Benefits of Using Cell Phone While Driving

The weight of the scientific evidence to date suggests that use of a cellular phone while driving does create safety risks for the driver and his/her passengers as well as other road users. The magnitude of these risks is uncertain but appears to be relatively low in probability compared to other risks in daily life. It is not clear whether hands-free cellular phone designs are significantly safer than hand-held designs, since it may be that conversation per se rather than dialing/handling is responsible for most of the attributable risk due to cellular phone use while driving.

The benefits of using this communications device while driving appear to be important. They include benefits to the users, households, social networks, businesses, and communities. Many of these benefits, which include public health and safety considerations, have not yet been recognized or quantified. Simple suggestions that drivers can “pull over” on the side of the road to make calls from cellular phones are unrealistic and, in certain situations, potentially dangerous. It is not known which of the benefits of cellular phone use would be foregone under various regulatory scenarios. 1

Cellular phone use while driving should be a concern of motorists and policymakers. We conclude that although there is evidence that using a cellular phone while driving poses risks to both the driver and others, it may be premature to enact substantial restrictions at this time. Indecision about whether cellular phone use while driving should be regulated is reasonable due to the limited knowledge of the relative magnitude of risks and benefits. In light of this uncertainty, government and industry should endeavor to improve the database for the purpose of informing future decisions of motorists and policymakers.

The Term Paper on Cell Phone Use While Driving

It cannot be denied that science and technology have been instrumental in helping us in our daily lives, especially with regards to the way we socialize with each other. Technology has inherently made the world a small village, with communication becoming a tool that is accessible almost everywhere at any given time. One of the most noticeable faces of current communications technology is the ...

In the interim, industry and government should encourage, through vigorous public education programs, more selective and prudent use of cellular phones while driving in order to enhance transport safety. 2 Table of Contents Executive Summary Table of Tables Table of Figures Preface Section 1. Introduction Section 2. Consumer Use of Cellular Phones Section 3. Risks of Using a Cellular Phone While Driving Driver Performance Studies Case Reports of Crashes Involving Cellular Phones Overall Trends in Crashes and Fatalities Epidemiological Studies Risk Comparisons Section 4.

Benefits of Using a Cellular Phone While Driving Personal Benefits Family/Household Benefits Social Network Benefits Business Benefits Community Benefits Summary Section 5. Benefits Foregone if Cellular Phone Use While Driving is Restricted Section 6. Cost-Effectiveness of Alternative Safety Measures Section 7. Legislative, Legal, and Policy Perspectives Section 8. Recommendations References Appendices Appendix A. Calculations to Determine Annual Fatality Risks Appendix B. Methods and Participant Characteristics for Focus Groups with Consumers and Emergency Services Personnel Appendix C.

State-by-State Review of Recent Legislation Appendix D. International Legislation 1 4 5 6 8 11 14 15 20 24 28 35 39 41 44 45 45 46 48 50 54 59 64 69 1 13 15 24 3 Table of Tables Table 1. Cellular phone call Duration While Driving. Table 2. Summary of Driver Performance Studies, 1991-1999. Table 3. Voluntary Risk Factors Affecting Driver Fatality Rates. Table 4. Involuntary Risk Factors Affecting Fatality Risks. Table 5. Cost-Effectiveness Ratios for Highway Safety Investments. Table 6. Cellular Phone Related Restrictions in the U. S. Table A-1.

Parameters Used to Compute Driver Fatality Risk Associated with Cellular Phone Use. Table A-2. Parameters Used to Compute Driver Fatality Risk Associated with Alcohol Consumption. Table A-3. Passenger Cars: Effect of 100-Pound Weight Reduction on All Fatalities. Table A-4. Passenger Cars: Effect of 100-Pound Weight Reduction on Driver Fatalities. Table A-5. Fatalities Among Individuals with BACs of Zero Caused by Individuals Driving with Elevated BACs. 13 16-17 37 38 57 60 2 4 6 7 11 4 Table of Figures Figure 1. Growth in Cellular Phone Subscriptions in Millions from 1985 to 2000. Figure 2.

The Term Paper on Cell Phones in the Hands of Drivers: a Risk or Benefit?

Cell Phones in the Hands of Drivers A Risk or a Benefit Ariana Laguna English 102 Professor Barnes/Walter 10 December 2012 Outline THESIS: Unless the risks of cell phones are shown to outweigh the benefits, we should not restrict their use in moving vehicles: instead, we should educate the public about the dangers of driving while phoning and prosecute irresponsible phone users under laws on ...

