A simple classification system was recommended for patients with restricted shoulder or neck range of motion.

OBJECTIVE To construct an empirical classification of patients with shoulder complaints, and then to investigate the relationship between the empirical classification and the setting in which the patients were recruited, their demographic and clinical characteristics, and the original diagnostic categories. STUDY DESIGN AND SETTING A latent class analysis was performed on the combined data of two previous studies. RESULTS Four clusters of patients emerged, one with patients who have a small chance of any restriction of the motion of the shoulder and a moderate chance of restriction of the motions of the neck, a second with patients who have a high chance of restriction of the motions of the shoulder and a moderate chance of restriction of motions of the neck, a third with patients who have a low to moderate chance of restriction of all motions, and a fourth with patients who have a high chance of restriction of all of the motions. Patients recruited from a clinic for rheumatology and rehabilitation and from the orthopedic clinic were more present in the second cluster. Patients in the third cluster were on average younger than those in the other clusters. There appeared to be little agreement between the clusters found and the classification of patients according to the original diagnostic categories. CONCLUSION Patients experiencing shoulder pain can be classified in a simple way into four categories, reflecting the distinction between problems of the shoulder and those of the neck. A simple classification rule is proposed with which almost all patients can be classified into the four clusters.