What matters: patient-determined outcomes and clinician/researcher outcomes

It’s easy to forget, sometimes, that when we choose an outcome measure, we need to seriously consider who will use the measures in the end.  Of course, I am assuming that we’re all using outcome measures – we are, aren’t we?  If anyone isn’t, shame on you – how on earth will you establish whether what you’re doing is having an effect? And don’t come at me with “oh but I just ask them” because I don’t want to have to list all the response bias, demand characteristics, and lack of consistency problems again. kthx.

Anyway, where was I? That’s right, the end-user in outcome measures.  Over the past few years, interest has risen in identifying the range of outcome measures that can be/should be used in research of treatments for chronic pain.  The IMMPACT project has worked to identify measures that researchers and clinicians agree are important – and more recently used focus groups to identify domains that people with pain thought were important.  I’ve blogged about this before here.

People with pain thought seven out of 19 outcome domains: enjoyment of life, fatigue, emotional wellbeing, physical activities, staying asleep at night, weakness, and difficulty concentrating, were so important they rated them at least 8 out of 10 (where 10 = extremely important).  Notably, these are all “subjective”, in other words, these are aspects of wellbeing that can’t be measured directly.

Beale, Cella and Williams (2011) were curious to see how well these domains were measured in 60 randomised controlled trials of pain rehabilitation.  These studies were included in a recent Cochrane review of psychological therapies for managing chronic pain (Eccleston, Williams & Morely, 2009).

Mapping the outcome measures used in each trial against the domains above, they found that physical activities were assessed in 93.8% of the 60 trials (over 5800 participants) and emotional well being was measured in 83.1% of trials.  Otherwise, the measurement of domains that people with pain thought were important was pretty poor.

None of the trials measured these important areas: difficulty concentrating, difficulty remembering things, hobbies, planning activities, and sex life; while only a few measured these domains : enjoyment of life, fatigue, weakness, staying asleep, falling asleep, relationship with family and friends, and taking care of family members.

The two most important areas to patients, enjoyment of life and fatigue, were not measured, or not measured well.

This result doesn’t entirely surprise me.  The practice of routinely measuring outcomes is still not well-established without (usually) some external agency requiring it.  In NZ, it’s often our national injury compensation agency that requires outcome measurement – as a result, outcomes are often those that have meaning to that agency.  Things like ability to return to work, mood, physical disability, and some cognitive variables such as catastrophising and fear of movement are often required.

Many medical practitioners question the practice of measuring outcomes at all – one doctor I know said “why can’t we just ask them?”  Hence my little grumble earlier in this piece.  Others want to limit the number of questions and pick and choose individual statements from psychometrically validated measures, like a pick and mix bag of sweets.  And it’s been my experience that some managers are loathe to spend time and money on collecting outcome measures “unless they’re needed by ACC”.

Why include outcomes that matter to patients?

Well, mainly because it’s their wellbeing clinicians are supposed to have at heart, isn’t it? It’s also about developing a greater ongoing relationship or dialogue between the person with pain and their treatment team.  By developing and including patient-centred measures, it’s far easier to have that conversation with the person who has pain.  The person can actively decide whether the treatment is helping – and if the person understands what’s in it for them, adherence to treatment is likely to be improved.

Where does this leave us?

One conclusion the authors of this study came to was that we simply don’t have a good range of outcome measures to reflect the areas patients think are important.  Actually, given the quite limited number of studies identifying what patients do think is important, maybe we haven’t got a good handle on the range of areas first!

I suggest we need to spend a little time considering patient-centred outcomes.  Maybe this is in the form of more patient focus groups talking about what is important to them.  Maybe it’s about using patient-derived measures like Goal Attainment Scaling.  Maybe it’s about working harder to develop measures that tap into things like quality of life.

