Eighty percent of all adults experience back pain at some time.For 1 adult in 10, back pain is a chronic problem. About 2.6million people are permanently disabled by back pain. Spinedisability is growing more than three times faster than thegrowth rate of the population. Back injuries account forone-third of all workplace injuries. Industrial back injuriescost $14 billion each year. The typical back pain episode resultsin 14 lost work days. A person off work for 6 months due to aback injury has only a 50 percent probability of returning towork; after 12 months off work the probability drops to 12percent; after 2 years of absence only 1 percent of injuredworkers return to work.

NONSURGICAL TREATMENT

Pain is a subjective experience; no more than half the peoplewith low back pain have identifiable structural abnormalities orclinical syndromes. Countless operations have been performed torelieve low back pain. Many of these procedures have beenineffective; that is the patient reports no improvement orincreased pain after healing from the procedure (Waddell, Kummel,Lotto, Graham, Hall, & McCulloch, 1979).

Because many people with chronic back pain are poor candidatesfor surgery or have failed to benefit from surgery, it isimportant to assess the efficacy of nonsurgical treatment.Chronic low back pain is the most costly problem for workers'compensation systems, and it is costly for health insurers.Nonsurgical chronic pain treatment can be expensive. Workers'compensation and health insurers can justify the high cost ofnonsurgical treatment only if it ultimately results in costsavings.

Under a grant from the National Institute on Disability andRehabilitation Research, researchers at the University of Miami(Cutler, Fishbain, Rosomoff, Abel-Moty, Khalik, & Rosomoff, 1992)conducted a review and meta-analysis of the back treatmentresearch literature to evaluate the impact of nonoperativetherapies. Nonsurgical treatments encompass any type ofconservative pain treatment, for example, physical therapy,occupational therapy, transcutaneous electrical nerve stimulation(TENS), individual or group psychotherapy, education, cognitiveretraining relaxation training, hypnosis, biofeedback, nerveblocks, medication management, either singly or in combination.

Meta-analysis had not previously been applied to the back paintreatment literature. The research team conducted a computerizedand manual search of the literature using the key words chronicpain, treatment, nonsurgical, and human subjects. The researchersreviewed 171 studies of which 37 fulfilled the criteria forinclusion.

Return to work was chosen as the indicator of a favorableoutcome. It is more objective and reliable than self-report dataand other outcome measures. Return to work is relatively welldefined, making comparison of results between studies possible.It is also an important outcome in the "real world" because ofthe personal and the public financial impact of not working.Working was defined as fulltime or part-time work for wages orsalary. Students and full-time homemakers were defined as notworking for the purpose of this research. To be included in themeta-analysis, articles had to provide a detailed definition ofpretreatment and followup work status for each participant. Someof the included studies reported part-time and full-time work atfollowup separately; others did not.

The 37 studies were divided into four groups according to thepretreatment and posttreatment employment status of theparticipants.

1. None working pretreatment, combined full-time and part-time work at followup.
2. Some working pretreatment, combined full-time and part-timework at followup.
3. None working pretreatment, full-time work at followup.
4. Some working pretreatment, full-time work at followup.

1. 0 percent working pretreatment, 42 percent working at followup(42% increase).
2. 20 percent working pretreatment, 54 percent working atfollowup (34% increase).
3. 0 percent working pretreatment, 32 percent working at followup(32% increase).
4. 12 percent working pretreatment, 60 percent working atfollowup (48% increase).

All four groups exhibited substantial improvement in work statusfollowing treatment. The mean proportion working at followup was41 percent; the median was 38 percent.

Duration of the effect of treatment is an importantconsideration--do people who return to work following treatmentremain at work? The studies analyzed varied with respect to howlong they followed up on those who had received treatment. Thefollowup periods ranged from 1 to 60 months. The researchersfound that employment rates during the first posttreatment yearwere similar to the rates in the second and third years. Dataindicated that the long-term employment outcome was as favorableas the short-term outcome. Employment did not increase ordecrease significantly over time.

To determine if the results could be explained on the basis ofspontaneous remission, placebo effects, or other factorsunrelated to the nonsurgical treatment, the return-to-work rateof the treatment group was compared to two naturally occurringcontrol groups. Six studies comparing treatment and controlgroups were subjected to meta-analysis. These studies usedcomparison data from people who were rejected for treatment andfrom people who dropped out of treatment. Members of thesecontrol groups returned to work at much lower rates than didmembers of the treatment group. These results suggest that thepositive return-to-work outcomes demonstrated for nonsurgicalchronic pain treatment cannot be attributed to factors other thanthe treatment.

