Quantitative dynamic wearable motion-based metric compared to patient-reported outcomes as indicators of functional recovery after lumbar fusion surgery.

BACKGROUND: Low back pain is a debilitating condition with poor patient outcomes despite the use of a wide variety of diagnostic and treatment modalities. A lack of objective metrics to support clinical decision-making may be a reason for these poor outcomes. This study aimed to compare patient recovery following lumbar fusion surgery using an objective motion-based metric (functional performance) and subjective patient-reported outcomes for pain, disability and kinesophobia. METHODS: A prospective observational study was conducted on 121 patients that received a lumbar fusion surgery. A wearable motion system was used to quantify three-dimensional multi-planar lumbar motion and benchmark each patient's lumbar function prior to surgery and post-operatively at follow-up time points for up to 2 years. Patient recovery profiles after surgery were evaluated using the acquired functional motion data and compared to patient-reported outcomes. FINDINGS: Our results found significant improvement after surgery in objective functional performance as well as patient-reported pain, disability, and kinesophobia. However, we found a delayed response in the objective metric, with meaningful improvement occurring only 6 months after fusion surgery. In contrast, we found significant improvement in all subjective scores as early as 6 weeks post-surgery. INTERPRETATION: Objective motion-based metric provides a unique perspective to assessing patient's functional recovery. While it is associated with dimensions of pain, disability and fear avoidance, it is also distinct and assesses a uniquely different dimension of functional health. This information can form the basis for the use of objective metrics to gauge patient recovery after lumbar fusion surgery.

Differences in lumbar spine measures as a function of MRI posture in low back pain patients and its clinical implications.

STUDY DESIGN: Observational Study. OBJECTIVE: The primary objective was to determine if there were differences in spine structure measures between experimental postures and standard supine posture MRIs. METHODS: Thirty-four low back pain patients were included. MRI was taken in 6 experimental postures. The dependent measures includes sagittal view anterior (ADH), middle and posterior disc heights, thecal sac width, left/right foraminal height (FH). In the axial view: disc width, left and right foraminal height. Measures were done L3/L4, L4/L5 and L5/S1. Each subject served as their own control. Spine measurements in the experimental posture were compared to the same measures in the standard supine posture. RESULTS: 94% inter-observer reliability was seen. In the sagittal and axial view, 55 of the 108 and 11 of the 18 measures were significantly different. In sagittal view: a) ADH was significantly smaller in the sitting flexed posture by 2.50 mm ± 0.63 compared to the supine posture; b) ADH in sitting neutral posture was significantly smaller than the standard posture by 1.97 mm ± 0.86; c) sitting flexed posture showed that bilateral FH measures were significantly different; d) Bilateral FH was larger in the sitting neutral posture compared to the standard supine posture by 0.87 mm ± 0.17. CONCLUSIONS: This research quantifies the differences in spine structure measures that occur in various experimental postures. The additional information gathered from an upright MRI may correlate with symptoms leading to an accurate diagnosis and assist in future spine research.

Psychosocial Factors and Low Back Pain Outcomes in a Pooled Analysis of Low Back Pain Studies.

OBJECTIVE: Assessment of possible relationships between work-related psychosocial measures and self-reported low back pain (LBP) outcomes in a large pooled dataset of 1929 participants from 82 facilities in the United States. METHODS: Pooled data from three prospective cohort studies were used to calculate odds ratios (OR) and 95% confidence intervals (95% CI) for relationships between psychosocial factors and the LBP outcomes. Personal and occupational confounders were controlled for in adjusted Logistic regression models. RESULTS: Supervisor support and job satisfaction were significantly (P < 0.05) related to all three LBP outcomes. Other psychosocial factors were significantly (P < 0.05) associated with at least one of the LBP outcomes. Adjusted ORs ranged from approximately 1.50 to 3.50 for most associations. CONCLUSIONS: There is a significant relationship between work-related psychosocial measures and LBP outcomes.

Prevalence of low back pain, seeking medical care, and lost time due to low back pain among manual material handling workers in the United States.

