More about payment by outcome…

I’m trying to read around the idea but I’ve found that the topic is not easy to locate in peer-reviewed journals. Perhaps I’m not using the right search terms, but I have found around 3 relevant articles so far… Clearly I’m no health informaticist, and given one of the papers I’ve found, I’m no health economist either!

What I have found today is an article written by Rory O’Connor and Vera Neumann, dating from 2006 and published in the Journal of the Royal Society of Medicine. The authors of this paper describe some of the history around measuring outcomes in health and relate it to changes in the NHS in the UK.

Some of the points they make are about the difficulty with defining ‘quality’, suggesting that there is a natural tension between attention to ‘technical management of illness’ and ‘prevention, rehabilitation, coordination and continuity of care’ – too much focus on the former and the latter suffer with ‘poorer outcome for the overall health community’.

I guess that means that we can design the best treatments – but unless they’re placed into context, they are simply the ambulance at the bottom of the cliff.

I don’t think any health care provider would disagree with the notion that biomedical measures such as laboratory results don’t provide a complete picture of treatment outcome – they need to be supplemented with other aspects of health that are important to the patient. O’Connor and Neumann identify that the problem with patient-centred constructs is that these are ‘fuzzy’ (OK that’s not their word, but you get what I mean!). As they say ‘compared with concrete measures such as blood pressure, constructs such as pain relief, walking ability or depression are complex to measure.’

I couldn’t agree more with their statement that ‘measuring outcome remains the ultimate validation of the effectiveness and quality of healthcare’. I also agree with their next sentence ‘payment by results will fail to meet …objectives… if the wrong outcomes are measured’. I think they need to add that outcomes are not dependent on the actions of only one party – not when it comes to self management, and not when the person is ambivalent about taking action.

The history of interest in outcome measurement in medicine is not new – outcomes have been recorded as far back as the treatment of scurvy by James Lind (1716 – 1794), and the steps taken by John Snow (1813 – 1858) to locate the source of the Broad Street pump cholera outbreak. Early health outcome measurement was limited to recording deaths – I’m glad we’ve moved a little on from here!

But how far have we moved? And who determines the appropriate outcomes to measure?
After all, there are four stakeholders involved in the management of a bout of back pain, especially if it has meant time off work for the injured worker… The person with the pain wants to measure how much his or her pain has dropped and how much he or she can do; the health provider wants to know whether pain has reduced and hopefully that the person is happy with his or her treatment; the employer wants to know whether the person can manage his or her tasks at work; and the funder? Hopefully, the funder wants to know that what is being paid for has been provided, and in New Zealand if it’s ACC, that the person is well on the way to recovering full function.

O’Connor and Neumann discuss the concept of ‘quality’ as being influenced by societal values – as I’ve mentioned, each one of the stakeholders has a slightly different idea of a quality outcome, depending on their own values. Once some common ground has been established on what ‘quality’ is – then comes an even more contentious problem: how to map that concept onto something that can be measured. Something that is robust in terms of psychometrics and can be readily used. Something that takes into account patient preferences, motivation and adherence to the intervention.

The timing of outcome measurement is important – there is a distinct post-intervention ‘honeymoon’ phase for many interventions, and as time goes on, the effects become less distinct. And as one clinician used to say time and again ‘two swallows do not a summer make’ – two measures of outcome won’t give a robust measurement over time, it’s important to have at least three to account for the temporal changes that occur especially with respect to improved mood.

Audits are pretty common across all health care services. These typically focus on the processes used to achieve an outcome. Certainly in the EQuIP4 process currently underway in New Zealand, the whole approach is process-driven. This approach is based on the idea that by making sure all the steps in delivery of a service are carried out well, the best outcomes should achieved – but has anyone checked that these steps are relevant? Are they associated with good outcomes or are they about doing things that are visible, able to be remedied – but are actually tangential to the outcome?

The remainder of this paper is focused on the NHS strategy. The points they make could well apply here in New Zealand too. Some pertinent points:

  • volume of activity and speed of service delivery can become surrogates for quality healthcare provision
  • limited resources within healthcare facilities to record outcomes make it difficult to obtain relevant measures
  • individual departments that do make improvements to service outcome need to receive the benefits of their improvements
  • payment by outcome is promoted as one way to increase the choice available to patients – but sadly, it often means that services that address chronic health problems such as chronic pain, made more complex by multiple other health problems or social disadvantages may lose funding.  Why? Because their patients don’t attain the kind of results that relatively straightforward patients and conditions may reach.  As has happened already with elective surgery funded by private insurers here in New Zealand, private providers choose low risk patients, perform relatively straightforward procedures, leading to reasonably ‘good’ outcomes.  Comparing apples with apples? or apples with pears?

