Impact and experiences of delayed discharge: A mixed‐studies systematic review

Abstract Background The impact of delayed discharge on patients, health‐care staff and hospital costs has been incompletely characterized. Aim To systematically review experiences of delay from the perspectives of patients, health professionals and hospitals, and its impact on patients’ outcomes and costs. Methods Four of the main biomedical databases were searched for the period 2000‐2016 (February). Quantitative, qualitative and health economic studies conducted in OECD countries were included. Results Thirty‐seven papers reporting data on 35 studies were identified: 10 quantitative, 8 qualitative and 19 exploring costs. Seven of ten quantitative studies were at moderate/low methodological quality; 6 qualitative studies were deemed reliable; and the 19 studies on costs were of moderate quality. Delayed discharge was associated with mortality, infections, depression, reductions in patients’ mobility and their daily activities. The qualitative studies highlighted the pressure to reduce discharge delays on staff stress and interprofessional relationships, with implications for patient care and well‐being. Extra bed‐days could account for up to 30.7% of total costs and cause cancellations of elective operations, treatment delay and repercussions for subsequent services, especially for elderly patients. Conclusions The poor quality of the majority of the research means that implications for practice should be cautiously made. However, the results suggest that the adverse effects of delayed discharge are both direct (through increased opportunities for patients to acquire avoidable ill health) and indirect, secondary to the pressures placed on staff. These findings provide impetus to take a more holistic perspective to addressing delayed discharge.

lost in 2013-14. 3 A Canadian study found that between 8 and 10% of beds in acute hospitals were occupied inappropriately by delayed patients. 2 Delayed discharge is recognized to be a system-level problem requiring effective team working within hospitals and coordination between health and social care. [6][7][8] However, an in-depth understanding of the impact of delayed discharge on patients and the health-care staff caring for them needs to be established so that managers and policymakers can make informed decisions about addressing the consequences of delays. The costs of delayed discharge to hospitals, the health and social care system, and patients and carers also need to be understood. This systematic review assesses the impact and experiences of delayed discharge at multiple levels, from the perspective of patients, health professionals and hospitals; and associated costs of delay.
This review systematically examines quantitative and qualitative studies to (i) quantify the impact of delayed discharge on health outcomes, (ii) qualitatively assess impacts on patients, health professionals and provider organizations, and (iii) evaluate the potential costs associated with delay. Studies conducted in OECD countries 9 were included to examine delayed discharge across health systems in countries with comparable economic development.

| METHODS
This review is reported using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 10 (Appendix S4), and the protocol is published in PROSPERO (CRD42016035256). 11

| Information sources and search strategies
Studies were identified using 6 biomedical databases (as below) which were searched in February 2016 for the period 2000-2016.
The searches were limited to publications dated from 2000 onwards to ensure the studies are relevant to contemporary health systems.
Specific search strategies were designed for Medline, Embase, CINAHL, PsycINFO, HMIC (Health Management Information Consortium) and NHS EED. The initial search was designed for Medline (via Ovid), which combined MeSH terms and keywords, and later adapted to other databases (Appendix S1), including search terms such as "delayed discharge," "timely discharge," "unnecessary days" and "inappropriate stays." This search was complemented with grey literature sources and consulting other systematic reviews and original papers. A bibliographic database was created to manage the references using EPPI-Reviewer 4. 12

| Study selection
Included studies addressed the impact and experiences of delayed discharge. Studies were included where they met one or more of the following inclusion criteria: (i) quantitative data on the impact of delayed discharge on health outcomes (eg quality of care, patient satisfaction, number of infections, mental health, mortality, morbidity, readmissions and functioning), (ii) qualitative data on experiences of delay from perspectives of patients (eg perceived impact on physical health or patient experience), health professionals (eg affect on staff role and working relationships) and hospitals (impacts at the organizational level, eg costs of managing delays and affect on culture), and (iii) information on costs of delay due to unnecessary bed-days.
Furthermore, only studies written in English, published since 2000 and conducted in the OECD were included.
The following exclusion criteria were also applied to the articles identified through the database search: research focusing on mental health, maternal and child and adolescent health, and palliative care was excluded; delays may occur in those settings for different reasons, for example relapse of mental health disorders, 13,14 and consequently, delays due to non-medical reasons are difficult to determine. [14][15][16][17] Abstracts, editorials, commentaries and book reviews were excluded because the review focused on primary research.

