Quality-Adjusted Life-Years and Helmet Use Among
Motorcyclists Sustaining Head Injuries
Hsin-Yi Lee, PhD, Yeh-Hsin Chen, MS, Wen-Ta Chiu, MD, PhD, Jing-Shiang Hwang, PhD, and Jung-Der Wang, MD, ScD
Motorcycles have become one of the most popular means of transportation in many de-veloping countries, particularly in Asia, be-cause of their convenience, affordability, and relatively high fuel efficiency.1,2However, rela-tive to car drivers, motorcyclists are more easily injured, and are more often killed as a result of even minor collisions with other, larger vehicles.
According to traffic accident data from Taiwan’s National Police Agency,3more than 1100 motorcyclists were killed on the country’s roads in 2007, accounting for 59.7% of all traffic accident deaths. Furthermore, 57.1% of all mo-torcycle-related fatalities involved injuries to the head, thereby indicating a distinct correlation between motorcycle-related injury and head in-jury.
In a recent review of traffic accidents in-volving motorcyclists, Liu et al.4concluded that if motorcyclists are wearing a helmet at the time of the crash, their risk of death is reduced by 42%, and their risk of head injury is reduced by 69%. These findings have been corroborated in many other studies indicating that, after imple-mentation and enforcement of laws requiring motorcyclists to wear helmets, there are clear reductions in motorcycle-related head injuries, the severity of such injuries, and the overall length of patient hospital stays.5–7Nevertheless, relatively few studies have set out to investigate the long-term effects of helmet protection among motorcyclists sustaining head injuries.
Many countries with national health insur-ance systems have been increasingly faced with a heavy financial burden, essentially as a result of an aging population and new forms of technology (e.g., gene therapy, positron emis-sion tomography). Most of these countries have adopted economic analyses as a means of containing costs associated with medical care and increasing overall cost-effectiveness by improving the health outcomes achieved per dollar spent.8One of the most basic methods used to achieve such improvements is that of evaluating the cost per quality-adjusted life-year
(QALY) gained from different health care ser-vices, an approach applied to ensure that pre-ventive measures are competitive by saving more lives than would treatment after an illness occurs. The fundamental concept involves adjusting the survival function with the mean quality of life at each time point t and then summing this adjusted value over a lifetime. The resulting measure, ‘‘quality-adjusted life expec-tancy’’ (QALE), is expressed in the following equation9–13:
ð1Þ QALE ¼ Z
E½QolðtjxÞ SðtjxÞ dt;
where S(t|x) denotes the survival function for condition x at time t and Qol(t|x) denotes the quality-of-life function for condition x at time t. If an age- and gender-matched referent can be simulated from the vital statistics for every case of condition x, one should be able to calculate the QALE for a general referent (in the present case, if a head injury had not occurred).
The difference between the QALE for head injury cases and referents just noted would be the expected loss of QALE for an average head injury case. In this study, we sought to quantify QALE and expected loss of QALE among motorcyclists in Taiwan who sustained head injuries while wearing or not wearing a helmet and to determine life-years saved as a result of helmet use in such cases.
METHODS
Our primary data source was the Head Injury Registry,14a system designed to monitor the epidemiology of traumatic head injuries in Taiwan. The registry includes in the definition of head injuries: brain concussions, contusions, skull bone fractures, brain damage with clear neuro-logical deficits, clinically observable cognitive deficits, posttraumatic amnesia, neurological se-quelae, and any evidence of intracranial hemor-rhage. Head-injured motorcyclists are defined as motorcycle riders or passengers who, after Objectives. We estimated loss of quality-adjusted life expectancy (QALE) among motorcyclists in Taiwan who sustained head injuries while wearing or not wearing a helmet.
Methods. Patients with head injuries (n = 3328) were grouped into categories representing good and poor outcomes (moderate disability or death) at dis-charge. After linkage with the National Mortality Registry, survival functions were determined and extrapolated over a 50-year period on the basis of the survival ratio between patients and age- and gender-matched reference pop-ulations, as calculated from available Taiwan vital statistics. Survival functions were then multiplied by scores from quality-of-life measures.
