patient-clinician interaction

Looking at you, looking at me: The effect of clinical encounters


ResearchBlogging.org
I’ve been browsing the databases for information on clinical encounters. I do this because while I’m well aware of many papers produced on the topic of patient-related predictors of outcome, the swing towards considering clinician characteristics as predictors of outcome means I hope to see more published on this incredibly important topic.

A wee while I ago I wrote about patient self-report measures of chronic pain clinical encounters (Stomski, MacKintosh & Stanley, 2010), in which I discussed the lack of reliable, valid measurement tools in which people with pain are asked about their clinical encounters – and worse still, the glaring omission in any of the reviewed measures of actually asking people with chronic pain whether the constructs being assessed actually mattered!

Today I came across this paper by Cvengros and colleagues, looking at the influence of congruence between what a patient prefers, and the actual dynamics of the clinical encounter.  As these authors point out: patients can vary in the degree to which they want patient-centred (active) or clinician-centred (passive) participation in a clinical encounter – and so do clinicians.  Their hypothesis was that if the preference of each clinical dyad was congruent, then the outcomes, particularly in terms of adherence to clinical recommendations, would be better.

Now, sadly, I couldn’t find a study like this in pain management.  This paper is drawn from the diabetes literature.  The relevance of this to people working in the field of chronic pain is pretty clear – both groups of patients ultimately need to self manage their chronic health condition, both groups of patients will have biomedical as well as psychosocial aspects of management to consider, and the consequences of failing to manage their condition well can lead to some fairly dire outcomes in both cases.  I guess the major differences are that the contribution and acceptance of psychosocial factors and self management is greater in diabetes management, and the messages in the community about the way to manage diabetes are somewhat more consistent than in the case of chronic pain (not too many ‘I can cure you!’ adverts for diabetes – actually, I tell a lie: I just took a quick look on a popular search engine and there are adverts like this “Simple 3 step natural approach heals diabetes within a month”)

Anyway, how did these researchers go about exploring this complicated area of health provider:patient interaction?

The first point they make is that rather than looking at a global indicator or trait measurement of attitudes, they chose to explore visit-specific behaviour.  As they point out, the content of each encounter can differ widely from when a provider first gives a diagnosis, through to treatment options, through to ascertaining whether the interventions have been effective.  They identify three main domains to examine – (a) information sharing and shared decision-making,
(b) encouragement of self-care, and (c) provision of socioemotional support.

They identified 621 patients who met their study criteria of language, diagnosis and at least two visits with their health care provider in the past six months (the latter in order to choose patients who had an established relationship with their provider).  A rather smaller group of 241 patients agreed to participate, with a further 23 patients excluded for various reasons.  Not exactly a great response rate!

Putting that aside for a moment, let’s take a look at the really interesting aspects of this study.  The authors didn’t believe there was an appropriate measurement instrument for assessing provider behaviour, so they developed one and it’s included in the paper.  It’s called the Provider Behaviour Questionnaire, and has nine items in three domains: Information sharing and shared decision-making; Behavioral involvement and self-care; and Provision of socioemotional support.  The items were closed statements with a yes or no answer such as “Dr. Smith gave you information about your health to read or to look up at home.” and “Dr. Smith suggested that you take an active role in managing your diabetes.” and “Dr. Smith asked about your personal life.”

Now I could argue that these statements may not exactly represent ideal provider behaviour, but they were developed from established measures such as the two subscales of the KHOS (Krantz Health Opinion Survey), and the caring scale of the PPOS (Patient Practitioner Orientation Scale).  The researchers also didn’t ask participants whether these were their preferred areas of support, and psychometric characteristics of the questionnaire are not reported.

Participants were asked to complete both pre- and post-consultation measures.

What did they find?

Cutting to the findings, Cvengros and Christensen found that symmetry in information sharing and shared decision-making, F(2, 208) = 3.71, p = .03 and behavioral involvement in health care, F(2, 205) = 3.60, p = .03, emerged as
significant predictors of patient satisfaction, with socioemotional support, F(2, 211) = 2.83, p = .06, emerging as a marginally significant predictor.

