Musich, Shirley PhD; Napier, Deborah MS; Edington, D. W. PhD
In1999, the National Safety Council estimated the total annual cost ofoccupational injuries at $125 billion: $62.0 billion for wage and productivitylosses, $19.9 billion in medical costs, $25.6 billion in administrativeexpenses, and $16.7 billion in additional employercosts. 1 An estimated $42.4 billionwas paid out under workers’ compensation (WC) insurance; the averagecost for all claims combined was $10,488 per injuredworker. 1 With widespread industrysafety programs, the incidence rates for occupational injuries (excludingfatal work-related injuries) have steadily declined from 8.3 cases per 100workers in 1990 to 6.2 cases per 100 workers in1998. 1,2
Inkeeping with these national trends, and in an attempt to make significantimprovements in quality of life for all US working people, Healthy People 2010established a goal of further reducing work-related injuries to 4.6 injuriesper 100 full-time workers (a 30% improvement over 1997 baseline levels of 6.6injuries per 100workers). 3
Comprehensiveinitiatives to manage the incidence and costs of occupational injuries ofteninclude medical case management (secondary prevention), safety/ergonomicprograms, and early return-to-workprograms. 4,5 Cost-containment strategies also used by corporations to control increases inWC costs include utilization review/managementprograms, 6,7 riskmanagement programs, 8 andrehabilitation programs for injuredworkers. 9 These programs focus oninjury prevention through worksite review and assessment of worker risk forinjury before injuries occur, and management of medical costs, rehabilitation,and return-to-work programs after an injury incident. Most programs showhighly favorable cost savings and reductions in the incidence rates forinjuries. 6–9
Thebenefits of primary prevention by improving worker strength and/or health havealso beeninvestigated. 10–13 In one of the earliest studies (1979), Cady etal 14 used five measures of fitnessand conditioning to classify firefighters into least-fit, middle-fit, andmost-fit categories. Their results showed a graded and statisticallysignificant protective effect for added levels of fitness and conditioning forback injuries (least-fit, 7%; middle-fit 3%; and most-fit 1%). It wasconcluded that physical fitness and conditioning were preventive for backinjuries in this population.
Since that study waspublished, other investigators have considered individual health risks thatwould predict the incidence of injury, including isometricstrength, 15 aerobicfitness, 16 cardiovascularrisks, 16 obesity, 17 smoking, 16,17 andpsychosocialvariables. 18–22 Results have varied by health risk. Isometric strength and cardiovascularrisks do not effectively predictinjury. 15,16 Only a fewstudies have reported an association with obesity, and then only among themost obese. 16,17 Although enhanced physical fitness has often been suggested to have aprophylactic effect on injury rate, the type of training required and thecritical aspects of fitness areunclear. 10,16,23 There is a consistent association between those who smoke and increased injuryrates, although the mechanism of this relationship needs furtherinvestigation. 16,17 Among psychosocial variables, job dissatisfaction, work-related stress, andlack of social support have been associated with higher injury rates. Thecomplex relationship, however, between psychosocial variables and the physicaldemands of work have made it difficult to reach definitive conclusions abouttheir relative importance to the risk ofinjury. 18–22
Primaryprevention programs focused on injury prevention by improvement of workerhealth/strength are less prevalent, and programs often include multifacetedapproaches. In an example of a back injury intervention among countyemployees, the program consisted of education, training, physical fitnessactivities, and ergonomic improvements. Savings in medical costs and reducedsick days resulted in a 2.79:1return-on-investment. 24 A healthrisk appraisal (HRA) assessed the likelihood for back injury with additionalquestions. After the intervention, there was a significant decrease in thepercentage of those employees at high risk for back injury. Overall healthstatus was not reported.
Another company initiated anintensive wellness program with incentives to join fitness centers andparticipate in education programs and health assessments (HRA). The number ofinjures and lost workdays steadily decreased over the 6 years of the program.The HRA was used to track improvements in fitness, nutrition, and cholesterollevels. Calculations of cost savings related to the program yielded areturn-on-investment of2.51:1. 25
Healthstatus has been related to medical costs. High-risk individuals have beenshown to have higher medical costs than low-riskindividuals. 26–35 Furthermore, people who change their risk status by improving their lifestylebehaviors have been shown to reduce theircosts. 36,37 It was ofinterest to know if these same relationships could also be applied to WCcosts.
