Cost and determinants of morbidity from work related disabling injuries in Taiwan

doi:10.1136/oem.52.2.138

1995;52;138-142

Occup. Environ. Med.

Y H Liu, M R Lin and J D Wang

related disabling injuries in Taiwan.

Cost and determinants of morbidity from work

http://oem.bmj.com/cgi/content/abstract/52/2/138

Updated information and services can be found at:

These include:

Rapid responses

http://oem.bmj.com/cgi/eletter-submit/52/2/138

You can respond to this article at:

service Email alerting

the top right corner of the article

Receive free email alerts when new articles cite this article - sign up in the box at

Notes

http://journals.bmj.com/cgi/reprintform

To order reprints of this article go to:

http://journals.bmj.com/subscriptions/

go to:

Occupational and Environmental Medicine

Occupational and EnvironmentalMedicine1995;52:138-142

Cost and determinants of morbidity from work

related

disabling injuries

in

Taiwan

Yi-Hung Liu,Mau-Roung Lin, Jung-Der Wang

Departments of Family Medicine and InternalMedicine, National Taiwan University Hospital Y-HLiu

J-D Wang

Center for The

Researchof Environmental and Occupation Diseases, Institute of Occupational Medicine and IndustrialHygiene, National Taiwan UniversityCollege of PublicHealth,Taipei, Taiwan, Republicof China Y-HLiu M-RLin J-D Wang Correspondenceto:

DrJung-Der Wang, Center

for the Researchof Environmentaland Occupational Diseases,

Institute ofOccupational Medicineand Industrial Hygiene, National Taiwan University, College of Public Health,No1, Section 1, Jen-Ai Road, Taipei, Taiwan, RepublicofChina

10016.

Accepted13October 1994

Abstract

Objectives-To estimate the cost and

determinants of morbidity from work

related disabilities.

Methods-114peoplewho lived in

metro-politan Taipei and who had received

dis-ability compensation from the Labor

Insurance Bureau from March to June,

1991 wererandomly selected. Therewere

77 workers interviewed through a

ques-tionnaire that inquired about possible loss of productivity including the dura-tionofmorbidity, the abilityto return to

work, andanychange in monthly income

uponreturningtowork.

Results-The mean (SD) duration of a

stayinhospitalwas29 (39) days, median:

15 days. The average duration between

discharge from the hospital and

return-ingto workwas 111 (146) days, median:

45 days. The main determinants of the

duration of the stay in hospitalwere the

number of stays in hospital and the

severity of the injury. Amultivariate lin-earanalysis showed that oldage and the severity of injury determinethedurations ofmorbidity. An ordinal logistic

regres-sion analysis showed that the severity of

injury, size of thefactory, and age

deter-mined the magnitude of future produc-tivity loss. Based on thesemodels, itwas

estimated that the total duration of

mor-bidity (in hospital and at home) due to

occupational disability was 660000

per-son-days eachyear.When thepercentage

ofthedecreaseinincomebecause of

per-manent disability was converted into a

loss of work days, the annual morbidity

costs were about 19 000-26 000 person-yearsbetween 1985-1990.

Conclusion-The morbidity cost was

aboutfive timesashighasthelumpsum

payment that a worker usually received fordisability compensation.Weconclude

that morbidity cost should be evaluated carefully in the future for the

establish-mentofaccurateand fairdisability com-pensationpayments.

(OccupEnviron Med1995;52:138-142)

Keywords: occupational disability; morbidity cost;

compensation

An average of one to two/thousand male

workers in Taiwan sustain work related

dis-abling injuries annually' 3; also, the annual

mortalityrateof male workers dueto

occupa-tional injuries is about 05/thousand.2I The risk of major occupational injuries among

female workers is aboutonesixth thatof male workers and has not shown any tendency to

decline in thepastsixyears.' Thetotal

mor-tality duetooccupational injuries in Taiwan is five to 10 times higher than that of more

industrialised countries such as Japan or the United States.2

Besides the above figures, there are other

reasons to call for continuous efforts and

moreallocation ofresourcesfor research and the prevention ofoccupational injuries. Most

occupational injuries canbe avoided ifsafety

mechanisms are clearly understood and

proper precautions followed.4 Furthermore, the social costs of occupational injuries are

muchgreaterthan thoseof other health prob-lems because theage groupofinjuredworkers is relatively young and they are economically

productive.3

There are two key elements of social cost.