BACKGROUND: Low back pain (LBP) is a common and costly problem throughout the United States. To achieve a greater understanding of the occupational risk factors, the National Institute for Occupational Safety and Health (NIOSH) funded a low back health effects consortium, which performed several surveillance studies throughout the United States. This study combines data from the consortium research groups resulting in a data set with nearly 2000 workers in various regions of the country. The purpose of this paper is to examine prevalence and personal risk factors of low back health effects among these workers. METHODS: There were three common questions regarding history of low back health effects in the past 12 months including 1) have you had LBP lasting 7 days, 2) have you sought medical care for LBP, and 3) have you taken time off work due to LBP. The questionnaire included demographic questions. There were five data collections institutions or sites including NIOSH, Ohio State University, University of Wisconsin-Milwaukee, Texas A&M University, and University of Utah. RESULTS: The 12-month period prevalence of low back pain lasting 7 days, seeking medical care, and lost time due to LBP were 25, 14 and 10%, respectively. There were no statistically significant differences in gender, age or weight between cases and non-cases for any prevalence measure. The height of workers was significantly greater in the cases compared to non-cases for all three prevalence definitions. There were significant differences among the sites on the prevalence of seeking medical care for LBP and lost time due to LBP. The Ohio State University had significantly higher prevalence rates for seeking medical care and lost time due to LBP than University of Wisconsin, University of Utah, or Texas A&M University. CONCLUSION: LBP, the least severe low back health effect studied, had the highest prevalence (25%) and lost time due to LBP, the most severe low back health effect studied, had the lowest prevalence (10%) among nearly 2000 US manual material handling workers. There was a significant site or regional influence in prevalence rates for seeking medical care and lost time due to LBP.

A Randomized Trial Evaluating the Prophylactic Activity of DSM265 Against Preerythrocytic Plasmodium falciparum Infection During Controlled Human Malarial Infection by Mosquito Bites and Direct Venous Inoculation.

Background: DSM265 is a selective inhibitor of Plasmodium dihydroorotate dehydrogenase that fully protected against controlled human malarial infection (CHMI) by direct venous inoculation of Plasmodium falciparum sporozoites when administered 1 day before challenge and provided partial protection when administered 7 days before challenge. Methods: A double-blinded, randomized, placebo-controlled trial was performed to assess safety, tolerability, pharmacokinetics, and efficacy of 1 oral dose of 400 mg of DSM265 before CHMI. Three cohorts were studied, with DSM265 administered 3 or 7 days before direct venous inoculation of sporozoites or 7 days before 5 bites from infected mosquitoes. Results: DSM265-related adverse events consisted of mild-to-moderate headache and gastrointestinal symptoms. DSM265 concentrations were consistent with pharmacokinetic models (mean area under the curve extrapolated to infinity, 1707 µg*h/mL). Placebo-treated participants became positive by quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and were treated 7-10 days after CHMI. Among DSM265-treated subjects, 2 of 6 in each cohort were sterilely protected. DSM265-treated recipients had longer times to development of parasitemia than placebo-treated participants (P < .004). Conclusions: This was the first CHMI study of a novel antimalarial compound to compare direct venous inoculation of sporozoites and mosquito bites. Times to qRT-PCR positivity and treatment were comparable for both routes. DSM265 given 3 or 7 days before CHMI was safe and well tolerated but sterilely protected only one third of participants.

Low back functional health status of patient handlers.

PURPOSE: The purpose of this study was to assess low back functional health among a group of nurses with a history of low back pain symptoms in a university hospital using a direct measure of low back functional performance and compare to traditional low back disability and pain questionnaires. METHODS: Fifty-two nurses and patient care associates volunteered for the study. The clinical lumbar motion monitor (LMM) was used to directly measure low back functional performance. The participants performed a series of standard tasks involving trunk flexion and extension at different asymmetries. The LMM measures the motion signature of the participant (range of motion, velocity and acceleration) in all three planes of the body. The clinical LMM evaluation documented objective assessment of low back function normalized for age and gender. The Oswestry Disability Index (ODI) was used to evaluate self-reported disability and the McGill Pain Questionnaire visual analog scale assessed pain symptom. RESULTS: The average functional performance probability was 0.49 with a standard deviation of 0.29, indicating that on average the functional performance was impaired. The average ODI score was 13.4 with a standard deviation of 11.6. The correlation between the functional performance probability and ODI was 0.046 (not statistically significant). CONCLUSIONS: The clinical LMM functional performance measure provides a direct measure of trunk function. The low correlation between the ODI and clinical LMM functional performance probability indicates that the direct functional performance measure adds another component to our understanding of low back health or impairment that traditional questionnaires lack.

Cumulative spine loading and clinically meaningful declines in low-back function.

