Racial bias unlikely in resuscitation attempts for premature births

A recent article in the New Zealand Herald suggested there is racial bias in decisions to resuscitate premature babies.

Pakeha babies are more likely to get greater attention than Maori, Pacific Island or Indian babies.

The article was based on a comprehensive report from the Perinatal and Maternal Mortality Review Committee (PMMRC). Here are two key passages in the PMMRC report.

The first states: "This analysis found that after adjustment for age, body mass index (BMI), socioeconomic status, parity, smoking, multiple pregnancy, baby sex, and year of birth, there remained a higher risk of death after birth at 20 to 24 weeks' gestation among babies of Maori, Pacific and Indian mothers, which suggests that there were other factors increasing risk for these women."

The second passage states: "Māori, Pacific and Indian live born babies were statistically significantly less likely to have an attempt at resuscitation than babies of other ethnic groupings, and Maori and Pacific babies were significantly less likely to survive compared to babies of 'other' ethnic groupings among all non-anomalous live born babies at 23 to 26 weeks' gestation ..."

"While the reasons for these differences by ethnicity have not been elucidated in the analyses in this report, previous analysis on inequities by ethnicity in New Zealand suggest institutional bias or implicit biases are likely to play at least some part."

For those alarmed by the findings I would like to note a few issues.

Implicit or institutional bias in the provision of medical care may well exist, but in this specific instance, the evidence may not lead exclusively to such a conclusion. In statistical terms these events are not "independent" — there is not an equal chance of babies of all ethnicities presenting with similar complications. Some babies may present greater complications than others.

So the headline could easily be, "Premature children from lower socio-economic backgrounds less likely to be resuscitated". This would be equally true and would focus attention on other social and systemic failures, which are probably more important than racial bias.

Doctors are often faced with agonising trade-offs where trying to save the life of one child may mean taking away time from other children (or other patients) who have better chances of survival. But in focusing on one critical case we often end up ignoring faceless others. It seems cruel and heartless to state it this way, but even with human lives, one often needs to make such trade-offs.

This is a very human reaction, we focus more on "identified" lives than "statistical lives", on problems that seem more immediate than those which are more diffuse. This is why we spend billions of dollars thwarting terrorist attacks which pose minuscule risk to human lives while devoting far less resources to combating scourges like drunk driving or easily preventable diseases like malaria or cholera, which claim thousands of lives.

It is why we are so engrossed with the fate of young soccer players trapped in a Thai cave yet do not pay the same level of attention to hundreds of deaths from a tsunami in Indonesia. A single death is a tragedy while numerous deaths become a statistic.

The PMMRC committee seems motivated, at least in part, by survey evidence suggesting institutional bias among medical students towards particular ethnicities.

But, it is well-established that there are often discrepancies between survey responses and actual behaviour. To go from implicit bias to denying life-saving treatment is a big step. While we may not be able to rule this out, it is not clear this is the only possible explanation.

The Herald article quotes researchers as saying that at times of emotional and/or physical stress we tend to fall back on pre-existing (and possibly unconscious) biases. This argument seems to have led to the PMMRC report's conclusions also.

Unconscious biases would possibly apply to decisions made instantaneously as opposed to more deliberately. It is possible that a police officer (suffering from conscious or unconscious bias) may be more inclined to shoot a fleeing dark-skinned suspect than a fleeing fair-skinned suspect when forced to make a split-second decision.

But the decision to provide or deny medical treatment in cases of premature birth is typically not a decision of this type. There is a long lead-up period with a sequence of events that have taken place in the interim. This is not a decision a doctor is making on the spur of the moment but one arrived at over a period of time and deliberation.

I am sure there are deficiencies in the quantity and quality of medical care in New Zealand that need to be rectified. But differences in care are likely driven primarily by socio-economic factors rather than racial bias.