Amani Lavefjord: Individualizing psychological treatment for chronic pain: Can network analysis bridge the scientist-practitioner gap?

Abstract

Psychological treatment for chronic pain can generally help people improve their functioning and well-being. Still, there is room for improvement in terms of overall effect sizes. There is also a need to address the gap between scientists, often focusing treatment on standardized manuals for treatment, and clinicians, often aiming to individualize treatment. One suggested step forward is to adopt the clinician’s agenda for individualization, but in a systematic fashion ensuring a scientific approach. The aim of this thesis was to examine the clinical utility of one method suggested to be useful for systematic individualization; individual level network analysis. Both study I and II tested the clinical utility of using so-called network centrality for treatment guiding, where centrality reflects how well connected one network node is to other nodes. Specifically, we tested whether interventions guided by the most central node were more beneficial for the individual compared with interventions guided by the least central node. In study I, we tested this by using networks based on participants’ one-time ratings of how aspects of psychological inflexibility and pain interference were perceived to cause each other. Results were promising in that individual level treatment effects most often occurred when treatment was guided by the most central node. Study II replicated study I, but networks were instead estimated using repeated measures of the network variables. Results did not indicate utility of guiding treatment based on centrality in such a data driven network. Study III demonstrated that individual level networks can give us information that may be lost in group level networks, but that there are group level connections that are generalizable to many individuals. Such generalizable connections can be used to generate initial hypotheses for individual level treatment, but may not be valid for everyone – prompting us to continue the evaluation of systematic methods for individualization. Further, individual level network connectivity was correlated with depression and pain interference, but did not contribute with unique explained variance in the outcomes when controlling for baseline levels of the outcomes. Finally, overall network connectivity did not predict pain interference or depression on the individual level.