Symmetry in information sharing and socioemotional support were both nonsignificant predictors of self-reported diabetes adherence. However, symmetry between patient preferred and reported behavioral involvement was a significant predictor of self-reported diabetes adherence, F(2, 195) = 3.39, p = .04.

Finally, the found that congruence in the information sharing domain was marginally associated with lower hemoglobin A1c values, indicative of better glycemic control.

Now while the hypotheses were only partially supported, the limited recruitment, and the dichotomous nature of the questions in the Provider Behaviour Questionnaire mean that the findings are not all that robust, this is an interesting study because it starts to pose questions about the nitty gritty of interactions between providers and patients.  I haven’t reported some of the more detailed outcomes, but these show that it’s not just the match between patient and provider communication behaviour, it’s the direction of that match Cvengros and Christensen suggest that people with greater preferences in the information sharing and behavioral involvement domains may be demonstrating a greater preference for control over the health care experience.

Now this is important because, as they argue, when patients perceive limited control over their health care encounter they may attempt to regain this control by being nonadherent with provider recommendations.

What does this mean for me?

I’m not sure how much I can rely on these findings to change clinical practice at this time, but this study starts me thinking that it’s time to go beyond simplistic explanations of clinical encounters and the effect on patient self-management.  It also highlights that we have held assumptions about how clinical consultations work and while some of us may have learned to moderate a ‘just do it because I say so’ approach, many health professionals either have no clue that this is ineffective in many cases, perhaps haven’t been aware that there are other options, or maybe haven’t been trained or supported to develop alternative behaviour.

A final point I take from this study is that the influence of interactions seem to have quite clear and measurable effects on clinical outcomes.  It’s not easy to ignore changes in hemoglobin A1c.  I only wish there were similar measures available in chronic pain management!

Cvengros, J., Christensen, A., Cunningham, C., Hillis, S., & Kaboli, P. (2009). Patient preference for and reports of provider behavior: Impact of symmetry on patient outcomes. Health Psychology, 28 (6), 660-667 DOI: 10.1037/a0016087
Stomski, N., Mackintosh, S., & Stanley, M. (2010). Patient Self-report Measures of Chronic Pain Consultation Measures: A Systematic Review The Clinical Journal of Pain, 26 (3), 235-243 DOI: 10.1097/AJP.0b013e3181c84e76

It’s not what you say, it’s the way that you say it?


ResearchBlogging.org
I have a bit of a theme happening – health care interactions. I think it’s because this week I’ve been talking about this with patients, and it seems to be something that either raises the hackles or fills them with gratitude! Anyway, I was glad to find this paper the other day on ‘perceived quality of doctor–patient interaction in rehabilitation’. This is slightly different from the usual focus, which is often on interactions in primary care.

It’s an interesting paper written by a group of researchers in Germany. They describe the use of a newly-developed rating instrument, the ‘P.A.Int-Questionnaire’, which stands for (in German) ‘Patient-Arzt-Interaktion’. I’m guessing that’s German for ‘patient-doctor interaction’ (correct me if I’m wrong!). By using this, they aimed to examine the relationship between perceived quality of interaction and long-term treatment outcomes.

The quality of interactions between patients and clinicians were rated on three dimensions – affective behaviour (ie warmth, empathy, ‘coherence’); instrumental behaviour (ie information exchange, structuring and reinforcement); and participation and involvement of patients.

I’ll just briefly describe the methodology, before cutting to the findings and then going onto how this might influence practice in chronic pain management.

Two different questionnaires were developed – tapping into the same domains, but one for patients and one for clinicians.
Seven rehabilitation clinics were involved in the study, with 61 doctors and their 470 patients reviewing their interactions with each other on admission, discharge and ward round. Patients also rated their health status on admission, discharge and six months later.

