The cost-effectiveness of compulsory bicycle helmets in New Zealand Paul Hansen

Paul A. Scuffham

Department of Economics, University of Otago, Dunedin, New Zealand

Injury Rewention Research Unit, University of Otago Medical School, Dunedin, New Zealand

Abstraa: This paper examines the cost-effectiveness for primary school children (age 5-12 years), sec- ondary school children (15-18 years) and adults (over 18 years) of the legislation enacted on I January 1994 requiring roadcyclists in New Zealand to wear helmets. The cost to cyclists not in possession of a helmet before they became compulsory of either obtaining one or quitting cycling was compared with the number of deaths and hospitalisations expected to be prevented over the average life of a helmet Corresponding to Victorian and United States estimates of the efficacy of cycle helmets at preventing serious head in‘uries, the cost per life saved was $88 379 to $113 744 for primary school children, $694 013 to $814 874 for secondary school children, and $890 041 to $1 014 850 for adults (New Zealand dollars = approximately 0.95 Australian dollars). The cost per hospitalisation avoided was $3304 to $4252, $17 207 to $20 278, and $49 143 to $56 035 respectively. These estimates are extremely sensitive to the estimated efficacy of helmets at protecting cyclists. Mainly anecdotal evidence for New Zealand suggests that they are not be very effective at preventing serious head injuries; future research into the change in injury patterns as a result of the helmet regulation would be valuable. Nonetheless, the rank- ing of the abovementioned estimates does not contradict the policy in some parts of the world requiring helmets for children and/or teenagers, but not adults. (AustJPublic Health 1995; 19: 450-4)

INCE the start of this decade many states and counties in the United States, Canada and S Australia have enacted or have pending legisla-

tion that requires pedal cyclists to wear bicycle hel- mets.’ New Zealand has followed this trend; since l January 1994 it has been mandatory for cyclists in New Zealand to wear approved cycle helmets while riding on the road.4 In New Zealand and Australia this type of regulation applies to cyclists of all ages; however, in parts of the United States and Canada, helmets are compulsory for children and/or teenagers only.’

Deaths and hospitalisations in New Zealand result- ing from head injuries associated with cycle use and the medical costs of the latter’s treatment have been extensively described elsewhere.= In 1990 prices, the average cost of hospital treatment for a skull fracture and intracranial injury respectively were $3760 and $955 (New Zealand dollars = approxi- mately 0.95 Australian dollars).S The incidence of serious head injuries, it has been asserted, ‘provide[s] a compelling argument for requiring the wearing of cycle helmets, both on and off the road’.5 However, two additional considerations are obviously neces sary for evaluating the desirability (from a public policy perspective) of compulsory cycle helmets. First, wearing a helmet does not guarantee the avoidance of head i n j ~ r i e s , ~ ’ ~ and second, requiring every cyclist to wear a helmet is not a costless regu- lation.

This paper brings these three strands together in a cost-effectiveness analysis of compulsory cycle hel-

Correspondence to Paul Hansen, PO Box 56, Dunedin, New Zealand. Fax 64 3 479 8174, e-mail phansen@commerce. otago.ac.nz.

met use in New Zealand. Before the legislation’s introduction, Land Transport New Zealand esti- mated the expected social costs and benefits to New Zealand cyclists as a single group. (Paul Graham, Land Transport New Zealand, personal communica- tion). However, because cyclist participation,16 injurf and voluntary helmet-wearing” rates differ across age groups in New Zealand, a less aggregated focus is appropriate. Hence, the current study esti- mates cost-effectiveness ratios for three groups of New Zealand cyclists: primary (age 5 to 12 years) and secondary (13 to 18 years) school children, and adults (over 18 years).

Methods The rwdel In this study the cost to society of the helmet regula- tion is defined to be the cost to cyclists not in pos session of a helmet before they became compulsory, of either: 1. obtaining one, or 2. quitting cycling. The latter category is included because, as a result of the legislation, some people will have chosen to quit cycling rather than purchase and wear a helmet, thereby losing the benefits to them of cycling. The benefit of the regulation is defined to be the num- ber of deaths and hospitalisations expected to be prevented over the average life of a helmet. For each of the three age groups, the ratio of this cost to the present value of each of these benefits is an estimate of the cost per life saved and hospitalisation pre- vented respectively. Other costs and potential bene- fits are discernible, for example, the cost of the regulation’s enforcement and the likely reduction in both the severity of hospital-treated head injuries and the incidence of head injuries that do not result

450 AUSTRALIAN JOURNAL OF PUBLIC HEALW 1995 va. 19 NO. 5

COMPULSORY BICYCLE HELMETS IN NZ

in death or hospitalisation. These were formally ignored in this study because suitable data were unavailable.