One is the directcostsuchasexpendituresfor

goods and services. The other is the indirect

costthat includes the lossofpotential earning

or loss ofproductivity.9 The indirect cost is usually several times higher than the direct

cost.'0 In occupational accidents, the indirect

cost canbe further divided into themortality

costofoccupationalfatalities and the morbid-itycostofoccupationalinjuries.9

Inthispaper we try toestimate thecostand determinantsofmorbidity duetowork related

permanent physical impairment-that is, occupationaldisabilities,in Taiwan.

Subjectsand methods

To draw a representative sample of workers with work relateddisabling injuries,wevisited the Labor Insurance Bureau weekly during the months of March to June, 1991. Every worker who lived in the metropolitan Taipei

areaandhad received hisorhercompensation in theprevious weekwasregistered and givena

number. Through the procedure of random sampling, 114 people were selected. Additional information such as their address,

the telephone numbers of their home and workplace,andthe details oftheirinjuries,etc were obtained from the Labor Insurance Bureau. To receive disability compensation resulting from occupational injuries, these

workershad to havefinishedtheir initial med-ical treatment, and the disability should be

relatively stable. All workers with an

occupa-tionally-related permanent disability will cer-138

Cost and determinantsofmorbidity from work related disablinginjuriesin Taiwan

Table 1 Comparison of the pattern of injuries and the degree of disabilities Intemiewed Notinterviewed

n 77 n=37 Pvalue Age (y): Mean(SD) 37-3(11.1) 33-6(11 3) 0-096 Range 8-62 16-57 SexM:F 58:19 34:3 0036t Injury:

Blindnessdue toeyeball injury 1 0 039t

Cervical and thoracic trauma 2 2

Lumbosacraltrauma 4 0

Lowerextremity 3 1

Bilateralupper extremities 5 0 Unilateral upper extremities 60 34

Finger(s) 2 0 Degreeof disability: 1-3j: 3(2) 1 068t 4-6* 2(1) 1 7-9t 16(3) 4 10-12t 27(1) 21 13-15* 29(5) 10 Disability compensation (NT$): Mean (SD) 113 725(114 950) 117938(155851) 0-82* Range 15 300-470 000 19 200-915 000 Median 71400 81600

Durationofstay inhospital:

Mean(SD) 23(36) 46(129) 0-69*

Range 0-252 0-738

Median 13 11

*Wilcoxon rank-sumtest.tX2test.*()No of workers not returned to work.

tainly obtain their compensation payment

because ofworkmen's compensation legisla-tion in Taiwan. In the years 1985 to 1990,

there were 34 416 workers who obtained

compensation.'

Afterperformingapilottest, wedesigneda

questionnaire"'- thatinquired aboutpossible loss ofproductivity including the duration of morbidity (in hospitalandathome),any

com-pensation payments or reimbursements

received, whether the subjects couldreturn to

work, and if there were any changes in their

incomes. We employed two groups of

inter-viewers with three members in each group: onemedicalprofessional (amedical doctoror

physical therapist) and two social workers.

Eachinterviewerreceived asix hourinterview

training course andeveryitem ofinquirywas

standardized. From July to November 1991, interviewersvisited the selectedworkers.