The purpose of this study was to investigatewhether the associations between health risks, as defined from the HRA, and WCcosts would be similar to those established for medical care costs.Specifically, the following were examined: (1) effect of individual healthrisks and overall health status on WC costs, (2) percentage of excess WC costsassociated with excess risks/non-participation, and (3) cost savingsassociated with programparticipation.
Thisproject was a 4-year case study (1996 to 1999) of the associations betweenhealth risks as measured on the HRA and WC costs and lost injury days amonglong-term employees at Xerox Corporation. The corporation initiated the XeroxHealth Management program in 1978 with the first round of HRAs in 1981. In1995, a more intensive program was instituted at the Rochester, New York,location, with health risks assessed every 3 years (1995 and 1998). Themission of the Xerox Health Management program is to provide integrated healthprograms that optimize health and personal productivity. HRA participantsreceived health risk-targeted awareness materials and the opportunity toparticipate in special, subsidized, on-site lifestyle management programs (eg,weight management, fitness, back care,stress).
Thestudy population consisted of 3338 employees who met the following criteriafor this 1996 to 1999 study: (1) continuously employed by Xerox Corporationfrom 1981 to 1999, and (2) selected Blue Choice Health MaintenanceOrganization as their medical plan. This pilot employee subgroup had beenpreviously selected by the corporation to track short- and long-term healthand economic (medical cost) effects of their health promotion program.Demographics from personnel records, including age, gender, and job status(exempt, hourly, non-exempt), characterized those with WC claims during the4-year period (n = 265) andthose without claims (n =3073).
WC Claims and Lost InjuryDays
Data for WC paid claims and days lost from injury from1996 to 1999 were available from the corporation benefits department. Data forclaims and lost injury days were received on a per-claim basis and thenaggregated to determine the total WC cost and lost injury days per employeeper year. Costs for each year were adjusted to 1999 dollars using publishedmedical inflationrates. 38
Lostinjury days were converted to a cost measure using an average daily wage perday of $150 (assigned by the corporation benefits department) multiplied bythe number of lost days for each employee. A combined outcome measure used inthis study summed the total costs from claims and lost days costs for the4-year period. This approach thus incorporates both direct costs (ie,insurance payments) and indirect costs (ie, lost work time) into theanalyses.
The HRAwas originally developed by the Centers for Disease Control/Carter Center andwas modified by the University of Michigan Health Management Research Centerfor the employer. In addition to self-reported age and sex, 15 health riskfactors were selected to establish healthstatus 36 :
* Fivelifestyle variables: smoking, physical activity level, alcohol use,drug/medication use, and safety beltuse.
* Four psychologicalvariables: perception of physical health, life satisfaction, job satisfaction,and stress.
* Six health/biologicalvariables: blood pressure, cholesterol, relative body weight, serious medicalproblems, illness absence days, and Health Age Index (a measure ofcontrollable health risks).
High-riskcriteria for each of the defined health risks are given in Table 1. Individualhealth risks for HRA participants were summed, and overall health status wasassessed as low-risk (0 to 2 health risks, 50th percentile); medium-risk (3 to4 health risks, between the 50th and 90th percentiles); and high-risk (5 ormore health risks, 90th percentile and above).
WCCosts by Individual Risk Status and by Overall HealthStatus
WC claims, costs associated with lost injury days,and total WC costs were assessed by individual risk status (high-risk vslow-risk) for each of the 15 individual health risks as reported on the 1998HRA (n = 943). Each individualhealth risk was assessed individually without adjustment for other healthrisks. (HRA participants without WC claims or lost workday costs were assigned$0.)
Frequency of claims, WC claims, costs associatedwith lost injury days, and total WC costs were also assessed by overall healthstatus (1998 HRA: low-risk, medium-risk and high-risk) and fornon-participants among those with claims over the time period(n =265).
Excess WC Costs Related to ExcessHealth Risks
Total WC costs were calculated for low-risk,medium-risk, and high-risk employees and for non-participants (1998 HRAparticipants n = 943;non-participants n = 2395).Excess WC costs related to excess health risks and for non-participation weredefined as excess WC costs greater than the base cost of the employees having0 to 2 health risks (low-risk status). The percentage of total costsattributable to excess risks/non-participation was thencalculated.
WC Cost Savings AssociatedWith HRA Participation
Participation status during the timeperiod was also investigated for an impact on WC costs. HRA participationstatus, including the 1995 and 1998 HRAs, was used in defining the followingparticipation groups: two-time HRA participants(n = 515); one-time HRAparticipants (either 1995 or 1998, n = 823); or HRA non-participants(n = 2000). The participationgroups were also considered among those with claims: two-time HRA participants(n = 28), one-time HRAparticipants (n = 46), and HRAnon-participants (n =191).
