Mortality among appropriately referred patients refused admission to intensive-care units

doi:10.1016/S0140-6736(96)10018-0

The Lancet

Volume 350, Issue 9070, 5 July 1997, Pages 7-11

https://doi.org/10.1016/S0140-6736(96)10018-0 Get rights and content

Summary

Background

The provision of intensive care is a perplexing issue for clinicians and the public. Concerns about the apparent lack of beds and the appropriateness of the patients admitted are tempered by the high cost of providing this service. As part of a study commissioned by the UK Department of Health, we tested the hypothesis that there is excess mortality among patients who are refused admission to intensive-care units.

Methods

All referrals to six intensive-care units with different numbers of beds were monitored during a 3-month period. Data on mortality 90 days after first referral were obtained from family physicians for all patients known to be alive at hospital discharge. We adjusted, where possible, for confounding, including for age, sex, appropriateness of referral, disease severity, surgery and emergency categories, and bed provision. We did multivariate analysis by multiple logistic regression to compare the adjusted 90-day mortality rates for patients who were refused admission and for those admitted.

Findings

480 patients were admitted and 165 were refused admission. 90 days after referral there had been 178 (37%) deaths among the admitted group and 75 (46%) among the refused group. After multivariate adjustment, 113 patients appropriately referred for intensive care but refused admission to their first-choice intensive-care unit had a relative risk of death of 1·6 (95% CI 1·0–2·5), compared with the group of appropriately admitted cases with medium APACHE II scores for disease severity. Age, the assessed need for treatment or monitoring interventions, and emergency status also contributed to differences in mortality among all referrals. Bed provision did not contribute significantly to excess mortality.

Interpretation

Although this study is observational and case-mix adjustment is incomplete, we found a higher rate of attributable mortality in patients who were refused intensive care, particularly for emergency cases. We question whether the provision of more beds alone would be a solution and conclude that there is an urgent need for more appropriate admission and discharge criteria.

Introduction

Most referrals to intensive-care units are emergencies or prebooked surgical cases. Some patients are refused admission, either because admitting clinicians deem them to be too well or too ill, and, therefore, inappropriately referred; because the units are full; or because there are insufficient trained nurses. We investigated how mortality rates were affected by refusals of admission to a unit, because there is concern that refusals have higher attributable risk of death. There are no published data that have addressed this issue, and this information may be useful for future research and policy.

We report on part of a study on the provision of intensive care in the UK, commissioned by the Department of Health because of concern that expenditure is low compared with other countries.¹ We investigated mortality rates in relation to admission or refusal in six intensive-care units. The other part of the study was on national data and provided background information about the provision of intensive care in England in May and June, 1993.

Section snippets

Methods

Three pairs of intensive-care units were chosen for a 3-month audit of referrals, and all consultants in charge agreed to participate. The units were stratified from areas with three different levels of provision of intensive care beds, derived from national data. The three categories had 0·81–1·85 beds/100000 population, 1·86–2·81 beds/100000, and 2·82–13·83 beds/100000 (based on staffed intensive-care beds per caput).¹ Each unit had to have at least four beds and to have been in operation for

Results

Details of the six hospitals and their intensive-care units are shown in table 1.

817 referrals (541 admissions and 276 refusals) were made during the 3-month study period. After exclusion of 105 repeat admissions and refusals, 24 patients referred during a 3-week period when one unit was closed to new admissions, and 37 children younger than 16 years at first referral, we had data on 483 admissions (including 44 patients first refused but later admitted) and 168 refusals (table 2). Patients

Discussion

Intensive-care beds should be available for patients with disorders from which they may recover, for whom these facilities are essential. We compared 90-day mortality among patients admitted to intensive-care units and among those assessed as appropriately referred but refused admission. In the final multivariate logistic regression model, for appropriately referred refusals compared with similar admissions with medium APACHE II scores, we found an excess adjusted mortality of 60% (of

References (14)

MA Metcalfe et al.
Study of intensive care in England
(1995)
WA Knaus et al.
APACHE II: a severity of disease classification system
Crit Care Med
(1985)
KM Rowan et al.
Intensive Care Society's APACHE II study in Britain and Ireland, I: variations in case mix of adult admissions to general intensive care units and impact on outcome
BMJ
(1993)
P McCullagh et al.
Generalised linear models
(1983)
Guidelines on admission to and discharge from intensive care and high dependency units
(1996)
DL Crosby et al.
Provision of postoperative care in UK hospitals
Ann R Coll Surg Engl
(1994)
A Kilpatrick et al.
A changing role for intensive therapy: is there a case for high dependency care?
Anaesthesia
(1994)

There are more references available in the full text version of this article.

