An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
. 2001 Sep 15;323(7313):629.
Mortality after discharge from intensive care
Only normalisation of physiology will reduce risk of mortality after discharge
Editor—The triage model described by Daly et al to identify patients at higher risk of death after discharge from intensive care seeks to address a number of important issues.1 Daly et al used five variables (patient's age, chronic health points, length of stay in intensive care, acute physiology score, and cardiothoracic surgery) to produce a predictive model that gave a relative risk of death of 9.44 in the developmental group (mortality 14% in those at risk, 1.5% in those not at risk according to this model) and 6.76 in the validation group (mortality 28% in those at risk on day of discharge, 4% in those not at risk in the 48 hours before discharge). This adds further statistical background to previous studies, which had highlighted four of these variables as risk factors at discharge from intensive care.2,3 The fifth variable, cardiothoracic surgery (57% of the developmental model), makes this group atypical of most intensive care units in the United Kingdom, although this point is acknowledged in the internet version of the paper.
Daly et al claim that if patients at risk on day of discharge stay an extra 48 hours in intensive care, mortality after discharge may be reduced by 39%. This piece of statistical fast footwork is given, although no prospective component to the study shows that an extra 48 hours in intensive care will reduce the risk of (any or most or all) patients. Of the five factors in the model, only normalisation of physiology will reduce the risk of mortality after discharge (as is noted in the internet version of the paper). It may be either not possible or take much more than 48 hours to reduce the risk in an individual patient; thus the extrapolation from a predictive triage model to conclusions regarding reduction in mortality and resource requirements for 48 hours longer stay is invalid. This may be what McPherson alludes to in his accompanying editorial.4
No consideration of the relative timing of deaths after discharge was made. “Early” deaths, within, say, 48 hours, may reflect precipitate discharge or communication problems whereas late deaths may reflect more the standard of ward care.
None of the 20 intensive care units in this study were in hospitals with high dependency units (at that time). The advantages of such stepdown care have been long and widely recognised.5 Further consideration is also merited of the cause of death of patients after discharge from intensive care. Although this paper excluded discharges deemed “not for resuscitation,” no numbers are given. Previous studies have found 25% of deaths after intensive care were “expected” at discharge.2
References
1.Daly K, Beale R, Chang RWS. Reduction in mortality after inappropriate early discharge from intensive care unit: logistic regression triage model. BMJ. 2001;322:1274. doi: 10.1136/bmj.322.7297.1274. . (26 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Wallis CB, Davies HTO, Shearer AJ. Why do patients die on general wards after discharge from intensive care units? Anaesthesia. 1997;52:9–14. doi: 10.1111/j.1365-2044.1997.003-az002.x. [DOI] [PubMed] [Google Scholar]
3.Ridley S, Purdie J. Cause of death after critical illness. Anaesthesia. 1992;47:116–119. doi: 10.1111/j.1365-2044.1992.tb02006.x. [DOI] [PubMed] [Google Scholar]
4.McPherson K. Safer discharge from intensive care to hospital wards. BMJ. 2001;322:1261–1262. doi: 10.1136/bmj.322.7297.1261. . (26 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Association of Anaesthetists of Great Britain and Ireland. The high dependency unit. London: Association of Anaesthetists of Great Britain and Ireland; 1991. [Google Scholar]
BMJ. 2001 Sep 15;323(7313):629.
Research in intensive care needs to find balance between scientific method and ethics
Editor—Using an observational dataset of 13 924 patients admitted to 20 intensive care units, Daly et al have shown a surplus mortality adjusted by case mix among patients discharged to the wards, using a risk index with a cut off point for risk of death of 60%.1-1 This suggests (if the model is correct) that a substantial proportion of patients must have been discharged from intensive care with a risk of death higher than this and indicates the need for prospective studies that include the circumstances surrounding death in the wards after discharge from intensive care units. We are planning such a study in the West Midlands.
McPherson says in his accompanying editorial, in which he calls for randomised studies, that provision of intensive care at the margin of possible benefit simply has to be assessed by random allocation like everything else, without providing suggestions about how this might be achieved in the context of emergency care.1-2 He implies that the intensive care community is reluctant to expose its practice to scientific evaluation, using as examples the albumin controversy and arguments against randomisation on ethical grounds.
Most intensive care practitioners show enthusiasm for evaluation of their practice but have concerns about methodological issues.1-3 Prospective randomised evaluation of a predictor of risk is certainly feasible between hospitals, although the potential for confounding in terms of variation in structures and processes is considerable. Random allocation of critically ill patients to different levels of care in an institution is complicated by ethical difficulties generated by the likely absence of equipoise1-4 and the Hawthorne effect, as anyone familiar with acute medical care will understand. Large observational databases containing validated information from many thousands of patients provide an important alternative in this context and may well be more robust than meta-analyses in terms of the validity of their output.1-5
A research method based on collaborative networks, observational databases to adjust for case mix, and agreed standards of care should form the basis for evaluating existing technologies, incorporating prospective randomised controlled trials where this is possible. As clinical researchers, we have a duty to find a balance between rigorous scientific method and the ethical problems associated with (in this instance) persuading patients or their relatives that they should be allocated to early discharge from intensive care to understaffed and overworked hospital wards.