| Assessment of eligibility
The title/abstract of references were screened for eligibility by 2 reviewers, and then, the full text of those references which fulfilled the inclusion criteria was assessed. Discussion with a third reviewer was used to resolve disagreements.

| Quality assessment
We determined the quality of the quantitative studies using a standardized tool for assessing the methodological quality of quantitative/ observational studies. 18 The focus of some questions was adapted to ensure their relevance to the topic, covering the following fields: control group, confounders, sample, measures, reliability and relevance in a health service context (Appendix S2).
The qualitative studies were quality-assessed using 6 criteria of the "weight of evidence" (with respect to reliability and usefulness) developed by the EPPI-Centre 19,20 (Appendix S2). A score (low, medium, high) was then allocated to each study. Reliability was based on assessment of rigour in study sampling, data collection, analysis and findings.
Usefulness was based on assessment of the breadth and depth of findings and the extent to which the perspectives of health-care professionals and patients/carers' perspectives were prominent in the studies.
We used the checklist for the critical assessment of economic evaluation 21 and the NICE guide on methods of technology appraisal 22 to select and appraise the quality of health economic studies.

| Data extraction and synthesis of the results
The following characteristics were summarized for each study: design, setting, year of publication, country, target population, sociodemographic characteristics, disease(s) and reason(s) for delayed discharge (Appendix S3). For the quantitative and health economic studies, results were classified into categories depending on the nature of the outcome. Experiences of delay reported in the qualitative studies were divided into 3 categories: (i) perceptions of patients, (ii) perceptions of health professionals, and (iii) experiences of delay for hospitals.

| Identification of the studies
The search retrieved 11 656 references. After conducting the title and abstract screening, 589 references were included for full-text assessment. A total of 552 studies had to be excluded mostly because they did not consider experiences, impact or outcomes of delay ( Figure 1), leaving 37 papers included in the review, reporting data on 35 studies.

| Characteristics of the studies
The study characteristics are summarized in Table 1. There were 10 quantitative, 8 qualitative and 19 health economic studies. More than half the studies were undertaken in the UK (14) and the United States (8). Half the studies analysed data across different service areas, and others focused on 1 type of service, for example trauma (5), acute (4) and intensive care (4). Thirteen studies examined elderly patients only.

| Quality assessment
Three of the quantitative studies were deemed to have high methodological quality, 23-25 4 moderate quality [26][27][28][29] and 3 low quality [30][31][32] F I G U R E 1 PRISMA flow chart of the selection process for the delayed discharge review. *Two studies provided data on costs and quantitative variables. 30 Table 2). Two of the eight qualitative studies were removed due to low reliability and usefulness, 33,34 as determined using the 6 criteria for assessing study quality (Table 3).
In this review, health economic studies refer to those studies reporting on cost of delay. These studies were quantitative and looked at cost implications of delayed discharge. There were 2 studies 30,32 reporting data on costs and health outcomes, both deemed with low methodological quality.

| Summary of the quantitative studies
The characteristics and methodological quality of the ten quantitative studies are summarized in Table 2. Seven cohort studies, either prospective (3) or retrospective (4), and 3 cross-sectional studies were identified. Eight studies used checklists (eg the Appropriateness Evaluation Protocol, AEP) or health professionals' criteria to identify patients who were delayed for non-medical reasons. 23,25,26,[28][29][30][31]

