Regulatory Proposal for Reducing Fire Risks from Cigarettes

Appendix D

Health Care Cost Values

This appendix describes the derivation of the health care cost (HCC) values employed to estimate the benefits of a reduction in cigarette-fire injuries.  The values are based on data originally presented by Miller, et al. in a 1993 study of the societal costs of cigarette-ignited fires in the United States.⁹⁶  The appendix discusses the limitations of these data and of the HCC approach in general, and addresses questions that may arise concerning the applicability of U.S. health care data to a Canadian context.

Injuries to Fire-Fighters

The benefits analysis employs an HCC estimate of $1,679 (2002 $CAD) to value a reduction in the number of non-fatal injuries to fire-fighters.  This value is obtained directly from Miller, et al.⁹⁷  The authors do not provide a detailed explanation for the derivation of this figure, which is much lower than the cost of treatment estimates they provide for serious injuries to civilians, nor do they provide more detailed estimates of the cost of treating fire-fighters for injuries of differing type or severity.  Fire-fighters, however, are less likely than civilians to sustain serious burns or other injuries as a result of a fire, since they approach the fire alert, well-protected, and well-trained.  As a result, it is reasonable to expect that average treatment costs would be relatively low.

Injuries to Civilians

The analysis of the benefits of reducing non-fatal injuries to civilians relies on a more detailed assessment of treatment costs.  The values employed vary depending upon the severity of the injury and, in some cases, the nature of the injury.  The derivation of these values is described below.

Serious Injuries Requiring Hospitalisation

Miller, et al. provide treatment cost data for three categories of non-fatal injuries requiring hospitalisation: burns, including burns in conjunction with anoxia; anoxia; and other injuries associated with cigarette fires.  These costs include post-admission payments for care in both hospitals and nursing homes.  In addition, Miller et al. provide a general cost estimate for initial treatment in an emergency room.  The analysis adds this figure to the post-admission values provided for each type of injury to develop the values applied in calculating the benefits of a reduction in serious injuries requiring hospitalisation.  Exhibit D-1 provides additional detail on the derivation of these values.

Exhibit D-1 Health Care Cost for serious Injuries to Civilians Injuries Requiring Hospitalisation (2002 $CAD)

Cost Element	Anoxia	Burns	Other Injuries
Emergency Room a	$459	$459	$459
Hospitalisation	$7,318 b	$78,279 c	$21,504 d
Total	$7,777	$78,738	$21,963

^a Miller, et al. (1993):  $299 (1992 $USD).  Excludes follow-up care.
^b Miller, et al. (1993):  $4,764 (1992 $USD).
^c Miller, et al. (1993):  $50,963 (1992 $USD).
^d Miller, et al. (1993):  $14,000 (1992 $USD).

Serious Injuries Requiring No Hospitalisation

The benefits analysis employs two estimates of the medical costs of treating serious civilian injuries that do not require hospitalisation.  The value for burns and anoxia, $1,072 per case (2002 $CAD), is based on Miller, et al.'s estimate of the average cost for the treatment of burns in an emergency room, including the cost of follow-up care.⁹⁸  Miller, et al. apply this value to anoxia as well as burn cases; the analysis of the benefits of the proposed ignition propensity cigarette standard follows this approach.  Miller, et al., however, do not estimate the cost of emergency room treatment and follow-up care for other types of injury.  Instead, they develop a general estimate of the cost of "other treatment," which equates to $791 (2002 $CAD).⁹⁹  The analysis applies this figure to estimate the benefits of reducing "other" injuries that do not require hospitalisation.

Minor Injuries

To estimate the benefits of reducing minor injuries associated with cigarette fires - including burns, anoxia, and other forms of injury - the analysis employs a uniform value of $151 per case (2002 $CAD).  This value is based upon Miller, et al.'s estimate of average treatment costs for fire-related injuries treated in a physician's office (i.e., injuries that do not require hospitalisation or emergency room treatment).¹⁰⁰

Limitations of the Health Care Cost (HCC) Approach

The health care costs associated with burns, anoxia, and other injuries likely underestimate individual willingness to pay to avoid these injuries.  As Chapter 5 notes, the HCC method does not take into account averting expenditures (such as fire detectors or fire extinguishers), lost productivity, or willingness to pay to avoid pain and suffering.  Excluding averting expenditures from the analysis is unlikely to have a large impact on the final benefit numbers, since fire detectors and extinguishers cost little compared to the cost of medical treatment.  Productivity losses, however, may be significant in the case of fire injuries, and willingness to pay to avoid the pain and suffering associated with such injures may be quite high.