Costs of compulsory cycle helmet use Two distinct groups of cyclists were identified on the basis of how often they rode a bicycle. The numbers in each of the three age groups that do so regularly (defined to be at least once in the four weeks prior to being interviewed) was estimated using popula- tion data from the 1991 census18 and cyclist partic- ipation rates from the 1991 Life in New Zealand Survey.I6 These estimates were adjusted to reflect the population structure at the start of 1994, using data from the Department of statistic^.'^ Thus, there are just over 800 000 ‘regular’ cyclists in New Zealand. However, in 1990 there were 1.2 million bicycles in the country, according to the New Zealand Travel Survey.20 Recognising the possibilities that some of these bicycles are not in use, that some cyclists own more than one, and that some cyclists share bicycles (but probably not helmets), it was assumed that there are a further 200000 people, separable into the three age groups, who ride a bicycle on an irreg- ular basis (less than once a month on average). The numbers of ‘regular’ and ‘irregular’ cyclists not in possession of a helmet were estimated according to the rate of voluntary helmet use for each age group observed in the Ministry of Transport’s Cycle Helmet Survey of September 1993 (three months before helmets became compulsory) . I 7 These people can be regarded as having been affected by the regulation, as, according to this survey, the rate of voluntary helmet use had levelled off over the previous two years.

It was assumed that all of the ‘regular’ and half of the ‘irregular’ cyclists not in possession of a helmet bought one when they became compulsory, while the remainder quit cycling in preference to buying the cheapest helmet available (being less than the police fine for not wearing one). Therefore, the number of helmets that had to be bought multiplied by their average retail price after deducting Goods and Services Tax (a transfer from helmet buyers to the government, and as such not a cost to society as a whole) gave an estimate of the expenditure on hel- mets. The value of cycling to the irregular cyclists assumed to have quit can be regarded as being less than the minimum cost of complying with the regu- lation (or else they would not have quit). Accordingly, the number of cyclists who quit multi- plied by the taxexclusive price of the cheapest hel- met available gave the upper bound of an estimate of the cost (that is, lost benefit) to this group. The total cost to cyclists in each age group of the regula- tion is the sum of the costs to regular and irregular cyclists respectively.

Benefits of compulsory cycle helmet use The Consumers’ Institute of New Zealand estimates that, as a result of natural deterioration in the mate- rials from which cycle helmets are constructed, the protective life of a helmet that is used normally and not damaged in an accident is, on average, about three years.*’ The number of serious head injuries expected to be prevented in New Zealand over this

period (1994-1996) was calculated as follows. The annual number of deaths and hospitalisations that would have occurred in the absence of the regula- tion was assumed to be equal to the average for 1989-91 and 1990-92 respectively (the most recent periods for which disaggregated data were avail- able). This assumption was justified on the expecta- tion that between these periods the increase in the population (and hence, number of cyclists) would have been partially offset by a natural (that is, in the absence of the helmet regulation) decrease in the rate of serious cycle accidents. Such a decrease was observed over the five years to 1992,44 and can possi- bly be explained as the result of improved bicycle design (especially the widespread adoption of mountain bikes) and increasing awareness by cyclists and motorists of road safety issues.

The number of cyclist fatalities and discharges from public hospitals with a head injury as the pri- mary diagnosis, from road (as opposed to off-road) accidents, were identified from the New Zealand Health Information Service (HIS) mortality and morbidity statistics using the ICDMAP definition of a head injury, defined as in Collins et al.5

The number of these events expected to be pre- vented as a result of the regulation was estimated according to the findings of two overseas studies.”J* A population-wide study in Victoria reported that an increase in helmet use from 31 per cent of cyclists in 1990 to 75 per cent in 1991 was associated with a 48 per cent decrease in the number of serious head injuries to cyclists.” A case-conuol study in Seattle reported that the risk of a serious head injury (defined as requiring emergency room treatment) to a cyclist who wore a helmet was 75 per cent (95 per cent confidence interval (CI) 57 to 85) less than the risk to a cyclist who did not.I2 Using the follow- ing approach these findings were used to predict two possible scenarios for New Zealand.