I won’t go through all the remainder of this article apart from these points made by O’Connor and Neumann.  To make a process like this work, they state the following need to be put in place:

  1. Data concerning healthcare interventions must be recorded accurately
  2. measures must be appropriately chosen for the outcome under consideration
  3. outcome measures must be recorded accurately at the appropriate times
  4. health care costs should be realistic.  Complex multidisciplinary care is expensive.  Trying to substitute simplified versions of best practice through fragmenting multidisciplinary teams will only result in inferior care for patients.
  5. adequate numbers of expertly trained staff are needed to collect and analyse these data.

I thoroughly agree with outcome measurement, and that providers need to be monitored to ensure the treatments are appropriately provided. At the same time, the various viewpoints of each of the stakeholders need to be taken into consideration. AND finally, this needs to be a consultative process, informed not just by the purchasers, not just by the patients, but also by health providers and academics knowledgeable about pain management (for example) who can help ensure measures and processes work well.

O’Connor, R., & Neumann, V. (2006). Payment by results or payment by outcome? The history of measuring medicine. Journal of the Royal Society of Medicine. 99:226-231.

The impact of pain management on quality of life


What exactly is quality of life? And what is pain management? This article, presented at a 2002 Roundtable on ‘The role of coxibs in successful pain management’ is written by a medical researcher, which should give you a bit of a clue about how pain management is defined for the purpose of this paper (i.e. pain control). Much of the discussion about quality of life measures, however, is useful, and the article itself is reasonably short – and it talks the language of doctors, which I thought gives an interesting slant.

Firstly, the point is made that quality of life involves assessment over multiple dimensions, with the World Health Organisation’s ‘Domains and factors of quality of life’ being used as an example of the domains that can be considered. For those who haven’t reviewed this recently, there are six domains:
1. Physical
2. Psychological
3. Level of independence
4. Social relationships
5. Environmental health
6. Spirituality
and a final ‘general’ facet that evaluates ‘overall perceptions of health and quality of life’.  This link to WHOQOL leads directly to the Seattle Quality of Life Group website.

The point is made that pain affects cognitive, motivational, affective, behavioural and physical components, and quality of life has a similarly all-encompassing nature. Katz points out that ‘quality of life … can be defined as an individual’s ability to perform a range of roles in society and to reach an acceptable level of satisfaction from functionoing in those roles’, but at the same time recognising that quality of life research is relatively new, and the effects of pain on quality of life is only just beginning to be investigated.

When we’re choosing any assessment tool, there are some practical things to consider – quite apart from the psychometric properties of the instruments we may think of. Katz nicely identifies some considerations:
1. Which is more applicable – a disease-specific or generic instrument? He identifies that there are both advantages in specific measures (for example, able to detect more subtle change) as well as disadvantages (difficulty comparing across patient groups). The SF-36 (Medical Outcomes Study Short Form 36) is suggested as a general health status measure that combines the ‘best’ of both worlds.
2. What dimensions of quality of life need to be measured? Katz acknowledges that quality of life measures are multi-dimensional since ‘an instrument that does not include several dimensions will make it impossible to determine the nature of a score change’. The areas of physical, psychological, social, somatic, and spiritual are thought to be important, and again the SF-36 is identified as a key contender because of the breadth of dimensions included.
3. How much responder burden is acceptable? This refers to the amount of work needed to complete the questionnaires – an important consideration when we recognise that many people don’t like paperwork, or have relatively low reading levels.
4. What administrative issues need to be considered? Ahhh, now that’s an excellent point! A very comprehensive database that I know of has been abandoned from time to time because of difficulty obtaining clerical time to score and enter questionnaires, and in the mists of time, some of the original scoring ‘rules’ have been lost… Questionnaires used in pain outcome measurement need to be applied at least twice, more appropriately three or four times – and ‘compliance’ or the number of people who complete all of the measures all of the time reduces over time…
5. Has the instrument been validated, and is it reliable? Well, this goes without saying… and should be a ‘given’ for people working in the area of pain management.

There are a number of issues that this paper does not cover well.
Firstly, there is an assumption that if quality of life is reduced when pain is experienced, an individual’s quality of life should improve simply by reducing pain intensity. While this intuitively makes sense, it assumes that pain intensity alone is responsible for loss of quality of life. Reports of pain intensity are influenced by things like psychological distress, fear of injury, fear of being out of control, low mood, health anxiety and so on (e.g. Severeijns, Vlaeyen, van den Hout & Weber, 2001). When analgesia is used, while it may reduce pain intensity it may not address underlying issues such as low mood, health anxiety, and most especially fear of injury. The latter is one reason some patients tell me they don’t want to take medication, because it ‘masks’ pain!

Katz suggests that ‘analgesic agents should be compared … incorporating the use of symptom distress scales, which may be the most sensitive way of discriminating among analgesics in effects of quality of life’. I suggest that distress scales may not be the most appropriate measure of effectiveness in quality of life, and in fact changes across several domains such as Physical, Psychological, Level of Independence, and Social Relationships should be observed before an intervention should be considered ‘successful’.