• One study compared treated people with people rejected due tolack of insurance: 65 percent of treated people were working atfollowup, compared with only 24 percent of people rejected due tolack of insurance. The weighted mean difference of 40 percent infavor of the treatment group is statistically significant(p<.001).
• One study compared treated people with people who dropped outof treatment: 77 percent of people who completed treatment wereworking at followup, compared to 24 percent of people who droppedout of treatment. The weighted mean difference of 56 percent infavor of the treatment group is statistically significantP<.001).
• Four studies compared outcome differences between peoplerejected due to lack of insurance with those who dropped out oftreatment. More of the people who were rejected for lack ofinsurance returned to work, but the difference was notstatistically significant.

Perhaps people who are rejected for insurance reasons andtreatment dropouts are not adequate controls. The type ofinsurance coverage a person has may in itself be related toimportant outcome variables predictive of return to work. Forexample, commercial health insurance does not routinely covernonsurgical treatment of chronic pain. On the other hand,workers' compensation carriers will often cover nonsurgical paintreatment but may refuse to do so on an individual basis due tolitigation problems or the perception that the injured worker isnot a good candidate for rehabilitation. Those who are rejectedbecause of insurance reasons may, therefore, not be comparable tothe treatment group and, consequently, may not be an appropriatecontrol group. Dropouts may also be an inappropriate controlbecause it is likely that this group includes a high proportionof people who are resistant to treatment. The authors suggestthat a better comparison group for future studies would be peoplewho inquire about treatment but do not enter for reasons otherthan insurance coverage.

The researchers observe that there was great variability withinthe studies analyzed regarding the percentages of participantsfollowed up. Those who respond to a followup contact often areself-selected and may differ in other ways from those who cannotbe contacted or do not respond. This may have influenced theaccuracy of the outcome measures.

Although the studies analyzed contained the deficiencies noted,Cutler and his colleagues concluded that their research clearlydemonstrated that nonsurgical treatment of chronic pain hadsignificant and lasting benefits. It more than doubled the numberof people who returned to work, and those who returned remainedat work. This desirable outcome was directly attributable to thenonsurgical treatment these people received.

WORK AND LOW BACK PAIN

Damkot, Pope, Lord, and Frymoyer (1984) analyzed the associationbetween work and present or past complaints of low back pain. Theresearchers conducted interviews with 303 men to determine whichaspects of their jobs and working environments were related tolow back pain. Respondents included men from 18 to 55 years old,of whom 96 percent were employed. Thirty-four percent of the mensaid that they had never had low back pain, 44 percent reportedmoderate low back pain at present or in the past, and 22 percentreported severe present or past low back pain. Sixteen percent ofparticipants worked in jobs involving manual labor (e.g.,carpenters, mechanics, machine operators); 15.2 percent worked insedentary jobs (e.g., electrical engineers and supervisors); 69percent worked in jobs involving both laboring and sedentaryactivities.

Representative work sites were visited to observe standardpatterns of physical activity. Observational data were obtainedfor 150 variables including task frequencies, weights lifted,lifting posture, pushing, pulling, carrying, type of floorsurface, type of chair support, and driving vehicles. These datawere used to construct a workplace interview questionnaire.Through the questionnaire the researchers gathered informationabout each participant's work activities and off-duty activitiessuch as lawnmowing, chainsawing, and lifting. The questionsregarded the frequency of various tasks, weights lifted, liftingpostures, floor surfaces, type of chair support, and requirementsfor stretching, bending, and twisting. Those who reported havinglow back pain were also asked to identify any specific incidentto which they attributed their symptoms and to describe thesituations that provoked their pain. Participants also underwentcomplete medical, psychological, and biomechanical assessments.