For a given population, let P be the probability of a serious head injury to a cyclist who does not wear a helmet and P(I-h), 0 I h I I , be the probability of such an event to a cyclist who does. Hence, the risk of a cyclist having a serious head injury is qP+(l-q)P(l-h), where q is the proportion of cyclists who do not wear helmets. For a given reduction in q (that is, an increase in the rate of helmet use) and the corresponding reduction in this risk an estimate of h, a measure of the efficacy of helmets at prevent- ing serious head injuries, can be derived. From the finding of the Victorian study reported above, it can be calculated that wearing a helmet reduces the risk to a cyclist of a serious head injury by 81.5 per cent; that is, h = 0.815. The Seattle study directly reports a point estimate of h = 0.75.14

For each estimate of h, and assuming that esti- mates apply to New Zealand cyclists (although P need not), and a given increase in the rate of helmet use (that is, decrease in q), the corresponding decrease in the risk of serious head injury to cyclists, and hence the number of such events, can be derived. Thus, two predictions were made of the number of these events that will be prevented in each age group as a result of helmet use increasing from the rates reported in the September 1993 Cycle Helmet Survey17 to 100 per cent. Discounting and

AUSTRALIAN JOURNAL OF waiic HEALTH 1995 VOL. I 9 NO. s 45 1

HANSEN ETAL

summing each of these (annual) estimates over three years (the life of a helmet, as noted above) gave the present value of the expected benefit of the regulation.

RSults Estimates of the costs and expected benefits of the regulation are presented in Tables 1, 2 and 3. Table 4 presents estimates of the costeffectiveness ratios, which are calculated by dividing the total cost of the policy (column ( I ) in Table 2) by the present value of each estimate of the number of events prevented over the three-year period (the figures in parenthe- ses in Table 3). All prices and costs are in New Zealand dollars as at January 1994.

As reported in Table 4, the cost of preventing the death or hospitalisation of a cyclist increases with the age group to which he or she belongs. Children (5 to 12 years old) are the least expensive group, fol- lowed by teenagers (13 to 18 years old), and then adults.

Discussion As acknowledged, several assumptions were neces sary to arrive at the estimates reported above. Probably the most controversial of these are: 1. the proportions of cyclists who bought helmets or quit cycling when helmets became compulsory, 2. the val- uation of the latter group's 'lost benefit', 3. the pro- tective life of a helmet, and 4. the reduction in the risk of a serious head injury as a result of wearing a helmet.

In the absence of data with which to confront these first two sets of assumptions, discussion is restricted to the sensitivity of the estimates of the cost per head-injury avoided to changes in the

Toble 1 : Estimoted expenditure on helmets for regular cyclists

Proportion Average Expenditure

Number helmets required price ($) ' helmets ($) not wearing Helmets helmet on

Aaearouo a b c=axb d e=cxd

5 to 12 82 180 0.145 1 1 916 26.67 317 803 13 to 18 184 402 0.371 68 413 44.44 3 040 280 19+ 557 435 0.534 297 671 44.44 13 228 478 Note. (a) Estimates of overage helmet prices (excluding Goods and Services Tax)

were provided by Geoff Todd of Browns for Bikes, a moior cycle retailer in Dunedin, on the basis of all helmets sold in the month either side of the helmet regulation coming into effect.

assumptions. Not surprisingly, increasing the pro- portion of cyclists who are assumed to have quit when helmets became compulsory has the effect of lowering the cost estimates for all ages. Also, the esti- mates are relatively insensitive to smaller valuations of the lost benefit to cyclists who quit as a result of the helmet regulation; for example, setting this cost (column (k) in Table 2) to zero reduces the esti- mates of the cost per serious head injury avoided by no more than 5 per cent.

The assumption that the protective life of a hel- met is three years is more open to discussion. Many cyclists probably keep their helmets for longer than this. The actual life of a helmet may also vary accord- ing to the age of cyclists; because children's heads grow as they get older they may replace theirs more frequently than adults do. The estimates reported in Table 4 are sensitive to the assumed average life of a helmet; for example, doubling the life to six years almost halves the costs per serious head injury avoided.