| Impact of delayed discharge on health outcomes
Ten studies explored the impact of delayed discharge on health outcomes. These studies typically carried out assessments at 2 time points (at baseline and at discharge or during the delay period), and some compared the results to non-delayed patients. [24][25][26]32 Two studies explored the factors associated with delayed discharges and inappropriate stays in hospital. 23,28 The potential impact of delayed discharge on mortality was examined in 2 studies with moderate methodological quality. One study found that 5 of the 58 patients suffering delayed discharge (8.6%) died in hospital after they were medically fit for discharge. 27 The other demonstrated a significant association between increased risk of mortality and inappropriate stay during the first year after discharge. 29 A prospective study conducted in a district general hospital in the UK which focused on patients over 65 years old, with moderate methodological quality, found that 7 of 58 cases of delayed discharge (12.1%) developed at least 1 medical complication prolonging their hospitalization. 27 Conditions included "urinary tract infection, recurrent dizziness, leg swelling, poor oral intake, lower respiratory tract infection, bronchopneumonia and Clostridium difficile diarrhoea." A retrospective cohort study conducted in Israel 29 with moderate methodological quality showed that among patients who had been medically fit for discharge, 9 (8.7%) suffered from pneumonia; 14 (13%) suffered from urinary tract infection; 9 had sepsis (8.7%); and 1 (0.96%) patient acquired Clostridium difficile during the inappropriate stay. Another UK study with low methodological quality assessed consecutive patients who sustained proximal femoral fracture over 60 years of age and found that nosocomial infection happened in 58% of patients (99 patients) when inappropriate stay lasted longer than 8 days. 31 Two studies, with high 24 and moderate 26 methodological quality, respectively, evaluated depression and anxiety, one of which found statistically significant differences in levels of depressive symptoms in patients with delays in discharge. 26 Five studies examined impact on daily living activities/mobility, 3 of which had high methodological quality. A UK study 24 found that patients with delayed transfer between hospitals presented worse scores on activities of daily living. In Canada, a cross-sectional study found that there was a significant difference in the Hierarchical Assessment of Balance and Mobility (HABAM) score when clinical stability was achieved the first year after the inappropriate hospital stay. 25 A prospective cohort study that took place in Switzerland found that delayed discharge patients became more impaired in daily living activities, either basic or instrumental, during the prolonged stay. 26 Two UK studies with moderate 27 and high 23 methodological quality showed that delayed discharge had a negative impact on mobility and daily living activities.

| Synthesis of results from qualitative studies
Features of the 6 qualitative studies, including quality assessment, are summarized in Table 3.  T A B L E 3 (Continued) of: (i) patient experience, (ii) patients' physical health, (iii) staff/health professionals, and (iv) hospitals.

| Impact on patient experience
The data on patient experience was derived from interviews with patients and health professionals. Delays in discharge affected patients' emotional state. Hospitals were considered poor environments for a protracted stay because wards could be noisy even at night, they lack personal privacy 35-37 and they limit patients' autonomy. 38 A knock-on effect of delayed discharge was pressure on hospitals to expedite other patients' discharge. Where discharge was rushed to free up beds, this could cause patients to worry and become dissatisfied with services, particularly when they felt unable to ask questions. 39 It also sometimes led to disengagement from discharge planning. 38 Patients expressed anxiety and other negative feelings about delays. Emotional outcomes of delay included tedium or boredom, depression and loss of independence. [35][36][37] One elderly patient awaiting assessment on a stroke rehabilitation ward communicated a sense of desperation and reported being "so low" due to not knowing when they could leave hospital. 38

| How experiences of delay affect physical health
Due to a lack of movement and loss of independence, patients expressed concern about deterioration in their general health while in hospital 38 and an increased risk of bed sores. [35][36][37] Patients also reported that pressure to discharge them due to bed shortages meant that they had not recovered sufficiently prior to discharge 40 and that their needs had not been addressed effectively. 39 In some cases, this led to avoidable readmissions to hospital. 41

| Experiences of staff
Discharge delays caused stress for staff for several reasons: they lengthened waiting lists (which the staff had responsibility to reduce) and created pressure for some patients to be discharged home, which in turn created frustration and guilt among staff who felt patients were being pressured to leave hospital. 40 The strong management focus on reducing delayed discharges made staff feel "under the cosh" [35][36][37] and adversely affected interprofessional relationships. The consultants and managers concerned with achieving government targets were most likely to pressure other staff to discharge patients and to become "disillusioned" about their care role 41 because they were preoccupied with discharging patients, rather than providing care to those in need. [35][36][37]41,42 Moreover, some procedures for addressing delayed discharge were perceived by staff to have "systematized" or dehumanized patients. 41 Finally, some health professionals reported negative reactions towards patients, including "blame" for contributing to delays. 20

| Experiences at the hospital level
A number of organizational effects of delayed discharge were described. Adverse effects on the hospital culture included "poor mood on the ward," which, in turn, had a knock-on effect on the mood of patients' receiving care. 43 At the local system level, delays in transfer between health and social care providers contributed to "blame" and mistrust in interorganizational relationships. 35-37 A hospital social worker described the "extra pressure" and "flak" they received from other staff in relation to delays. [35][36][37] Information sharing between health and social care was also undermined by delays. [35][36][37] Furthermore, use of "Section 5" in England-which gives notice of the proposed discharge date to social services and can trigger fines-had a negative effect on relations between health and social care, demotivating staff and causing tensions where this measure was used to pressure social workers to find placements. 35-37

