Predicting outcomes: Adults vs Babies

An interesting article in Archives of Internal Medicine (Chan PS, Spertus JA, Krumholz HM, Berg RA, Li Y, Sasson C, Nallamothu BK, Investigators. GWTG-RR: A validated prediction tool for initial survivors of in-hospital cardiac arrest. Arch Intern Med 2012, 172(12):1-7. http://archinte.jamanetwork.com/article.aspx?articleid=1162169); using a prospective registry the authors analyzed 43,000 adults who had an in-hospital cardiac arrest and were resuscitated with initial success of the resuscitation, that is, return of a perfusing rhythm. They then produced a clinical prediction algorithm, by in-putting various data points from a randomly selected 2 thirds of the cohort. Using the other third they validated that the algorithm gave good prediction for survival without serious neurologic morbidity.

The authors use the Glasgow Outcome Scale to determine the degree of neurologic morbidity, in that scale a moderate outcome (level 2) is “some disability such as dysphasia, hemiparesis or epilepsy and/or deficits of memory or personality but are able to look after themselves, do shopping and travel by public transport. They may be able to work when special arrangements are made”severe outcome (level 3) is defined as “a conscious patient who is dependent for daily support from another person by reason of mental or physical disability, usually a combination of both ” and the more severe level of disability, level 4, is a persistent vegetative state or coma (level 5 being death or brain death).

Back to the study by Chan. The authors were able to construct a scale which predicted outcomes quite well, from 70% chance of a good outcome in the lowest decile of scores to 2.8% chance of good outcome in the worst decile.

I like to compare such things to the outcomes of preterm babies. A 2.8% chance of a survival without profound disability is equivalent to a 401 g male infant born at 23 weeks without steroids (using the Tyson calculator). Obviously the definitions of disability are not the same.

Two things that are quite different in this study compared to any similar neonatal study. Firstly it is impossible from the paper or even the supplementary data tables to find out how many severely disabled (or worse) survivors there are at each level of prediction. The presentation is just about relatively good survival and its likelihood. Much of our considerations in neonatology are to do with the survival of “handicappped” infants and our attempts to avoid that outcome. I think if this were a neonatal study we would have an indication at each prediction level of the number of deaths, the number of “severely” or “profoundly” impaired infants, as well as the numbers without impairments. We do however know that overall there was a 30% long term survival, and 6% were level 3 or worse, so 24% were level 2 or level 1 (level 1 is no significant impairment). But within each decile we do not know how many survivors were level 3 or worse, we are only informed of the proportion who were level 1 or 2.

The second thing to note is that the authors are able to calculate the impact of already being in a persistent vegetative state or brain dead on the neurological status post-resuscitation! (hint: it isn’t a positive risk factor). There were 2,223 of the patients who survived an arrest who were already in one of these 2 categories before the arrest. I find that scary, that so many patients in this cohort were already in a persistent vegetative state or worse, and still had cardiac massage. (I think I had better go and write my advance directives now).

The other thing that is startling is the commentary in the accompanying editorial: (Huszti E, Nichol G: Prediction of “mostly dead” vs “all dead” after in-hospital cardiac arrest: Comment on “a validated prediction tool for initial survivors of in-hospital cardiac arrest”. Arch Intern Med 2012, 172(12):8. http://archinte.jamanetwork.com/article.aspx?articleid=1162172). You might think that a rate of intact survival which is so low would lead the commentators to discuss the appropriateness of instituting a code when good outcomes are very rare. But no. The comments are of a quite different nature. The last few sentences of that editorial I quote here:

“We note that the easiest way to reduce the large regional variation in outcome after the onset of cardiac arrest are to not attempt resuscitation of any patient or to withdraw care from all patients who seemingly have a poor prognosis. But that strategy would obviously be unacceptable to most of the public and health care providers. Given the limitations described herein, we urge caution to those who consider applying the rule prospectively to guide clinical practice. As Miracle Max noted in Rob Reiner’s film, The Princess Bride (1987): “There’s a big difference between mostly dead and all dead. Mostly dead is slightly alive.” Most members of the public would want health care providers to persevere in caring for a patient who is slightly alive.”

I don’t know if that final comment is accurate, but if so, why do we get such heat for caring for babies who are “slightly alive”? Especially when “slightly alive” for a preterm baby is almost always dramatically better than a 2.8% chance of a good outcome. Maybe we should be lobbying for adults to be treated in the same way as we treat premature babies.

About keithbarrington

I am a neonatologist and clinical researcher at Sainte Justine University Health Center in Montréal
This entry was posted in Neonatal Research and tagged , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s