Cited by (170)

Intensive Care Unit prioritization: The impact of ICU bed availability on mortality in critically ill patients who requested ICU admission in court in a Brazilian cohort
2021, Journal of Critical Care
Citation Excerpt :
Another study performed in the United States showed a significantly increased mortality rate than expected among patients who were initially denied admission to ICU because of lack of beds; deaths were either due to early mortality outside ICU or delayed ICU admission [22]. It is essential to admit patients to the ICU to implement early interventions and a close monitoring before their condition deteriorates beyond the point where recovery is possible [8-14,17,22-26]. Therefore, according to our findings, the shortage of ICU resources in increased demand periods may cause a delay in transferring the patient to the ICU, thus affecting outcomes and leading to an increased mortality [9-14,20,24].
To assess hospital mortality in patients who requested ICU admission in court due to the scarcity of ICU beds in the Brazilian public health system and the consequences of these judicial litigations.
Retrospective cohort study that included adult patients from the public health system of the Federal District, Brazil, who claimed ICU admission in court from January 2017 to December 2019.
Of the 1752 patients, 1031 were admitted to ICU (58.8%). Hospital mortality was 61.1% (1071/1752). Of the requests, 768 (43.8%) were made by patients with priority levels III or IV, resulting in the ICU admission of 33.9% of these patients. Denial of ICU admission (p < 0.001) increased mortality. ICU admission reduced hospital mortality in patients classified as priority level I (p < 0.001), priority level II (p < 0.001), and priority level III (p < 0.001), but not as priority level IV (p = 0.619).
A large proportion of patients was denied ICU admission and it was associated with an increased mortality. A considerable portion of the ICU-admitted patients were classified as priority level III and IV, impairing the ICU admission of patients with priority level I which are the ones with the greatest benefit from it.
The problem of the last bed: Contextualization and a new simulation framework for analyzing physician decisions
2020, Omega (United Kingdom)
Faced with a full Intensive Care Unit (ICU), physicians need to decide between turning away a new patient in need of critical care and creating a vacancy by prematurely discharging a current occupant. This dilemma is widely discussed in the medical literature, where the influencing factors are identified, the patient discharge process described and the patient health consequences analyzed. Nevertheless, the existing mathematical models of ICU management practices overlook many of the factors considered by physicians in real-world triage decisions.
This paper offers a review of the medical and mathematical literature on patient discharge decisions, and a proposal for a new simulation framework to enable more realistic mathematical modeling of the real-world patient discharge process. Our model includes a) the times at which discharge decisions are made and setup times for patient transfer from the ICU to a general ward and preparation of an ICU bed for an incoming patient, in order to capture the impossibility of an immediate switch of patients; b) advance notice of the number of patients due to arrive from elective surgery requiring intensive postoperative care and potentially triggering the need for early discharges to avoid surgery cancelations; and c) patient health status (to reflect the dependency of physicians’ discharge decisions on health indicators) by modeling length of stay with a phase-type distribution in which a medical meaning is assigned to each state.
A simulation-based optimization method is also proposed as a means to obtain optimal discharge decisions as a function of the health status of current patients, the bed occupancy level and the number of planned arrivals from elective surgery over the following days. Optimal decisions should strike a balance between patient rejection and LoS reduction.
This new simulation framework generates an optimal discharge policy, which closely resembles real decision-making under a cautious discharge policy, where the frequency of early discharge increases with the ICU occupancy level. This is a contrast with previous simulation models, which consider only the triage of the last bed, disregarding the pressures on physicians faced with high bed occupancy levels.
Impact of triage-to-admission time on patient outcome in European intensive care units: A prospective, multi-national study
2019, Journal of Critical Care
Citation Excerpt :
For critically ill patients, care in an ICU affords a survival advantage over non-ICU settings [1-8].
Ubiquitous bed shortages lead to delays in intensive care unit (ICU) admissions worldwide. Assessing the impact of delayed admission must account for illness severity. This study examined both the relationship between triage-to-admission time and 28-day mortality and the impact of controlling for Simplified Acute Physiology Score (SAPS) II scores on that relationship.
Prospective cross-sectional analysis of referrals to eleven ICUs in seven European countries between 2003 and 2005. Outcomes among patients admitted within versus after 4 h were compared using a Chi-square test. Triage-to-admission time was also analyzed as a continuous variable; outcomes were assessed using a non-parametric Kruskal-Wallis test.
Among 3175 patients analyzed, triage-to-admission time was 2.1 ± 3.9 h. Patients admitted within 4 h had higher SAPS II scores (33.6 versus 30.6, Pearson correlation coefficient −0.07, p < 0.0001). 28-day mortality was surprisingly higher among patients admitted earlier (29.6 vs 25.2%, OR 1.25, 95% CI 0.99–1.58, p = 0.06). Even after adjusting for SAPS II scores, delayed admission was not associated with higher mortality (OR 1.08, CI 0.83–1.41, p = 0.58).