Datawerethen coded into acomputerand

analysed by a statistical package SAS/STAT

PC 6-04."4 Weapplied multiple linear

regres-sionanalysisfor theestimation of the duration of totalmorbidity includingthe duration ofa

stay in hospital and rest at home after acute

injuries. Then, an ordinal logistic regression

analysis'5 was also performed to obtain the

adjusted odds ratio (OR) of all the

determi-nants of future loss ofproductivity. We have classified the decrease inmonthlyincome (in

NT$) into four ordinal categories:

<5000, 5000-9999, 10 000-24 999, and >25000. The independent variables we put

in tofit the model included age, sex, severity

ofinjury,educationallevel,size of thefactory, relationshiptotheemployer, etc.

To estimate the total morbidity cost of work relateddisabilityinTaiwan, wecanonly

use theseverityofinjury andageasthe

inde-pendent variables because no information on the other determinants is available at the

national registration data. We have

con-structed a model topredict the magnitude of theperson's future decrease inworking

capa-bility. In thismodel,theproportion ofchange

inmonthly income was assumed to be a per-centage of loss of working capability. The resulting models from the multiple regression were used to estimate the productivity lost

due tomorbidity fortheyears 1985 to 1990.

Each worker's future loss ofproductivity was calculated as the percentage of loss multiplied by his or her potential working years. To

esti-matethepotential working years, we assumed that each worker would have worked until 65 years of age had he or she not been injured. Thus, we obtained the person's potential working years bysubtracting the person's age

from 65. Both parts of the morbidity cost-thatis, the durationofthestay in hospital plus convalescence at home and the future loss of

productivity-sharedthe same time unit: per-son-daysorperson-years.

Results

Among the 114 workers randomly selected, there were 92 men and 22 women. Their

mean (SD) age was 36-1 (11-3) with a range from 16-62. According to the regulations of the Labor Insurance Bureau, the degree of

disability was classified, from severe to mild,

on ascaleof1-15 (1 being the most

severe).'6

Most of the disabilities were mild-that is,

76-3% of the workers fell between 10-15 on the severity scale. The age and sex

distribu-tionof these 114workers,aswellasthecause, agent, and severity of theinjury, were similar

to those of all disabled workers in Taiwan in 1990.3

We successfully interviewed 77 workers;

theinterviewratewas67-5%.Themean(SD)

durationfrom the onset ofinjuryto the date of the interview was 10 1 (6-7) months. Sixty five ofthem (84%) had returned to work by

the timeofourstudy. Therewere 12workers who could not be interviewed because of diffi-culties in making an appointment. Other

common reasons for a lack of an interview

wereeither a change inplace of employment

and moving out ofmetropolitan Taipei, or a

loss of contact with their families (21/37).

Three people refused to be interviewed and

one person died. Other than the distribution of sex, there was no significant difference in the characteristics of those interviewed and those whowere not(table 1).

The mean (SD) duration ofastayin

hospi-talfor the 77 peopleinterviewedwas 29 (39) days, median: 15 days. The number of stays inhospital ranged from0 to 6. The degree of disability was further

categorised

into severe

(degree 1-3), moderate (degree

4-9),

and mild (degree 10-15). The duration of

conva-lescence at home (after discharge from the hospital and before returning to work) was

aboutfive timeslongerthan thatof the stayin

hospital, and the variability of the

convales-cence was also greater (table 2). The total durationofmorbiditywasthesumof the stays inhospitalandathome.

Both the number of times a patient was

admitted to hospital and the severity of the disability (butnot age) were important deter-minants in the multiple linear regression 139

Liu, Lin, Wang

Table 2 Duration ofmorbidity (in hospital or at home) among disabled workers (stratifiedby the severity of thedisability)

Degree of Workers Durationof Durationofstay Durationof disability n Morbidity inhospital rest athome

1-2* 2 Mean (SD) 9680 190(88) 9490 Range 7428-11932 128-252 7300-11 680 3-6 3 Mean(SD) 250(142) 52(18) 198(156) Range 90-360 35-71 19-309 Median 300 53 265 7-9 16 Mean(SD) 209(174) 35(29) 174(157) Range 18-565 6-112 5-565 Median 165 25 133 10-12 27 Mean (SD) 58(75) 11(10) 47(78) Range 0-360 0-31 0-360 Median 30 10 19 13-15 29 Mean(SD) 93(152) 12(16) 81(141) Range 7-767 0-82 0-685 Median 40 7 26

*Two workerswho had beenbedridden fortwoyears after theirinjurywereassumed to be unable to work until theretirementageof65.

analysis of the duration of the stay inhospital.

