Upper Limb and Low Back Mucsuloskeletal Disorders: State and National Estimates based on Workers' Compensation Accepted Claims

Barbara Silverstein (A), Elizabeth Grossman (B),
John Kalat (A), Nancy Nelson (A)

(A) Safety & Health Assessment and Research for Prevention (SHARP) Program
Department of Labor & Industries
PO Box 44330
Olympia WA 98504-4330

(B)* Occupational Safety and Health Administration
US Department of Labor
200 Constitution Avenue NW
Washington DC 20210

ABSTRACT

Uniform national occupational health surveillance systems for work-related musculoskeletal disorders (WMSDs) are lacking. The BLS survey most approximates a national occupational health surveillance system but has not had a category that encompasses all work-related musculoskeletal disorders. Worker's compensation data tend to underestimate the magnitude of occupational illnesses more than injuries, but there is evidence of underestimation of both (Silverstein et al, 1997, Fine et al, 1986). Additionally, workers compensation systems differ between states in terms of exclusions, who is covered, type of insurance provider, number of time loss days before a claim is compensable, lack of denominators to calculate standardized incidence rates, and classification. Most states classify their occupational injuries and illnesses according to ANSI z16.3, ICD-9 or some similar classification system. This usually includes the type, nature and source of injury. A combination of these codes can be used to isolate "work-related musculoskeletal disorders." This data can be used to estimate what can be considered the lower bounds of work-related musculoskeletal disorders.

METHODS

We used Washington State Fund accepted claims data to estimate the incidence of upper limb and back disorders because Washington has the most complete data available including 1) claims resulting in medical bills only as well as those resulting in indemnity payments (lost time) and 2) has denominator data for calculating incidence rates (payroll hours can be translated into full-time equivalent employees: 2,000 hours =1 FTE). Combinations of ANSI z16.2 codes were used to identify work-related musculoskeletal disorders. We combined nature AND type AND body part. Nature codes included (190) dislocation/pinched nerve/sciatica (with body part of back or neck); (260) inflammation/irritation to joints, tendons, muscles, bursa (tendonitis, tenosynovitis, bursitis); (310) sprains and strains; (562) diseases of peripheral nerves and ganglia (CTS, ganglionic cysts). These were combined with Type codes and included (081) leaning, kneeling, sitting (for back and lower extremity claims); (082) rubbed or abraded by objects being handled, (083) rubbed or abraded by vibrating objects; (085) repetition of pressure; (086) repetitive motion; (100) bodily reaction; (120) overexertion; (121) lifting objects; (122) pulling or pushing objects); (123) wielding or throwing objects; (124) carrying objects; (129) overexertion NEC). The body parts for the upper limb included neck to fingers (codes 200-398 and 450) and the back (code 420). Data was originally collected for 1988-1992. It was supplemented with data for 1993-1995.

Additionally, we used specific ICD-9 and CPT procedure codes on medical bills to identify 3 diagnostic conditions: rotator cuff syndrome (ICD:7261, 72611,8404; CPT:23410, 23412,23415,23420 ), epicondylitis (ICD 72631, 72632, CPT: 24350); carpal tunnel syndrome (ICD 354.0, CPT:64721). Average and median costs for these claims were calculated for claims in which the z16.2 relevant body part was identified to minimize misclassification of costs associated with other conditions the claimant might have had. The same was done with average and median lost days per compensable claim. These were compared to the more generic ANSI z16.2 definition by relevant body part. Data from 1987-1995 claims were included in this analysis.

To assess the representativeness of Washington State Fund data for national comparisons, data from the BLS Supplemental Data System (1985-1988) for the private sector was used. A different number of states participated in each of the years and supplied a sample of records to the SDS, thus a weighted sample was used. We used estimates of employment from the US Department of Commerce, Bureau of the Census County Business Pattern data as denominators to calculate rates because it included the private sector. Employment is based on the mid March payroll. For 1987, state employment was reported in County Business Patterns in thousands of workers. Cases were similar to those used in the z16.2 data used by Washington State. States were rank ordered for each year and upper limb and back incidence rate for compensable cases.