Table 2: Estimated expenditure on helmets and lost benefit for irregular cyclists, and total cost to all cyclists

Proportion Average Expenditure irregular Minimum Lost benefit Total not wearing Helmets helmet on cyclists helmet to irregulor cost to

Age group f b g=fxb/2 d h = g x d i=fxbR I k=ix i l=e+h

5 t o 12 19946 0.145 1 446 26.67 38 567 1 446 13.29 19 218 375 588 1 3 t o 1 8 44757 0.371 8 302 44.44 368 960 8 302 13.29 110 339 3 519 578 19+ 135 297 0.534 36 124 44.44 1 605 364 36 124 13.29 480 092 15 313 934

Number helmets required price ($)" helmets ($) who quit price ($) cyclists ($) 011 cyclists($)

Noks: (a) Estimates of average helmet prices (excluding Goods and SeM'ces Tax) were provided by Geoff Todd of Browns for Bikes, a maior cycle retailer in Dunedin,

(b) Minimum helmet prices were specials obsermd to k an offer at Deka New Zealand Ltd and The Warehouse, twa large general retail stores with branches on the bas is of all helmets sold in the month either side of the helmet regulation corning into effect.

throughout New Zeoland

Table 3: Estimates of the annual number of deaths and hospitalisations expected in the absence of the helmet regulation, and under two scenarios, the reductions expected in risk and number of these events as the result of the regulation

Annual head injuries Yo reduction in Annual in the absence risk of serious Annual deaths hospitalisations of regulation head iniuries prevented prevented

Deaths Hospitalisations Victoria Seattle Victoria Seottle Victoria Seattle Aae a r o w m n O 0 P O q = m x o/lOO r = m x pll00 s=n x o/lOO t=nx pll00

5 to 12 4 107 39.0 30.3 1.6 (4.3)' 1.2 (3.3) 41.7 (113.7) 32.4 (88.3) 13 to 18 3 121 62.1 52.7 1.9 (5.1) 1.6 (4.3) 75.1 (204.5) 63.7 (173.6) '19+ 9 163 70.2 61.6 6.3 (1 7.2) 5.5 (15.1) 114.4 (31 1.6) 100.4 (273.3) N o h : (a) These percentages are calculated from the estimates of h=0.815 and 0.75 from the Victorian and Seattle studies respectively, according to the method explained

(b) Figures in porentheses are the sum of the pmsent wlues of the annual benefits over three yeors (the average life of a helmet) using a discount rate of 5%. in the text.

452 AUSTRALIAN JOURNAL OF PUBLIC HEALTH 1995 va. 19 NO. 5

COMPULSORY BICYCLE HELMETS IN NZ

Table 4: Estimates of the cost (NZ$) per (discounted) life saved and hospitalisation prevented, by age group

Cost per Cost per hospitalisation life saved prevented

h e a r o m Seattle Victoria Seattle Victoria

5 to 12 88 379 113 744 3304 4252 13 to 18 694 013 817874 17207 20 278 19+ 890 041 1 014 850 49 143 56 035

Finally, the estimates of h, the parameter repre- senting the reduction in the risk of a serious head injury as a result of wearing a helmet, are the subject of debate."15 It has been suggested that the Seattle study's estimates reflect, in part, the greater aversion to risk of voluntary helmet wearers, rather than the efficacy of helmets per ~ e . ~ , ' This hypothesis is s u p ported by another American study's finding that helmet wearers are less likely than nonwearers to suf- fer in'uries to parts of the body other than the head?' Similarly, the strength of the Victorian study's reported correlation between the increase in the rate of helmet use and the decrease in serious head injuries to cyclists is weakened by the observed contemporaneous decrease both in cycle use and all road accidents."

