INTRODUCTION

In the last few years, progressive reductions in hospital beds, growing social and medical frailty that impedes hospital discharge, and an inadequate availability of community healthcare services have led to a severe lack of hospital beds. Consequently, emergency physicians are forced to admit patients to clinically inappropriate wards.

The so-called outlier, out-lying hospital in-patient, overflow, sleep-out, or boarder1,2,3 is a patient who, because of unavailability of hospital beds in his/her clinically appropriate ward, is admitted wherever an unoccupied bed is. In such a case, clinical management is provided by the medical staff of the clinically appropriate ward (generally, internal medicine), but care is delivered by nursing staff of the hosting ward. An example is a patient with pneumonia who, because for unavailability of beds in internal medicine, is admitted to a surgical ward.

About 7–8% of all admissions every year are outlier patients.2 The phenomenon is common, particularly in countries with a public health system, and could pose a serious threat for quality and safety of patient care.

The aim of the present work is to systematically review literature evidences about such a phenomenon that is another face of hospital overcrowding.

METHODS

We performed a systematic review of studies investigating outliers, published in peer-reviewed journals with no time and language restrictions.

We searched Medline/PubMed and EMBASE using the following terms: ((“Outlier” OR “out-lying hospital in-patient” OR “overflow” OR “sleep-out” OR “boarder” OR “bed-spaced patient” OR “clinically inappropriate ward” AND “mortality”, “length of stay”, “satisfaction”, “adverse event”, “medical error”, “patient safety”)).

The search and screening were conducted by two independent researchers (MLR and ER). Studies were considered potentially eligible for this systematic review if aimed to assess the quality and/or the safety of care for patients admitted to clinically inappropriate units. Our search was supplemented by a hand search of references of included studies. Among them, we found some bed management policies available on hospital websites. They provide recommendations for a safe management of outliers. The search on Medline/PubMed and EMBASE using terms ((“bed management” AND “policy” OR “healthcare policy” OR “hospital utilization”)) did not produce useful results, so we decided not to include them in our review. Figure 1 shows the process and the results.

Figure 1
figure 1

Algorithm of study identification and selection.

Full size image

Initially, we considered pooling some outcomes (mortality, length of stay, and readmission rates) but abstracted data yielded alarmingly high degrees of heterogeneity (I2 > 95%), so we decided to analyze our results thematically. Study characteristics were examined to explain differences in findings (Table 1). We used PRISMA guidelines to report our findings.

Table 1 Characteristics of the Included Studies
Full size table

RESULTS

Our research retrieved 17 eligible papers, mainly studies conducted on medical patients. We divided them in six thematic categories according to the investigated outcome (details in Table 2a–g).

Table 2 Results of eligible studies grouped in six thematic categories
Full size table
  1. a.

    Mortality

    The impact of outlier status on in-hospital mortality was reported in eight studies. Perimal-Lewis et al.9 found that being an outlier patient increases the risk-adjusted risk of in-hospital mortality by over 40% (50% of deaths happened in the first 48 h after admission). Bai et al.4 reported similar findings: the risk of in-hospital mortality was three times higher among “bed-spaced patients” in the first week just when patients need more interventions. They also suggested several possible reasons for this: less patient contact with physicians on the clinically appropriate ward; inadequate communication between physicians and host-allied health team members; different skills and experience of the allied health team on host ward. Santamaria et al.8 reported a mortality increase among outliers in general and Perimal-Lewis et al.6 among outliers affected by dementia. These data were refuted by Stowell et al.10 and by Stylianou et al.1 on large numbers (over 70,000 admissions in 3 years of observation) and by Alameda et al.13 among outliers with heart failure. Stowell et al.10 and Stylianou et al.1 examined also 30-day mortality without finding any increase; Perimal et al.6, 9 instead revealed a nonsignificant increase in outliers. Serafini et al.7 investigated 3828 consecutive patients hospitalized in medicine and geriatrics in 2012 and, after adjustment for age and sex, the risk of death was about twice as high for outlier patients admitted to surgical area versus the medical one (hazard ratio 1.8, 95% CI 1.2–2.5).