Even after accounting for quantifiable parameters of illness severity, delayed admission did not negatively impact outcome. Triage practices likely influence outcomes. Severity scores may not fully reflect illness acuity or trajectory.
Intermediate care to intensive care triage: A quality improvement project to reduce mortality
2017, Journal of Critical Care
Medical patients whose care needs exceed what is feasible on a general ward, but who do not clearly require critical care, may be admitted to an intermediate care unit (IMCU). Some IMCU patients deteriorate and require medical intensive care unit (MICU) admission. In 2012, staff in the Johns Hopkins IMCU expressed concern that patient acuity and the threshold for MICU admission were too high. Further, shared triage decision-making between residents and supervising physicians did not consistently occur.
To improve our triage process, we used a 4Es quality improvement framework (engage, educate, execute, evaluate) to (1) educate residents and fellows regarding principles of triage and (2) facilitate real-time communication between MICU residents conducting triage and supervising physicians.
Among patients transferred from the IMCU to the MICU during baseline (n = 83;July–December 2012) and intervention phases (n = 94;July–December 2013), unadjusted mortality decreased from 34% to 21% (p = 0.06). After adjusting for severity of illness, admitting diagnosis, and bed availability, the odds of death were lower during the intervention vs. baseline phase (OR 0.33; 95%CI 0.11–0.98).
Using a structured quality improvement process targeting triage education and increased resident/supervisor communication, we demonstrated reduced mortality among patients transferred from the IMCU to the MICU.
Occupancy of the Departments of Intensive Care Medicine in Catalonia (Spain): A prospective, analytical cohort study
2015, Medicina Intensiva
Ante la demanda por aumentar el número de UCI o su capacidad, o por regionalizar servicios, es conveniente conocer su efectividad actual.
Analizar la ocupación diaria de los Servicios de Medicina Intensiva (SMI) de Catalunya y con qué frecuencia pacientes críticos no pueden ingresar en UCI y/o son derivados a otros centros.
Prospectivo observacional multicéntrico.
Treinta y cinco de los 40 (87%) SMI de Catalunya.
Registro diario.
Camas abiertas y camas libres, pacientes no altados de UCI por falta de camas en planta, pacientes críticos no ingresados en UCI por falta de cama y cuántos fueron derivados a otros centros. El análisis estadístico es descriptivo de cohortes.
Los promedios diarios fueron: 383 camas de UCI abiertas, 58 (15%) camas libres y 16 pacientes no dados de alta por falta de cama en planta. Cada día 6 pacientes en promedio no ingresaron en UCI por falta de camas y uno (16%) fue derivado a otras UCI. La ocupación media fue del 83 ± 19% con un 100% de ocupación en el 35% de los registros. El subanálisis de los 24 hospitales públicos mostró una ocupación ligeramente superior (87 ± 16%) con 100% de ocupación en el 38% de los registros.
La ocupación media de las UCI de Catalunya podría parecer correcta, pero existe sobresaturación muy frecuentemente en ciertas unidades lo que dificulta el ingreso precoz y además provoca que una cantidad significativa de pacientes críticos no puedan ingresar en UCI.
Before deciding increases in the number or capacitance of Intensive Care Units (ICUs), or the regionalization of Units, it is essential to know their present effectiveness.
To analyze the daily occupancy rate of ICUs in Catalonia (Spain) and the frequency of denied admission due to lack of capacity.
A prospective, observational multicenter study was carried out.
A total of 35 out of 40 ICUs of Catalonia (87%).
Daily registry.
Open beds and free beds, patients not discharged due to unavailability of ward beds, critically ill patients not admitted due to a lack of ICU beds, and rate of transfer to other ICUs. Statistical analysis: A descriptive cohort analysis was made.
Daily averages were 383 open ICU beds, 58 available beds (15%), and 16 patients not discharged due to unavailability of ward beds. Each day 6 patients on average were not admitted due to a lack of ICU beds, and one of them (16%) was transferred to another ICU. The mean occupancy rate was 83 ± 19%, and a 100% occupancy rate was reported in 35% of the registries. A subanalysis of the 24 public hospitals demonstrated slightly higher occupancy (87 ± 16%), with a 100% occupancy rate reported in 38% of the registries.
The mean occupancy rate of Catalonian ICUs may appear correct, but in some Units over-occupancy very often precludes early ICU treatment and even ICU admission for a significant number of critically ill patients.
Racial & ethnic disparities in geographic access to critical care in the United States: A geographic information systems analysis
2023, PLoS ONE

View all citing articles on Scopus

View full text

ArticlesMortality among appropriately referred patients refused admission to intensive-care units

Summary

Background

Methods

Findings

Interpretation

Introduction

Section snippets

Methods

Results

Discussion

Study of intensive care in England

APACHE II: a severity of disease classification system

Crit Care Med

Intensive Care Society's APACHE II study in Britain and Ireland, I: variations in case mix of adult admissions to general intensive care units and impact on outcome

BMJ

Generalised linear models

Guidelines on admission to and discharge from intensive care and high dependency units

Provision of postoperative care in UK hospitals

Ann R Coll Surg Engl

A changing role for intensive therapy: is there a case for high dependency care?

Anaesthesia

Articles
Mortality among appropriately referred patients refused admission to intensive-care units