Results. Percentages of good and poor outcomes were 87.2% and 12.8%, respectively, in the helmeted group and 66.4% and 33.6% in the nonhelmeted group. The mean QALE for helmeted motorcyclists, calculated by weighting percentages of good and poor outcomes, was 31.7 quality-adjusted life-years (QALYs), with an average loss of 5.8 QALYs. For nonhelmeted motorcyclists, the mean QALE was 25.9 QALYs, with a loss of 10.7 QALYs.
Conclusions. Helmet use could save approximately 5 QALYs among motor-cyclists sustaining head injuries. Future cost-effectiveness analysis can calculate the incremental cost-effectiveness ratio for regulation of helmet use. (Am J Public Health. 2010;100:165–170. doi:10.2105/AJPH.2008.159004)
having sustained direct or indirect trauma to the head, exhibit one or more of the aforementioned diagnoses.
Data on head injuries were recorded by experienced neurosurgeons from each of the hospitals included in the study and extracted by a research assistant from the Injury Pre-vention Center at Taipei Medical University to ensure consistent data quality. The procedure for ascertaining each case involved a careful review of inpatient medical records, examina-tion of surviving motorcyclists with head in-juries admitted to the hospital, review of death certificate data to determine the number of patients who died during the course of their treatment, and review of the deaths that oc-curred outside the hospital (e.g., at the scene of the accident) through the National Mortality Registry. Registry certificates were issued once a forensic examination indicated that brain injury was the primary cause of death.
Data on 3328 patients hospitalized as a re-sult of motorcycle-related head injuries were collected from 22 major hospitals in Taipei City and 4 hospitals in Hualien County between 2001 and 2007. The Glasgow Outcome Scale15 was used to categorize head injury outcome at the time of hospital discharge as follows: death, persistent vegetative state, severe disability (con-scious but dependent), moderate disability (dis-abled but independent), or good recovery.
Although our primary aim was to investigate the difference between helmeted and nonhel-meted motorcyclists in terms of loss of QALYs, the outcome of the patient at discharge was considered to be the most important prognostic factor. Thus, in accordance with the Glasgow Outcome Scale, we divided our head injury cohort into 2 categories: those with good out-comes (good recovery) and those with poor outcomes (ranging from moderate disability to death). We used these categories to subse-quently estimate and extrapolate survival functions and quality-of-life functions.
Extrapolation of Long-Term Survival
After verifying the survival status of patients by cross linking them with the updated data-base of the National Mortality Registry at the end of the follow-up period (December 31, 2007), we used the Kaplan–Meier method to estimate survival functions based on 2001 to 2007 follow-up data. Because a high
percentage of these patients were still alive at the end of follow-up (survival rates were 92.6% and 85.5% for the helmeted and non-helmeted groups, respectively), we estimated the lifetime survival function by incorporating life expectancy data from the general popula-tion of Taiwan.11–13,16
We assumed that head injuries might pro-duce excess mortality, which could be quanti-fied from following-up the cohort for a short period of time (5 years). If this excess mortality appeared to be a constant hazard in the chronic stage (1 or 2 years after the head injury), the long-term survival of our cohort could be projected from a survival function for an age-and gender-matched reference population.11,12 The survival function of this reference popula-tion can be generated, via the Monte Carlo method, from national vital statistics life tables.12
Also, we were able to obtain expected and mean quality-of-life values at each study time point by conducting a cross-sectional survey among the patients; these data served to ad-just the survival curve.13In previous studies involving simulation11and mathematical16 methods, we have shown that this is a valid means of predicting life expectancy, even when the censored rate is above 50%; it has also been corroborated by several real-world ex-amples.16–19
Quality-of-Life Values
We used the EuroQol 5-dimensional (EQ-5D) questionnaire, a preference-based, generic, self-reported instrument,20,21to estimate qual-ity-of-life values for helmeted and nonhelmeted motorcyclists who sustained head injuries. This instrument has been extensively used as an outcome measure22–24; the Chinese version was developed in Taiwan, along with the value categorization used in this study.25
The 5 dimensions encompassed in the questionnaire are mobility, self-care, usual ac-tivities, pain–discomfort, and anxiety–depres-sion, with 3 levels of severity (no problems, some or moderate problems, and severe or extreme problems). The resulting utility value is coded from 0 to 1 (on the basis of the 5-dimensional classification), where 1 indicates optimal health.21
Using computer-generated random num-bers, we selected a cross-sectional sample of 600 surviving patients from the Head Injury
Registry. Telephone contact numbers were obtained from the members of this sample when they were admitted to the hospital; survivors were subsequently contacted by telephone after their discharge to determine their willingness to participate in our study. In cases in which patients were unable to answer the questionnaire, their family members were invited to provide the necessary information.