Except for the two most severely injured workers (degree of disability 1-2 on table 2,

namely, confined to bed) the 75 workers we

visited had an average duration of convales-cence athome of111 (146) days, median: 45 days. Resultsofthemultipleregression

analy-sis ofconvalescenceathome anddurations of morbidity of these 75 workers showed that

Table3 Modelsconstructedfrom multiple linear regressiontopredict the durationof stay inhospital,rest athome, and totalmorbidity*

Dependent Independent Parameter

variable variables estimate Pvalue R2 Durationofstay Intercept 66-7 0-0001 0-66 inhospital Degreeof -5-8 0-0001

severity (1-15)

Number ofstays 19-0 0-0001

inhospital

Durationof Intercept 272-5 0 0050 0-35 rest athome Degree of -28-6 0-0001

severity (1-15) Age 4-4 0-0085 Durationof Intercept 392 6 0-0002 0-38 morbidity Degreeof -37-3 0-0001 severity (1-15) Age 4-1 0 0160

*Other independentvariables,suchaseducational level(P=0-63),thesize ofthefactory judged

by numberof workers (P=0-92), relation withtheemployer (P=0-67)were notincluded in

the model becausetheywerenotsignificant. The twoworkers withpermanentdisabilitywere excludedfrom the model ofthe duration of rest at home andmorbiditytodowngrade their influence on themodels.

Table4 Decreasein monthly income stratified bydemographicandothercharacteristics ofthedisabled workers

Decrease inmonthly income

5000- 10 000-<4999NT$ 9999NT$ 24999NT$ >25OOONT$ Pvalue* Sex M:F 45:14 4:4 5:1 4:0 0-664 Age: 16-44 46 5 2 3 0-072 >45 13 3 4 1 Education level: <Highschool 39 7 6 4 0040 >Highschool 20 1 0 0 Married 15 2 2 1 0-833 Unmarried 44 6 4 3 Size offactory judged bynumber of workers:

<29 41 4 6 2 0-894

>30 18 4 0 2

Relative of theemployer:

No 35 6 6 3 0-070

Yes 24 2 0 1

Injuredpart:

Hands 54 5 4 2 0 003

Others 5 3 2 2

Severityof thedisability:

Mild 48 4 3 1 <0001 Moderate 11 3 3 1

Severe 0 1 0 2

*Mantel-HaenszelX2testfor trends.

these two

dependent

variables shared the

same determinants that is, the age of the worker andtheseverityof theinjury (table

3).

In exploring each person's future loss of

productivity, we calculated the decrease in income after thedocumentation of permanent

disability.

The decrease in monthly income

was stratified and analysed by the Mantel-Haenszel

X2

testfortrends. Table 4 showsthat there is a trend in the decrease of monthly

income associated with a more severe

injury

(P< 0-0001), an injury that involved parts other than the hands (P= 0-003), or a lower educational level (P= 004). Also, an ordinal

logistic registic regression analysis was per-formed to explore the determinants of the decrease in monthly income. The result showed that after adjustmentfor various fac-tors, an increase in the severity of injury, a

decrease in size of thefactory,or anolderage

are independently associated with an

increased OR of greater loss in monthly

income (table5).

