Three studies of spine disorders: cost-utility of cognitive behavioral therapy; classification of incident cases; and distribution of treatment cost
Norton, Giulia Distefano
MetadataShow full item record
We performed three distinct studies of low back pain treatment using a generalizable database of medical and pharmaceutical claims paid by multiple commercial payers. Objectives: (1) Evaluate the cost-utility, from the US commercial payer perspective, of Cognitive Behavioral Therapy (CBT) to manage chronic non-specific low back pain (LBP); (2) Identify unique subgroups ofpeople with incident LBP based on health care utilization; (3) Specify the distribution of costs for people with intervertebral disc herniation (IDH). Methods: (1) We constructed and validated a Markov model with 1-year and 10-year time horizons. We applied rates of service utilization and recovery from a randomized controlled trial assessing CBT to treat chronic LBP, estimated standardized costs from the claims database, and derived utilities and recurrence rates from the literature. (2) Using Latent Class Analysis, we identified four subgroups (classes) ofpeople with LBP based on utilization over 27 months (9 quarters). We modeled increasing numbers of classes and service combinations to build the final model. (3) We compared empirical cost data to four candidate distributions. Results: (1) The incremental cost-utility of CBT, compared to education, was $7197 per QALY in the first year and $5833 per QALY over ten years. (2) We identified four mutually exclusive subgroups ofpatients. Two groups were very likely to receive services in each quarter (Pr>0.70 for 7% and >0.40 for 33%), implying continuing symptoms or frequent relapse. The extremely high utilization group (8%) had >85% likelihood of receiving services, >70% painkillers, in each quarter after the first. Differences in types of service received imply the groups were characterized by pain or disability. Just over half the patients recovered, not receiving services after the frrst quarter. (3) Lognormal distribution fit costs of iDH more closely than Weibull, Gamma, or Exponential. Results for all studies were robust across numerous sensitivity analyses. Conclusions: (1) CBT is a cost-effective approach to managing NSLBP among commercial health plans members, regardless of plan structure. (2) Almost half of people with low back pain are at high risk of continuing symptoms or repeated relapse over 27 months. (3) Costs for IDH follow a Lognormal distribution.
Thesis (Ph.D.)--Boston University