  2. b.

    Length of stay (LOS)

    LOS was explored in seven studies. Stowell et al.10 and Stylianou et al.1 found a longer LOS among outlier patients (8 vs 7 days and 7 vs 3 days, respectively), consistent with findings by Alameda et al.13 among outliers affected by heart failure (11.8 vs 9.2 days). Perimal-Lewis et al.9 registered a significantly shorter length of stay among outliers (110.7 h vs 141.9 h). No difference was found by Serafini et al.,7 either for medicine or geriatrics (10 vs 9.8 days and 13 days for both, respectively) or by Bai et al. (5.31 vs 5.97 days; p = 0.1119).4

  3. c.

    Readmissions

    Readmissions have been studied by five studies;1, 6, 7, 9, 10, 13 Perimal9 reported that readmission rates within 7 or 28 days were substantially lower in the outlier group (2.1 vs 1.2% and 2.1% vs 4.9%). Alameda13 found an insignificant increase in readmissions with the same DRG at 30 days among outliers affected by heart failure (15% vs 10%). While a univariate analysis suggested increased hospital admissions, adjustment for various patient characteristics found that outlier status did not affect readmission.1 On the other hand, two studies found increased readmission rates;7, 10 the latter found this to be true in both geriatric (29.9% vs 7.2%, p < 0.0001) and general medicine patients (23.7% vs 16.3%, p = 0.01).

  4. d.

    Other indicators

    Additional investigated variables include rates of VTE prophylaxis and test ordering, finding that outliers had lower rates of VTE prophylaxis,10 though no difference in blood or imaging tests.

    ED stay was longer in patients eventually admitted to outlying wards;6 respiratory patients were less likely to be outliers than other diseases.7

    One study11 found that the “time burden” from visiting patients on outlying wards was significant, nearly doubling the total time spent with patients, though most of this was due to travel time. In addition to taking more time, another study found that elective operations were reduced by almost 15% in presence of outliers boarding on the surgical wards. Another study13 measured a composite outcome, called “in-hospital morbidity” (intra-hospital infection (urinary, respiratory, bacteremia, or others beginning 48 h after admission), intra-hospital hemorrhages (digestive, urinary, or others), and intra-hospital venous thromboembolism). Anyway, in-hospital morbidity was found not statistically different between outliers and inliers (24% vs 18%, p = 0.254).

    On the other hand, outliers were more likely to miss medications12 and resulted in increased rates of calls for in-hospital emergency teams.8

  5. e.

    Perceived quality and safety of care

    Goulding et al. explored quality and safety issues from two—the provider and patient—perspectives,2, 3 finding that both groups were worried. Healthcare providers were concerned about five threats to patient safety: (1) increased workload; (2) poor communication between the two wards; (3) less experience about these patients on clinically inappropriate wards; (4) unsuitable ward environment; (5) characteristics of outlying patients. In addition, patients on inappropriate wards may be perceived as less important and moving patients between wards could disorient older and cognitively impaired patients.3 Patients were worried about not belonging, possible communication deficiencies, medical staff availability, nurses’ experience, and resource availability.2

  6. f.

    Safety issues and solutions

    Four studies evaluated the impact of organizational changes on outliers’ risks. One study15 suggested solutions such as active discharge planning from the admission, increase of transfers from general internal medicine to geriatrics in another building, and implementation of a consultant-led ward round 7 days a week. Another study instituted a “physician of the week”16 to review outlying patients and improve continuity of care, and added a discharge facilitator and a short stay ward for patients and acutely unstable patients who required a high level of medical care. The study by Lepage et al.14 identified five domains of potential failure in the management of outliers: care in emergency department, transfer to the outlying wards, first day of hospital care, care from second day to discharge, day of discharge. They then implemented the following solutions: a doctor, in the clinically appropriate wards, who is in charge of outlying patients each day, a nurse coordinator who facilitates communication between the emergency department, specialty wards, and outlying wards and ensure that the location of outlying patients is known and their medical needs adequately coordinated, and standardized medical records in order to ease the transfer of information between departments and aid health professionals.