OBJECTIVE: The objective was to assess the role of cumulative spine loading measures in the development of a clinically meaningful decline in low-back function. BACKGROUND: Cumulative spine loading has been a suspected risk factor for low-back pain for many years, yet the measures that characterize risk have not been well delineated. METHODS: A total of 56 cumulative exposure measures were collected in a prospective field study of distribution center workers. An individual's risk for a clinically meaningful decline in low-back function (true cases) was explored with daily, weekly, and job tenure cumulative exposure measures using univariate and multivariate statistical modeling techniques. True noncases were individuals with no decline in low-back function. RESULTS: An individual's risk for a clinically meaningful decline in low-back function (true cases) was predicted well versus true noncases (sensitivity/specificity = 72%/73%) using initial low-back function (p(n)), cumulative rest time, cumulative load exposure, job satisfaction, and worker age. CONCLUSIONS: Cumulative rest time was identified as an important component for predicting an individual's risk for a clinically meaningful decline in low-back function. APPLICATION: This information can be used to assess cumulative spine loading risk and may help establish guidelines to minimize the risk of a clinically meaningful decline in low-back function.

Are workers who leave a job exposed to similar physical demands as workers who develop clinically meaningful declines in low-back function?

OBJECTIVE: The objective is to quantify differences in physical exposures for those who stayed on a job (survivor) versus those who left the job (turnover). BACKGROUND: It has been suggested that high physical job demands lead to greater turnover and that turnover rates may supplement low-back disorder incidence rates in passive surveillance systems. METHOD: A prospective study with 811 participants was conducted. The physical exposure of distribution center work was quantified using a moment monitor. A total of 68 quantitative physical exposure measures in three categories (load, position, and timing) were examined. Low-back health function was quantified using the lumbar motion monitor at baseline and 6-month follow-up. RESULTS: There were 365 turnover employees within the 6-month follow-up period and 446 "survivors" who remained on the same job, of which 126 survivors had a clinically meaningful decline in low-back functional performance (cases) and 320 survivors did not have a meaningful decline in low-back functional performance (noncases). Of the job exposure measures, 6% were significantly different between turnover and cases compared to 69% between turnover and noncases. Turnover employees had significantly greater exposure compared to noncases. CONCLUSION: Turnover employees had similar physical job exposures to workers who remained on the job and had a clinically meaningful decline in low-back functional performance. Thus, ergonomists and HR should be aware that high turnover jobs appear to have similar physical exposure as those jobs that put workers at risk for a decline in low-back functional performance.

Spine kinematics predict symptom and lost time recurrence: how much recovery is enough?

PURPOSE: The purpose of the study was to determine thresholds for low back kinematic measures for the amount of functional recovery necessary to reduce the risk of recurrent pain symptoms or lost time. METHODS: Low back kinematic ability measures were collected at baseline when the workers returned to work for full duty. The range of motion, velocity, and acceleration were collected using the lumbar motion monitor. RESULTS: Follow-up data was collected on 196 of the 206 workers. Workers with sagittal extension velocity of <40 deg./s at baseline were twice as likely to report recurrent low back pain symptoms. Workers with sagittal flexion velocity <34 deg./s were 3 times more likely to report lost time. CONCLUSIONS: Kinematic functional performance measures may be used as recovery criteria in low back pain patients to minimize recurrence risk.

Shoulder muscle fatigue during repetitive tasks as measured by electromyography and near-infrared spectroscopy.

OBJECTIVE: The objective of this study was to quantify shoulder muscle fatigue during repetitive exertions similar to motions found in automobile assembly tasks. BACKGROUND: Shoulder musculoskeletal disorders (MSDs) are a common and costly problem in automotive manufacturing. METHOD: Ten subjects participated in the study. There were three independent variables: shoulder angle, frequency, and force. There were two types of dependent measures: percentage change in near-infrared spectroscopy (NIRS) measures and change in electromyography (EMG) median frequency. The anterior deltoid and trapezius muscles were measured for both NIRS and EMG. Also, EMG was collected on the middle deltoid and biceps muscles. RESULTS: The results showed that oxygenated hemoglobin decreased significantly due to the main effects (shoulder angle, frequency, and force). The percentage change in oxygenated hemoglobin had a significant interaction attributable to force and repetition for the anterior deltoid muscle, indicating that as repetition increased, the magnitude of the differences between the forces increased. The interaction of repetition and shoulder angle was also significant for the percentage change in oxygenated hemoglobin. The median frequency decreased significantly for the main effects; however, no interactions were statistically significant. CONCLUSIONS: There was significant shoulder muscle fatigue as a function of shoulder angle, task frequency, and force level. Furthermore, percentage change in oxygenated hemoglobin had two statistically significant interactions, enhancing our understanding of these risk factors. APPLICATION: Ergonomists should examine interactions of force and repetition as well as shoulder angle and repetition when evaluating the risk of shoulder MSDs.

