Statistics from Altmetric.com
We read with great interest the two recent articles on mortality in meningococcal disease.1,2 While we would agree with the message contained in both articles, namely that the mortality associated with this condition has decreased with time, we have serious concerns regarding the presentation of the data in the paper from the St Mary’s group.
Booy and colleagues report a crude mortality of 2% for the year 1997,2 a figure that has generated considerable media interest. Several reasons are cited for this falling mortality: the provision of mobile intensive care, meticulous attention to stabilising the patient whilst in the district hospital, and the existence of a specialist “sepsis” intensive care unit. However the way in which the mortality data are presented demonstrate several contradictions. Booy and colleagues purport that intensive care begins from the time the retrieval team is contacted, indeed they calculate PRISM mortality risk from this time, yet mortality is calculated only from those patients who physically arrive back in their own unit. It is well known that mortality from meningococcal disease is greatest in the first 6 hours, primarily from myocardial failure.3 To cull non-survivors before PICU admission thus creates a self fulfilling prophesy.
To illustrate this, we present data from our own unit (Guy’s Hospital), which cover a similar geographical catchment area to St Mary’s. The table shows mortality data for the period January 1998 to November 2001, illustrating time of death after arrival of the retrieval team. For the sake of completeness, we have also included the four deaths that occurred before the arrival of the team. Over this period we have undertaken 183 retrievals on patients with severe meningococcal disease; 147 (80%) of these required mechanical ventilation and/or inotropic support. 12% of these patients presented with meningitis alone, the remainder with septic shock. The overall crude mortality is 8.2% (15/183), which includes four patients who died before arrival of the retrieval team. Our death rate becomes comparable to that of St Mary’s if we exclude patients who die within 6 hours of the retrieval team’s arrival, producing a mortality of 4/176 (2.3%).
Our unit policy is one of rapid stabilisation before transfer, as evidenced by a median time spent out of the PICU (the sum of the time spent in the district general hospital and the transit time back to PICU) of 2 hours 35 minutes. This resulted in only one death in the district general hospital, none during transfer, but a considerable proportion in the early hours following PICU admission.
It is our impression that the St Mary’s retrieval process is a considerably longer one, which may artifactually reduce PICU mortality. We would therefore ask that the St Mary’s group present their data in a similar fashion, including retrieval times. Specifically, were the 29 deaths before physical admission to the PICU occurring whilst under the management of the retrieval team at the local hospital (and thus under PICU management, by their own definition)? If so, mortality should be adjusted accordingly. Second, has this trend continued in subsequent years? This disease attracts media and public attention par excellence. It is therefore vital that outcome data which are accessible to the public and may be used to influence service reorganisation be presented in a transparent manner.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.