Identifying heavy health care users among primary care patients with chronic non-cancer pain.

All remaining variables (sex, comorbidity, depression levels, attitude toward medical cure) were useful in predicting heavy health care use, but their respective ORs did not reach statistical significance after entering the variable “type of health care user prior to recruitment” into the model.

also in 111,027 other papers

However, the associated OR failed to reach statistical significance, most likely because its variance was largely accounted for by patients’ health care use in the prerecruitment year.

also in 6,035 other papers