Prioritizing Back Injury Risk in Hospital Employees: Application and ...

JOEM

l

Volume 42, Number 6, June 2000

645

Prioritizing Back Injury Risk in Hospital Employees: Application and Comparison of Different Injury Rates Rose H. Goldman, MD, MPH Michael R. Jarrard, MD, MPH Rokho Kim, MD, DrPH Susan Loomis, MSN, RN Elisha H. Atkins, MD, MPH

B

To ident@ high risk areas for back injury in a large teaching hospital, we calculated standard injury rates and newly developed composite statistics for nursing and non-nursing work groups. Data were extracted from the hospital’s workers’ compensation database. The hospital-wide total injury rate was 4.6 reports per 100 full-time equivalents (FTE); Compensation Case Rate, 1.4 cases per 100 IT?; Compensation Severity Rate, 76 days lost per 100 FTE; and the Cost Rate, $3742 per 100 FTE. The Total Injuy Reports Rate for nursing varied from 14.2 per 100 FTE for Intensive Care Unit (ICU) Nursing to 3.8 per 100 FTE for Pediatric Nursing. Non-nursing areas also demonstrated increased rates for back injury. Individual statistical rates ranked areas differently in risk, whereas composite statistical measures consistently ranked ICU Nursing, Buildings and Grounds, and Orthopedics/Neurolo@cal Nursing as the top three. Patient handling was the precipitating event in the majority of nursing back injuries, indicating the need for ergonomic intervention. The use of combined statistical measures provided a more integrative measure for describing and following back injury risk over time.

From the Harvard School of Public Health, Harvard Medical School (Dr Goldman); Burlington Northern Santa Fe Railway, Fort Worth, Texas (Dr Jarrad); the Department of Environmental Health, Seoul National University School of Public Health, Seoul, Korea (Dr Kim); Partners Corporate Manager of Occupational Health Services, Boston (MS Loomis); and the Occupational Health Service, Massachusetts General Hospital (Dr Atkins). Address correspondence to: Dr Rose Goldman, Occupational and Environmental Health Center, Cambridge Hospital, 1493 Cambridge Street, Cambridge, MA 02139. Copyright 0 by American College of Occupational and Environmental Medicine

ack injuries account for about one fourth of workers’ compensation claims,’ and total costs have been calculated to be between 10 to 25 billion dollars per year in the United States.2-4 Health care workers, particularly nurses, have been identified as one of the highest-risk groups. In the Occupational Health Supplement to the 1988 National Health Interview Survey, four health care occupational groups were listed among the top 15 high-risk occupations for back pain in female employees, on the basis of self-reported back pain and national estimates of workers at high risk.5 Nursing aides, orderlies, and attendants were the highest-risk group (prevalence 18.8%), and licensed practical nurses were second (prevalence 16.3%). Additional studies have also confirmed the increased prevalence of back injuries among nurses.6’7 Other hospital workers, such as building and grounds and maintenance workers, have also been reported to be at risk for back injuries.s A large teaching hospital in Boston noted similar problems with back injuries, which accounted for 30% of its direct workers’ compensation costs. To identify high-risk areas for an initial intervention program, we calculated various injury rates, including newly developed composite statistics for both nursing and nonnursing work groups.

Methods This project was performed through the Employee Health Ser-

Back Injury Rates in a Hospital

646

TABLE 1 Worker’s Compensation TPA Database: Outcome Measures and Their Definitions* Injury reports WC cases+

Lost days costs

All hospital incident reports that record injuries involving the back are included. This is the broadest measure of injury from either source. A WC case is defined for this study as one in which ANY money was spent, whether for medical, indemnity, or other purposes. Cases with lost days but no money spent represent denied claims and were not included. The sum of all lost days of the WC cases. This is based on calendar days, not workdays. The sum of all money actually paid as of the date the database was published (Aug 1994 for fiscal years 1992 and 1993; Feb 1995 for fiscal year 1994). It does not include reserves that have been set aside to pay for future bills.

