Statistics from Altmetric.com
Editor,—I read with interest Goodacre's review of chest pain observation unit (CPOU) experience in the United States.1 While the title raises a critical question “Should we establish chest pain observation units in the UK?” the subsequent review is unable to help us answer this question. This is because the alternatives to CPOU are likely to vary greatly in the two countries. In the United Kingdom many patients judged to be at low risk will be discharged from the accident and emergency department compared with the more common “routine” inpatient observation in the United States. Indeed in the three randomised studies identified, CPOU was compared in Farkouh's study with monitored cardiology beds and in the studies of Roberts and of Gomez with inpatient telemetry monitoring and hospital admission respectively. This strategy was despite the fact that in the latter two cases the subjects were at “low risk of myocardial infarction”. This definition refers to a less than 7% risk using the computer protocol of Goldman et al.2 The conclusion drawn in the abstract is that “there is no strong evidence that a CPOU will improve outcome if routine practice is good” but it would be my contention that it is far from likely that current practice in the UK has been shown so to be.
Unfortunately the title and abstract are what grab the eye and indeed Minerva announces in an ensuing edition of the British Medical Journal that “Dedicated units sound like a good idea but there's little evidence that they save lives or prevent inappropriate discharge.”3
I whole heartedly agree with Goodacre that further studies should be done to determine if CPOU units should be used in the UK.
The author's reply
It is true that my conclusion regarding outcomes was not based upon the cost studies listed in table 2. From these studies I concluded that the chest pain observation unit (CPOU) is cost saving in the United States but this may not necessarily be reproduced in the United Kingdom. If the introduction of a CPOU leads to increased rates of referral to coronary care or for angiography, or to CPOU assessment of patients who would otherwise be directly discharged, it is possible that costs may be increased. Therefore we must either demonstrate that cost savings are reproduced in the UK or demonstrate that a CPOU will improve outcomes.
Examination of emergency department disposition patterns provides a theoretical mechanism by which the CPOU may improve outcomes but does not in itself constitute strong evidence. Historical evidence of missed myocardial infarction can be compared with modern practice in US CPOUs to conclude that they improve such outcomes (reference 5 above) but the limitations of this analysis are discussed in my review.
Had I concluded that “there is strong evidence that the CPOU will not improve outcomes” I would indeed have required a power calculation to assess the possibility of a (false negative) type 2 error. I did not. The distinction is important; lack of evidence of benefit should not be confused with evidence of lack of benefit. It is indeed possible that the CPOU will improve outcomes in the UK but evidence is required.
I share the concerns of both correspondents regarding the quality of acute chest pain assessment in the UK. The conclusions of my review should not be taken as supporting present practice in any way. Indeed, as I stated, descriptive studies show that CPOUs are a safe and practical means of assessing patients with chest pain. No such evidence exists to support our present approach.
Evaluation of the role of the CPOU in the UK will be challenging but offers an excellent opportunity to develop a cost effective, evidence-based service for our patients.