Prognostic factors for neuroreflexotherapy in the treatment of subacute and chronic neck and back pain: a study of predictors of clinical outcome in routine practice of the Spanish National Health Service.

Prospective cohort follow-up study. To refine the indication criteria for neuroreflexotherapy (NRT) in the treatment of subacute and chronic neck (NP), thoracic (TP) and low back pain (LBP) in the Spanish National Health Service (SNHS), by identifying prognostic factors for clinical outcome. NRT consists of the temporary subcutaneous implantation of surgical devices in trigger points. Previous randomized controlled trials have shown its efficacy, effectiveness, and cost-effectiveness for treating subacute and chronic LBP. Clinical audits in routine practice have shown similar results in NP, TP, and LBP patients. All 1514 patients from the SNHS in the Balearic Islands in which a NRT intervention was performed and who were discharged between January 1, 2004, and December 31, 2005, were included in this study. Treatment failure was defined as a baseline score equivalent to or lower than the corresponding one at discharge for local pain, referred pain, or LBP-related disability. Multivariate logistic regression models were developed for each of those variables. Maximal models included reason for referral (NP, TP, or LBP), age, sex, baseline values for each variable, number of days in which the surgical devices used in NRT were left implanted, duration of the current episode, time elapsed since the first episode, and previous failed surgery for the current episode. Calibration of the models was assessed through the Hosmer-Lemeshow test, while discrimination was assessed through the area under the ROC curve and the Nagelkerke R test. When referred to NRT, patients' median (IQR) duration of the episode was 210 (90, 730) days. Failure rates ranged between 9.9% for local pain and 14.5% for disability. Variables associated with a worst prognosis for local pain, referred pain, and disability were surgical devices remaining implanted for a shorter duration and, especially, a longer pain duration. Patients referred for NP were more likely to improve than those referred for TP or LBP. Regarding the evolution of local and referred pain, lesser improvement was observed in the least severe complaint at baseline. Models showed a good calibration. The area under the ROC curve ranged between 0.719 and 0.804, and R ranged between 0.101 and 0.255. A longer duration of the current episode is the worst detected prognostic factor for response to NRT, but prognostic models are inaccurate for predicting the clinical outcome of a given patient. In order to improve the success rate of this technology, these results only support earlier referral for patients complying with current indication criteria.