TPA, third party administrator. + WC, workers’ compensation. l

Total Injury Report Rate (TIRR): Number of Injury Reports/100 FTE Total Injury = (Number of reports filed/areal~r~x~200.000 hours worked/100 FTE) number of hours worked/w&year Report Rate

Compensation Case Rate (CCR): Number of WC cases/l00 FTE Compensation Case = (Number of WC $$ Casesiarealvr) x (200,000 hours worked/l00 FTE) number of hours worked/am/year Rate

Compensation Severity Rate (Compensation SR): Number of Days Lost 1100 FTE Severity = (Number of WC lost davsiarealvr) x (200.000 hours worked/l00 FTE) Rate number of hours workediareaiyear

Cost Rate: Dollars Actually Spent.000 FTE Cost = ($$ Suent/area/vr) x (200.000 hours worked/l00 FTE) number of hours worked/area/year Rate

Composite Risk Indicator (CRI) camp ml =

TIRRw x CCRx x CompSR, x CostRateL 1,000,000

Average Relative Risk (ARR)

Justified Average Relative Risk (JARR)

Fig. 1. Third party administrator database: individual and composite rate calculations.

vice of a large, academically affiliated hospital with about 900 beds during the study period. Injury and illness data for back injuries from fiscal years October 1, 1992 to September 30, 1994 were examined for

all 8000 of the general hospital employees, including those working in the on-site Ambulatory Care Center as well as three affiliated Neighborhood Health Centers. It did not inelude employees who worked in the

l

Goldman et al

research areas or staff physicians who received workers’ compensation coverage through another source. The hospital is self-insured, with workers’ compensation case management performed by the Employee Health Service, but data management and billing are done by an outside third party administrator (TPA). The TPA served as the source for data on injury reports, workers’ compensation cases, lost days, costs, accident type, and agent. The definitions of these measures are listed in Table 1. The hospital payroll system was the source of the actual hours worked in each of the working group categories. A full description of the database is contained in a previous article.” The standard injury rate statistics, summarized in Fig. 1, including Total Injury Report Rate (TIRR), Compensation Case Rate (CCR) and Compensation Severity (or days lost) Rate (Comp-SR), were applied across all hospital departments. The Comp-SR was averaged over 2 fiscal years because data were available for that time period only. The benchmark was computed as a target based on a 25% reduction per year. Although government-based injury epidemiological studies have traditionally avoided inclusion of cost data, decision-making and priority setting within a company invariably includes discussion of costs. We attempted to include the cost variable in some of the statistics and to examine its impact on ranking, risk assessment, and prioritization. The costs used in these calculations were direct costs accounted for by the TPA, including medical care, wages, and selected legal charges. We estimated the 95% confidence intervals (CI) for the observed rate statistics on the basis of Poisson or binary distribution as appropriate. In addition, composite statistics, including Compensation Composite Risk Indicator (Comp-CRI), Average Relative Rate (ARR), and Justified Average Relative Risks (JARR)

JOEM

l


were also used’ and were compared over time. As noted previously, difficulties arose when following relative rates over time, leading to the development of the JARR, which uses a benchmark for comparison.” In this case, the benchmark was arbitrarily set at a 25% reduction, so that all relative rates could be compared with that goal, including the hospital-wide average. To assess the temporal trends and variation in the ranks of the back injury for 3 years, the JARRs of different worksites were compared. For these composite statistics, we could not provide the 9 5 % CIs because of the lack of information on the covariance among the standard injury rate statistics mentioned above. Although, to some degree, back injuries can be viewed as a cumulative trauma disorder because repeated small traumas frequently produce pain and impairment, analyzing the precipitating event can prove to be useful in ultimately designing preventive intervention strategies. We therefore examined the accident types recorded in the accident report to assess common causes of back injuries in the hospital.

