Estimating the prevalence of low back pain in the general population. Evidence from the South Manchester Back Pain Survey.

STUDY DESIGN: This report gives the results of a population-based cross-sectional mailed questionnaire, with prospective follow-up of survey responders and nonresponders. OBJECTIVE: To determine the 1-month period prevalence of low back pain in an adult population in the United Kingdom and to estimate the effect of nonresponse bias. SUMMARY OF BACKGROUND DATA: Previous United Kingdom population studies have reported a 1-year period prevalence of low back pain of 37%. However, the definitions of low back pain have varied, and the influence of nonresponse rarely has been reported. METHODS: The study population was made up of all 7669 adults (18 to 75 years old) registered with two family practices in a sociodemographically mixed suburban area. The questionnaire, including a pain drawing to identify the site of any pain, was mailed to the entire study population. Two repeat mailings were sent to nonresponders. Family practice consultations about low back pain by individuals from the study population were monitored over the following 12 months using computerized records of all surgery contacts. RESULTS: Of the study population, 4501 (59%) responded. The 1-month period prevalence of low back pain was 39% (35% in males, 42% in females). The age distribution was unimodal, with peak prevalence in those aged 45 to 59 years old. Responders to the first mailing had a small but nonsignificant increase in prevalence compared with those who responded to the second or third mailing. Nonresponders had a subsequent consultation rate for low back pain that was 22% lower than that for the survey responders. CONCLUSIONS: After considering potential differences in nonresponders, the estimated 1-month prevalence of low back pain was between 35% and 37%. Prevalence figures in survey responders may overestimate the true population prevalence by a modest amount.