Improved estimation in negative binomial regression

The influence of pain-related comorbidities on pain intensity and pain-related psychological distress in patients presenting with musculoskeletal pain

BACKGROUND

Musculoskeletal pain (MSP) is the largest contributor to chronic pain and frequently occurs alongside other medical comorbidities.

OBJECTIVE

Explore the relationships between the presence of pain-related comorbidities, pain intensity, and pain-related psychological distress in patients with MSP.

METHODS

A longitudinal assessment of individuals 18-90 years old in the Midwestern United States beginning a new episode of physical therapy for MSP. Electronic medical records were assessed the full year prior for care-seeking of diagnoses for pain-related comorbidities (anxiety, metabolic disorder, chronic pain, depression, nicotine dependence, post-traumatic stress disorder, sleep apnea, and sleep insomnia). Pain intensity and pain-related psychological distress (Optimal Screening for Prediction of Referral and Outcome - Yellow Flags tool) were captured during the physical therapy evaluation. Generalized linear models were used to assess the association between pain intensity, psychological distress, and pain-related co-morbidities. Models were adjusted for variables shown in the literature to influence pain.

RESULTS

532 participants were included in the cohort (56.4% female; median age of 59 years, Interquartile Range [IQR]:47, 69). Comorbid depression (beta coefficient (β) = 0.7; 95%CI: 0.2, 1.2), spine versus lower extremity pain ((β = 0.6; 95%CI: 0.1, 1.1), and prior surgery (β = 0.8, 95%CI: 0.3, 1.4) were associated with higher pain intensity scores. No pain-related comorbidities were associated with pain-related psychological distress (yellow flag count or number of domains). Female sex was associated with less pain-related psychological distress (β = -0.2, 95%CI: -0.3, -0.02).

CONCLUSIONS

Depression was associated with greater pain intensity. No comorbidities were able to account for the extent of pain-related psychological distress.

Improved and computationally stable estimation of relative risk regression with one binary exposure

In medical statistics, when the effect of a binary risk factor on a binary response is of interest, relative risk is often the preferred measure due to its direct interpretation. However, statistical inference on this quantity is not as straightforward as for other measures of association, especially when further explanatory variables have to be taken into account. Starting from a review of available methods for inference on relative risk, this paper deals with small and moderate sample size settings for which we show that classical approaches can be problematic. For this reason, we propose the use of improved estimation procedures, aiming at mean or median bias reduction of the maximum likelihood estimator. In particular, these methods are developed for a new alternative specification of a model recently proposed by Richardson et al, where higher computational stability of the estimation methods is achieved. A real-data example and extensive simulation studies show that the proposed methods perform remarkably better than the standard ones.