RESULTS

Table 1 presents estimates of upper limb and back disorders per 1,000 FTEs. In general, both upper limb and back disorders increased in incidence gradually from 1988 to 1992 and then began a gradual decrease. Back disorder rates are roughly double those of the upper limb. Back disorder incidence rates have been gradually decreasing since 1990. In part, this decrease is reflective of overall claims rates which have decreased at a more rapid pace than work-related musculoskeletal disorders, Figure 1. At the same time, the overall claims rejection rate increased from roughly 11% in 1992 to 12.5% in 1995. The apparent decrease in claims rates may reflect more of an administrative change than actual change in underlying incidence rates.

Figure 2 shows the incidence rates for upper limb musculoskeletal injuries (sudden and gradual onset) and WMSDs (gradual onset only) by body area. With the exception of hand wrist injuries (gradual and sudden onset combined) the slopes for the other conditions are relatively flat.

Table 2 shows the demographic characteristics of claimants as well as the direct medical and indemnity costs and the average time loss per compensable claim. Claimants with carpal tunnel syndrome as well as all hand/wrist WMSDs tended to be female more than for shoulder and elbow disorder claimants. The percent of claimants with carpal tunnel syndrome who were female increased significantly over time. The increase was greater than the estimated percent increase in the female working population in Washington State based on census data. Although there was also an increase in the proportion of female claimants for epicondylitis, this was not significant. The proportion of female claimants with rotator cuff was flat over the time period. Median age of claimants did not increase significantly over the time period.

Both claim cost and lost time days for specific diagnoses are limited to those cases which also had a z16.2 body part code appropriate for the diagnosis (ICD9). This was done to reduce possible inflated costs due to multiple other problems that a claimant may have had. In general, higher costs and greater time loss are associated more with specific diagnoses than with overall WMSD categories. The percent of claims that were compensable were greater for specific diagnoses as well. For all upper limb disorders in Table 2, the percent compensable, average time loss for compensable claims and average costs per claim decreased substantially over time. For example, the average cost of a compensable hand/wrist disorder went from $7,283 in 1987 to $5,422 and the cost of a compensable carpal tunnel syndrome case went from $15,066 to $11,615. Median costs did not change as much. Median lost time days for carpal tunnel syndrome decreased from 104 in 1987 to 82 in 1995. Rotator cuff syndrome was the most costly of the diagnoses based claims. Median costs per compensable claim increased from $3,570 in 1987 to $9,410 in 1992 to $6,462 in 1995.

Based on the analysis of BLS Supplemental Data System data, states were rank ordered based on incidence rates. In 1985, Washington ranked 19th out of 23 for upper limb WMSDs and 20th for back WMSDs. In 1986, it ranked 18th and 19th respectively out of 22 states, in 1987 it ranked 18th and 18th out of 24 states, and in 1988, it did not participate in the 14 state sample.

CONCLUSIONS

Washington State data provides a good assessment of overall risk of work-related musculoskeletal disorders because 1) all claims are included (including approximately 61% which are "medical only" claims, 2) all accepted claims are included (minimizes uncertainty about work-relatedness), 3) actual denominator data associated with the actual population at risk was available (payroll hours rather than census or per dollars of premium paid). Compared to other states with comparable reporting criteria for compensability (4 days), Washington State rates were similar to others in the SDS data. Because Washington allows firms to self-insure, it might be reasonable to assume that the State Fund insures the more hazardous employers in the state, leading to overestimation of the magnitude of WMSDs. However, the rates at which claims are filed with the State Fund and self-insured firms are similar. In 1988, the rate of claims processed by the State Fund was 153.4 per 1,000 person years. By 1992, that rate had declined to 136.6 per 1, 000 person years. The self-insured rate increased from 112 to 133.7 per 1,000 person years in the same time.

The true risk of WMSDs due to workplace risk factors is probably higher than that estimated by Washington State workers compensation data. However, this data provides a reasonable estimate for the country as a whole. Thus the lower bounds on national estimates would range from 9.7-11.4 per 1,000 person years for upper limb WMSDs and 15.3-21.0 per 1,000 person years for back WMSDs.

DISCLAIMER. The opinions expressed in this paper do not represent those of the Occupational Safety and Health Administration.

REFERENCES

Silverstein B , et al. Work-related musculoskeletal disorders: comparison of data sources for surveillance. AJIM 31:600-608 (1997)
Fine LJ, et al. Detection of Upper Extremity CTDs. JOM 28 (8):674-680 (1986)

Table 1. Number and rate of work-related musculoskeletal disorders per 1,000 person years in Washington State, Washington State Fund accepted claims

Table 2. Washington State Fund workers compensation claims for specific body areas and diagnoses, 1987-1995.