Results Table 2 displays the consolidated back injury data, annual rates, and benchmark rates for selected work areas. Table 3 uses the data and rates from Table 2 to contrast the rank orders derived from the different rates over the same time periods. Per 100 full-time equivalents (FTE), the overall hospital-wide injury Total Injury Report Rate (TIRR) was 4.6 injury reports (95% CI, 4.3 to 4.9); Compensation Case Rate (CCR), 1.4 workers’ compensation cases (95% CI, 1.3 to 1.6); Compensation Severity Rate (Comp-SR), 76.0 lost days (95% CI, 74.6 to 77.3); and the Cost Rate $3742. Overall rates for nursing were about double that of the overall hospital-wide rates. For all nursing areas combined, the TIRR was 9.8 reports per 100 FTE (95% CI, 9.0 to 10.7). The TIRR for nursing subdi-

visions varied widely, from a high of 14.2 (95% CI, 11.8 to 16.8) for Intensive Care Unit (ICU) Nursing to a low of 3.8 (95% CI, 2.1 to 6.3) for Pediatric Nursing. The CCR for all nursing was 2.3 compared with the hospital-wide rate of 1.4. The combined risk and the relative risk indicators (Average Relative Rate [ARR] and Justified Average Relative Risks [JARR]) ranked ICU Nursing as the highest-risk area. Non-nursing work areas also demonstrated increased risk for back injury. Buildings and Grounds ranked fourth in TIRR (10.0 injury reports with 95% CI, 7.5 to 13.1); first in CCR (4.5 cases with 95% CI, 3.5 to 5.7); second in Comp-SR (259 lost days with 95% CI, 242.9 to 276.6); and ninth in Cost Rate ($4230 per 100 FTE). When using the combined risk indicators, Buildings and Grounds ranked second overall. Use of the traditional injury, compensation case, and compensation severity rates gave different rankings for the working groups, sometimes with a marked reordering as shown in Table 3. For example, Orthopedics/Neurological Nursing ranked second with TIRR, fifth with CCR, and eighth with Comp-SR. In addressing cost issues, Table 4 displays different cost assessments and rankings for 1 fiscal year. For All Nursing, direct costs were $4096 per case for 1 fiscal year. However, some cases are active for more than 1 year, so the actual cost per case would be much higher if followed over all years that payments were made. This point is confirmed by noting that the mean yearly cost for all hospital back injuries occurring during fiscal year 1992 was $284,872. However, the total amount paid for back injuries that year, including both fiscal year 1992 injuries as well as ongoing payments for past injuries, was $1,059,484. The composite statistics, CompCR1 and JARR, which both assign equal weight to all four injury rates, show a similar ranking of working groups. Removing the Cost Rate

647 from the JARR calculation resulted in only small changes to the overall ranking of the working groups. Figure 2 shows the reported types of accidents (mechanism of injury) for the hospital as a whole and for selected work areas for 1 year. Handling patients was the reported event in over 50% of the hospital’s injuries. Lifting and handling materials accounted for an additional 20%. The distribution of accident type varied across the work areas. Patient handling was the major cause of injury in all nursing areas and radiology, accounting for over 80% of the accidents in all nursing areas except Operating Room Nursing. Materials Management also reported patient handling injuries, probably because of its Patient Transport Department. Figure 3 displays the JARR for different nursing services over a 3-year period. The graphs show wide variation from year to year with some groups, reflecting the impact of one or more serious injuries on relatively small numbers. A more consistent trend is seen in others.

Discussion and Conclusions Our study demonstrated that the application of different standard statistics generated different results in ranking the risk of back injury among various work areas. Some areas, such as ICU Nursing, Building and Grounds, and Orthopedics/ Neurology Nursing, ranked among the highest in all four individual rates and would be deemed high-risk regardless of which rates were used. Other areas, however, seemed to be high-risk with some rates and lowrisk with others. Pharmacy, for example, had the lowest TIRR but the highest Cost Rate. We found that using combined statistical measures provided a more integrative measure for describing the overall risk of back injuries and for targeting high-risk areas for potential intervention. These combined statistics also provided the opportunity to view the work area risk in


646

TABLE1 Worker’s Compensation TPA Database: Outcome Measures and Their Definitions* Injury reports WC cases+

Lost days costs

All hospital incident reports that record injuries involving the back are included. This is the broadest measure of injury from either source. A WC case is defined for this study as one in which ANY money was spent, whether for medical, indemnity, or other purposes. Cases with lost days but no money spent represent denied claims and were not included. The sum of all lost days of the WC cases. This is based on calendar days, not workdays. The sum of all money actually paid as of the date the database was published (Aug 1994 for fiscal years 1992 and 1993; Feb 1995 for fiscal year 1994). It does not include reserves that have been set aside to pay for future bills.