As an indication of the productivity losses associated with fire injuries, Miller, et al. estimate that the average length of hospital stay for burns ranges from 9.7 days to 34 days (depending on the data source).¹⁰¹  For burn victims requiring additional long-term care, the average nursing home stay is approximately two years.¹⁰²  Pindus, et al. (1993) interviewed nine burn survivors and found that respondents were initially hospitalised from two weeks to nine months; moreover, most were subsequently hospitalised for reconstructive surgery.¹⁰³  Even individuals not initially treated in the hospital (i.e., the less seriously injured) typically must pay more than one visit to a physician, and many burn victims require physical therapy after their wounds have healed.  During the time individuals are receiving medical treatment, they are not working.  In some cases, physical impairment prevents burn victims from ever returning to previous levels of productivity;  Miller, et al. estimate that 15 percent of hospitalised burn patients and one percent of those treated in emergency rooms and released experience permanent decreases in earning power as a result of their injuries.  Spouses or other family members may also become less productive as they take leave or quit work to help care for the injured.¹⁰⁴

Individual willingness to pay to avoid the pain and suffering associated with a given injury/illness will vary depending on the severity of the symptoms.  Interviews with burn survivors conducted by Pindus, et al. indicate that pain and suffering is significant for burn victims.  "Intense pain was mentioned repeatedly as a most significant memory of the hospitalisation experience."¹⁰⁵  In addition, "several respondents particularly noted the need for professional counselling to cope with disfigurement," and "six respondents have permanent functional capacity losses as a result of their burn injuries."¹⁰⁶  A study comparing total damage awards in product liability suits - including damages awarded for pain and suffering - to the portion of the award based strictly on financial damages found that the greatest ratio of total to financial damages was associated with burn injuries.¹⁰⁷  Thus, excluding consideration of pain and suffering from the analysis is a source of significant downward bias in estimating the benefits of a reduction in fire injuries.¹⁰⁸

A number of studies have explored the relationship between individual willingness to pay to avoid an illness and the medical costs of treating the illness.  In general, these studies have found that willingness to pay to avoid an illness exceeds the cost of treatment.  Stieb et al. (2002), for example, found that willingness to pay to avoid a cardiorespiratory illness, as estimated by analysis of the "total cost" of such illnesses, is on average 1.3 to 1.9 times the cost of treatment.¹⁰⁹  The ratios that these studies develop provide a basis for adjusting HCC values to better approximate willingness to pay to avoid illness.  Most of these studies, however, concern illnesses with characteristics significantly different from those associated with cigarette fire injuries, and none have examined the relationship between willingness to pay to avoid burn injuries and the cost of treating such injuries.  Given the high degree of pain and suffering associated with burn injuries, the application of a generalised ratio of willingness to pay values to HCC values would provide a poor approximation of willingness to pay to avoid such injuries.  The analysis therefore makes no such adjustment.

Limitations of the Health Care Cost Data

In addition to the fundamental limitations of the HCC approach, the treatment cost estimates provided by Miller et al. may not incorporate the full cost of treating cigarette fire injuries, and are based on data obtained from the U.S. rather than Canada.  The effect of these limitations on the results of the analysis are discussed below.

Potential Exclusion of Long-Term Medical Costs

A potential limitation of the treatment cost estimates obtained from Miller, et al. is that they may exclude some direct medical costs.  In particular, Pindus, et al. note that some burn victims require multiple reconstructive surgeries, and that in some instances this treatment occurs long after the injury occurred (in one case, surgery was scheduled more than 25 years after the initial injury).  It is unclear whether the Miller, et al. analysis fully captures these long-term costs, but Pindus, et al. conclude that their case studies "provide evidence in support of the belief that the costs of burn injuries are underestimated."¹¹⁰  To the extent that this is the case, reliance on values from Miller, et al. may understate the costs of treating those injured in cigarette fires, and thus understate the benefits of reducing such injuries.

Comparison of Canadian and United States Health Care Costs

Economists have established that nation-to-nation comparisons of health care costs are problematic because nations account for costs differently and institutional conventions vary.¹¹¹  Differences in health care costs between Canada and the United States extend beyond the effects of variation in currency exchange rates.  The following discussion identifies key differences between the Canadian and U.S. health care systems and describes how these differences may affect the costs of treatment in the two countries.  The impact of these differences on the benefit estimates derived from the 1993 U.S. injury cost data is then considered.

Private insurance companies fund the majority of health care services in the United States.  This coverage, provided for the most part through employers, is not universal, and approximately 15 percent of Americans are without insurance.¹¹²  In contrast, the Canadian system is universal, administered by the provinces, and funded via taxes on income.  Each year, provincial government authorities negotiate annual budgets with hospitals based on past budgets, clinical performance, projected changes in services, and projected changes in input costs; hospitals then receive lump-sum payments on the basis of these budgets.¹¹³  Although insurance and funding in Canada are public, health care providers are private, not-for-profit institutions.