Between January and March 2008, 190 patients were successfully interviewed (re-sponse rate: 31.7%). Excluding the 20 patients who refused to be interviewed, most of the others who did not respond had changed their telephone number or could not be reached after 3 calls. To examine the representativeness of the interviewed sample, we conducted a supplementary analysis in which we com-pared selected characteristics (e.g., age, gender, severity of injury) of those who were inter-viewed, those who could not be reached, and those who refused to be interviewed.
Also, we asked patients who had sustained their injury in 2007 to recall their health condition at the time of their discharge from the hospital, given that their injuries had occurred within a short period before their interview (less than 1.2 years) and recall bias would thus be minimal. As a result, these patients could contribute 2 or more data points of quality of life to our study. We used smoothing methodology to estimate differences in duration of time after occurrence of head injury, and we assumed that the utility value for all hypothetical referents was 1 throughout the survival of the sample patient.
Calculation of Quality-Adjusted Life Expectancy
After calculating lifetime survival functions for the groups with good and poor outcomes, we adjusted these values according to the corresponding quality-of-life function and the period of time after the occurrence of the head injury to calculate QALE on the basis of an 80-month follow-up period and 50 years of extrapolation.13Given that the mean age of our cohort was approximately 37 years and the average life expectancy of the Taiwanese pop-ulation in 2007 was 78.4 years, we believed that a 50-year extrapolation could approximate life-long utility loss among patients sustaining a head injury.
To facilitate our analyses, we used the software program MC-QAS (http://www.stat. sinica.edu.tw/jshwang), built on the R statistical package. The validity of this program in ex-trapolating long-term survival data has been established in previous studies, including those focusing on permanent disability from occu-pational injuries,17HIV infection,16and cancer.19 We calculated average QALE values for hel-meted and nonhelhel-meted motorcyclists by weighting the respective percentages of good and poor outcomes and then summing these per-centages.
RESULTS
The basic characteristics of the 3328 mem-bers of our cohort are summarized in Table 1. There were no significant age or gender dif-ferences between helmeted and nonhelmeted motorcyclists. However, the rate of helmet use was much higher in urban Taipei City (71.3%) than in rural Hualien County (28.7%).
Nonhelmeted motorcyclists had longer average intensive care unit stays and worse outcomes at discharge than helmeted motor-cyclists. No significant demographic or helmet use differences were observed among patients who were interviewed, those who could not
be reached, and those who refused to be interviewed (Table 2 ), although the proportion of good outcomes at discharge appeared to be slightly lower among respondents than among nonrespondents. Finally, Taipei City residents made up a higher percentage of the interviewed group (70.5%) than they did of the other groups (approximately 50%).
An independent-sample t test showed that the mean utility score for poor outcomes was 0.7 (SD = 0.3; P < .01), significantly lower than the mean utility score for good outcomes (0.9; SD = 0.2). Patients with good outcomes were also found to have a longer QALE than those with poor outcomes, as illustrated in Figure 1.
Life expectancies and QALE values, by discharge outcome and helmet use, are sum-marized in Table 3. After the proportions of the 2 outcome categories were weighted among helmeted and nonhelmeted motorcyclists (Ta-ble 1), the QALE of the typical helmeted motorcyclist after 50 years of extrapolation was 31.7 QALYs, representing a loss of 5.8 QALYs relative to a hypothetical age- and gender-matched reference group. Among non-helmeted motorcyclists, the estimated QALE after 50 years of extrapolation was 25.9 QALYs, with a greater loss (10.7 QALYs) relative to the reference group.
