Raising the rates of circumcision
Measures designed to improve public health have to serve several masters: you need to change the behaviour of the largest group of people possible, and you need to do it in a way that’s cost effective and time efficient. Traditionally, providers have adopted largely a one size fits all approach but research by Sema Sgaier suggests that, when you do this, the message is lost on large sectors of society. Instead, she’s investigating the use of a process called “psychographic-behavioral segmentation” that uses machine learning to divide the population up into groups that need to be targeted a specific way. She’s looking at how to maximise the uptake of circumcision among men to prevent HIV infection. She explained the strategy to Chris Smith...
Sema: Voluntary medical male circumcision is an HIV prevention intervention, a pretty effective one that has been prioritised in 14 eastern and southern African countries. It’s a one-time intervention and it’s highly cost-effective in settings where HIV prevalence is high and male circumcision prevalence is low. By now, the community had circumcised about 12 million men. This was 8 million short of the 20 million circumcisions that was set as the global target. And so really, the goal here was how do we bridge the 8 million gap? The typical approach is that we were using, building awareness, and having a one-size-fits-all approach was not working.
Chris: Because there's a lot of diminishing returns, isn’t there? You end up to get the last few percent, you have to spend more and more, and more to get there. But you could make the intervention strategy far more cost-effective and far more rapid if you know who to go after, who’s most susceptible to the message and target them first.
Sema: Absolutely. You're absolutely correct. So, we were getting what we called the low-hanging fruit – those were that were already committed, that were ready, that were younger, easier to circumcise. If we look at the trajectory, the trends of circumcision, basically, it’s been getting harder and harder. And so, we have to be much more nuanced and we have to be much more targeted in what we do to be able to spend our dollars in the best way.
Chris: Of course, one point about this is that you need to know who those groups are or you need to know who the people are who are susceptible to which messages so that you know who to go after.
Sema: That’s right. So, we did a large scale study in 2 of the 14 countries, specifically in Zambia and Zimbabwe to really be able to determine these psychographic behavioural segments. Essentially, the first phase was a qualitative phase. The goal of that phase was really to understand the full set of internal factors that drive men to either circumcise or not. After that, we did a large scale quantitative study of 4,000 men in each countries where the factors that we identified in the qualitative study were measured extensively. And we used machine learning algorithms on this data to be able to develop groupings or segments of men.
Chris: And why is this a breakthrough in this particular space because insurance companies know this, people selling chocolate bars on television know that there will be subsets of the audience that they're going for? So, why is it a big step forward in this context?
Sema: In this context, we have been so-called segmenting people. I mean, there is the notion that people are different but it’s mostly been on demographic factors. So, we think of people as young versus old or urban versus rural, or rich versus poor. This idea of really being nuanced on what's driving choices and behaviours has not been common practice.
Chris: What were the segments that you highlighted in the two countries and were they indeed the same?
Sema: So, in Zimbabwe, we identified six segments and in Zambia, we identified seven segments. They're were quite different especially on the internal factors such as the beliefs and motivations. I’ll give you an example to illustrate that. So in Zimbabwe, two segments: one is the ‘enthusiast’ and this is just a name we gave them and another one is the ‘scared rejecters’. The ‘enthusiasts’ are quite motivated when it comes to circumcision. They believe in all of the benefits. They emotionally associate with a sense of achievement when they get circumcised. But there's just one thing that’s holding them behind. That’s understanding how to manage the healing process and the pain. When we look at ‘scared rejecters’, these have very low motivation. They're mostly older, they're married. They believe in the benefits but they just don’t see that it’s relatable to them. They actually aren't much influenced by friends. So really, the people that they're mostly looking up to are healthcare providers. And the they're really concerned about the pain, the possible infections from the procedure, and really want much more additional information.
Chris: If one looks at where is the HIV risk, which of those segments has the highest risk? And therefore, when you do a sort of cost-benefit analysis, which segment is worth targeting the most?
Sema: That’s a really good question. So, we were able to measure and estimate the risk of HIV and one of the things we saw is the risk to acquire HIV actually did not align with the likelihood of getting circumcised.
Chris: What does your result suggest we should be doing then?
Sema: First of all, it provides choices to the programme and also, insights for development of strategies to generate demand. So, more nuanced strategies if we wanted to get the ‘scared rejecters’. The typical approaches that we’re using now which is really around raising awareness and mass media is not going to work because these people are already pretty aware. What we need to do is really get them in touch with healthcare providers and have healthcare providers really do that interpersonal communication. In the case of ‘enthusiasts’, those men are mostly concerned with pain. And so, we need to really explain to them the pain aspect and how to deal with pain. The programme today does not talk about pain because there's been this understanding that maybe talking about pain has actually pushed men away. But our findings suggest actually quite the opposite.