* TPA, third party administrator. + WC, workers’ compensation.

rotal Injury Report Rate (TIRR): Number of Injury Reports/l00 FTE rotal Injury = mumber of reports filedlareaivr~x~200.000 hours worked/l00 FTE) number of hours worked/area/year Report Rate

Compensation Case Rate (CCR): Number of WC cases/l00 FTE hnpensation

Rate

Case

= (Number of WC $$ Casesiarealvr) x (200.000 hours worked/l00 FTE) number of hours workedlareaiyear

Compensation Severity Rate (Compensation SR): Number of Days Lost /I00 FTE Severity = (Number of WC lost davsiarealvr) x (200.000 hours worked/l00 FTE) number of hours workedlareaiyear Rate

Cost Rate: Dollars Actually Spent/l00 FTE Cost = I$$ Soent/area/vr) x (200,000 hours worked/l00 FTE) Rate number of hours worked/area/year

Composite Risk Indicator (CRI) camp CRI =

TIRRx x CCRx x CompSR, x CostRateL 1,000,000

Average Relative Risk (ARR)

Justified Average Relative Risk (JARR)

I

Fig. 1. Third party administrator database: individual and composite rate calculations.

vice of a large, academically affiliated hospital with about 900 beds during the study period. Injury and illness data for back injuries from fiscal years October 1, 1992 to September 30, 1994 were examined for

all 8000 of the general hospital employees, including those working in the on-site Ambulatory Care Center as well as three affiliated Neighborhood Health Centers. It did not inelude employees who worked in the

l

Goldman et al

research areas or staff physicians who received workers’ compensation coverage through another source. The hospital is self-insured, with workers’ compensation case management performed by the Employee Health Service, but data management and billing are done by an outside third party administrator (TPA). The TPA served as the source for data on injury reports, workers’ compensation cases, lost days, costs, accident type, and agent. The definitions of these measures are listed in Table 1. The hospital payroll system was the source of the actual hours worked in each of the working group categories. A full description of the database is contained in a previous article.’ The standard injury rate statistics, summarized in Fig. 1, including Total Injury Report Rate (TIRR), Compensation Case Rate (CCR) and Compensation Severity (or days lost) Rate (Comp-SR), were applied across all hospital departments. The Comp-SR was averaged over 2 fiscal years because data were available for that time period only. The benchmark was computed as a target based on a 25% reduction per year. Although government-based injury epidemiological studies have traditionally avoided inclusion of cost data, decision-making and priority setting within a company invariably includes discussion of costs. We attempted to include the cost variable in some of the statistics and to examine its impact on ranking, risk assessment, and prioritization. The costs used in these calculations were direct costs accounted for by the TPA, including medical care, wages, and selected legal charges. We estimated the 95% confidence intervals (CI) for the observed rate statistics on the basis of Poisson or binary distribution as appropriate. In addition, composite statistics, including Compensation Composite Risk Indicator (Comp-CRI), Average Relative Rate (ARR), and Justified Average Relative Risks (JARR)

J v C

f t C

L I t t I 1 t 1 1 ,

TABLE 2 Consolidated Back Injury Data, Annual Rates, and Benchmark Rates for Selected Work Areasa

Work Area: Super Groups

Average Annual* Total Injury 3-Year Total Injury Report Rate: FY 92-94 Reports Reports/l00 FTE (W

1,833,613 ICU nursing OR nursing 1,808,046 Pediatric 740.280 nursing Medical 3,250,807 nursing Surgical 1,324,844 nursing Ortho/neuro 1,456,500 nursing 70,597,602 A// nursing 2,245,727 Dietary Materials 1,118,705 mgmt. Pharmacy 650,698 Radiology/ 3,049,051 radiation Environmen- 2,793,573 tal services Buildings 8 1,077,064 grounds Hospital45,035,416 wide