Another issue is that although Katz indicates that taking repeated measures is important, he makes little mention of the need for longer-term followup. At least part of the initial response from any intervention is likely to be due to the ‘meaning response’, or expectancies or placebo response that people have simply from having been given a treatment. As a result, it’s important to measure changes in quality of life (and any other outcome measure) some months or even years later after the treatment is first initiated. The risk otherwise is that an initial lifting of mood, sense of hope, even physiological changes secondary to placebo can subside over time – and in the end, it’s the effects that are sustained that are important.

However, it’s difficult to argue with Katz on this: quality of life measures should be included as a key variable in future pharmacological research. Personally, I think it should be included as an outcome measurement of any pain management intervention.

For further information on a broader range of outcome measures that are being considered for pain management, this article by Dworkin and colleagues provides some insights into the areas that the IMMPACT group recommend: (1) pain; (2) physical functioning; (3) emotional functioning; (4) participant ratings of improvement and satisfaction with treatment; (5) symptoms and adverse events; and (6) participant disposition.  Worth a read – and they’re continuing to publish more on this area.

Katz, N. (2002). The Impact of Pain Management
on Quality of Life. Journal of Pain and Symptom Management, 24(1S), S38-S47.

Severeijns, R., Vlaeyen, JWS., van den Hout MA., & Weber, JWA., (2001)Pain catastrophizing predicts pain intensity, disability, and psychological distress independent of the level of physical impairment. The Clinical journal of pain vol. 17, no2, pp. 165-172

Working with goals

I mentioned in my post yesterday that it’s not easy to help people work out goals.  Most people have a fairly general idea of how they’d like their life to be, despite pain, but have lost sight of the possibility of how life might be in the future. I spoke about this with a person this morning who told me that he has lost confidence that life can be anything different, so when he was asked ‘what do you want from pain management’, he was floored!  He told me that he has lived ‘from day to day’ for so long that he had lost sight of considering where he was going in his life. I don’t think this is uncommon for people living with long-term pain, before they start accepting pain as chronic and start ‘living’ with their pain.

So the first step may be to use a menu approach as I suggested yesterday, but somehow this needs to lead in to longer-term ‘life change’ goals, so the new knowledge can be integrated and retained.  My vision for people learning to live with pain is that they move from thinking of themselves as ‘patients’ and move to thinking of themselves as ‘people’ – and more importantly, that they are people who are doing things. That way their pain condition becomes no more important a part of them than their height, ability to do maths, or interpersonal skill: just part of who they are.

We’ve discussed the menu approach, which helps people to look at things they’re interested in.  Then quite often we help the person develop the pain management skills are linked to the clinical hypotheses we, as clinicians, developed about the factors that we have assessed as influencing their pain presentation.  These are may be things like activity management, relaxation training to reduce physiological arousal, communication skills and so on…

Then the part that I think is vitally important, but not always recognised: change management in ‘real life’.  Some people call this ‘goal setting’!

The ‘wish list’ (I promise I’ll feature this tomorrow!) asks people to identify general areas they’d like to see change in their lives.  This goal-planning worksheet is one way to help people develop specific ‘next steps’ to take, and ways to identify and measure the actions and resources they need to take to make it happen.

Click on it, and you’ll be able to see that in the centre is the ‘goals area’. I’ve described this as ‘I want to…’ with lots of space to draw or write exactly what the person wants. If they can’t be very specific, that’s fine because details are refined as part of this worksheet process.

Starting at the top, working clockwise, are a series of headings that the person can complete that will help shape up the way to achieve the ‘I want to’ part. Going in order is important (from top, clockwise), because each section builds on the others.

Step one and probably most helpful for retaining motivation is ‘Why is this important to me?’. We know that internal motivation is drawn from values, or reasons that something is important. By helping the person think through why they want this change, we help them generate their own reasons for change, rather than reasons we might generate.

Step two is about time frame: often this depends on the level of confidence the person has for acheiving their goal.  The time frame also may determine the steps the person may take, and reflect other people’s actions or proceses (eg case manager). It may be helpful to revisit the time frame once other areas are completed and adjust, but it’s useful to work out what the person believes is possible first.

Step three is about where the goal will be achieved – for example, at home, at work, in a clinic… This may help determine the strategies that can be employed.  Some activities can be used in one setting, but not another, while it may be important to help the person see that they can achieve in one setting before generalising to another (eg using relaxation at home to enable them to complete vacuuming, before developing the skill to self-regulate at work, where it can be much more difficult).