With theproportion of decrease inmonthly

income as the dependent variable, a multiple linear regression model was constructed to

explore the importance of the age of the worker and theseverity of the injury. The pro-portion of decrease in monthly income for those over45 years of age showedadramatic increase (14-6%). Moreover, the severity of theinjurysustainedby the worker also signifi-cantlydetermined the magnitude of decrease inthe person'smonthly income (table 6). The total duration of morbidity due to occupa-tionaldisability, inferred from themodel, was about 660 000 person-days, or 1808 person-years, annually for the past six years. The average morbidity cost for each disabled

TableS Adjusted ORs ofdifferent independentvariables

forlossofproductivity calculatedfromanordinallogistic regression analysis (the dependent variablesare thefour

differentlevelsof decreaseinmonthly income-istrowof

table4)

Independentvariables AdjustedOR(95%CI) Pvalue Severityofinjury:

Mild 1.0

Moderate 0-19(0 70to0054) 0 01 Severe 0-004(0-10to00002) <0 001 Size offactory (bynumber ofworkers):

<29 0-16(0-80to0029) 0 03 30 1 0

Age:

16-44 1.0

,_45 0-29(094 to 0086) 0 05 Relatives of theemployer:

No 0-27(1 19 to0-058) 0-08 Yes 1.0

Education level:

<Highschool 0 26 (2-38 to0-028) 0.23

>Highschool 1.0

Table6 Proportionofdecrease inmonthly income predicted bythemultipleregression model stratifiedbyage andseverity oftheinjury

Decrease inmonthly Age(y) Severity income(%)

<44 Mild(10-15) 7.34 Moderate(4-9) 24-73 Severe(1-3) 87-37 >45 Mild(10- 15) 21-93 Moderate(4-9) 39-32 Severe(1-3) 100 0 140

Cost and determinantsofmorbidityfrom work relateddisabling injuries in Taiwan

Table7 Morbiditycosts (inperson-years)ofoccupational disabilitiesinTaiwan 1985-90, estimated by our modelsfrom

tables 3 and 6

Morbidity Calendar Durationof Productivity Total morbidity costs

year Severityofinjury Victims(n) morbidity (y) loss(xlOOy) cost(xlOy) person(y)

1985 Total 5358 186 206 224 4-2 Severe(1-3) 156 220 24 26 16-5 Moderate (4-9) 1042 790 70 78 74 Mild(10-15) 4160 840 112 121 2-9 1986 Total 6002 1870 223 242 4-0 Severe (1-3) 114 160 18 20 1746 Moderate(4-9) 955 670 70 77 8-1 Mild(10-15) 4932 1030 135 145 2-9 1987 Total 6496 2140 244 265 4-1 Severe (1-3) 167 240 24 26 15-7 Moderate(4-9) 1078 770 77 85 78 Mild (10-15) 5243 1120 143 154 29 1988 Total 6163 1980 227 247 4-0 Severe (1-3) 123 180 20 21 174 Moderate(4-9) 1003 710 73 81 8-0 Mild(10-15) 5022 1100 134 145 2-9 1989 Total 5617 1760 206 223 40 Severe(1-3) 118 170 19 21 179 Moderate(4-9) 853 600 63 69 8-1 Mild(10-15) 4634 990 123 133 29 1990 Total 4780 1520 177 192 4-0 Severe(1-3) 101 140 19 20 20-2 Moderate(4-9) 757 530 55 61 8-0 Mild(10-15) 3910 850 102 110 2 8

worker wasequaltoalossof4 1 person-years.

After stratifying the morbidity cost with the

severity of the disability, we found that the average losses ofworkingyearsfor themildly, moderately, and severely disabled workers

were 2-9, 7 9, and 17-6 person-years,

respec-tively. During the past six years the annual

morbidity costfor alldisabled workersranged

from 19000 to 26000 person-years (table 7).

Discussion

In the estimation of the morbidity cost of

occupational disability, difficulties and limita-tions do exist and need to be assessed. We could only focus on the metropolitan Taipei

area and the number of workers selected was

also limitedto 114duetotheamountof time and effortwe wereabletooffer for home visits.

Nevertheless, the metropolitan Taipei area

includes a wide variety of industries that are also widespread throughoutthe country, and the random selection we used

provided

us

witha

representative

sample aswas shownby

the lack of obvious difference between the

demographicdataof the workers in ourgroup and workers in the rest of the

nation.'

