In 1980 little was known about the epidemiology  of  low back disorders  in the United States. Jennifer Kelsey’s studies of lumbar disc herniations strongly implicated vehicular vibration  and cigarette smoking as  risk factors. Population- and industrial-based studies in England, Jamaica, Sweden and Denmark had identified biomechanical exposures (lifting, bending, twisting), psychosocial factors (depression,  job dissatisfaction),  and cigarette  smoking  as potential risk factors. At the University of Vermont, we were interested in primary and secondary prevention of low back disorders, as well as improving the quality of care for back pain sufferers.  It made sense to assemble information from a large cohort, and then, using random selection techniques, study this population in detail to further understand and quantify specific risk factors and how they might be modified.  We chose a model, large academic primary care practice which used the Problem Oriented Medical Record, and therefore had a reliable data base with regards  to medical encounters for back pain, as well as an extensive general medical record. The result of this analysis was the article published in Spine.  We clearly acknowledged such a study design could not establish causality, but it did give us a basis for further prospective radiographic, psychological, biomechanical and work place studies.

The results largely confirmed the Swedish epidemiologic data.   People reporting LBP had more occupational exposures to lifting, vehicular and non-vehicular vibration, and work place stresses, more complaints of other medical ailments and depression, and were more often smokers. Women with multiple pregnancies were at greater risk. More importantly, we now could do more sophisticated prospective studies of  these multiple risk factors. For example, we carefully analyzed industrial exposures and developed tools to measure and model the effects of vehicular vibration.  This information then provided a basis for rethinking how the seating pans of automobiles and trucks are configured as a method  for primary and secondary prevention of LBP. As another example, we looked at the predictive value of spinal radiographs, including flexion-extension laterals.  This study clearly demonstrated no predictive value for virtually all of the common radiographic findings such as narrowed lumbar disc, paradiscal osteophyes, end plate sclerosis, Schmorls nodes, and segmentation abnormalities (there was a minor relationship between L4-5 disc space narrowing and the presence of traction spurs, thought to connote instability).  What is amazing after the passage of 31 years is how these findings have stood the test of time despite some significant methodologic flaws.  What amazes me even more is the attention this study has attracted over the 31 years since it was written. I am heartened that people remain  interested in identifying risk factors  from which preventive strategies can be designed - a far better strategy than spinal surgery. I would also gratefully acknowledge that this study represented the combined efforts of a dedicated team of bioengineers, psychologists, PT and OT, doctors, and statisticians.