Development of a clinical prediction rule to identify patients with neck pain likely to benefit from thrust joint manipulation to the cervical spine.

STUDY DESIGN Prospective cohort/predictive validity study. OBJECTIVE To determine the predictive validity of selected clinical examination items and to develop a clinical prediction rule to determine which patients with neck pain may benefit from cervical thrust joint manipulation (TJM) and exercise. BACKGROUND TJM to the cervical spine has been shown to be effective in patients presenting with a primary report of neck pain. It would be useful for clinicians to have a decision-making tool, such as a clinical prediction rule, that could accurately identify which subgroup of patients would respond positively to cervical TJM. METHODS Consecutive patients who presented to physical therapy with a primary complaint of neck pain completed a series of self-report measures, then received a detailed standardized history and physical examination. After the clinical examination, all patients received a standardized treatment regimen consisting of cervical TJM and range-of-motion exercise. Depending on response to treatment, patients were treated for 1 or 2 sessions over approximately 1 week. At the end of their participation in the study, patients were classified as having experienced a successful outcome based on a score of +5 ("quite a bit better") or higher on the global rating of change scale. Sensitivity, specificity, and positive and negative likelihood ratios were calculated for all potential predictor variables. Univariate techniques and stepwise logistic regression were used to determine the most parsimonious set of variables for prediction of treatment success. Variables retained in the regression model were used to develop a multivariate clinical prediction rule. RESULTS Eighty-two patients were included in data analysis, of whom 32 (39%) achieved a successful outcome. A clinical prediction rule with 4 attributes (symptom duration less than 38 days, positive expectation that manipulation will help, side-to-side difference in cervical rotation range of motion of 10° or greater, and pain with posteroanterior spring testing of the middle cervical spine) was identified. If 3 or more of the 4 attributes (positive likelihood ratio of 13.5) were present, the probability of experiencing a successful outcome improved from 39% to 90%. CONCLUSION The clinical prediction rule may improve decision making by providing the ability to a priori identify patients with neck pain who are likely to benefit from cervical TJM and range-of-motion exercise. However, this is only the first step in the process of developing and testing a clinical prediction rule, as future studies are necessary to validate the results and should include long-term follow-up and a comparison group. LEVEL OF EVIDENCE Prognosis, level 2b.