Although about30% of the workers were not

interviewed,there was asignificantdifference inonly the distribution ofsexbetween those interviewed and thosenotinterviewed. Assex was not found to be a main determinant in

morbidity costs, wethink thatourresults are

acceptable.

The design of thisstudywas retrospective,

whichwas different fromthe fouryear follow

Table 8 Comparison ofcosts(inperson-years) dueto

occupational mortalityanddisabilityinTaiwan,

1985-90

Calendaryear Mortalitycost Morbiditycostof disability

1985 32 500 22400 1986 32500 24200 1987 36 700 26 500 1988 43 100 24700 1989 44 300 22 300 1990 43 800 19 200

uprecommendedby Bakeretal.9In our coun-tryit usuallytakes along time (an average of

10.1

months) to receive disability compensa-tion and there was about a six to 12 month time lagbetween the compensation payment andourinterviews. Moreover, workers

gener-allysettletheircompensation onlyif their dis-abilities have stabilised to avoid any future

medical costs. Thus, it is reasonable to

assume that the severity of the disability will

notchangeappreciablyafterourinterviews. Theremightbesomedoubt about the legit-imacyof theuseof theproportion of decrease inmonthlyincomeas anindextoestimate the

degree of loss of

productivity.

We think that this is a reasonable method of estimation in Taiwanorinotherdevelopingcountries with-out a comprehensive system of workmen's

compensation. Alarge

proportion

of

employ-ers in Taiwan do not take full responsibility

for workers withoccupational disabilities,and thosewho donotpaymuch attentionto

occu-pational safety and health are more likely to havevictims ofinjury. Instead ofbeing

trans-ferred to another job with a similar income,

workers whoareinjured mightreceivealower

salarywhentheirjobs change, orare evenlaid off because of the change in their working capability. The disability

compensation

pro-vided bythe Labor Insurance Bureauis often the onlypayment theyreceive. Therefore our indexisareasonableestimate because the loss of

productivity

ofthedisabledworkers is

usu-ally reflectedintheirmonthlyincome. Our ordinal logistic regression analysis

showed that the three most

significant

factors for the decrease inmonthlyincomewere: the

severity of the injury, the size ofthe factory,

and age (table 5). The severity ofthe injury

and age were probably the most important

factors, because both of them are also the mainpredictors of the duration ofmorbidity (table 3). Fortunately, the modelsweused to

predictnational loss also contained these two factors(tables3 and6), whichprobablymade ourestimation of nationalmorbiditycost due

to

occupational

disability relatively accurate.

Liu,Lin, Wang

Although we have tried to obtain more infor-mation from the Labor Insurance Bureau, it does not regularly collect data on the size of the factory. Failure to control for the determi-nants other than the severity of theinjuryand age, as well as the small number ofsubjects

interviewed, might limit the generality of our conclusions. We think that our results of mor-bidity cost are unbiased but have a wide confi-dence interval (CI), and thus recommend a larger sample for future studies.

We have computed the working years of potential life lost (WYPLL) as an alternative way to estimate the indirect costs of occupa-tional mortalities in ourprevious article.' We found that the WYPLL ofoccupational mor-talities (inTaiwan) wasabout 38 800 person-years annually. In this study, we have estimated the WYPLL of occupational dis-ability in Taiwan and obtained a figure of about 23 200person-years annually. Thus the ratio of WYPLL between occupational

mor-tality and disability was about 2:1 (table 8). Theannualmortality cost showed a tendency to increase, a result that may be due to the large proportion of occupational fatalities' from traffic accidents and an increase in the

frequencyofmajor occupational accidents at construction sites. On the contrary, the annualmorbiditycostofoccupational

disabil-itydecreasedyearly. The decrease in the pro-portion of workers in manufacturing

industries as well as the migration of haz-ardous industries to neighbouring developing countriesmightcontribute to the result.