Traffic Fatalities per Billion Vehicle Miles Traveled and U. S. Cellular Phone Subscribers in Millions, 1970-1999. Figure 3. Traffic Fatalities and U. S. Cellular Phone Subscribers, 1970-1999. 11 26 27 5 Preface This study was commissioned by AT&T Wireless Services (Redmond, Washington) through a research grant made to the Harvard Center for Risk Analysis of the Harvard School of Public Health (Boston, Massachusetts).

The terms of the grant relationship protected the intellectual freedom of the Harvard researchers to determine the direction of the study as well as to make independent conclusions and recommendations.

This Phase 1 report will be supplemented by a Phase 2 report, scheduled for completion in 2001, that provides additional quantitative information on the benefits and risks of the use of cellular phones while driving. The authors thank Candy Castle and her colleagues at AT&T Wireless for their support of this project. The information in this report does not provide a definitive resolution of the risk-benefit issue concerning use of cellular phones while driving.

The objective of the report is to stimulate greater scientific and public policy discussion of this issue. The intended audiences for the report include policymakers in government and industry, scientists interested in cellular phone safety, and any motorist or citizen interested in this matter. Given this broad audience, we have sought to minimize technical jargon throughout the report. A shorter summary of this report has also been issued as the June issue of Risk in Perspective, a periodic publication of the Harvard Center for Risk Analysis.

In order to assure the technical quality of this report, a rigorous process of independent peer review was applied. The authors thank the following reviewers who offered constructive comments on an earlier draft of this report: Alasdair Cain (University of South Florida), Nancy A. Dreyer (Epidemiology Research Institute), John Evans (Harvard School of Public Health), Susan A. Ferguson (Insurance Institute for Highway Safety), Michael Finkelstein (Association 6 for the Advancement of Injury Control), James Hammitt (Harvard School of Public Health), A.

The Essay on Using a Cell Phone While Driving: It Can Wait

In recent years, a new and extremely dangerous epidemic has plagued our roads. The use of cell phones while operating vehicles is an ongoing controversy that is resulting in more and more deaths each year. Consequently, we are now dealing with much more than the occasional switch of a radio station, or a casual conversation with a passenger; we are faced with constant preoccupation while driving. ...

James McKnight (consultant), Donald Redelmeier (University of Toronto), Donald Reinfurt (University of North Carolina Highway Safety Research Center), Malcolm MacClure (Harvard School of Public Health), J. Scott Osberg (AAA Foundation for Highway Safety), Jonathan Wiener (Duke University Law School), and Milton Weinstein (Harvard School of Public Health).

The authors of the report remain responsible for its technical quality as well as its findings and recommendations. The authors also thank Axiom Research Company, LLC (Cambridge, Massachusetts) for their assistance with the focus group research.

The contributions of focus group participants are greatly appreciated. 7 Section 1. Introduction Cellular phones1 were first introduced into the U. S. market in the mid-1980s, and have since experienced dramatic growth. Over the past decade in the U. S. alone, cellular phone subscription sales have increased seventeen-fold (CTIA, 2000a), ranking it among the fastestgrowing industries in the U. S. Once a luxury available to only the affluent and a small segment of business users, cellular phone use has become increasingly commonplace among American families as well as businesses.

Although cellular phones are functional in a variety of situations, they are a particularly useful technology to people on the move, including people operating motor vehicles. The majority of cellular phone owners report that they use the technology while driving. Indeed, a substantial percentage of the total calls initiated from cellular phones were by drivers of motor vehicles. Concerns have been raised that use of a cellular phone while driving increases the risk of traffic collisions, property damage, injuries, and fatalities.

A variety of groups, including the wireless communications industry and transportation safety groups, have initiated educational campaigns that encourage the prudent use of cellular phones while driving. These safety concerns have also led policymakers to consider whether the use of a cellular phone while driving should be regulated or even prohibited. Such bans, at least with respect to use of handheld phones by drivers, have already been enacted in some foreign countries, prior to the widespread diffusion of the technology. Many states and localities in the U.

The Essay on Using a Cell Phone While Driving

Have you ever used a cell phone while driving? Is it dangerous or not? Using a cell phone while driving is common, but widely considered dangerous. We can’t imagine our life without using a cell phone. It is a part of our daily activities. We use a cell phone to make calls, E-mails, text massages, surf the Internet, to listen music’s and many other daily activities. Many of these activities take ...