Therefore, the estimates used in this study of the number of serious head injuries expected to be pre- vented in New Zealand as a result of compulsory hel- met use may be excessive. This possibility is s u p

5 - 12 years

13 - 18 years

over 18 years

I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I

I I I

10

9

8

7

0

P

5 s

6 , 2 3 - -.

m a - 3

4 1 3 In v

3

2

1

0

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

h (helmet efficacy)

Figure 1 : The relationship between cost per life saved and helmet efficacy

Table 5: Estimated cost per live saved for bicycle helmet regulation compared with other recent regulation a

Enactment Target Cost per a r o w life saved Policy date

Compulsory 1994 5-12 years 96 780 to 124 556 cycle helmets 13-1 8 years 759 984 to 895 620

Fencing of private swimming pools 1987 under 6 years 824 423 to 1 047 924 Compulsory motor cycle helmets 1973 all riders 702 130 Compulsory troctor safety all drivers frames 1970 and possengers 350 01 1 to 468 087

Note: (a) See Propper's Table 1 for the assumptions used in calculating her

estimates. 25

(b) Calculated using a 10 per cent discount rate.

19+yean 974647101 111320

ported by statistics from the Land Transport Fatal Accident Report.*' According to the traffic officers attending the accident, only 3, 1 and 2 of the 16, 17 and 9 cyclists not wearing helmets who were killed (all causes, notjust head injuries) in New Zealand in each of the years 1989, 1990 and 1991 might have survived if they had been wearing a helmet, com- pared to 8 , l l and 2 for whom a helmet would prob- ably not have made a difference. Furthermore, a recent New Zealand study found that after control- ling for the decline in all other injuries to cyclists,

5 - 12 years

13 - 18 years

over 18 years

- _ _ _ _

1

0.9

0.8

0 0.7 8

2 '23

T 0

0.6 $.

%.

- c 0.5 2

3

_. ln

0

m 0.4

a

0.3 $ - 3.

ln - 0.2

0.1

0

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 h (helmet efficacy)

Figure 2: The relationship between cost per hospitalisation pre- vented and helmet efficacy

AUSTRAUAN JOURNAL OF WBUC HEALTH 1995 va. 19 NO. 5 453

HANSEN ET ~ l .

the voluntary increase in helmet use over the four years prior to the introduction of the helmet regula- tion had little effect on the decrease in public hos pital admissions for head injuries (Scuffham and Langley, 'Trends in cycle injury in New Zealand under voluntary helmet use', in preparation).

On the other hand, h, the parameter representing the reduction in the risk of a serious head injury as a result of wearing a helmet, does not capture the likely reduction in the severity of injuries that are treated in hospital (in particular, long-term disabil- ity). As acknowledged above in the Methods section, to the extent that this qualitative effect is ignored, the expected benefits of the helmet regulation are underestimated.

Not surprisingly, the calculated cost-effectiveness ratios are extremely sensitive to changes in the esti- mate of helmet efficacy. Figures 1 and 2 show the effect on the cost-effectiveness ratios of reducing the value of h below the estimate derived from the Victorian study ( h = 0.815); the cost per serious head-injury avoided increases exponentially as the efficacy of helmets decreases.

It is also interesting to compare the results with cost-per-life estimates for other regulations that may be regarded as similar, in the sense that their objec- tive is also to reduce serious injuries and the cost of complying with the regulation falls predominantly on private individuals (as opposed to the health care system). Table 5 presents Propper's estimates, updated to 1994 prices, and those of the current study (using a 10 per cent discount rate, as she did).= Such a crude comparison does not enable policyariented priorities to be established because of the heterogeneity of both the lives saved (espe- cially with respect to age) and the other benefits pro- duced by the regulations, such as injuries prevented, which are not recorded in the table. (For a review of league tables as policy instruments, see Gerard and M~oney.~') Rather, the estimates presented in the table serve to put the present study's estimates in a context.

Conclusion The cost per serious head-injury avoided of requir- ing cyclists to wear helmets is lowest for children, fol- lowed by teenagers and then adults. This qualitative result does not contradict the policy, noted in the introduction, in parts of the United States and Canada, of mandatory helmets for children and/or teenagers, but not adults. Although several studies have found helmets to be highly efficacious,"*l4 anecdotal evidence from New ZealandZ3 and the results of our exploratory study (in preparation) suggest that they may be considerably less so. A clearer picture of the efficacy of helmets and there- fore the present policy's desirability should emerge as time passes. This is of more than retrospective aca- demic interest. Given that the recommended effec- tive life of a helmet is about three years, it will soon be necessary for New Zealand policy makers to revisit the issue of whether, for all age groups of cyclists, the expenditure on cycle helmets and the other costs associated with the regulation are worth incurring for the number of deaths, hospitalisations and other injuries that are prevented.

