PURPOSES: We used two different methods to classify low back pain (LBP) in the general population (1) to assess the overlapping of individuals within the different subgroups in those two classifications, (2) to explore if the associations between LBP and some selected bio-psychosocial factors are similar, regardless which of the two classifications is used.
METHOD: During 1 year, 49- or 50-year-old people from the Danish general population were sent fortnightly automated text messages (SMS-Track) asking them if they had any LBP in the past fortnight. Responses for the whole year were then classified into two different ways: (1) In relation to the number of days with LBP in the preceding year (0, 1-30, and >30), (2) In relation to the frequency and duration of episodes of LBP (more or less never pain, episodic, and more or less constant pain). Some bio-psychosocial factors, collected with a questionnaire at baseline 9 years earlier, were entered into regression models to investigate their associations with the subgroups of the two classifications of LBP and the results compared.
RESULTS: The percentage of agreement between categories of the two classification systems was above 68 % (Kappa 0.7). Despite the large overlap of persons in the two classification groups, the patterns of associations with the two types of LBP definitions were different in the two classification groups. However, none of the estimates were significantly different when the variables were compared across the two classifications.
CONCLUSION: Different classification systems of LBP are capable of bringing forth different findings. This may help explain the lack of consistency between studies on risk factors of LBP.