| Summary of the health economic studies
The features of the 19 health economic studies are summarized in At least 4 studies referred to the cost of inappropriate admissions for specific health conditions of surgical procedures. [45][46][47][48] An Irish study 49 estimated an average of extra 8653 bed-days per year by elderly patients that were waiting to be placed in long-term care facilities. These extra days represent a loss in terms of opportunity cost, as they could be used for other interventions or to avoid overnight stays in A&E. 50 Delayed discharge seems to be positively correlated with social isolation and referral to a public-funded rehabilitation unit, whereas being admitted from an institution appears to be a protective factor for older patients presenting with hip fracture.
According to a recent prospective study, 51  . Therefore, to improve efficiency in a trauma system, it will be necessary to implement changes from acute care through to rehabilitation.
Discharge delays can have an impact not only on other admissions, but also on many other hospital services, including staff workload, physiotherapy, medical or surgery review, radiology, laboratory, pharmacy, transport, social and therapy services. In a prospective study conducted in London, they estimated that the repercussion of delayed discharge on other services can cost £0.5 million annually. 44 In the Netherlands, 58 an intervention to improve the discharge process reduced almost 50% of the inappropriate hospital stay, with a consequent improvement in trauma care quality and financial efficiency.  A major weakness of our findings is the lack of comprehensive evidence on delayed discharge from a single country's health system.

| Strengths and limitations
This might have more easily allowed exploration of structural or policy related explanations. In the absence of sufficient data for intranational examination, we followed the well established route of examining research from OECD countries. 59 The OECD has been a prime source of international comparative data on health systems for many decades.
A second potential weakness is the low methodological quality: we

| Implications for practice, policy and research
Recent policies on delayed discharge advocate "system-level" approaches to addressing delays, for example encouraging shared leadership and integration across health and social care. 6,60,61 Previous research has found examples of joint working between health and social care that improved working relationships and facilitated ownership of delays. [35][36][37] This review confirms the importance of system-level approaches that address the effects of delay at multiple levels. It is recognized that delayed discharge is a contested concept, due to differing interpretations in policy and among health and social care providers on the reasons for, and measurement of, delays. 62 For the purposes of this review which included studies from different countries, we have defined delayed discharge as a stay for a patient who is beyond being deemed "medically fit" to leave hospital, but is unable to do so for non-medical reasons.
However, we acknowledge that the nature of delayed discharge is not fixed, but varies across health systems and is locally negotiated by health and social care providers in response to different policy and organizational environments. 63 For example, NHS England's definition of a patient that is ready for discharge includes the safety and appropriateness of the discharge destination as well as clinical considerations, whereby a discharge may only occur when (i) a clinical decision has been made that the patient is ready for transfer, (ii) a multidisciplinary team decision has been made that the patient is ready for transfer, and (iii) the patient is safe to discharge/transfer. 64 The variety of reasons for delay, including those linked to local and national policy, may add to the self-perceptions of patients, particularly frail older people, of being "bed blockers" that contribute to delays, which may adversely affect their health by causing further stress and anxiety linked to feeling at fault for delays. [35][36][37]43 Even where there is variation in defining delays in discharge there is broad acceptance among professionals working in the health service that it continues to be a significant problem that impacts the provision of care. 65 Our findings provide renewed emphasis for the need to standardize the approach to measuring delays and invest in delayed discharge the issue as a priority given its impact not only on patients' health and experiences of care, but also on staff well-being, interprofessional relationships and information sharing; and on distal (in addition to proximal) costs. 66 We have highlighted that the real cost of delayed discharge must include unit level (eg LOS or infection costs or cancelled operations), organizational and local system-level impacts. However we should also consider other costs that have not been quantified yet (eg the impact on staff morale, staff turnover, agency fees, cost of social care, nursing homes), but could have a huge economic impact. Delayed discharge represents an opportunity cost that is not necessarily equal to the forgone margin from a new admission. There are many repercussions on other services, such as staff, physiotherapy, radiology, pharmacy, surgery, occupational therapy, laboratory and lack of downstream beds that should be considered when assessing the costs of delayed discharge. 44 Further attention should also be placed on societal costs related to productivity losses due to delay discharge of patients in working age or their caregivers, transport costs to visit delayed patients and impacts on other sectors. 67 Finally, to assess the longer term impact of delays, prospective cohort studies are required that combine routine data from health and social care databases and supplement this with additional process and outcome data.