DISCUSSION
Although previous studies have documented the protective effects of helmets in reducing mortality and improving short-term outcomes among motorcyclists,4–7,26relatively few of these studies have focused on the long-term health benefits of wearing a helmet. One of the major challenges faced in such research is that most patients who have sustained head injuries survive for long periods. If a head injury cohort is not followed for a sufficiently long period, the result is often a high censored rate (above 50%), and the research team would not be able to obtain the lifetime survival function for estima-tion of QALE required by the convenestima-tional method.27To the best of our knowledge, ours is the first study of its kind to estimate the lifelong utility losses of helmeted and nonhel-meted motorcyclists sustaining head injuries. To determine the validity of our basic assump-tion, a constant excess mortality rate associated with head injuries, we initially estimated the QALE of patients with head injuries according to outcome at discharge. This strategy helped ensure the homogeneity of our subcohorts, with a stable survival ratio relative to the refer-ence population after the first several years of follow-up.
The percentage of good outcomes was higher among helmeted motorcyclists (87.2%) than among nonhelmeted motorcyclists (66.4%), and thus the weighted average QALE loss in the helmeted group was below that of the nonhelmeted group (5.8 versus 10.7). In other words, use of a helmet can save an average of approximately 5 QALYs among individuals sustaining head injuries as a result of motorcycle accidents.
In an earlier study, Tsauo et al.28found that the average QALE loss for a single head injury case was 4.8 QALYs, lower than the values we found (Table 3). The difference in these findings may be attributable to a pair of factors. First, there was a 12-year gap in the data periods used in the 2 studies. The QALE loss in our study relative to the reference population might have been slightly increased because there was a 2-year increase in the average life expectancy of the general popula-tion29at the time of our study as compared with the average at the time of the Tsauo et al. study.
TABLE 1—Frequency Distributions in the Helmeted (n = 2879) and Nonhelmeted (n = 449) Patient Groups and Between-Group Comparisons: Taiwan, 2001–2007
Characteristic Helmeted, % or Mean (SD) Nonhelmeted, % or Mean (SD) Between-Group Comparison P Age, y 36.6 (16.9) 38.4 (19.6) .06a
No. of days in intensive care unit 1.8 (4.8) 2.9 (5.7) <.01b
No. of days in general hospital ward 8.6 (12.5) 9.7 (13.9) .31b
Male 60.9 65.0 .09c Location of accident <.001c Taipei City 71.3 45.7 Hualien County 28.7 54.3 Outcome at discharge <.001d Good recovery 87.2 66.4 Moderate disability 4.6 9.6 Severe disability 2.6 9.3 Vegetative state 1.0 3.8 Death 4.6 10.9 aBy the t test. b
By the Mann–Whitney test. cBy thec2test. d
Second, Tsauo et al.28included only cases occurring within Taipei City, where the rate of good outcomes at discharge (88.2%) tends to be higher than that of our cohort from Taipei and
Hualien. Although they did not report the percentage of helmet use in their study popula-tion, the rate of good outcomes in their study is comparable to the rate among helmeted
motorcyclists in our study (87.2%) and much higher than that among nonhelmeted motorcy-clists (66.4%), whose life expectancy would pre-sumably be shorter.
Study Limitations and Strengths
One of the major limitations of our study was the low response rate (31.7%). As mentioned, to examine the representativeness of the sample, we compared the characteristics of respondents and nonrespondents. Although nonrespondents were more common in Hualien County, there were no significant between-group differences in terms of age, gender, number of days spent in intensive care units or general wards, helmet use rates, and distribution of years of admittance (Table 2). Most important, there were no dif-ferences with respect to the major determinant of both survival and quality of life: outcome at time of discharge. Thus, our sample appears to be relatively representative.
Another possible limitation is the validity of the quality-of-life measure among patients with severe head injuries, given that such patients suffer from cognitive impairment.30,31During the interviews, only 5 patients were unable to answer the questionnaire because they had been in a vegetative state since their accident. In these instances, we invited patients’ family members (caregivers) to answer the questionnaire if they agreed, and they usually assigned the lowest score in all 5 EQ-5D dimensions.
Under such conditions, it was difficult to accurately validate the quality of life per-ceived by patients themselves and compare the results with those reported by caregivers. Nonetheless, differences may have been rela-tively small given that the EQ-5D included only 3 levels of severity (no problems, some or moderate problems, severe or extreme prob-lems). Moreover, considering that these pa-tients accounted for only 2.6% of the papa-tients interviewed, we believe that the potential overall bias is small.