Developing physical exposure-based back injury risk models applicable to manual handling jobs in distribution centers.

Using our ultrasound-based "Moment Monitor," exposures to biomechanical low back disorder risk factors were quantified in 195 volunteers who worked in 50 different distribution center jobs. Low back injury rates, determined from a retrospective examination of each company's Occupational Safety and Health Administration (OSHA) 300 records over the 3-year period immediately prior to data collection, were used to classify each job's back injury risk level. The analyses focused on the factors differentiating the high-risk jobs (those having had 12 or more back injuries/200,000 hr of exposure) from the low-risk jobs (those defined as having no back injuries in the preceding 3 years). Univariate analyses indicated that measures of load moment exposure and force application could distinguish between high (n = 15) and low (n = 15) back injury risk distribution center jobs. A three-factor multiple logistic regression model capable of predicting high-risk jobs with very good sensitivity (87%) and specificity (73%) indicated that risk could be assessed using the mean across the sampled lifts of the peak forward and or lateral bending dynamic load moments that occurred during each lift, the mean of the peak push/pull forces across the sampled lifts, and the mean duration of the non-load exposure periods. A surrogate model, one that does not require the Moment Monitor equipment to assess a job's back injury risk, was identified although with some compromise in model sensitivity relative to the original model.

Biomechanical, psychosocial and individual risk factors predicting low back functional impairment among furniture distribution employees.

BACKGROUND: Biomechanical, psychosocial and individual risk factors for low back disorder have been studied extensively however few researchers have examined all three risk factors. The objective of this was to develop a low back disorder risk model in furniture distribution workers using biomechanical, psychosocial and individual risk factors. METHODS: This was a prospective study with a six month follow-up time. There were 454 subjects at 9 furniture distribution facilities enrolled in the study. Biomechanical exposure was evaluated using the American Conference of Governmental Industrial Hygienists (2001) lifting threshold limit values for low back injury risk. Psychosocial and individual risk factors were evaluated via questionnaires. Low back health functional status was measured using the lumbar motion monitor. Low back disorder cases were defined as a loss of low back functional performance of -0.14 or more. FINDINGS: There were 92 cases of meaningful loss in low back functional performance and 185 non cases. A multivariate logistic regression model included baseline functional performance probability, facility, perceived workload, intermediated reach distance number of exertions above threshold limit values, job tenure manual material handling, and age combined to provide a model sensitivity of 68.5% and specificity of 71.9%. INTERPRETATION: The results of this study indicate which biomechanical, individual and psychosocial risk factors are important as well as how much of each risk factor is too much resulting in increased risk of low back disorder among furniture distribution workers.

Musculoskeletal disorder risk during automotive assembly: current vs. seated.

Musculoskeletal disorder risk was assessed during automotive assembly processes. The risk associated with current assembly processes was compared to using a cantilever chair intervention. Spine loads and normalized shoulder muscle activity were evaluated during assembly in eight regions of the vehicle. Eight interior cabin regions of the vehicle were classified by reach distance, height from vehicle floor and front to back. The cantilever chair intervention tool was most effective in the far reach regions regardless of the height. In the front far reach regions both spine loads and normalized shoulder muscle activity levels were reduced. In the middle and close reach regions spine loads were reduced, however, shoulder muscle activity was not, thus an additional intervention would be necessary to reduce shoulder risk. In the back far reach region, spine loads were not significantly different between the current and cantilever chair conditions. Thus, the effectiveness of the cantilever chair was dependent on the region of the vehicle.

Musculoskeletal disorder risk as a function of vehicle rotation angle during assembly tasks.

Musculoskeletal disorders (MSD) are costly and common problem in automotive manufacturing. The research goal was to quantify MSD exposure as a function of vehicle rotation angle and region during assembly tasks. The study was conducted at the Center for Occupational Health in Automotive Manufacturing (COHAM) Laboratory. Twelve subjects participated in the study. The vehicle was divided into seven regions, (3 interior, 2 underbody and 2 engine regions) representative of work areas during assembly. Three vehicle rotation angles were examined for each region. The standard horizontal assembly condition (0° rotation) was the reference frame. Exposure was assessed on the spine loads and posture, shoulder posture and muscle activity, neck posture and muscle activity as well as wrist posture. In all regions, rotating the vehicle reduced musculoskeletal exposure. In five of the seven regions 45° of vehicle rotation represented the position that reduced MSD exposure most. Two of the seven regions indicated 90° of vehicle rotation had the greatest impact for reducing MSD exposure. This study demonstrated that vehicle rotation shows promise for reducing exposure to risk factors for MDS during automobile assembly tasks.