Multiple studies indicate that hospital care in Canada costs approximately 40 percent less per patient than in the United States.¹¹⁴  Some of this difference is attributable to Canada's lower relative wages.¹¹⁵  There is also a greater proliferation of technology in the United States that leads to higher costs.¹¹⁶  The primary reason, however, that the Canadian system is less expensive is that providers are reimbursed by a single source: the government.  This significantly reduces administrative costs; Canadian health care providers do not need to determine patients' insurance coverage or prepare detailed bills for private insurers, and charges need not be attributed to individual patients.¹¹⁷

The considerations noted above suggest the possibility that the treatment cost values derived from Miller et al., which reflect U.S. medical costs, may overstate the cost of medical treatment for fire injuries in Canada.  The potential magnitude of such bias, however, is likely to be tempered by the exclusion of certain costs from the Miller, et al. values.  As noted above, administrative costs account for a significant share of the difference between U.S. and Canadian health care costs.  The values obtained from Miller, et al. do not include these administrative costs; instead, they reflect only the direct cost of patient care.  As a result, the potential for bias in transferring these values from the U.S. to Canada is greatly reduced.

---

⁹⁶ D. Ray, W. Zamula, T. Miller, P. Brigham, M. Cohen, J. Douglass, M. Galbraith, D. Lestina, V. Nelkin, N. Pindus, and P. Smith-Regojo, 1993, Societal Costs of Cigarette Fires, Report 6, U.S. Consumer Product Safety Commission. Appendix A of this document, "Estimating the Costs to Society of Cigarette Fire Injuries: Final Report," presents the health care cost data.

⁹⁷ Miller, et al. (1993) estimate that the cost of treating a fire-fighter for a non-fatal injury averages $1,093 (1992 $USD). This value is converted from U.S. to Canadian dollars using the PPPI, and inflated to 2002 dollars by applying the Canadian Medical Cost Index (MCI). The analysis follows the same procedure to convert other values from the source document to 2002 Canadian dollars.

⁹⁸ Miller, et al. (1993) estimate average costs of $698 (1992 $USD).

⁹⁹ Miller, et al. (1993) estimate average costs of $515 (1992 $USD).

¹⁰⁰ Miller, et al. (1993) estimate average costs of $98 (1992 $USD).

¹⁰¹ Miller, et al. (1993), pp. A-11 to A-12.

¹⁰² Miller, et al. (1993), p. A-14.

¹⁰³ Pindus, N., V. Nelkin, and T. Miller, 1993, "Experiences of Burn Survivors: Case Studies," Appendix B in D. Ray, W. Zamula, T. Miller, P. Brigham, M. Cohen, J. Douglass, M. Galbraith, D. Lestina, V. Nelkin, N. Pindus, and P. Smith-Regojo, 1993, Societal Costs of Cigarette Fires, Report 6, U.S. Consumer Product Safety Commission.

¹⁰⁴ Pindus, et al. (1993), pp. B-11, B-18, B-21, B-26, B-38, and B-43.

¹⁰⁵ Pindus, et al. (1993), p. B-4.

¹⁰⁶ Pindus, et al. (1993), p. B-4.

¹⁰⁷ W. K. Viscusi, 1988, Pain and suffering in product liability cases: systematic compensation or capricious awards? International Review of Law and Economics 8: 203-220.

¹⁰⁸ This conclusion is supported by a U.S. Environmental Protection Agency (2000) review of multiple studies comparing willingness to pay to avoid an illness to the cost of treating the illness. This review found that willingness to pay to avoid an illness increases relative to treatment costs as pain, suffering, and severity increase. U.S. Environmental Protection Agency, 2000, Appendix B, Handbook for Non-Cancer Health Effects Valuation, prepared by the Social Sciences Discussion Group with assistance from Industrial Economics, Incorporated.

¹⁰⁹ In this case, "total cost" is defined as the combined value of averting expenditures, pain and suffering, costs of treatment, and lost productivity. David M. Stieb et al., 2002, Economic evaluation of the benefits of reducing acute cardiorespiratory morbidity associated with air pollution, Environmental Health: A Global Access Science Source.

¹¹⁰ Pindus, et al. (1993), p. B-5.

¹¹¹ Aaron, H.J., 2003, "The costs of health care administration in the United States and Canada - questionable answers to a questionable question," New England Journal of Medicine 349(8): 801.

¹¹² Weil, T.P., 1995, "How do Canadian hospitals do it?" Hospital Topics 73(1): 10.

¹¹³ Woolhandler, S., T. Campbell, D.U. Himmelstein, "Costs of health care administration in the United States and Canada," New England Journal of Medicine 349(8): 773.

¹¹⁴ Hurley, D., 1993, "Comparing costs north of the border: Canadians spend more time in hospital, but U.S. costs are higher," Medical World News 34(4): 62; Weil, T.P., 1995, "How do Canadian hospitals do it?" Hospital Topics 73(1): 11.

¹¹⁵ Woolhandler, et al., p. 774.

¹¹⁶ For example, there are approximately 11.2 magnetic resonance imaging (MRI) units per million people in the United States, compared to 1.1 units per million people in Canada .

¹¹⁷ Woolhandler, et al., p. 773.