It should be kept in mind that speed limits in Taipei City and Hualien County were 50 km/h and 60 to 70 km/h, respectively, and motor-cycles were prohibited on freeways. Therefore, percentages of poor outcomes in our study areas may have been lower than those in areas with higher speed limits. Higher speeds can produce more severe injuries32as well as greater loss of health-related utility.
TABLE 2—Comparison of Characteristics Among Patients With Head Injuries Who Were Interviewed (n = 190), Who Could Not Be Contacted (n = 390), and Who Refused to Be Interviewed (n = 20): Taiwan, 2001–2007 Characteristic Interviewed, % or Mean (SD) Not Contacted, % or Mean (SD) Refused, % or Mean (SD) Between-Group Comparison P Age, y 36.7 (17.9) 39.4 (19.5) 42.7 (20.1) .19a
No. of days in intensive care unit 3.5 (7.2) 3.3 (7.7) 3.5 (5.7) .94a No. of days in general hospital ward 8.6 (12.9) 8.4 (19.1) 10.9 (23.5) .81a
Male 59.5 62.3 55.0 .68b Location of accident <.001b Taipei City 70.5 51.3 50.0 Hualien County 29.5 48.7 50.0 Outcome at discharge .22b Good recovery 76.8 80.8 70.0 Moderate disability 9.5 5.6 5.0 Severe disability 5.8 5.6 5.0 Vegetative state 2.6 2.3 0.0 Missing 5.3 5.6 20.0 Helmet use 85.5 79.6 89.5 .2b Year admitted .2b 2001–2002 5.8 3.8 5.0 2003 21.6 23.8 15.0 2004 15.3 16.9 30.0 2005 20.0 16.2 15.0 2006 26.8 30.3 10.0 2007 10.5 9.0 25.0 a By analysis of variance. bBy thec2test.
FIGURE 1—Quality-adjusted life expectancy among patients with head injuries and the reference population after 50 years of extrapolation, by outcome at discharge: Taiwan, 2001–2007.
Conclusions
Use of a helmet can save an average of approximately 5 QALYs among individuals sustaining head injuries as a result of motor-cycle accidents. Economic evaluations (e.g., cost-effectiveness and cost-utility analyses) generally involve a comparative analysis of alternative courses of action in terms of their costs and consequences.8By quantifying con-sequences, we successfully estimated the expected utility, in terms of QALYs, that would have accrued to motorcyclists wearing a helmet during an accident in which they sustained a head injury. Such results can be applied in future assessments of the value of mandatory helmet legislation for motorcycle riders.
The next step involves identifying the dif-ference in costs (e.g., to the health care sector, to patients and families) associated with man-datory legislation and lack of regulation. Future cost-utility or cost-effectiveness analyses can calculate the incremental cost-effectiveness ra-tio for regulara-tion of helmet use (i.e., the cost per QALY gained by wearing a helmet). If the human capital or productivity saved by pre-vention can be converted to monetary values, cost-benefit analyses can also be performed by
summing direct medical costs and indirect productivity losses.33j
About the Authors
Hsin-Yi Lee, Yeh-Hsin Chen, and Jung-Der Wang are with Institute of Occupational Medicine and Industrial Hygiene, College of Public Health, National Taiwan University, Taiwan, Republic of China. Wen-Ta Chiu is with the Institute of Injury Prevention and Control, Taipei Medical University, Taiwan. Jing-Shiang Hwang is with the Institute of Statistical Science, Academia Sinica, Taiwan. Jung-Der Wang is also with the Department of Internal Medicine and the Department of Environmental and Occupational Medicine, National Taiwan University Hospital.
Correspondence should be sent to Jung-Der Wang, MD, ScD, Institute of Occupational Medicine and Industrial Hygiene, College of Public Health, National Taiwan Uni-versity, Room 719, No. 17, Xu-Zhou Road, Taipei 10055, Taiwan (e-mail: jdwang@ntu.edu.tw). Reprints can be ordered at http://www.ajph.org by clicking the ‘‘Reprints/ Eprints’’ link.
This article was accepted May 6, 2009.
