What does a ‘modifiable risk factor’ mean?

The Cambridge Dictionary defines the term ‘modify’ as ‘to change something such as a plan, opinion, law or way of behaviour slightly, usually to improve it or make it more acceptable’ [1]. In epidemiology, the term ‘modifiable’ is often poorly defined, and frequently used without explaining how a factor is designated as such. In statistics it would usually be used to describe a variable, factor, or parameter that can be changed or manipulated in some way.

Classification of what constitutes a modifiable risk factor varies across the medical and social science literature. For example, some classify age, race, and chronic health conditions as ‘non-modifiable’ [2]. However, in certain contexts we can argue that age as a risk indicator for health can be modified [3] -for example, the age at which women become pregnant. In other contexts, age may be closely related to frailty but independent from it, and therefore may not be considered modifiable.

Race and/or ethnicity are social constructs which are closely related to other determinants of ill health such as poverty, isolation, and discrimination -mechanisms that produce poorer health outcomes and are potentially modifiable [4]. Race and ethnicity as non-modifiable factors are commonly used as proxy for underlying biological processes or as a genetic category: in doing so, the identification of structural causes of ill health may become obscured [5]. Even when considering genetics, with the emerging evidence around the role of epigenetics in shaping chronic disease risk, we can perhaps say that this is at least partially modifiable [6].

Some long-term conditions are arguably modifiable. For example, remission of type 2 diabetes may be achieved if people lose weight, hypertension can potentially be reversed with behavioural changes, and for some people kidney function is variable, with chronic kidney disease appearing to ’remit’ [7, 8].

What counts as ‘modifiable’ depends on the context, but can we agree basic universal criteria for definition?

In some contexts, the purpose of classifying a risk factor as modifiable is a call for action at the individual or group level. In others, a wider definition of ‘modifiable’ may be warranted such as when considering aetiological associations with health outcomes, as well as accounting for moral, ethical, and human rights dimensions of causality.

The potential for modification could be considered at the individual, community, national, or global levels. This includes the social, economic, political, commercial, and environmental determinants of health. Clearly articulating what is meant by ’modifiable’ gives us the ability to develop common understanding, consensus, definition, and application of the term, while fully recognising that such consensus may very much be context- and purpose-dependent.

Are there any general criteria to consider when specifying factors as ‘modifiable’ in the context of health research?

The following questions can help us define what we are considering as modifiable risk factors of health and disease outcomes (Fig 1):

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Criteria to consider before classifying health risk factors or determinants as ‘modifiable’.

https://doi.org/10.1371/journal.pgph.0002887.g001

Conclusion

We conceptualised the above questions for application in our own empirical research. The next step is to establish consensus. In the absence of an agreed definition, transparent criteria to define what can be considered as modifiable determinants of health are desperately needed. An explicit consideration of what can be considered “modifiable” can facilitate more comprehensive public health planning and make the concept of modification more inclusive of the wider determinants of health. This is urgently needed if we want to tackle health inequities both on the global and within-country scales.