3-Year Total Average Annual* Compensation Comp Case Rate: Cases Cases/l 00 FTE

d-Years Total Costt

Average Annual* Cost Rate: Dollars/l00 FTE

P-Year Total FY92-93 (hr) Lost Days*

Average Annual* WC Severity Rate: Days/l 00 FTE

130 61 14

14.2 6.7 3.8

35 17 3

3.8 1.9 0.8

76,398 83,357 1,359

8,333 9,221 367

1,269,304 1,212,705 518.569

2,303 1,093 0

362.9 180.3 0.0

150

9.2

32

2.0

94,377

5,806

2,151,458

1,012

94.1

70

10.6

17

2.6

51,863

7,829

893.879

541

121.0

93

12.8

18

2.5

80,033

10,990

979,071

577

117.9

519 31 48

9.8 2.8 8.6

122 13 20

2.3 1.2 3.6

387,387 22,599 6,486

7,315 2,013 1,160

7,146,396 1,560,183 751,401

5,526 631 167

154.7 80.9 44.5

11 68

3.4 4.5

6 20

1.8 1.3

39,712 27,464

12,206 1.801

455,357 2,043,888

448 440

196.8 43.1

58

4.2

26

1.9

99,973

7,157

1,933,630

1,399

144.7

54

10.0

24

4.5

22,778

4,230

710,929

922

259.4

1036

4.6

320

1.4

842,578

3,742

30,559,872

11,608

76.0

3.5

1.1

2,807

57.0

rates” a FY, fiscal year; FTE, full-time equivalent; WC, workers’ compensation; ICU, intensive care unit; OR, operating room. * The basic formula for all rate calculations is: (no. of cases/areaIyr) x (200,000 hours worked/l 00 FTE)/(no. of hours worked/area/yr). + Total costs includes all costs for cases occurring within the 3-year period. This does not include payments for cases before FY 92, nor does it include money set aside for future payments. * Lost day data was available only for FY 92-93. 5 25% reduction in all average annual rates. This value was an empirically chosen goal.

JOEM

l


649

TABLE 3 Comparison of Rank Order Using Individual Average Annual Rates and Composite Statisticsa TIRR

#/loo #/loo Super Group Work Area FTE Rank FTE ICU nursing Buildings & grounds Ortho/neuro nursing Pharmacy Surgical nursing All nursing OR nursing Medical nursing Environmental services Materials mgmt. Hospital-wide Dietary Radiology Pediatric nursing Benchmarksn

Comp-SR*

CCR

ARR Comp CRI (Hosp-wide)

Cost Rate

#I1 00 $1100 Rank FTE Rank FTE Rank Score+ Rank Score* Rank

14.2 10.0 12.8 3.4 10.6 9.8 6.8 9.2 4.2 8.6 4.6 2.8 4.5 3.8

1 4 2 14 3 5 8 6 11 7 9 15 10 12

3.8 4.5 2.5 1.8 2.6 2.3 1.9 2.0 1.9 3.6 1.4 1.2 1.3 0.8

2 1 5 10 4 6 8 7 9 3 11 13 12 15

362.9 259.4 117.9 196.8 121.1 154.7 180.3 94.1 144.7 44.5 76.0 80.9 43.1 0.0

3.5

13

1.07

14

57

1 2 8 3 7 5 4 9 6 13 11 10 14 15

$ 8,333 $ 4,230 $10,990 $12,206 $ 7,829 $ 7,315 $ 9,221 $ 5,806 $ 7,157 $ 1,160 $ 3,742 $ 2,013 $ 1,801 $ 367