Contributors
H. -Y. Lee conducted the data analyses and contributed to the writing of the article. Y. -H. Chen assisted in the quality-of-life assessment and conducted the initial sta-tistical analyses. W. -T. Chiu contributed to the collection of quality-of-life data. J. -S. Hwang supervised the statis-tical analyses. J. -D. Wang originated the study, super-vised the analyses, and led the writing.
Acknowledgments
Partial support for this study was provided by grants from the Bureau of Health Promotion, Taiwan Depart-ment of Health (DOH96-HP-1701), and from the National Science Council, Taiwan (NSC97-2811-B-002-064).
We express our sincere appreciation to Shu-Fen Chu and all of the staff in the 26 collaborating hospitals for their full support in providing complete and invaluable data.
Human Participant Protection
This study was approved by the National Taiwan Uni-versity institutional review board. Participants provided verbal informed consent before their interviews.
References
1. Hsu T. Comparative study on motorcycle ownership forecasting: model of Asian Conutries–Taiwan, Malaysia and Vietnam. J East Asia Soc Transport Stud. 2007; 7:105-115.
2. Lai W. Modeling the working mode choice, owner-ship and usage of car and motorcycle in Taiwan. J East Asia Soc Transport Stud. 2007;7:869-885.
3. National Police Agency. Police Policy Report: Taipei. Taiwan, Republic of China: Ministry of the Interior; 2007.
4. Liu BC, Ivers R, Norton R, et al. Helmets for preventing injury in motorcycle riders. Cochrane Data-base Syst Rev. 2008;1:CD004333.
5. Chiu WT, Huang SJ, Tsai SH, et al. The impact of time, legislation, and geography on the epidemiology of
TABLE 3—Life Expectancy and Quality-Adjusted Life Expectancy (QALE) in Years, by Outcome at Discharge and Helmet Use: Taiwan, 2001–2007
Life Expectancy QALE
80-Month Follow-Up, Mean or Mean (SE)
50-Year Extrapolation, Mean or Mean (SE)
80-Month Follow-Up, Mean or Mean (SE)
50-Year Extrapolation, Mean or Mean (SE)
Good outcome at discharge
Head injury patients 6.4** (0.03) 36.7* (6.2) 5.9** (0.5) 35.3** (5.8)
Reference population 38.0* (0.02) 38.0** (0.02)
Difference 1.3 2.8
Poor outcome at discharge
Head injury patients 4.1** (0.2) 9.6* (5.6) 2.9** (0.2) 7.6** (4.6)
Reference population 34.0* (0.03) 33.9** (0.7)
Difference 24.4 26.3
Helmeted motorcyclists with head injury
87.2% good outcome + 12.8% bad outcome 6.1 33.2 5.5 31.7
Expected life loss relative to reference population 4.3 5.8
Nonhelmeted motorcyclists with head injury
66.4% good outcome + 33.6% bad outcome 5.7 27.6 4.9 25.9
Expected life loss relative to reference population 9.1 10.7
*P < .01; **P < .001, for between-group comparison of good versus poor outcome at discharge. Between-group differences were assessed via the log-rank test (80-month follow-up) and the z statistic (50-year extrapolation).
traumatic brain injury. J Clin Neurosci. 2007;14(10): 930-935.
6. Chiu WT, Kuo CY, Hung CC, Chen M. The effect of the Taiwan motorcycle helmet use law on head injuries. Am J Public Health. 2000;90(5):793-796. 7. Keng SH. Helmet use and motorcycle fatalities in Taiwan. Accid Anal Prev. 2005;37(2):349-355. 8. Drummond MF, Torrance GW, O’Brien BJ, Stoddart GL. Methods for the Economic Evaluation of Health Care Programmes. 3rd ed. New York, NY: Oxford University Press; 2005.
9. Glasziou PP, Cole BF, Gelber RD, et al. Quality adjusted survival analysis with repeated quality of life measures. Stat Med. 1998;17(11):1215-1229. 10. Shen LZ, Pulkstenis E, Hoseyni M. Estimation of mean quality adjusted survival time. Stat Med. 1999; 18(12):1541-1554.
11. Hwang JS, Tsauo JY, Wang JD. Estimation of expected quality adjusted survival by cross-sectional survey. Stat Med. 1996;15(1):93-102.
