Utility of features of the patient's history in the diagnosis of atraumatic shoulder pain: a systematic review

Abstract

Background: Whereas physical examination tests for shoulder disorders have numeric values that describe the utility of the test and its effect on the probability of having a diagnosis, this information is lacking for elements of the history. The purpose of this study was to conduct a systematic review of the literature to determine numeric data (sensitivity, specificity, predictive values, and likelihood or odds ratios) for elements of the history with regard to diagnoses in patients with chronic atraumatic shoulder pain.

Methods: We performed a systematic review to extract information from the existing literature regarding the numeric utility of different features of the patient history as they pertain to chronic atraumatic shoulder pain. Data sources were MEDLINE through PubMed (1946-January 2012) and EMBASE through Ovid (1980-January 2012).

Results: Twenty-one studies met inclusion criteria. A diagnosis of rotator cuff tear was more likely with a history of hypercholesterolemia, having a relative with rotator cuff disease, excessive lifting, above-shoulder work, hand-held vibration work, or age older than 60 years. Acromioclavicular arthritis was more likely in weightlifters. Glenohumeral arthritis was more likely if the patient has a history of prior dislocation, age >75 years, or a diagnosis of knee osteoarthritis. Adhesive capsulitis was more likely with a history of diabetes or thyroid disorder. Posterior labral tear was more likely in football players.

Conclusions: The numeric values for the utility of these history features will help establish numeric probabilities for diagnoses in patients with shoulder pain.

Keywords: Shoulder pain; diagnosis; history; likelihood ratios; predictive values; probability; sensitivity; specificity.