Beale, M., Cella, M., & de C. Williams, A. (2011). Comparing patients’ and clinician-researchers’ outcome choice for psychological treatment of chronic pain PAIN DOI: 10.1016/j.pain.2011.06.007

Eccleston C, Williams ACdeC, Morley S. Psychological therapies for the
management of chronic pain (excluding headache) in adults. Cochrane
Database Syst Rev 2009;2:CD007407.


  1. Perhaps your last paragraph goes to the nub of the issue; as there is little agreement on what is a good outcome, practitioners like myself find choosing one vehicle out of the myriad out there daunting and when we do, the interpretation can be just as confusing. I’m afraid I have a great deal of sympathy with the Doc who says ” Why cant we just ask” even if I can see that there is huge bias in that approach.
    I have found that the Oswestry, Roland Morris and Bournemouth all OK, the SF36 difficult and have finally decided that a simple verbal 0 -10 out of 10 score to be the most helpful for both myself and the patient.
    Your article though makes me think I need to revisit this issue again. Thank you

    1. Keith, if all I’ve been able to do is prompt someone like you to revisit this incredibly challenging part of providing healthcare, then I think I’ve done something worthwhile! I’d like to see if I can unpack some of the difficulties around choosing a good selection of outcome measures – I’ve mentioned in some of my earlier posts the range of measures we use in the facility I work in. They include the Pain Anxiety Symptoms Scale, Tampa Scale for Kinesiophobia, Pain Disability Index, Depression Anxiety Stress Scale, Chronic Pain Acceptance Questionnaire and some practical outcomes like return to work, sleep and use of health care. I’ll take a stab at the topic again in a short while – so come on back and I’ll see what I can come up with!
      Thanks for taking the time to comment, it’s so good to know people are reading what I write and thinking about how they do things in chronic pain management.

  2. I did my msaters research on mapping the domains of the outcome measures suggested by the IMMPACT group to The ICF and the results were very poor suggesting we are really leaving out environmental factors and need to consider a lot more. I now use the ICF checklist for chronic widespread pain to guide me in thinking about what the variables affecting my patient could be and also using this as an outcome measure itself for charting purposes. It does take a little time getting used to the terminology but once you use it awhile you get accustomed to it. You can also identify variables which you might want to use other outcome measures. I agree we have a way to go however I feel we need to work on classifying chronic pain better ( i feel the definition of chronic pain is too broad)in order to get a better idea of how in depth we need to go in our evaluation.

    1. Thanks Lesley, I agree that very few of the measures identified by IMMPACT actually map onto ICF domains and this is a gap in our outcome measurement technology that badly needs more work (PhD anyone?!).
      The different perspective of ICF language can be quite difficult for people to understand initially, so I think that has hampered the use of these domains in mainstream pain management research. Probably the more the constructs are used the more people will ‘get it’ but right now it’s a challenge.
      Thanks for reminding me of this – I can feel a blog post coming on!

  3. I think it is a good idea as well for new outcome measures that matter to the patient but when I read up on gas scaling it was very involved and time consuming. maybe we need to see what the goals are for the individual patient maybe three-four goals and set a 1-5 scale of how they are measuring up maybe once every 2-3 weeks. I know this is subjective but if it is their goal then they are the ones who can measure how close they are to achieving this goal and we can have outcome measures we can use to ID why they are unable to achieve these goals such as catastrophization or whatever.

    1. Initially the GAS does take a while, but the process becomes a lot easier and quicker as the clinician gets used to it. The main problem with this, and any individualised scale is carrying out grouped data comparisons.

      1. I’m an OT and a Canadian so I a tad biased, but I really like the Canadian Occupational Performance Measure (COPM) as an individualised measure. It is client centred and helps me to get to the heart of what the client wants to work on fairly quickly. Once I know the client and he/she understands the role of OT, it takes about 20 to 30 minutes to administer.
        Like most measures it is not perfect but in the field of relevant measures it is the best I have found. You don’t have to be an OT to use it.

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 )

Google photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.