Chronic pain is a major medical challenge, with the number of Americans seeking treatment exceeding the number affected by diabetes, heart disease, and cancer combined. The treatment of chronic pain is provided almost exclusively in outpatient settings by standard-care providers, allied health professionals, and complementary and alternative medicine practitioners. Treatment is challenging because chronic pain can lead to depression, hopelessness and drug dependence, as well as reductions in creativity, performance, motivation, and treatment compliance. It creates a large financial drain on social services, insurance providers, and family resources. Lacking independent tests for pain intensity, diagnosis depends on the information exchange between the patient and the health professional. Therefore, informatics can play a pivotal role in creating and disseminating evidence-based medicine technologies within the practice of pain management. To provide evidence-based medicine, the reservoirs of evidence-generating medicine must be tapped and used effectively. To enable this, there is a critical need to map 'person in pain' phenotypes to effective treatments and to disseminate that knowledge with decision support tools. The fundamental hypothesis of this work is that secondary care, pain specialist practices can serve as rich reservoirs of evidence-generating information. To explore this possibility, Michigan State University (MSU) has partnered with Michigan Pain Consultants (MPC) to build a phenotype-to-outcome model for pain treatment, utilizing MPC's rich data repositories of administrative data, proprietary patient survey data, and detailed progress notes from approximately 80,000 visits per year. In this proof-of-concept study, we plan to identify, extract, and organize the informatio content of the MPC data reservoir and then create, test, and validate a model uniting person-in-pain phenotypes with treatments and optimal outcomes. In future work we plan to expand this study to include more areas of pain management and to design decision support tools that will allow enhanced treatment for this largest and most challenging cause of chronic morbidity in the US. Full completion of this project will require the powerful tools of natural language processing, factor analysis, structural equation modeling, active learning methodologies, and clinical decision support software designs. The most important benefit of this work is that a health IT solution will be derived from, and implemented back within, community clinical medicine. Thus, it represents its stakeholders in the closest manner possible. Secondarily, the database generated by this work will enable future research; the process of working with specialist practices will provide a template for other such initiatives; and, the architecture of the solution will create a model for other community-based informatics solutions. The focused goal of this funding request is to establish a proof-of-concept infrastructure to enable these various endpoints.
The proposed research is relevant to public health and the mission of AHRQ because it aims to apply the principles of evidence-based medicine to pain management decisions in community practice. To accomplish this, the rich data repositories of secondary care pain clinics will be extracted, analyzed, and consolidated for eventual use in a queryable decision support tool that will link symptoms, histories, findings, diagnoses, treatments, and outcomes for the various pain phenotypes typifying chronic pain.
| Juckett, David A; Davis, Fred N; Gostine, Mark et al. (2015) Patient-reported outcomes in a large community-based pain medicine practice: evaluation for use in phenotype modeling. BMC Med Inform Decis Mak 15:41 |