4 9 2 1 5 6 3 8 7 14 10 12 13 15

12.8 7.0 6.4 3.9 5.1 5.1 4.6 3.2 2.8 1.3 1.4 0.7 0.7 0.0

12 $ 2,800 11

0.77

JARR (Benchmark) Scores

Rank

1 2 3 7 4 5 6 8 9 11 70 13 14 15

3.2 2.5 2.3 2.0 2.0 7.9 1.9 1.6 1.5 1.3 1.0 0.8 0.7 0.4

1 2 3 4 5 6 7 8 9 10 11 12 14 15

4.3 3.3 3.0 2.6 2.6 2.6 2.5 2.1 2.0 1.8 1.3 1.0 1 .o 0.5

1 2 3 4 5 6 7 8 9 10 11 12 14 15

12

0.75

12

1 .oo

12

a TIRR, Total Injury Report Rate; FTE, full-time equivalent; CCR, Compensation Case Rate; Comp-SR, Compensation Severity (or days lost) Rate; ARR, Average Relative Rate; JARR, Justified Average Relative Risks; ICU, intensive care unit; OR, operating room; Comp-CRI, compensation composite risk indicator. l Comp-SR is for fiscal years 1992-l 993 only because lost day data were available only for those years. + Comp-CRI = @RR X CCR X Comp-SR x Cost Rate. 1 ,ooo,ooo

* Comp-ARR=

area TIRR area CCR area Cost Rate area Comp-SR + hospital Cost Rate 1, ( hospital TIRR )+( hospital CCR b( hospital SR--I(

,comp~JARR~(~)+(~)+(are~~~~““)+i:,~~~~~~~~) 4 n 25% reduction in all average annual rates. This value was an empirically chosen improvement goal.

comparison with the overall background (Average Relative Rate) or to a benchmark level (Justified Average Relative Risks). The combined statistics seemed to be useful measures for following risk or risk reduction over time. However, we also noted, as might be expected with following any statistical measures over time, that small numbers in the different working groups can lead to considerable variation, making actual trend analysis difficult. Our study also supports previous studies that have demonstrated an increased risk of back injuries among nurses and produced similar rates of injury in similar settings. For example, a large west coast hospital re-

ported 25.5 “lost time injuries per year” among 733 nurses, which is an approximate rate of 3 per 100.‘” Our Comp CR, which is similar to lost time injuries, was 2.3 per 100 FTE for all nursing. The varying risks of back injury observed among the nursing work areas (Table 3) reflected different types of nursing patient care demands, in particular, the amount of patient lifting required. ICU, Surgical, and Orthopedics/ Neurology units, which demonstrated the highest risks for back injury, have the greatest number of non-ambulatory patients and therefore require more total body transfers and lifts. Nurses working in the Operating Room and Medical Units

were found to have a midrange risk. Patient lifting in the operating room is usually well coordinated and done with several people. General medical units have more ambulatory patients. Pediatrics Nursing, which falls well below All Nursing and hospital-wide levels, cares for patients who weigh less than adults. The proposed lifting guidelines of the National Institute for Occupational Safety and Health recommend a maximum lift of 51 pounds, under ideal conditions. l1 The weight of adult patients consistently exceeds this recommendation, whereas that of pediatric patients seldom does. Adult patient lifting is therefore likely to be the greatest single factor explaining the differ-

650


l

Goldman et al

TABLE4 Three Ways to

Analyze Cost* Cost to Date of FY92 New Cases

SuDer GrouD Environmental services OR nursing Medical nursing Surgical nursing KU nursing Buildings & grounds Pharmacy Ortho/neuro nursing Dietary Materials mgmt. Radiology/radiation med A// nursing Hospitalwide

FTE

ComD Cases

482 307 537 220 317 210 120 249 394 185 507 1793 7606

8 4 12 6 9 2 9 4 5 3 38 82

Rank

$78,245 $47,687 $66,826 $18,741 $18,929 $11,849 $2,292 $3,456 $5,022 $2,030 $1,902 $155,639 $284,872

10 11

Avg Cum Cost/ Case

Cost Rate: $/IO0 FTE

cost

Rank

Rate

Rank

$9,781 $11,922 $5,569 $3,124 $2,704 $1,317 $1,146 $384 $1,256 $406 $634 $4,096 $3,474

2 1 3 4 5 6 8 11 7 10 9

$16,220 $15,538 $12,443 $8,510 $5,978 $5,631 $1,914 $1,385 $1,276 $1,097 $375 $8,681 $3,745

