Decision aids for women with a previous caesarean section

Abstract

Focusing on women's preferences improves decision making

Rates of caesarean section are a cause of concern worldwide, although the problems vary according to the setting. In many poor countries, mostly in Africa, where average rates are 2%, caesarean section is underused because of lack of facilities and trained personnel.¹ In other developing countries such as ones in Latin America and eastern Asia, incidence is 30% of all births or higher, even though large sections of the population lack access to basic obstetric care, while in developed countries it has steadily risen to about 20–25%.¹ Despite such big differences between countries, the modifiable causes of rising caesarean section rates and what to do about them are unclear.

In this week's BMJ, a randomised controlled trial by Montgomery and colleagues looks at the effect of two computer based decision aids compared with usual care in pregnant women who have had a previous caesarean section.² One aid provided structured information about possible outcomes and their probabilities associated with different modes of delivery and left women's preferences implicit; the other was a decision analysis model that required women to define their preferences, while information about probabilities was concealed.

Importantly, one of the outcomes measured in the trial was the actual birth method, which usefully separates how choices are experienced from the option chosen. The trial found that both aids significantly improved the subjective experience of women about their choices compared with usual care. However, rates of caesarean delivery were similar in the information group, and lower in the decision analysis group compared with usual care.

Unlinking the experience of decision making from its outcome brings a refreshing perspective to the problem of overuse of caesarean section. In light of these authors' findings, it is tempting to conclude that the rise in caesarean rates is due to delivery being seen as purely a medical problem, and the solution being guidelines and recommendations. In 1985, representatives of a study group convened by the World Health Organization wrote, “there is no justification for any region to have caesarean section rates higher than 10–15%.”³ At the time, such levels were considered high but acceptable in developed countries. However, now that caesarean rates in many countries exceed 20%, the recommendation has been dramatically overtaken by events. Notably, rates continue to rise despite evidence showing that caesarean delivery may increase the risk of maternal death.^{1 4 5}

Surprisingly little research exists on determinants of caesarean section, at either the aggregate or the individual level.^{6 7} The few randomised trials that have been published found no effect of decision aids on caesarean section rates.^{8 9} This is despite evidence in other areas of medical care showing that decision aids such as pamphlets and videos can improve people's knowledge of the options, create realistic expectations of their benefits and harms, improve decision making, and increase participation in the process.¹⁰ Against this background, Montgomery and colleagues may have opened up a promising new avenue for research.

As the study was underpowered to measure an effect on birth method reliably, this finding requires confirmation. That information alone had no impact on rates of caesarean birth is consistent with the results of previous trials.^{8 9} That decision analysis did have an effect may have two important corollaries.

Firstly, in this study women seem to have been part of the decision making process regarding mode of delivery. Previous trials of decision aids may not have shown a correlation between women's preferences for birth method and the actual birth method because women lacked this decision making power.⁸

Secondly, the result seems to confirm the psychological principle that people do not reliably make decisions involving choice under uncertainty, in the sense that, depending on how the uncertain options are presented, their choices systematically contradict their aims.¹¹ Reasons for this include widespread avoidance of negative outcomes (loss aversion) and difficulties in reasoning about probabilities.

Although this principle is less well recognised in medical decision making, it poses profound challenges for conventional notions of informed choice in medical care. Although a definitive answer must await further research, the present study suggests that women with a previous caesarean section make better choices about mode of delivery when the purely cognitive demands of reasoning about the probabilities of uncertain birth outcomes are separated from their preferences about the outcomes. Interestingly, the study also suggests that this improvement in decision making is possible even when women's subjective experience of the decision making process is less positive. If this hypothesis can be confirmed, it could help bridge the gap between mere knowledge about the outcomes of decisions and effective decision making.