worst case scenario (FG6/1 - Pharmacist/Primary Care) They can be processed by, you know, reception staff, who have really limited clinical knowledge. And that’s the concern, is if there’s not sufficient amount of detail in there, that there has been an AKI, and these meds have been stopped for that reason, and they need to get back on, it’s quite, it is possible that a non-clinical person might be making, taking those medications off that patient’s repeat, and it doesn’t get, then, ever seen by a clinician, worst case scenario (FG6/1 - Pharmacist/Primary Care) To counter this, GPs recognised that a holistic clinical review of the patient to optimise follow-up and commence or titrate medication to protect kidney health may reduce the risk of readmission and progression of kidney disease (quote C9). Integration into existing long-term condition monitoring systems was suggested to ensure regular, structured monitoring and medication optimisation (quote C10). GPs recognised that altering medication in order to achieve better long-term outcomes also carried short-term risks and reported needing to trade-off between these competing priorities. Some described a cycle of titrating medication to improve long term kidney health only for the patient to have a further AKI with admission, temporary withdrawal of medication followed by re-titration after discharge (quote C11). Indeed, concerns were expressed that some patients were worried about restarting medication which had been stopped in hospital because it had contributed to their AKI. This made it challenging to optimise medication following discharge (quote C12). GPs questioned if they had the knowledge to make these decisions (quote C13) and at one site described that having access to advice from nephrology services through a dedicated clinical mailbox was useful in decision making (quote C14). Two of the sites employed specialist AKI nurses and this was considered an effective way to optimise medication and monitoring to reduce readmission and progression of renal disease; however, secondary care systems did not have capacity to follow up all AKI patients and so prioritised more severe AKI (stages 2 and 3). Older, frailer patients with multiple long-term conditions were considered less likely to be able to attend these reviews and so the trade-off was that their follow-up was often left to primary care teams (quote C15). Discussion Summary This study applied two complex systems methods to illuminate key activities and challenges involved in supporting patients following an admission complicated by AKI. Senior oversight of discharge planning emerged as essential, yet inconsistently available. Staff described pressure when completing discharge documents and balancing competing priorities. Patients reported anxiety and vulnerability yet despite this, were expected to be active participants in their care. Finally, enacting the post-AKI discharge plan in primary care was complicated by inadequate communication across the secondary/primary care interface, imperfect systems with limited capacity and difficulties balancing short-term stability and long-term condition management. Strengths and Limitations Data from across the system was analysed and different models of care included. Importantly, patient accounts were the starting point for our analysis, placing patients at the centre of the care system to explore secondary care and community-based professionals influence on the patient journey. Using two complementary systems-based methods strengthened our ability to understand AKI management as a dynamic, adaptive process rather than a linear sequence of tasks. FRAM modelled the structure of the care system, identify key subsystems, and examine how variability in one function influenced others, revealing dependencies or vulnerabilities in post-AKI care. STEW offered an analytic lens focused on the realities of work, illuminating the challenges staff and patients encountered and the adaptations they made to maintain care. Together, these approaches revealed how the system is organised and how it is actually enacted in practice, highlighting the adjustments, trade-offs, and informal strategies that underpin safe and effective post-AKI management. Their combined use generated a richer account of system performance than traditional qualitative data analysis methods and strengthened the explanatory power of our findings and their relevance for the design of future improvement interventions. FRAM has previously proved valuable in various healthcare settings to understand how work is shaped by system conditions and to identify targets for improvement [27, 28], as well as to illuminate challenges faced by patients and staff during care transitions [29, 30]. Only individuals whose AKI was identified during admission and who were discharged to their own homes were included. Patients with unrecognised AKI or those discharged to residential settings may have different needs and recovery trajectories. Health literacy was not assessed, despite its potential influence on patients’ ability to engage with follow-up care. Comparison with existing literature Our finding align with evidence that effective discharge requires senior leadership, yet clinicians in our study reported having to make decisions without senior support [31, 32]. While high workloads, inadequate supervision, and systemic pressures are known to contribute to stress during discharge planning [33], our study adds insight into the additional stress generated when having to trade-offs between different goals at discharge. Consistent with earlier qualitative work, patients with AKI frequently reported anxiety, vulnerability, limited understanding of their diagnosis and felt poorly supported during transitions of care [34]. It is recognised that patients work to improve and maintain their health and that care systems must support and adapt to patient needs and capability [35]. Indeed, the FRAM has previously been used to provide valuable insights into the complexities of hospital discharge and highlighted the importance of supporting patient-facing functions to support the ‘invisible work’ that patient perform [29]. Despite recommendations stressing the importance of structured patient engagement and shared decision-making, these practices were inconsistently delivered [18]. Effective discharge documentation is a key factor in ensuring safe care transitions, yet often communication is sub-optimal [36]. Strategies such as standardised electronic templates have demonstrated effectiveness in enhancing discharge communication [37]. However, in the context of workforce pressures, even where these were implemented, clinicians described workarounds to avoid the extra steps needed to complete templates. Challenges in primary care mirrored existing evidence of variable adherence to AKI guidance and the difficulty of balancing short-term stability with long-term condition management, particularly for individuals with multiple long-term conditions [38, 39]. Emerging evidence suggests that coordinated, multidisciplinary follow-up may improve outcomes, but robust interventional studies remain limited [40, 41]. Implications for future research and policy AKI imposes a substantial economic burden on the NHS largely driven by prolonged hospitalisation, complications, and preventable episodes. This reinforces the need for more reliable post-discharge pathways and better-coordinated primary care follow-up [42]. The challenges identified — variable communication, inconsistent follow-up, reliance on patient self-management, and competing clinical priorities — are common across international health systems, suggesting wider relevance of our findings. In England, the NHS Patient Safety Strategy has shifted to recognising the need for a systems approach to improve safety management [43]. Our use of FRAM and STEW aligns with this shift in policy, with a need acknowledge the complexity of care across the primary/secondary care interface and promote co-designed and rigorously evaluated improvement interventions to support discharge decision-making and procedures. Such interventions need to reconcile work-as-imagined (by those designing interventions) and work-as-done (by those who use them) [25]. Ways to meaningfully involve patients, families and carers need to be strengthened and the input from AKI nurse specialist in hospital and after discharge appeared promising. Future guidance should reflect the variable systems used in GP practices, consider patient priorities and acknowledge the trade-offs clinicians make between competing priorities. Conclusion This study highlights the complexity of care for patients following an admission complicated by AKI. Key challenges have been identified including senior clinical oversight of discharge planning, the pressure felt by those making discharge decisions, patient anxiety, lack of meaningful patient involvement in care and the variable systems for enacting discharge plans in the community. Additionally, the trade-offs inherent at discharge and for primary care teams implementing recommendations after discharge requires ongoing refinement of policies and guidance. Future work should focus on co-design of interventions to optimise care transitions and ensure sustained patient and staff wellbeing. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1