1 2 3 4 5 6 7 a 9 10 11

l Cumulative costs of new cases occurring during fiscal year (Fy) 1992. All reported expenses through July 1994 are included. FTE, full-time equivalent; OR, operating room; ICU, intensive care unit.

ences in back injuries among the various nursing services. Examination of the data for accident types (Fig. 2) provides additional evidence for the importance of patient handling as the major risk factor for back injury among nurses. In addition to the patient weight factor, health care workers frequently move patients while using awkward postures that create additional back stress. Our results, along with other studies of the biomechanics of patient transfers, support the need for ergonomic lifting methods, such as the use of lifting devices to perform patient transfers and lifts.12P’” The use of trained lifting teams is one way to organize the use of the lifting devices and to ensure that the equipment will be used and handled appropriately. ” The use of a lifting team removes the majority of lifting from nursing staff and concentrates it among a few trained individuals who use assisting devices. Thus far, Charney’s reports of lifting teams has demonstrated a marked decrease in back injuries in hospital employees without reported injuries among the trained lifting technicians.16 The use of expensive beds that change position into chairs or stretchers is another modality implemented by hos-

pitals to decrease the need for patient transfers. Reduction in the monetary as well as human cost of back injuries is another incentive for hospitals to invest in equipment and staff to lower the risk. Costs frequently become one of the determining factors in deciding where and how to intervene. Different cost statistics, however, can give different information. Table 4 demonstrates how work areas can appear to be high- or lowrisk depending on the cost statistic used. A frequently used statistic is cost per case. To gain an accurate cost per case assessment, costs must be followed over time, because back injury costs frequently exceed 1 year’s duration. Tracking the costs per case over many years can be cumbersome. In addition, cost per case has other limitations. Higher paid employees will have more expensive cases compared with lower paid workers, even in some instances in which the severity of the injury may have been less. Also, cost per case does not reflect the impact of the numbers of injuries in the work area. There may be many back injuries in a work area with lower paid employees and a lower cost per case, but the overall cost for that area may be high. In our study, only three

hospital areas had above-average cost per cases (as determined for only 1 year): two nursing areas and Environmental Services. If only the cost per case were used, other areas identified as risky by the other statistics would be designated low-risk. Using another statistic, Cost Rate (or expense per standard number of employees), facilitates comparisons among work areas. The results are also in a format (cost per year per 100 FTE) that is comparable with the other illness and injury statistics. For example, a small work area with low wages could have many low-cost cases and thus a low cost per case, but the total cost per 100 workers could be high. This is illustrated by the Buildings and Grounds Department, which had the highest CCR over the 3-year period (Table 3). In 1 year (fiscal year 1992) nine cases had a total cost of $11,849, or $13 17 per case (Table 4). Because this is a rather small department (210 FTE), the cost per 100 FTE becomes $5631. Our study demonstrated that nonpatient care areas in the hospital also experience high rates of back injury (Table 3). For example, the Buildings and Grounds Department was 2.47 times the hospital-wide rate, not far behind the highest nursing group,

JOEM

l


651


i-,Q PushinglPullmg i q Slip/Fall 93 Miscellaneous ‘Bs Handling Material b Lifbng/Bendmg [U Handling Patlent

Work Area

Fig. 2. Accident types for selected work areas for fiscal year 1993

ICU Nursing, which was 3.19 times the hospital-wide rate. Pharmacy and Environmental Services were also found to have elevated levels of back injury above the hospital-wide and benchmark levels. Therefore, it is important to include non-patient care areas in the process of ergonomic evaluations and interventions in the hospital setting. In these analyses, we did not specifically control for age, gender, and previous prior back injury because of the limited availability of such information. Although, historically, ICU nurses may have been somewhat older because experience had been a

prerequisite for work in this setting, recent changes in health care employment have influenced the retention of nurses in all settings at this hospital. For this reason, we perceived no major age differences among the various nursing groups. There is a potential for bias in the reporting of non-lost time injuries, TIRRs. Nursing areas with employees who were familiar with the hospital incident report system had more injuries with no lost time or payment than did areas such as Environmental Services, where familiarity with incident reports is thought to be more limited. The number of compensa-

tion cases, days lost, and costs include more objective measures that describe more serious injuries. The fact that these rates would be less influenced by reporting biases is another reason to use different types of injury rates to describe the risk. Cases with few lost days were underrepresented in this TPA database. When the TPA cases were compared with the Occupational Safety and Health Administration 200 log, several cases with fewer than 5 lost days reported on the log had 0 days recorded in the TPA database. The workers’ compensation wage replacement benefits begin