Quantitative biomechanical workplace exposure measures: distribution centers.

Physical work exposure characteristics assessed in most previous epidemiologic studies have been described mostly in gross categorical terms (e.g. heavy work, lifting and forceful movements, etc.) and have resulted in relatively moderate associations with low back pain risk. We hypothesized that it was necessary to characterize work demands in a much more quantitative fashion so that the precise biomechanically meaningful measures of exposure were available for risk analysis. In this study, we used sophisticated instrumentation to continuously document 390 physical exposures during lifting (in four types of distribution centers) throughout work. This study profiles these exposures and shows how these exposures vary as a function of the type of distribution center and compares the exposures to (previously documented) manufacturing exposures. Static load and load moment measures were found to greatly under-represent true (dynamic) load and load moment exposures to workers. Lift durations averaged 11-12% of the cycle time in distribution environments. This study indicates that distribution workers are commonly exposed to greater extreme loads and move much more rapidly than manufacturing employees. The information provided here can serve as a basis for low back pain risk assessments.

Quantitative dynamic measures of physical exposure predict low back functional impairment.

STUDY DESIGN: Prospective field study of work exposure and changes in back function. OBJECTIVE: Quantify dynamic physical exposures in the workplace and their association with decreases in kinematic back function (indicative of low back pain [LBP]). SUMMARY OF BACKGROUND DATA: Previous epidemiologic studies of work have measured gross categories of exposure and found moderate relationships with LBP. More precise quantitative measures of exposure and spine function were hypothesized to increase the chances of identifying any significant associations. METHODS: Three hundred and ninety real-time physical exposure measures were collected from distribution center workers performing repetitive manual materials handling tasks. Low back health effect measures were quantitatively measured prospectively for workers performing each of the jobs using a kinematic measure of function. RESULTS: Significant decreases in spine function were observed in workers associated with 40% of the jobs sampled. Numerous significant univariate odds ratios were identified that indicated an association between physical exposure and decreased function. A multivariate model including right lateral trunk velocity, timing of the maximum dynamic asymmetric load moment exposure, and the magnitude of the dynamic sagittal bending moment predicted reduced spine function well. The model resulted in excellent sensitivity (85%) and specificity (87.5%) as well as excellent positive predictive value (89.5%) and negative predictive value (82.4%). CONCLUSION: This study suggests that with proper quantification of job exposure and spine function, it is possible to identify which dynamic physical exposures are associated with reduced spine function and increases in LBP.

Instrumentation for measuring dynamic spinal load moment exposures in the workplace.

Prior research has shown the load moment exposure to be one of the strongest predictors of low back disorder risk in manufacturing jobs. However, to extend these finding to the manual lifting and handling of materials in distribution centers, where the layout of the lifting task changes from one lift to the next and the lifts are highly dynamic, would be very challenging without an automated means of quantifying reach distances and item weights. The purpose of this paper is to describe the development and validation of automated instrumentation, the Moment Exposure Tracking System (METS), designed to capture the dynamic load moment exposures and spine postures used in distribution center jobs. This multiphase process started by obtaining baseline data describing the accuracy of existing manual methods for obtaining moment arms during the observation of dynamic lifting for the purposes of benchmarking the automated system. The process continued with the development and calibration of an ultrasonic system to track hand location and the development of load sensing handles that could be used to assess item weights. The final version of the system yielded an average absolute error in the load's moment arm of 4.1cm under the conditions of trunk flexion and load asymmetry. This compares well with the average absolute error of 10.9cm obtained using manual methods of measuring moment arms. With the item mass estimates being within half a kilogram, the instrumentation provides a reliable and valid means for assessing dynamic load moment exposures in dynamic distribution center lifting tasks.

Quantification of a meaningful change in low back functional impairment.