The key study questions were

  1. whether there were similarities between the patient rating of the experience, as compared with the clinicians rating;
  2. what the relationship was between the perceived quality of the interactions and long-term patient outcomes;
  3. whether there were any differences between those perceiving good interactions and those who perceived interactions as poor; and
  4. whether depression, anxiety or pain intensity influenced those ratings or outcomes.

The findings?

  • Patients and clinicians both rated affective behaviour in  interactions in a similar way, and quite positively.  So people were happy with the warmth demonstrated in the consultations.
  • Instrumental behaviour (information and structuring) wasn’t viewed as positively by patients as it was by clinicians – it seems that clinicians believe they are better at information exchange and guiding an interaction than what the patients experience is.
  • Patients who were happy with the interactions they had on admission were more likely to respond well to treatment in terms of pain reduction, as well as to anxiety when they were discharged – and this was maintained even to six months after discharge.

The final conclusions from the authors of this study relate to both the questionnaire itself and the findings about quality of interactions and treatment outcomes.

Firstly they indicate that the questionnaire has adequate internal consistency and is  sensitive to different situations and subjects (with recognition that the instrument needs further work in terms of factor structure and validity/reliability).

Secondly, they identify that improvements with respect to pain intensity and pain frequency at discharge and six months after discharge were stronger and more sustainable in the patient group reporting high quality of interaction with their clinician.  Again, the authors point out thatthat there are some limitations with respect to interpreting these findings:

Patients rating their interactions as high quality also reported higher pain intensity at admission. This could mean their pain scores ‘regressed to the mean’ (ie, reduced from an artificially high level to a more ‘usual’ level simply as a measurement artifact), or that these patients obtained better interactions from clinicians because they had higher pain intensity. Another drawback to the study is the 40% of patients who were unavailable for the 6 month followup measure.
A further point the authors make is that although pain, anxiety and mood improved, functional measures did not change (good point!). I note also that information about demographics and type of patient (chronicity of problem, previous health status and so on) wasn’t available, so we can’t determine whether there were differences between clinics.

What can we make of these findings?
It seems that these clinicians were really nice people – they conveyed warmth and empathy, and their patients liked this. They didn’t, however, do so well in terms of information exchange, guiding the discussion, or ‘making it work’ (actions). This might mean that although these clinicians were nice people they maybe hadn’t developed the specific skills involved in structuring a patient interaction.

While the clinicians involved in this study were physicians, I wonder whether occupational therapists, physiotherapists, nurses or even social workers could also be relatively under-skilled in the specifics of instrumental behaviour. I’m hoping that psychologists have some skill in this area – but I’m not entirely sure.


What was interesting for several reasons was the specific finding that people who rated the interactions very positively experienced better outcomes even in the long-term.
Now I don’t want to over-interpret this finding, but it does make me wonder whether this study is tapping into one aspect of the so-called placebo effect, or ‘meaning response’. After all, one of the main differences between ‘alternative’ health care providers and those of us working in, for example, a public health system, is the time available for a consultation. At the very least, the longer time available means more opportunity to demonstrate caring and empathy, and for information exchange.

At the same time, ‘alternative’ therapists positively engage with their patients to involve them in their (sometimes) complex treatment regimes (ever tried homeopathic treatments? you often have to stop using peppermint toothpaste, eating peppermints, stop drinking tea, coffee, chocolate…and I don’t even WANT to think about colonic treatments!). Of course, we’re not given the number of patients who drop out from ‘alternative’ treatments either, so I guess I’m not comparing apples with apples here!


I think I’ll go about reviewing my interactions in terms of these three dimensions for a while – and then see what I need to learn to improve my communication with patients.

If you’ve enjoyed this post, and want to read more, you can subscribe using the RSS feed link above. Or you can bookmark and visit again! I love comments,respond to them quickly, and if you’ve got a topic you’d like me to write about – drop me a line via the ‘About’ page.

Susanne Dibbelt,, Monika Schaidhammer, Christian Fleischer, Bernhard Greitemann (2009). Patient–doctor interaction in rehabilitation: The relationship between perceived interaction quality and long-term treatment results Patient Education and Counseling : 10.1016/j.pec.2009.07.031