The University of Sydney -

PRO is a generic term that (as the name suggests) refers to the perspective of the patient him/herself. PROM is therefore used to include not only questionnaires designed to measure HRQoL but also the full gamut of self-report instruments developed to measure constructs such as symptom burden, psychological distress, survivorship needs, unmet needs and satisfaction with care. Because the elastic nature of HRQoL means it can at different times include any or all of these constructs, the terms PRO and HRQoL are sometimes used interchangeably. However, for the sake of accuracy, this practice is discouraged by most authorities, including the FDA.

A model of PROs

The term PRO has the advantage over HRQoL that it needs little explanation to define its meaning even to a layperson. However, this apparent straightforwardness masks the complex, multidimensional nature of the constructs it includes. One way to unpick this complexity is to use an influential taxonomy of different ‘levels’ of health outcome measure put forward by Wilson and Cleary ¹ in a JAMA special communication in 1995. To view the model, click HERE.

Wilson and Cleary are careful to point out that the uni-directionality of arrows in their model does not imply that relationships are not reciprocal between different levels; nor does the absence of arrows between certain components preclude the possibility that there may be relationships between them. Rather, the sequence of boxes is intended to highlight dominant causal associations and to indicate an increase in biological, social and psychological complexity.

Wilson and Cleary's model is useful because it helps us to see that measures can focus either on specific components of the PRO repertoire (e.g., symptoms, functional status or psychological support) or can seek to summarise these in the form of an HRQoL questionnaire. Further, given the range and complexity of factors influencing HRQoL, it is easy to see how its definition is likely to be operationalised in different ways according to which questionnaire one chooses. In fact, we might think of the job of an HRQoL questionnaire as one of sampling among pieces from a puzzle with the aim of estimating the overall picture. The important thing to note, however, is that while different questionnaires may vary in their approach to sampling, those that have been well designed should ultimately provide an accurate estimate of health-related quality of life. On the whole, this potential is evidenced by the close correlations seen among HRQoL questionnaires, despite significant differences between the items they include.

While Wilson and Cleary’s model provides a useful overview of the components of HRQoL, it does not immediately offer advice as to which level of PROM one should choose to use. More helpful in this respect is a property implicit to the model – namely, that as well as representing a continuum from simple to complex, the levels in the model also represent a continuum between measures that are ‘proximal’ (on the left) and ‘distal’ (on the right) to the disease itself. The distinction between proximal and distal is concerned with the degree to which measures are removed from the medical manifestations of disease and include elements of life experience that are less and less related to the most direct form of experience, symptoms. This distinction is important when choosing a PROM because it leads us to expect that:

1.       the effects of treatment will be smaller the more distal the measure becomes;

2.       successful treatment will have greater distal effects when the illness is severe;

3.       if distal measures are at high levels to begin with, it will be more difficult to show an effect (i.e., a ceiling effect will occur);

These principles highlight the need to base choice of PROMs on close consideration of one’s research objectives, estimations regarding the effects of treatments being studied, the populations who will be taking part, and - in the light of all three of these variables - the degree to which factors external to the immediate presentation of the disease can and should be controlled for. Needless to say, choosing between the array of PROMs can be difficult; indeed, it constitutes one of the chief skills in the art of research design and can make all the difference between a positive and negative finding.

The question of what HRQoL measures add to symptom measurement within cancer trials is discussed within the context of Wilson and Cleary’s model by Ferrans (2007).³ For more detailed information about assessing other PROs, see How do I measure psychological distress in people with cancer? and How do I measure symptoms in people with cancer?

References

1. Wilson IB, Cleary PD. Linking clinical variables with health-related quality of life. A conceptual model of patient outcomes. JAMA. Jan 4 1995;273(1):59-65.

2. Brenner MH, Curbow B, Legro MW. The proximal-distal continuum of multiple health outcome measures: the case of cataract surgery. Medical Care. Apr 1995;33(4 Suppl):AS236-244.