On Outcomes

I have to admit I get disproportionately agitated when others show they don’t understand what an outcome is and why it matters. So here’s my summary version, for the record. 

An outcome is a condition of wellbeing for whole populations: children, adults or families. It’s framed in the positive: all young people are of a healthy weight. It’s about the  result of a social system that directly affects people. I tend to avoid the word result (in the US I think this is often used instead of outcome) because it seems to indicate a fixed end point, whereas in social systems there is no end point. This is the next principle.

Social systems are complex, which means they are characterised by dynamic properties – including ‘emergence’ – and therefore uncertainty. The meaning of emergent here is that when actions are taken within the system they change the system, however subtly, and as a result new properties emerge from the new connections and interaction of its parts. 

This is the next aspect; by a complex system we mean a set of parts with often unknowable connections with and between each other, and that act in ways we often can’t see, understand or predict. 

Obesity is the obvious example here. It’s easy to assume being obese is an equation between calories in and out, but this is to reduce the challenge to one of how much you eat and how much you move. We know that there are a vast (unknown) range of factors that also impact, from how much money you have to your genetic make up, from the quality of your neighbourhood to the stability of your home life, from heritable traits to daily stress. 

I therefore see an outcome as the emergent property of a complex system. In other words, it is what the system produces, and so in this way is a subtle addition to the original definition. 

As a result we cannot have short, medium or long term outcomes. This is one of my biggest frustrations. An outcome is what it is; we cannot directly control it if we agree it is a product of a complex system. The current outcome produced by a system may well be different tomorrow or next month or for the next generation. We can’t however set ‘outcome targets’ across a temporal plane. 

We answer this by tracking indicators. An indicator is a way of understanding the outcome. If we are working towards an outcome where everyone lives in a decent home, we might track levels of homelessness as well as housing quality and energy efficiency. Its crucial that we understand the time series involved (and not compare our data with others, which doesn’t work, because the context will be different). Where have we come from, what are the trends? By looking at this data we start to build the narrative – what story is this telling us? And what additional data might we want to try and get hold of?

What we can have is short-medium-long term outputs, the things you control because you’re producing them. But now we’ve shifted from the system – whole populations – to the programme – the intended beneficiaries and targeted cohorts. The art of the policy-maker and funder is to work with actors from across the system to identify those programmes – interventions – that offer a reasonable likelihood of success. Perhaps you are about to start a family cooking programme to help address obesity. You agree that your story about the data suggests that poor cooking skills and knowledge is a barrier to healthy eating and that changing this might have an impact on childhood obesity levels. 

You are only working with a small cohort of the total population of obese children – those that attend your programme. Imagine a tiny circle (your programme) within a much larger one (the total population of obese children). As the person running the programme, you can’t be held to account for the levels of obesity in the total population, nor the obesity levels of those on your programme (for, as we have seen, so many other variables are at play). What you can be held to account for is how well you deliver your programme.   

And to do this we track performance measures. These are the agreed metrics for your programme, and they fall into three categories:

  • How much did you do? # Of courses, sessions, attendees etc
  • How well did you do it? % attending all classes, % positive feedback etc   
  • Is anyone better off? % People feeling more confident, losing weight etc

That’s it: population level outcomes and indicators, programme level performance measures. The task of the commissioner is to work with those across the system to understand it as much as possible using indicators to illustrate the way the system is performing and collectively agree a series of interventions. Together the interventions should have a reasonable chance of positively impacting the system, tracked through the indicators. 

Due credit here for the foundations of these ideas from Mark Friedman and to David Burnby who showed me an alternative way. 

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 )

Facebook photo

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

Connecting to %s