The interview data were subjected to univariate and multivariateregression analyses to determine if the three pain groups--nopain, moderate pain, severe pain--differed regarding workactivities. Occupational factors only partially discriminatedbetween the presence or absence of pain, or among degrees of painreported. A statistically significant relationship was foundbetween the average weekly hours of heavy truck driving and thethree levels of back pain. Driving heavy trucks an average 2.1hours per week was associated with no pain, 2.2 hours per weekheavy truck driving was associated with moderate pain, and 6hours per week heavy truck driving was associated with severepain. No other significant relationships were identified betweendiscrete occupations and specific pain levels. Lifting, pushing,and pulling were found to be important risk factors for low backpain which, when combined with vehicular vibration, increased therisk exposure for truck drivers who were required to do their ownloading.

Across occupations, a significant relationship was found betweenstretching and reaching and severe low back pain--59 percent ofthe men reporting severe pain were required to stretch and reach.Those with severe low back pain were also more likely to reachwith arms fully extended than were those in the two other paingroups. The total physical work demands for the men who reportedmoderate or severe low back pain were found to be greater thanwere the work demands for those in the no pain group. Thefrequency and amount of lifting, however, were not significantlydifferent for the three pain groups. Prolonged sitting was notassociated with severe back pain. A clear difference wasidentified between the moderate and severe pain groups regardingattributing the onset of pain to a specific occurrence: 70percent of the severe pain group but only 37 percent of themoderate pain group reported that their pain had a sudden acuteonset. Surprisingly, no relationship between report of suddenonset and compensable injury was identified.

VERMONT LOW BACK PAIN REC

The Vermont Rehabilitation Engineering Center (REC) for Low BackPain has been continuously funded by the National Institute onDisability and Rehabilitation Research since September 1983. Therecognition that low back pain is the most costly and disablingmusculoskeletal impairment has increased during that time. Aprior Rehab BRIEF (Vol. 10, No. 9) reported the work of theVermont REC. The Vermont REC continues to make meaningfulprogress on the prediction, prevention, and treatment of low backpain (Vermont Rehabilitation Engineering Center for Low BackPain, Progress Report, 1991).

Predicting Back Disability

Rowland G. Hazard, M.D., leads a team at the Vemmont RECconcerned with identifying the most highly predictive andcost-effective set of tests that would be included in an optimalworker capacity evaluation. As a part of this project, aconsensus process involving 30 experts from many disciplines hasdetermined the 13 factors considered most predictive ofoccupational back injury:

• severity of previous back problems,
• previous work loss due to low back problems,
• prior compensation for back pain,
• anticipated difficulty with sitting/standing on the job,
• previous spinal surgery,
• perception of fault for previous back problem,
• previous use of health care services for back problems,
• history of back pain,
• lifting capacity,
• retaining attorney for health-related claims,
• current worker compensation status,
• physical demands of work to date, and
• previous work satisfaction.

Vehicular Vibration

Damkot et al. documented a dose-related correlation between lowback pain and exposure to vehicular vibration. Prior research atthe Vemmont REC has also shown that long-term exposure to vehiclevibration is associated with above average rates of herniateddisc. Operating vehicles with little or no springing, such asfork-lift trucks, increases the risk of back disorders. David G.Wilder, Ph.D., directs a project that includes studies of workersin many occupations that require the use of vibrating equipmentor vehicles.

Vibration levels for different types of vehicles being drivenover the road have been recorded and measured. These signals canbe played back and reproduced in a vibration simulator thatsimultaneously produces vertical and horizontal vibrationcoinciding with that of a specified vehicle. By reproducing thesevibration characteristics in the laboratory, researchers are ableto more accurately measure physical and behavioral responses tothe vibratory environment. Computer software has been developedto collect, analyze, and display the vibrational data. Acomputer-based animated "movie" can depict the center of pressurepath, the reaction force vector, reaction torque, seat position,and muscle activity.

One of the goals of this project is to develop specifications foran optimal seat that lessens subjective fatigue and pain forpeople whose work involves prolonged exposure to vibrationthrough the operation of vibrating equipment or vehicles. Theeffects of various back supports, arm rests, foot rests, and seatshock absorption systems are being investigated. The effect ofsitting improperly--poor posture--in a properly designed andadjusted seat is also being examined. Volunteers with low backpain will be age- and gender-matched with pain-free controlvolunteers for this study.

Workload Assessment System

Preventing industrial back injuries and accommodatinglow-back-impaired workers requires detailed information aboutbiomechanical stresses on the job. Gerald Weisman, M.S., directsa project at the Vermont REC to develop an inventory ofbiomechanical job stresses. In the past, efforts to assessworkload have relied on observations by an evaluator (typicallyan ergonomist, occupational therapist, engineer, orrehabilitationist) watching a worker doing the job. One evaluatormust spend several hours observing one job. The WorkloadAssessment System can record detailed objective data aboutseveral workers or jobs at the same time and does not require thecontinuous presence of the evaluator.