Cost trends among those with claims over the4-year period were compared for the three participant groups. Total costsavings were then calculated from the differences in the slopelines.
Categorical variables were statistically testedusing the chi-squared test. Comparisons of continuous variables were testedusing the analysis of variance procedure with post-hoc Tukey’sStudentized range test for multilevel comparisons. Because WC costs werehighly skewed, a log transformation was performed before statistical testingof cost variables. Slope trends over time were tested using regressionanalyses testing for significance of independent variable interactions (HRAparticipation vstime).
Duringthe 1996-to-1999 study period, 265 (7.9%) employees incurred WC claims. Amongthose with claims, 26% also had lost injury days. Compared with employeeswithout WC claims, employees with claims were more likely to be female,hourly, and younger (53.2 years vs 54.8 years)(Table 2).
The distribution of WC claims and lost injury days(and the summed cost measure) were highly skewed. The median cost for thesummed cost measure for the 4 years, shown in Fig. 1, was $527;the mean cost was $8887. The top 10th percentile of employees with costsaccounted for 54.4% of the total of WC costs. Reflecting national trends andimplementation of a corporate injury policy, the annual percentage ofemployees with claims steadily declined from 2.7 injuries per 100 workers in1996 to 2.0 injuries per 100 workers in 1999.
WCCosts by Health Status and by Individual Health Risks
Therewere 943 HRA participants in 1998 (28.3% participation rate). The individualhealth risks that were most highly related to high WC costs (summed measure)were Health Age Index greater than 4 years, smoking, poor perception ofphysical health, low physical activity level, and life dissatisfaction(P < 0.05)(Table 3). Althoughindividual risk status and overall risk status as assessed by the 1995 HRAwere also tested, the risk-cost relationships were not as strong. Concurrentrisk status (1998 HRA) was more highly associated with WC costs (1996 to 1999)than past risk status (1995 HRA) with future costs.
Thepercentage of employees with WC claims increased with increased risk status(1998 HRA): 4.9% among low-risk, 5.4% among medium-risk, and 8.2% amonghigh-risk employees (P = 0.26)(Table 4). Overall,HRA participants had a significantly lower percentage (5.6%) with claimscompared with non-participants (8.9%, P = 0.002). WC claims, costsassociated with lost workdays, and total WC costs increased with increasingrisk status. Total WC costs increased from $2178 per person among low-riskemployees to $15,162 per person among high-risk employees. Overall,participants had lower total WC costs ($6506) compared with non-participants($9482, P < 0.001). Multilevelpost-hoc testing indicated that non-participants’ costs weresignificantly greater than low-risk (P < 0.10) and medium-risk (P <0.15) participants’ costs.
ExcessWC Costs Related to Excess Health Risks
Among 1998 HRAparticipants and non-participants, the total WC costs (claims + lostinjury day costs) were $2,354,044. Low-risk employees had the lowest averagecost ($106: $105 claims + $0.60 lost injury day costs) compared withmedium-risk ($288: $191 claims + $97 lost injury day costs), high-risk($1,241: $729 claims + $512 lost day costs), and non-participants ($839:$541 claims + $298 lost day costs)(Fig. 2). If thosecosts above the low-risk baseline cost are defined to be “excess”costs, the total excess WC costs for medium- and high-risk participants andfor non-participants would be $2,000,494 (see calculations under Fig. 2). The percentage of total costs associated with excess risks/non-participation is 85%.
Topreclude an influential effect from extreme values on our conclusions, theeffects of outliers were investigated. There were two extreme observations(one greater than $100,000 and one greater than $200,000). The calculationswere repeated removing the most extreme observation (greater than $200,000),and then both observations. Both of these observations were categorized asnon-participants, and neither observation was unduly influential in the excesscost conclusions. Thus, no observations were excluded from ourcalculations.
WC Cost Savings AssociatedWith HRA Participation
Among those employees with claims,the slopes of the two HRA participant groups were statistically tested fordifferences. Because they were not statistically different(P > 0.20), these two groups werecombined into one group of HRA participants. The cost trend for this group wasthen compared with the cost trend for the non-participants. The slope of thecost trend among HRA participants (combined) decreased at an annual rate of$1352 per year from the 1996 baseline, whereas the slope amongnon-participants decreased at a rate of $114 per year(P < 0.02)(Fig. 3). Theincidence rate of injury claims among HRA participants was also significantlylower compared with the rate among non-participants (5.5% vs 9.6%, P < 0.001). The total savings associated with the difference in the slope lines for participation amongthose with claims was $366,448 for 74 participants, or $1238 per person per year (see Fig. 3 for calculations).