Acknowledgments This research was undertaken while the second author was a recipient of an Accident Rehabilitation and Compensation Insurance Corporation (ACC) Training Fellowship based in the Injury Prevention Research Unit. The unit is fundedjointly by the ACC and the Health Research Council of New Zealand. The authors wish to thank the Land Transport Safety Authority for the provision of the helmet-wearing data and Nancy Devlin, Paul Graham, John Langley and an anonymous referee for their useful com- ments on earlier drafts of this paper. The analysis and conclusions drawn in this paper are the respon- sibility of the authors. References

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Hadlines-the tuwslefter ofthe WHO helmet initiative. Vol. 2(3). Atlanta: Centre for Iniurv Control. Emory Universitv School of Public Health, 1995. ' Traffic regulations 1976, amendment no. 24. Wellington: New Zezand "Government, -14 June 1993. Phillips DE, Langley JD, Marshall SW. Injury: the medical and related costs in New Zealand 1990. NZ Med J 1993; 106 21S18. Langley JD, Phillips DE, Marshall SW. Inpatient costs of injury due to motor vehicle traffic crashes in New Zealand. Accid Anal An, 1993; 25: 585-92. Collins BA, Langley JD, Marshall SW. Injuries to pedal cyclists resulting in death and hospitalisation. NZ Medj1993; 1 06514-17. Trippe, HR Helmets for pedal cyclists. EMJ 1992; 305 843-4. McCarthy M. Cycling without helmets. BMJ 1992; 305:

Illingworth I. The argument for helmets. BMJ 1992; 305:

Williams M. The protective performance of bicyclist helmets in accidents. Accid Anal An, 1991; 23: 119-31. Spaite DW, Murphy M, Criss FA, Valenzuala TD, Meislin HW. A prospective analysis of injury severity among hel- meted and nonhelmeted bicyclists involved in collisions with motor vehicles. JTrauma 1991; 31: 1510-16. Finch CF, Newstead SV, Cameron, MH, Vulcan AP. H a d injuv reductions in Victoria two years a@ the introduction of mandatory biqcle helmet use. Report no. 51. Melbourne: Monash University Accident Research Centre, 1993. Thompson RS, Rivera FP, Thompson DC. A case-control studv of the effectiveness of bicvcle safea helmets. N End I

I

881-2.

882-3.

_ _ I

Med'l989; 320: 1361-7. Thomas S, Acton C, Nixon 1, Battistutta D, Pitt WR, Clark R. Effectiveness of bicycle helkets in preventing head injury in children: casexontrol study. EMJ 1994; 308: 173-6. Baker SP, Li G, Fowler C, Dannenberg AL. Znjurics to biqclists: a national pmpctiue. Baltimore: Johns Hopkins University Injury Prevention Centre, 1993. Pitt WR, Thomas S, Nixon J, Clark R, Battistutta D, Acton C. Trends in head injuries among child bicyclists. BMJ 1994; 308: 177. Russell DG, Wilson NC. Lye in New Zmlond Commission nport volume I : aunc&vcouovicu Dunedin: University of Otago, 1991. Cycle helmet suruq. Wellington: Land Transport Safety Authority, Ministry of Transport, 1993. 1991 census of population and dwellings, New Zcaland 5 popula- tion sttucture. Wellington: Department of Statistics, 1991. Kq statistics, December 1993. Wellington: Department of Statistics, 1993. New Zcaland HourehoM Travel Suruq, J u ~ 198GJune 1990. Wellington: Land Transport Division, Ministry of Transport, 1992. Head first choosing the best cycle helmet for you. Consumer (Wellington) 1992; 301: 1618. Motos accidents in New Zcaland. Statistical statement fm the calen- dar year 1992. Wellington: Land Transport Safety Authority, Ministry of Transport, 1992. Land Transport fatal accident report Wellington: Land Safety Transport Authority, 1989, 1990, 1991. Gerard K, Mooney G. QALY league tables: handle with care. Health Econ 1993; 2: 59-64. Propper C. Valuing human 15' in New Zmlank a tuw methodology and an economic eualuation of pmn'ously used techniques. Wellington: Road Research Unit, National Roads Board, 1988.

AS4 AUSTRALIAN JOURNAL OF WBUC HEALTH 1995 VOL. 19 NO. 5