STUDY DESIGN: Repeated measures study design. OBJECTIVE: Determine a meaningful change in low back functional impairment as measured with the lumbar motion monitor. SUMMARY OF BACKGROUND DATA: A quantitative functional performance probability (P(n)) measure has been developed and is scored from 0.00 to 1.00. Previous research has shown that a 0.5 cut-off provides excellent sensitivity and specificity for identifying impaired and healthy low back function. However, a meaningful change in the P(n) measure has not been defined. METHODS: The lumbar motion monitor was used to repeatedly measure P(n) in 3 groups of subjects including (1) asymptomatic, (2) recovering low back pain (LBP) and, (3) nonrecovering LBP. The asymptomatic group had 20 subjects. The recovering and nonrecovering LBP had 18 and 8 subjects, respectively. The asymptomatic group was tested 5 times at 1-week intervals. The 2 LBP groups were tested every 2 weeks for 3 months (6 evaluations). RESULTS: The P(n) in the asymptomatic group did not significantly change over the observed period. On the basis of the variability in the asymptomatic group it was hypothesized that a meaningful change in P(n) was 0.14. The defined meaningful change was evaluated in 2 patient with LBP populations. The P(n) in the recovered LBP group significantly improved during the 3 month observation period and there was a corresponding reduction of symptoms. In the recovering LBP group the within subject standard deviation was 0.14 and all patients had at least 1 visit to visit change greater than 0.14. Furthermore, 11 of the 18 recovering patients with LBP had a meaningful change between the first 2 visits. In contrast, none of the nonrecovering LBP group had a meaningful change between the first 2 visits. CONCLUSION: A meaningful change in P(n) was defined as 0.14.

Loading along the lumbar spine as influence by speed, control, load magnitude, and handle height during pushing.

BACKGROUND: Low back loading and risk associated with pushing activities have been poorly understood. Previous studies have demonstrated that increases in anterior/posterior shear forces are primarily initiated by antagonistic coactivity within the torso. Yet, few studies have considered the range of activities that might contribute to the antagonistic coactivation and subsequent spine loading. METHODS: Twenty subjects were tested to examine how various physical factors might influence spine loads during pushing tasks that workers might experience in industrial settings. Load magnitude, speed of push, required control, and handle height were varied while pushing both carts and overhead suspended loads. A biologically-assisted biomechanical model was used to assess compression, anterior/posterior shear, and lateral shear over the various levels of the lumbar spine. FINDINGS: Anterior/posterior shear loads were greatest at the upper levels of the lumbar spine and of a magnitude that would be of concern. Anterior/posterior shear was influenced by all experimental factors to varying degrees except for the nature of the load (cart vs. suspended). INTERPRETATION: This study confirms the notion that pushing and pulling is not as simple a task as once believed since it entails a complex biomechanical activity. Spine shear forces result from a complex coactivation of trunk muscle activities and spine orientations that are influenced by several occupational factors. This study may help explain why low back pain rates in some work environments associated with lifting may not be reduced even when lifting interventions (that change the task from lifting to pushing) are employed.

Low back pain recurrence in occupational environments.

STUDY DESIGN: Prospective assessment of return to work after low back pain. OBJECTIVE: To determine which factors or combination of factors best predict recurrence of low back pain (defined 4 different ways) when returning to full-duty work. SUMMARY OF BACKGROUND DATA: Recurrent back pain is one of the more costly health problems facing industry today. Few systematic evaluations of the various factors suspected of exacerbating low back pain have been reported in the literature. METHODS: A total of 206 workers who reported low back pain were evaluated as they returned to full-duty work. Five types of assessments were performed including: 1) a low back kinematic functional assessments, 2) evaluation of job physical demands, 3) psychosocial assessment of the job environment, 4) self-reported impairment including perception of symptoms and psychological measures, and 5) personal (individual) factors. One year after return to full duty workers were interviewed to assess who had a recurrence of low back pain according to 4 different definitions of low back pain (symptom reports, medical visits, self-reported lost days, and employer-reported lost days due to back pain). Multiple logistic regression models were developed to assess the best combinations of predictors. RESULTS: The most liberal definition of recurrence, recurrent symptoms, had a significantly greater recurrence rate at 58% than all other outcome measures (P = 0.0001). The medical visit recurrence rate of 36% was significantly greater than the more conservative lost time measures (P = 0.0001). The recurrence rate for self-reported lost time was 15%, whereas the more conservative employer confirmed lost time measure was significantly lower at 10% (P = 0.0077). Multivariate predictive models associated with the various recurrence definitions yielded sensitivities varying between 78% and 80% and specificity between 73% and 80%. CONCLUSION: Recurrence is greatly dependent on how one defines recurrence with symptom reporting yielding 5.5 times as many recurrences compared with employer confirmed lost time. In general, more quantitative measures of worker musculoskeletal function yielded the best predictions of recurrence when predicting the more restrictive definitions of recurrence (employer confirmed lost time).