References
1-1.Daly K, Beale R, Chang RWS. Reduction in mortality after inappropriate early discharge from intensive care unit: logistic regression triage model. BMJ. 2001;322:1274. doi: 10.1136/bmj.322.7297.1274. . (26 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
1-2.McPherson K. Safer discharge from intensive care to hospital wards. BMJ. 2001;322:1261–1262. doi: 10.1136/bmj.322.7297.1261. . (26 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
1-3.Vella K, Goldfrad C, Rowan K, Bion J, Black N. Use of consensus development to establish national research priorities in critical care. BMJ. 2000;320:976–978. doi: 10.1136/bmj.320.7240.976. [DOI] [PMC free article] [PubMed] [Google Scholar]
1-4.Morris AD, Zaritsky AL, LeFever G. Evaluation of ethical conflicts associated with randomized, controlled trials in critically ill children. Crit Care Med. 2000;28:1152–1156. doi: 10.1097/00003246-200004000-00039. [DOI] [PubMed] [Google Scholar]
1-5.LeLorier J, Gregoire G, Benhaddad A, Lapierre J, Derderian F. Discrepancies between meta-analyses and subsequent large randomized, controlled trials. N Engl J Med. 1997;337:536–542. doi: 10.1056/NEJM199708213370806. [DOI] [PubMed] [Google Scholar]
BMJ. 2001 Sep 15;323(7313):629.
Only community debate on appropriate end of life care will limit ballooning budget
Editor—McPherson's assertion that intensive care is outside the evidence based paradigm does not stand up to the evidence.2-1 The management of patients in intensive care units is now more evidence based than ever before, and many decisions made by specialists in intensive care have been tested by the rigour of a randomised controlled trial.2-2 Far from dismissing the Cochrane review, the intensive care community has engaged in vigorous debate on this issue. In Australia we are about to start recruitment to what will become the largest intensive care trial ever conducted—a double blind placebo controlled trial of albumin in fluid resuscitation under the auspices of the Australian and New Zealand Intensive Care Society Clinical Trials Group.
Yet it seems that it is the use of intensive care itself that McPherson considers not evidence based. Again, he is wide of the mark. We contend that, more than any other discipline, the specialists in intensive care and their specialty have been subject to trial.2-3 It is true that often these studies have not been randomised, but their results are supported by the optimisation studies in surgical patients at high risk, in which randomisation has taken place.2-4
McPherson represents a traditional view in the United Kingdom that is counterproductive to change. As Daly et al show in the accompanying paper,2-5 there are too few intensive care beds in the NHS. Trials to prove the effectiveness of intensive care will not influence spending since a life lost on the ward costs less than one saved in intensive care. There should be enough intensive care beds for those who need them, and available beds must be used wisely.
Nevertheless, where intensive care is recognised as highly effective—for example, in Australia—the demand for further beds may become insatiable. Unfortunately, as a profession we are getting worse at saying “no,” and wise use of beds in intensive care is already passing out of the specialist's control. To put the onus of justifying expenditure for intensive care units on to intensive care specialists is unfair, as only a major general community debate about appropriate end of life care can limit the ballooning budget. We doubt that this will happen because the community, in Australia and the United Kingdom, just does not have the stomach for it.
References
2-1.McPherson K. Safer discharge from intensive care to hospital wards. BMJ. 2001;322:1261–1262. doi: 10.1136/bmj.322.7297.1261. . (26 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
2-2.Scheinkestel CD, Davies AR, Bristow PJ. Evidence-based medicine in intensive care. Med J Aust. 2001;174:526–527. doi: 10.5694/j.1326-5377.2001.tb143406.x. [DOI] [PubMed] [Google Scholar]
2-3.Vincent JL. Need for intensivists in intensive-care units. Lancet. 2000;356:695–696. doi: 10.1016/S0140-6736(00)02622-2. [DOI] [PubMed] [Google Scholar]
2-4.Wilson J, Woods I, Fawcett J, Dibb W, Morris C, McManus E. Reducing the risk of major elective surgery: randomised controlled trial of preoperative optimisation of oxygen delivery. BMJ. 1999;318:1099–1103. doi: 10.1136/bmj.318.7191.1099. [DOI] [PMC free article] [PubMed] [Google Scholar]
2-5.Daly K, Beale R, Chang RWS. Reduction in mortality after inappropriate early discharge from intensive care unit: logistic regression triage model. BMJ. 2001;322:1274. doi: 10.1136/bmj.322.7297.1274. . (26 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