Consistentwith previousstudies, 39–41 the distribution of WC costs was highly skewed, with the top 10th percentileof costs accounting for 54.4% of total costs. Over the 4-year period, 7.9% ofemployees in the study group filed WC claims. Most (87.1%) filed single-yearclaims, although 12.1% had 2 years of claims and 0.8% had claims for all 3years. Among those filing claims, 74% received WC without loss of work time.The nature of WC costs was similar to medical care costs, highly skewed with asmall number of individuals accounting for a large percentage of the overallcosts.
The annual trend for the number of injuries per100 workers decreased steadily from 2.7 in 1996 to 2.0 in 1999. A Zero-Injuryprogram initiated in 1997 at Xerox perhaps accounted for some of thereduction. The study population is a long-term subgroup of the Xerox employeepopulation, with an average age of over 50 years. A decreasing injury trendmay not be surprising in a middle-aged group of employees; the trend wasconsistent with national trends in WCcosts. 1,2,41 Xerox Corporation has met and exceeded Healthy People 2010 goals of 4.6injuries per 100 workers and improved their injury rate by 25.9% during the4-year period, 1996 to 1999, in this subgroup ofemployees.
Similar to medical care costs, WC costs arerelated to individual health risks measured by the HRA, although the specifichealth risks most highly associated with high medical costs differ from thoseassociated with high WC costs. Unlike the relationship with medical costs,however, concurrent health risk status was more highly associated with WCcosts than previous risk status with future costs. The individual risks mosthighly related to high WC costs included Health Age Index (a difference inappraised age and achievable age from the HRA), smoking, poor physical health,physical inactivity, and life dissatisfaction. Other studies have reported therelationship of higher injuries with the individual health risks ofsmoking, 16,17 physicalinactivity, 16 and psychosocialvariables. 18–22 In contrast, the individual risks most highly related to high medical costsinclude absence due to illness, drug/medication use, medical problems, highblood pressure, and poor physicalhealth. 29,35,36
Thisstudy is unique in assessing the association of injury claims with riskstatus: low-risk (0 to 2 risks), medium-risk (3 to 4 risks), and high-risk (5or more risks). As health risk status increased, the amount of WC costs(claims and lost injury day costs) also increased. The rate of injuryoccurrence also increased as risk status increased: 4.9% among low-riskemployees, 5.4% among medium-risk employees, and 8.2% among high-riskemployees. Non-participants, however, had an 8.9% injury rate, which wassignificantly higher than participants. Although the rates of injury amongrisk status levels were not statistically different(P = 0.26), the increasinginjury trends with increasing risk status were clear. Non-participant rates ofinjury were significantly greater than participant rates(P < 0.002). The power of theseassociations for health promotion managers is that WC costs and lost injurydays are influenced by specific health behaviors and participationstatus.
In the assessment of excess WC costs associatedwith excess risks, we considered low-risk as the baseline; any risks abovethis baseline level were considered “excess risks.” The totalexcess WC costs for medium- and high-risk employees and for non-participantsaccounted for 85% of the total costs. Theoretically, if all participantschanged to low-risk status and WC costs followed this change in risk status,this would be the maximum savings. In other unpublished analyses that usemedical costs to calculate excess costs associated with excessrisks/non-participation, excess costs typically account for about 30% of totalcosts. This indicates that health risks/behaviors associated with medicalcosts even more highly influence WC claims and that improving the health ofthe corporate workforce could result in substantial savings in WC claims andlost workdays.
Our data indicated that changes in WCcosts and lost injury days were also associated with changes in health riskstatus. This risk–cost relationship has been well documented for medicalcare costs. 36 The data were notshown here because the number of injury claims among two-time HRA participants(n = 515) was very low(n = 28). Nevertheless, theassociation was consistent with trends observed with medical costs: as riskstatus increased from low-risk to high-risk, costs/lost days increased, and asrisk status decreased from high-risk to low-risk, costs/lost daysdecreased.
