REPORTING OF ACUTE LOW-BACK-PAIN IN A TELEPHONE INTERVIEW - IDENTIFICATION OF POTENTIAL BIASES

Study Design. This was a survey of 235 individuals with and 132 individuals without documented low back pain. Objectives. To approximate the magnitude of potential reporting biases in estimates of prevalence of and medical care use in low back pain. Summary of Background Data. The use of survey techniques presents several possible biases in the reporting of acute symptoms. These biases are especially pertinent in musculoskeletal symptoms, which often are recurrent and not life-threatening. Methods. Two-hundred-thirty-five patients with acute low back pain were contacted by telephone 4-16 months after their physician visit and surveyed regarding the presence and date of back pain episodes. One-hundred-thirty-two patients who had no functionally disabling back pain on physician interview were interviewed. Results. Of the patients who had sought care for back pain, 21% indicated they had not had back pain when interviewed 4-16 months later. Episodes of pain that occurred more than 8 months before the interview tended to be recalled as occurring more recently than they actually occurred, confirming ''forward telescoping'' Of the illness episode. Only 3% of the individuals without functionally impairing pain reported such pain on a separate interview. Conclusions. Lack of recall occurs regarding acute low back pain, usually a self-limited illness. This potential under-estimate of back pain prevalence may be balanced by forward telescoping of the dale of illness occurrence.