The focus of the development and use of patient-reported outcomes has previously been to contribute to clinical research, particularly in assessing the effectiveness of interventions. With a primary focus on generating a large dataset, the collection of patient-reported outcomes has often been fraught with difficulty. It has been costly to achieve and the data have been challenged in terms of their utility.
I would argue that the primary use of patient-reported outcomes should be to enhance the communication between a patient and healthcare professional with the aim of a better shared understanding of context and goals leading to shared decision-making, a visual representation of trajectory of illness over time, and the ability to make access to healthcare more flexible (particularly in the management of chronic disease). Context is important. Tracking patient-reported outcomes longitudinally will have quite a different meaning for people living with dementia compared with people living with diabetes, for example. One size doesn’t fit all, either in terms of the method of communication, or in how we use the information gathered to support care.
In the past, the purpose of the exercise of measuring outcomes may have been opaque and people may not have seen the relevance to their care. This unsurprisingly led to problems with low response rates, particularly as data have often disappeared into a black hole with little access for either patient or clinician. Low response rates increase the likelihood of selection bias and call into question assumptions based on interpretation of the aggregated data.
Being clear about the purpose of collection and how data are to be used (and by whom) is key. The “why” is the most important factor as this informs what instruments are used, the method of collection and the analytics needed later.
This information can be of huge benefit to patients and clinicians in the consultation. Work done with patients in South East Wales demonstrates that the act of filling out a PROM questionnaire could bring structure to a subsequent consultation, allowing patients to prioritise the issues that matter to them most. It also facilitated entry into more difficult or sensitive topics or other issues which the patient and/or their carer may not have identified as relevant. The very act of considering the questions in a PROM began to affect the balance of power in the consultation, giving the patient a tool to lead. This observation is interesting when we consider the counter argument, which is that PROM’s are too narrow and restrictive and that a free-flowing narrative is more important than a score. What we observed was that the former enabled the latter. We are also a long way from analysing patient narratives, so arguably standardised communication tools have a place.
That said, there are limitations to think about. It is tempting for organisations to make inappropriate assumptions on the basis of aggregated PROM data. PROM scores cannot reliably be used in isolation to make judgments about service effectiveness, given that each individual will have very different preferences and goals. For example trade-offs between longevity and quality of life become common in certain contexts and some tools will not be sensitive to pick up some quite significant changes from a patient’s perspective. However, aggregated data can offer a detailed needs assessment and be very useful in understanding how to meet need.
Tracking symptom scores over time could support flexible access to care by providing remote information about current health status. It could prompt us to acknowledge a declining trajectory and identify opportunities to talk about anticipatory care planning and improve end of life care.
Patients have told us that they would like more information to support their decision making. They want to know what patient-reported outcome data tell them about the likely trajectory of their illness and the life decisions they may make as a result They want to know what is likely to happen to somebody like them with all of their characteristics if they choose a particular course of action? For most people in most situations, these questions are difficult to answer. Optimal medical therapy is often talked about in terms of adherence to single disease clinical guidelines. While the guidelines are evidence-based, we know that the evidence is of varying quality and applicability. Therefore there is intense interest in generating “real world evidence” of the effects of treatment on those whose traits may well have excluded them from the trials from which the evidence is synthesised. This seems particularly important when PRO data are collected during trials, but frequently not included in the final analysis. We also have to help each other navigate the trade-offs, for example between longevity and quality of life. This approach is especially needed where evidence is already often scanty, such as in the world of medical devices, orphan and ultra-orphan medicines. We have to be certain that treatment benefits are delivered from patients’ perspectives when each new technology adoption has an opportunity cost somewhere else, for someone else.
There is a lot of work still to do in order that we learn to analyse and visualise outcome data in ways that are meaningful to all those who need to use it, including those who cannot or do not wish to use technology.
The measurement and use of patient-reported outcome data is a helpful tool in enhancing care, but much depends upon the attention to detail in defining the purpose of the endeavour at every stage and accessibility of the information produced as a result.