HRA program participation was associated witha higher annual rate of decrease of WC costs compared with non-participants.The injury incidence rate among participants was also significantly lowercompared with that of non-participants. In the Xerox Health Managementprogram, the HRA program serves as a gateway to health awareness materials andlifestyle management (risk reduction) programs; hence the measurement of HRAparticipation is a surrogate measure that includes the opportunity toparticipate in other programs. Cost savings associated with programparticipation amounted to $366,448 for 74 employees, or $1238 per person peryear among those with claims over the time period. These results indicate theimportance of encouraging employee participation and then measuringparticipation as an economic outcomemeasure.
Thestudy group is a subgroup of long-term employees selected by XeroxCorporation. The injury trends may not be generalizable to the entire employeepopulation. Given an average age over 50, both the nature of job descriptionsand attitudes toward safety may be more conservative than those among youngeremployees. Thus, the savings associated with low-risk status and programparticipation may be overstated. Nevertheless, the study reflects the impacton at least one group of employees within thecorporation.
The changes in WC costs with changes inhealth status were not included because of the low injury incidence rate amongtwo-time HRA participants. Documenting the changes in costs relative tochanges in health status in a larger employee population would strengthen thebenefits of risk reduction and low-risk maintenance. Nevertheless, we showedthat low-risk employees have the lowest WCcosts.
Theassociation of health risks with WC costs provides an important strategy forhealth promotion programs. Risk status does influence WC costs. Programparticipation status is also associated with cost savings. The strategies thathave been used to reduce medical care costs can also be used to reduce WCcosts.
1. NationalSafety Council. Injury Facts, 1999 Edition. Itaska, IL: NSC; 1999.
3. USDepartment of Health and Human Services. HealthyPeople 2010 (Conference Edition in Two Volumes). Washington, DC:USDHHS; Jan2000.
4. McGrail MP, Tsai SP, Bernacki EJ. A comprehensive initiative to manage the incidence and cost ofoccupational injury and illness. J Occup Environ Med. 1995; 37: 1263 –1268.
5. Green-McKenzie J, Parkerson J, Bernacki E. Comparison of workers’ compensation costs for two cohorts ofinjured workers before and after the introduction of managed care. J Occup Environ Med. 1998; 40: 568 –572.
6. Calhoun RJ. Workers’ compensation payers turn to cost-containmentstrategies. Occup Health Saf. 1990; 59: 84 –87.
7. Wickizer TM, Lessler D, Franklin G. Controlling workers’ compensation medical care use and coststhrough utilization management. J Occup Environ Med. 1999; 41: 625 –631.
8. Melhorn JM, Wilkinson L, Gardner P, Horst WD, Silkey B. An outcomes study of an occupational medicine intervention programfor reduction of musculoskeletal disorders and cumulative trauma disorders inthe workplace. J Occup Environ Med. 1999; 41: 833 –846.
9. Caradoc-Davies TH, Wilson BD, Anson JG. The cost benefit of rehabilitation of injured workers in NewZealand. N Z Med J. 1991; 104: 245 –247.
10. Frank JW, Kerr M, Brooker A, et al. Disability resulting from occupational low back pain. PartI: What do we know about primary prevention? Spine. 1996; 21: 2908 –2917.
11. Gundewall B, Liljeqvist M, Hansson T. Primary prevention of back symptoms and absence fromwork. Spine. 1993; 18: 587 –594.
12. Feuerstein M, Berkowitz SM, Huang GD. Predictors of occupational low back disability: implications for secondary prevention. J Occup Environ Med. 1999; 41: 1024 –1031.
13. Forrester BG, Weaver MT, Brown KC, Phillips JA, Hilyer JC. Personal health risk predictors of occupational injury among 3415municipal employees. J Occup Environ Med. 1996; 38: 515 –521.
14. Cady LD, Bischoff DP, O’Connell ER, Thomas PC, Allan JH. Strength and fitness and subsequent back injuries infirefighters. J Occup Med. 1979; 21: 269 –272.
15. Battie MC, Bigos SJ, Fisher LD, Hansson TH, Jones ME, Wortley MD. Isometric lifting as a predictor of industrial back painreports. Spine. 1989; 14: 851 –856.
16. Battie MC, Bigos SJ, Fisher LD, et al. A prospective study of the role of cardiovascular risk factors andfitness in industrial back pain complaints. Spine. 1989; 14: 141 –147.
17. Deyo RA, Bass JE. Lifestyle and low-back pain. The influence of smoking andobesity. Spine. 1989; 14: 501 –506.
18. Bigos SJ, Battie MC, Spengler DM, et al. A prospective study of work perceptions and psychological factorsaffecting the report of back injury. Spine. 1991; 16: 1 –6.