An electrogoniometer was designed and fabricated to measurethree-axis trunk motion. EMG data are being collected to providea graphic record of the contraction of spinal muscles whenperforming real work. Analysis of the EMG data will furnishinformation on task symmetry, gross frequency of movement, andenergy output. A device called a Limb Load Monitor (developed bythe REC at Moss Rehabilitation Hospital) is being used to assessthe gross load on the body. The Limb Load Monitor can be set toemit a tone when force exceeds a predetermined level. Each pieceof equipment has been tested in the laboratory to be certain thatit will not interfere with task performance.

The Workload Assessment System will collect concurrent data fromthe goniometer, the EMG, and the Limb Load Monitor for 2 to 3hours at a time. These data will then be uploaded to a computeranalysis program to provide a profile of the biomechanicalstresses experienced by a worker performing a specific job. Thesystem has been fieldtested in public- and private-sector worksites. Data are being gathered to determine the WorkloadAssessment System's utility in discriminating among differentoccupations known to have different low back injury rates.

Sit-Stand Workstation

More than 50 percent of all Americans work at desk jobs. Asincreased use of computers has resulted in more workers remainingin static, seated positions for prolonged periods, low backcomplaints among sedentary workers have been increasing.Meanwhile, many people are precluded from otherwise suitable jobsbecause of the back pain that they experience during requiredsustained sitting. Changes in posture can relieve discomfort andmay reduce the risk of back injury. Steven M. Reinecke, M.S.,directs a project to test the efficacy of a workstation at whicha person can comfortably alternate between a sitting and astanding posture without jeopardizing productivity. Futureresearch will determine whether workers who normally sit all dayprefer to alternate between sitting and standing, whether chairtype affects the percent of time sitting and standing, andwhether the ease of adjusting the workstation affects the percentof time sitting and standing.

IMPLICATIONS

Chronic back pain is less dramatic than many catastrophicdisabilities. Because it is so common, we may think that it isnot serious. Nevertheless, over 5 million Americans are disabledby low back pain. For half of them the disability is permanent.Implementation of the findings in the research briefed above mayhelp to curb the alarming growth rate of disabling back pain.Application of the knowledge gained could enable many people toreturn to work.

Questions are being raised as to whether people with chronic backpain are covered by the Americans With Disabilities Act (ADA).The ADA extends protection to individuals who have a physical ormental disability that substantially limits one or more majorlife activities. Listed major life activities include caring foroneself, performing manual tasks, walking, seeing, hearing,speaking, breathing, learning, and working. Although people withchronic back problems often have some limitations on walking,they are usually much more limited in their ability to stand,sit, and lift. Are these major life activities? They are neitherlisted nor excluded by the ADA regulations. The determinationwill depend in part on whether the courts decide that chronicback pain is really serious.

Cutler, R.B., Fishbain, D.A., Rosomoff, H.L., Abel-Moty, E.,Khalik, T.M., & Rosomoff, R.S. (1992). Does non-surgical paincenter treatment of chronic pain return patients to work? Areview and meta-analysis of the literature. Paper presented atthe American Pain Society 11th Annual Meeting, New Orleans.

Damkot, D.K., Pope, M.H., Lord, J., & Frymoyer, J.W. (1984). Therelationship between work history, work environment and low backpain in men. Spine, 9, 395-399.

Vermont Rehabilitation Engineering Center for Low Back Pain.(undated). Back facts. Burlington, VT: Vermont RehabilitationEngineering Center.

Vermont Rehabilitation Engineering Center for Low Back Pain.(1990). Progress report. Burlington, VT: Vermont RehabilitationEngineering Center.

Vermont Rehabilitation Engineering Center for Low Back Pain.(1991). Progress report. Burlington, VT: Vermont RehabilitationEngineering Center.

Waddell, G, Kummel, E.G., Lono, W.N., Graham, J.D., Hall, H., &McCulloch, J.A. (1979). Failed lumbar disc surgery and repeatsurgery following industrial injuries. The Journal of Bone andJoint Surgery, 61, 201-207.