652

Back Injury Rates in a Hospital Back Injury Rates in a Hospital

6.00

+ ICIJ Nursing -D-OR Nursing -A- Surgical Nursing -X- Ortho/Neuro Nursing -~-Pediatric

Nursing

-A- Medical Nursing -+-Nursing

Combined

*Hospitalwide Combined 93 Fiscal Year

Fig. 3. Nursing services Justified Average Relative Risks trend.

only after 5 workdays are lost. If a worker returns to work 4 or fewer days after an injury, that information would not be likely to be recorded in the database because the TPA does not pursue information on cases for which no lost wages are paid. The lack of these data could affect the Severity Rating of a work area. The overall ranking schemes would be affected only if certain areas had many cases with fewer than 5 lost days. Because our observed statistical outcomes for back injury reports and costs are similar to those reported in other hospital settings,8.‘0 we hope our approach to describing and ranking back injury risks among hospital

workers would be applicable to other hospital and industrial settings. The use of composite statistics merits further use and study, particularly in terms of prioritizing risk areas and following trends over time.

References 1. Klein B, Jensen R, Sanderson L. Assessment of workers’ compensation claims for back strains/sprains. J Occup Environ Med. 1984;26:443-448. 2. Frymoyer JW, Cats-Baril WL. An overview of the incidences and costs of low back pain. Orthop C’Iin North Am. 1991; 22:263-21 I. 3. Snook SH. The costs of back pain in industry. Occup Med State Art Rev. 1988; 3: 1-s. 4. Webster BS, Snook SH. The cost of 1989

l

Goldman et al

workers’ compensation low back pain claims. Spine. 1994;19:11 I l-l 116. 5. Guo H-R, Tanaka S, Cameron LL, et al. Back pain among workers in the United States: national estimates and workers at high risk. Am J Ind Med. 1995;28:591602. 6. Fuortes LJ, Shi Y, Zhang M, Zwerling C, Schootman M. Epidemiology of back injury in university hospital nurses from review of workers’ compensation records and a case-control study. J Occup Med. 1994;36:1022-1026. 7. Agnew J. Back pain in hospital workers. Occup Med State Art Rev. 1987;2:609616. 8. Kaplan RM, Deyo RA. Back pain in health care workers. Occup Med State Art Rev. 1988;3:61-73. 9. Jarrard MR, Goldman RH, Loomis SC, Atkins EH. Methods of prioritizing and measuring occupational health risks utilizing hospital back injury data. J Occup Environ Med. 1997;39:882-888. 10. Charney W, Zimmerman K, Walara E. The lifting team: a design method to reduce lost time back injury in nursing. Am Assoc Occup Health Nurs. 1991;39: 23 l-234. 11. Waters TR, Putz-Anderson V, Garg A, Fine LJ. Revised NIOSH equation for design and evaluation of manual lifting tasks. Ergonomics. 1993;36:749-776. 12. Garg A, Owen BD, Carlson B. An ergonomic evaluation of nursing assistants’ job in a nursing home. Ergonomics. 1992;35:979-995. 13. Garg A, Owen B. Reducing back stress to nursing personnel: an ergonomic intervention in a nursing home. Ergonomics. 1992;35:1353-1375. 14. Pheasant S, Stubbs D. Back pain in nurses: epidemiology and risk assessment. Appl Ergon. 1992;23:226-232. 15. Venning PJ, Walter SD, Stitt LW. Personal and job-related factors as determinants of incidence of back injuries among nursing personnel. J Occup Med. 1987; 29:820-825. 16. Charney W. The lift team method for reducing back injuries: a 10 hospital study. Am Assoc Orcup Health Nurs. 1997;45:300-304.