    Novati et al.5 significantly reduced outliers (from 6.3 to 5.4%) by implementing an algorithm, supporting rational outward allocation of patients and difficult discharges.

DISCUSSION

To our knowledge, this is the first systematic review on outliers on medicine wards. The literature suggests a possible trend towards increased mortality and hospital readmissions among outliers, though the data was too heterogeneous to pool. The majority of the studies4, 6,7,8,9 found a significant increase of in-hospital mortality rate or risk, especially in the first 2 days when patients are medically more active. Data about 30- or 90-day mortality are sparse. Readmissions were evaluated at different intervals (from 7 to 90 days after discharge) in the collected studies. Three out of five documented a larger proportion of 28-day readmissions among outliers; the fourth study documented an increased risk of readmission, but only at univariate analysis. Data about length of stay (LOS) were too inconsistent across the studies to reach any meaningful conclusions.

In addition to being too heterogeneous for pooling, most of the study designs among the included papers were poor, mainly monocentric, retrospective, based on administrative data, and underpowered.1, 13 On the other hand, the inconsistency of results can be due also to different contexts. For example, the habit of moving stable patients outside to admit unstable ones or planning early the discharge, different availability of community facilities, health services, and social support can contribute to discordance. Nevertheless, delay between admission and medical evaluation, discontinuity of care, errors or delay in tests request/execution, inadequate communication between ward teams, less familiarity with monitoring and treatment by hosting team, and nosocomial complications can variously affect mortality, length of stay, and readmission rate. Worrisome is the literature that suggests specialized wards lead to better outcomes from some conditions, such as stroke, renal failure, burns, asthma, gastrointestinal bleeding, trauma, and cancer.18,19,20,21,22,23

Evidence about other indicators such as proportion of elective surgeries canceled,17 thrombo-prophylaxis, in-hospital infections or in-hospital bleedings, number and appropriateness of investigations, calls to intra-hospital emergency team, and missed medications is limited, but there are other possible drawbacks to being boarded. Moving patients has been shown to increase the risk of healthcare-associated infection (HCAI).24

Two studies exploring patient and provider satisfaction both suggest a perception of reduced quality and safety.2, 3 This can be due to travel time, to lack of established relationships between providers and nurses on the outlying wards, and to worry about patients that are not immediately accessible. The hosting nursing team also feels a sense of inadequacy due to less expertise in the management of outlier’s health problems. Patients feel they do not belong to any ward, feel forgotten, are worried about errors due to staff inexperience, miscommunication, or resource unavailability, and dislike transfers between wards.

All suggested solutions5, 14,15,16 are multi-component as the problem is complex and needs a system approach and have not been rigorously studied, yet. The “best” solutions are likely to be tailored to the specifics of the individual systems.

CONCLUSIONS

Though literature evidence is quite limited and heterogeneous, the outlier status may be associated with worse outcomes. Certainly, patients and health professionals are dissatisfied. The reported solutions are targeted to locally identified problems and have not been rigorously studied.

There is a need to reach a universally accepted definition of outlier, to adequately measure the effect of outlier status on clinical and safety outcomes, and to develop validated tools to analyze and manage a phenomenon that could negatively impact on care and organizational outcomes.

To this aim, FADOI (the Federation of the Associations of Hospital Internists) has planned a multicenter, prospective, well-sized study comparing mortality rate and adverse event rate in outliers and non-outliers, named “Safety Issues and SurvIval For medical Outliers” (SISIFO) study (NCT03651414) that will start at the end of 2018.