Step four asks about general resources the person may need. This can be revisited also, but is often the real action phase of goal setting.  It’s OK to not know all the resources required – this may, in fact, be one of the sub-tasks required to achieve a goal.  So it’s fine to put down ‘I don’t know what resources are available – I’ll need to identify these as my ‘next best step” – which is one of the other steps in this process.

Step five identifies the ‘people’ resources the person has to help them achieve their goal. This can range from speaking to their partner, family or friends, right through to asking for a referral to see a specialist career counsellor.  If the person is receiving compensation, and has a case manager, then the case manager must be one of the ‘people’ identified in this step.  You, as health provider, should also be listed.

Step six is about what the person can do to recruit help from people – because it is all about self management and self responsibility.  What does this person need to do to get the help they want? Do they need to write letters? Meet with someone? Leave notes? Arrange regular phone calls?

Step seven identifies ‘what are my next best steps?’. This can be about finding out resources, contacting people who can support them, maybe identifying where the goal can be carried out, or even refining the time frame.

Finally, (and next to ‘why’, the most important part of this process) is ‘how will I know I’ve achieved my goal’. This is often the most challenging part of this whole process, because it provides the person with a specific way to measure achievement. This may require additional input from you as health care provider.

It’s not very easy for someone to think about how they will know whether they have achieved a goal – quite often the person hasn’t thought about the goal in more than general or vague terms.

For example, a goal may be ‘to sleep better’.  This isn’t precise enough for anyone to decide whether the goal has been achieved or not.  Remember that subjective feelings ‘sleeping better’ are often determined by proximal experiences (a bad night’s sleep or two close to the day the person is asked about their sleep will likely influence them to say it is no better), and by demand characteristics of the person asking (a treating clinician will obtain a different answer than if the person’s partner asks), as well as a number of other response biases.

It’s far more effective to ask the person ‘how will you (know) measure that your sleep is better?  Is it that you sleep all the way through the night, every night of the week?  That you wake up feeling like you’ve had a good sleep, five days out of seven?  That you no longer use any sleeping medication to help you get off to sleep?’

These latter measures help identify the different ways that ‘success’ can be achieved, and as a result both reduce the chance for measurement biases as well as determine just what the ‘end point’ will be (at what point will the person be personally satisfied with their progress?).  It’s much more client-centred, as well as more easily recorded and reproduced.

Have a go at using this worksheet for a person goal for yourself.  What would you like to see different in your life?  Then work out how you’ll achieve it using this step-by-step process.  Let me know how you go!

Outcomes and effectiveness and other wonderful things…

The IASP publishes a huge range of books and other resources, but one of the ones I really like getting is their Clinical Updates. Each month a brief update on an area of pain is published, and this month discusses an issue really close to my heart: outcomes, or ‘will this treatment help?’

This paper argues the use of patient-centred outcomes: outcome measures that are important from the patient’s point of view. Patient-centredness increases: patient’s satisfaction with care, adherence to treatment, and the ongoing relationship with a provider.

Treatment effectiveness, or the outcome of treatment to ‘ordinary’ patients in ‘ordinary conditions’ by ‘ordinary health care providers’ is equally important.

Measurement instruments are now available to identify which treatments

  • reduce pain
  • improve health-related quality of life
  • improve function

In fact, the problem is that there are too many measurement instruments! This paper strongly urges some coherence on the type of outcome measure that should be used, to reduce the enormous number of different instruments, and therefore increase the ability to compare treatment across populations and settings. The World Health Organisation’s International Classification of Functioning, Disability and Health (ICF) is suggested as a framework – perhaps helping ensure that pain is never assessed as a single dimension, or merely a symptom, but rather a subjective experience and perception.

Quality of life (QOL) measures that may fit with the ICF framework currently don’t include aspects of pain that are included in separate measures – things like the responses of others, coping ability, fear avoidance and disability arising from pain. QOL measures need to be augmented with a disease-specific instrument, that is capable of detecting change over time while not being a burden for the person to complete!

Core data sets are recommended – one area that I’ve really wanted to encourage in New Zealand for ages! But it’s not easy to get any group of clinicians, let alone clinicians and administration people from different hospitals in different cities in different islands to actually AGREE on anything – well, you know what I mean! Despite the work that has been done by organisations such as the European Organisation for Research and Treatment of Cancer Quality of Life to develop core measures, it still seems an incredibly difficult thing to come up with a set of areas to measure that cut across all pain types.

For more information on outcome measurement, try:
Workcover’s set of resources readily available online,
Dworkin’s 2005 paper for the IMMPACT group,
WorkSafe’s set of measures, also available online,
and just for some background reading on evidence-based health care, try Bandolier’s website for a whole range of topics to learn from.

And remember, if you’re keen to keep up with my posts – you can subscribe using RSS (click the box above!), or you can bookmark this blog and pop in from time to time. I love receiving comments and if you’ve got any area of pain or chronic health management you’d like to find out more about – drop me a line!