19. Lancourt J, Kettelhut M. Predicting return to work for lower back pain patients receivingworker’s compensation. Spine. 1992; 17: 629 –640.
20. Dempsey PG, Burdorf A, Webster BS. The influence of personal variables on work-related low-backdisorders and implications for future research. J Occup Environ Med. 1997; 39: 748 –759.
21. Bongers PM, de Winter CR, Kompier MAJ, Hildebrandt VH. Psychosocial factors at work and musculoskeletaldisease. Scand J Work Environ Health. 1993; 19: 297 –312.
22. Niemcryk SJ, Jenkins D, Rose RM, Hurst MW. The prospective impact of psychosocial variables on rates ofillness and injury in professional employees. J Occup Med. 1987; 29: 645 –652.
23. Ready AE, Boreskie SL, Law SA, Russell R. Fitness and lifestyle parameters fail to predict back injuries innurses. Can J Appl Physiol. 1993; 18: 80 –90.
24. Shi L. A cost-benefit analysis of a California county’s back injuryprevention program. Public Health Rep. 1993; 108: 204 –211.
25. Maniscalco P, Lane R, Welke M, Mitchell JH, Husting L. Decreased rate of back injuries through a wellness program foroffshore petroleum employee. J Occup Environ Med. 1999; 41: 813 –820.
26. Bly JL, Jones RC, Richardson JE. Impact of worksite health promotion on health care costs andutilization: evaluation of Johnson & Johnson’s Live for Lifeprogram. JAMA. 1986; 256: 3235 –3240.
27. Bowne DW, Russell ML, Morgan JL, Optenberg SA, Clarke AE. Reduced disability and health care costs in an industrial fitnessprogram. J Occup Med. 1984; 26: 809 –816.
28. Erfurt JC, Foote A, Heirich MA. The cost-effectiveness of work-site wellness programs forhypertension control, weight loss and smoking cessation. J Occup Med. 1991; 33: 962 –970.
29. Foote A, Erfurt JC. The benefit to cost ratio of work-site blood pressure controlprograms. JAMA. 1991; 265: 1283 –1286.
30. Yen LT, Edington DW, Witting P. Associations between health risk appraisal scores and employeemedical claims costs in a manufacturing company. Am J Health Promot. 1991; 6: 46 –54.
31. Yen LT, Edington DW, Witting P. Prediction of prospective medical claims and absenteeism costs for1284 hourly workers from a manufacturing company. J Occup Med. 1992; 34: 428 –435.
32. Leutzinger J, Hawes C, Hunnicutt D, Richling D. Predicting the ratio of benefit to cost in a cardiovasculardisease-prevention program. Manage Employee Health Benef. 1995; 3: 1 –10.
33. Goetzel RZ, Jacobson BH, Aldana SG, Vardell K, Yee L. Health care costs of worksite health promotion participants andnon-participants. J Occup Environ Med. 1998; 40: 341 –346.
34. Goetzel RZ, Anderson DR, Whitmer RW, et al. The relationship between modifiable health risks and health careexpenditures. J Occup Environ Med. 1998; 40: 843 –854.
35. Ozminkowski RJ, Dunn RL, Goetzel RZ, Cantor RI, Murnane J, Harrison M. A return on investment evaluation of the Citibank, N.A. Am J Health Promot. 1999; 14: 31 –43.
36. Edington DW, Yen LT, Witting P. The financial impact of changes in personal healthpractices. J Occup Environ Med. 1997; 39: 1037 –1046.
37. Musich SA, Adams L, Edington DW. Effectiveness of health promotion programs in moderating medicalcosts in the USA. Health Promot Int. 2000; 15: 5 –15.
38. USBureau of Labor Statistics. Trends in medical care costs. Stat Bull Metrop Insur Co. 1999; 80 (4): 32.
39. Hashemi L, Webster BS, Clancy EA, Volinn E. Length of disability and cost of workers’ compensation lowback pain claims. J Occup Environ Med. 1997; 39: 937 –945.
40. Hashemi L, Webster BS, Clancy EA, Courtney TK. Length of disability and cost of work-related musculoskeletaldisorders of the upper extremity. J Occup Environ Med. 1998; 40: 261 –269.
41. Hashemi L, Webster BS, Clancy EA. Trends in disability duration and cost of workers’compensation low back pain claims (1988–1996). J Occup Environ Med. 1998; 40: 1110 –1119.
© 2001 Lippincott Williams & Wilkins, Inc.