12. Hwang JS, Wang JD. Monte Carlo estimation of extrapolation of quality-adjusted survival for follow-up studies. Stat Med. 1999;18(13):1627-1640. 13. Hwang JS, Wang JD. Integrating health profile with survival for quality of life assessment. Qual Life Res. 2004;13(1):1-10.
14. Chiu WT, Yeh KH, Li YC, et al. Traumatic brain injury registry in Taiwan. Neurol Res. 1997;19(3):261-264. 15. Jennett B, Bond M. Assessment of outcome after severe brain damage: a practical scale. Lancet. 1975; 305(7905):480-484.
16. Fang CT, Chang YY, Hsu HM, et al. Life expectancy of patients with newly-diagnosed HIV infection in the era of highly active antiretroviral therapy. QJM. 2007; 100(2):97-105.
17. Ho JJ, Hwang JS, Wang JD. Estimation of reduced life expectancy from serious occupational injuries in Taiwan. Accid Anal Prev. 2006;38(5):961-968.
18. Ho JJ, Hwang JS, Wang JD. Life expectancy estima-tions and the determinants of survival after 15 years of follow-up on 81,249 workers with permanent occupa-tional disabilities. Scand J Work Environ Health. 2006;32(2):91-99.
19. Chu PC, Hwang JS, Wang JD, Change YY. Estimation of life expectancy and the expected years of life lost in patients with major cancers: extrapolation of survival curves under high-censored rates. Value Health. 2008;11(7):1102-1109.
20. EuroQol Group. EuroQol: a new facility for the measurement of health-related quality of life. Health Policy. 1990;16(3):199-208.
21. EuroQol Group. EQ-5D Value Sets: Inventory, Com-parative Review and User Guide. Dordrecht, the Nether-lands: Springer; 2007.
22. Brunner-La Rocca HP, Kaiser C, Bernheim A, et al. Cost-effectiveness of drug-eluting stents in patients at high or low risk of major cardiac events in the Basel Stent KostenEffektivitats Trial (BASKET): an 18-month anal-ysis. Lancet. 2007;370(9598):1552-1559.
23. Burstrom K, Johannesson M, Rehnberg C. Deterio-rating health status in Stockholm 1998–2002: results from repeated population surveys using the EQ-5D. Qual Life Res. 2007;16(9):1547-1553.
24. Scuffham PA, Whitty JA, Mitchell A, et al. The use of QALY weights for QALY calculations: a review of in-dustry submissions requesting listing on the Australian Pharmaceutical Benefits Scheme 2002–4. Pharmacoeco-nomics. 2008;26(4):297-310.
25. Hsieh CLWJ. Estimating utility values for health states of persons in Taiwan using the EQ-5D. Paper presented at: Symposium of Quality of Life, December 2007, Taiwan, Republic of China.
26. Coben JH, Steiner CA, Miller TR. Characteristics of motorcycle-related hospitalizations: comparing states with different helmet laws. Accid Anal Prev. 2007; 39:190-196.
27. Muennig PA, Gold MR. Using the years-of-healthy-life measure to calculate QALYs. Am J Prev Med. 2001;20:35-39.
28. Tsauo JY, Hwang JS, Chiu WT, Hung CC, Wang JD. Estimation of expected utility gained from the helmet law in Taiwan by quality-adjusted survival time. Accid Anal Prev. 1999;31(1):253-263.
29. Abridged Life Tables of Taiwan. Taipei, Republic of China: Ministry of the Interior; 2005.
30. Riemsma CA, Glanville JM, Eastwood AJ, et al. General health status measures for people with cognitive impairment: learning disability and acquired brain injury. Health Technol Assess. 2001;5:1-100.
31. von Steinbuechel N, Richter S, Morawetz C, et al. Assessment of subjective health and health-related qual-ity of life in persons with acquired or degenerative brain injury. Curr Opin Neurol. 2005;18(6):681-691. 32. Richter M, Otte D, Lehmann U, et al. Head injury mechanisms in helmet-protected motorcyclists: prospec-tive multicenter study. J Trauma. 2001;51(5):949-958. 33. Kenkal D. Cost of illness approach. In: Todley G, Kenkel D, Fabian R, eds. Valuing Health for Policy: An Economic Approach. Chicago, IL: University of Chicago Press; 1994:42-71.
