A recently updated detection model for low back pain was found to be significantly efficient in predicting the possible recovery of patients with acute low back pain.
Evident from the findings of a recently published study in the European Journal of Pain, a clinical prediction model with five easily collected variables present a potential efficacy in the possible prognosis of low back pain (LBP).
A total of 737 participants with a pain
score of ≥2/10 and duration of current episode of ≤4 weeks were
selected for the investigation. Days to pain recovery were considered
as the primary outcome. Before refitting the current model, some of
the variables from the development dataset were re-classified. The
calibration and discrimination of the prediction model were later
examined in the validation dataset. The predictor variables involved
in the analysis were the number of previous episodes, pain intensity
change over the first week, duration of current episode, pain
intensity, and depression.
Three variables went through the
re-classification and performed well in the development dataset. In
one month, the calibration for the validation sample was accepted.
The predicted proportions in the quintiles led to overestimate the
noticed recovery proportions at one week and underestimate at three
months. The renewed prediction model illustrated good external
validity and beneficial in practice; however, further impact studies
and validation in related populations is required.
The European Journal of Pain
Predicting pain recovery in patients with acute low back pain: Updating and validation of a clinical prediction model
T da Silva et al.
Comments (1)