The morbidity costs that we have calcu-lated should not be considered to be the only loss to the victims of occupational injuries

simplybecause effects suchaspain,grief, and

family or social disruption cannot be mea-sured in these terms.9 Workers in developing countries like Taiwan still haveextremelyhigh

risks in the working environment. Injured workers should be properly treated and receive enough compensation not only for their future but also the futureof theirfamily

members. At the same time, a greater effort

toward thepreventionofoccupational injuries

mustbe

implemented.

Theamountof

disabil-ity

compensation

payment

given

tothose who sustained

injury

was so limited that it

only

covered about one fifth the morbidity cost thatwecomputed.'6We recommend thatour

people,

government, and

employers

make a

joint

effort to establish a more reasonable compensation payment.

1 LiuYH, Wang JD. Death duetooccupationalaccidents andcost tosociety:ananalysisofcumulativemortality

rates and years of potential life loss. Journal of the

NationalPublicHealthAssociation (RepublicofChina) 1992;11:89-101.

2 The BureauofLabor Insurance in Taiwan. Statistics of laborinsuranceinTaiwan 1986-91.Taipei:TheBureau

ofLaborInsuranceinTaiwan,1992.

3 LiuYH, Wang JD.Determinants andtrends of

occupa-tional disability rates in Taiwan during 1985-1990.

JournaloftheNational Public HealthAssociation(Republic

ofChina) 1993;12:354-67.

4 Waller JA. Occupational injuries. In: WallerJA. Injury

control: aguide to the causes andprevention oftrauma.

Lexington,MA:Lexington Books,1985:421-46. 5 Feyer AM, Williamson AM. A classification system for

causesofoccupational accidents foruse inpreventive

strategies. Scand 7 Work Environ Health 1991;17:

302-11.

6 Occupational injuries.In:DavidB,WilliamF,DennisM,

eds. The NationalCommittee forInjury Preventionand

Control.(US)Injury Prevention: meetingthechallenge.

Am_JPrev Med 1989:177-91.

7 Smith GS.Injuriesas apreventabledisease:the control of

occupationalinjuriesfromthemedicalandpublic health

perspective.Ergonomics1987;30:213-20.

8 DepartmentofHealth,ExecutiveYuan, Taiwan Provincial Health Department, Taipei City Health Department.

Kaohsiung City Health Department. Health and vital statistics:1. General health statistics TaiwanArea,Republic of China, 1985-90. Taipei: Division of Statistics, DepartmentofHealth ExecutiveJuan. 1985-90.

9 Baker SP, O'Neill B, Karpf RS. The injury fact book. Lexington, MA: Lexington Books,1984.

10 Harlan LC, Harlan WR, Parsons PE. The economic

impactofinjuries:amajorsourceofmedicalcosts.AmJ7

Public Health1990;80:453-59.

11 Rice DP, MacKenzie EJ, Jones AS, Kaufman SR, deLissovoy GV, SmithGS, et al. Cost of injury in the

United States:areport to congress.San Francisco:Institute for Health and Aging, University of California and

Injury Prevention Center, The Johns Hopkins University, 1989:37-85.

12 MacKenzieEJ, Siegel JH, Shapiro S, Moody M, Smith

RT. Functionalrecovery and medical costs of trauma:

an analysis bytype and severity of injury. J Trauma

1988;28:281-97.

13 MacKenzie EJ, Shapiro S, Siegel JH. The economic

impactof traumatic injuries: one-yeartreatment-related expenditures.JAMA1988;260:3290-6.

14 SAS Institute.SASISTAT User's Guide,release 6 03edition. Cray, NC: SAS Institute, 1988.

15 SAS Institute. SASISTATSoftware: CALISand LOGIS-TIC procedures, release 6-04. Cray, NC: SASInstitute, 1990:175-230. SAS TechnicalReport P-200. 16 The BureauofLaborInsurancein Taiwan.Theregulations

of labor insurance. Taipei: The Bureau of Labor InsuranceinTaiwan,1989:38-79.