S. are now considering restrictions or bans on use of a cellular phone while driving. 1 The term “cellular phone” is used throughout this document to refer to all types of wireless communication including cellular and digital technology. It includes both hand-held and hands-free models. 8 The purpose of this report is to elucidate what is known about the risks and benefits of using a cellular phone while driving, including a discussion of public policy issues relevant to whether use of a cellular phone while driving should be restricted or prohibited.

We conclude that it is currently difficult for policymakers to reach an informed conclusion for three primary reasons. First, the risks of using a cellular phone while driving, though real, are not large enough to be detected in overall crash/fatality statistics but are potentially large enough to be a legitimate concern of motorists and policymakers. Second, the benefits of using a cellular phone while driving have been the subject of much less study and attention than the risks. This report begins to address this imbalance in the literature with data from several focus groups with cellular phone users.

Finally, the cost-effectiveness of saving lives through restricting cellular phone use does not appear to be very attractive compared to other traffic safety measures. We argue that a targeted and intensive program of scientific inquiry and policy discussion would promote the development of wise regulatory policy in this field. The report is organized as follows. Section 2 presents data on the rapid penetration of this technology throughout the U. S. Section 3 assembles the available evidence on risk and provides “ballpark” estimates of the magnitude of the risk of death from use of cellular phones while driving.

We consider both risks to the cellular phone user as well as risks to other road users whose safety may be endangered when a driver uses a cellular phone while the vehicle is in motion. The magnitude of risk is compared to other voluntary and involuntary risks that people face on a daily basis. Section 4 examines the benefits (non-economic as well as economic) of using a cellular phone while driving. Section 5 discusses the benefits that might be foregone if cellular phone use is restricted while driving.

The Essay on Texting While Driving 2

... driving while ‘phoning’. The results were unsettling; “We found that using a cellular telephone was associated with a risk of having a motor vehicle ... school bus, and drunk driving are obvious examples; talking on the phone and texting in a moving vehicle should be no ... 2011. Web. 12 Jan. 2012. Haughney, Christine. “Taking Phones out of Drivers’ Hands. ” Washington Post 5 April. 2009: A8. Print. ...

Section 6 compares the cost-effectiveness of restrictions on the use of cellular phones to the cost-effectiveness of other lifesaving measures 9 adopted by state and federal policymakers. Section 7 reviews recent legislative and legal developments, illustrating the extent of uncertainty among policymakers about the wisest course for public policy regarding this issue. Section 8 describes our recommendations concerning future research and risk management. 10 Section 2. Consumer Use of Cellular Phones Cellular phone service was introduced in the United States in 1983 (CTIA, 2000a).

As of June 2000, the number of subscriptions in the U. S. has exploded to 94. 2 million (CTIA, 2000b), with 27% of American households reporting in a recent survey that at least one member owns a cellular phone (PCIA, 1999).

Figure 1 reports data on the growth of cellular phone subscriptions in the U. S. from 1985 to 2000. In the United States about 75% of the cellular phones in use are of the hand-held design (National Highway Traffic Safety Administration [NHTSA], 1997a), though hands-free and voice-activated designs are also in use2. Figure 1. Growth in Cellular Phone Subscriptions in Millions from 1985 to 2000. 00 80 60 40 20 0 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 00 Use of cellular phone technology was initially confined to the affluent and a small segment of business users. However, usage patterns have changed substantially in recent years as the costs of owning and using a cellular phone have declined. The percentage of owners who use these devices primarily for personal/family purposes increased from 40% to 61% between 1990 and 1998 (Cain and Burris, 1999).

Cellular phone ownership is correlated with education 2

A hand-held phone is a portable model, generally small and lightweight, that must be held to the ear and mouth for use. It may be transportable, mobile, or pocket size. In contrast, a hands-free phone is a model that can be used while mounted in a vehicle or placed in a bracket. It may be operated with a remote speaker or microphone to improve performance (NHTSA, 1997a).

The Term Paper on Cell Phone Driving Phones Ground

... to pull over or use hands-free technology, and should avoid emotional or stressful conversations." (Cell Phone Use While Driving Increases Crash Risk, Gay ... automobile accidents while using cellular phones. The study also revealed that the risk was the same when drivers utilized "handsfree" phones. They also stated ...

11 level. Only 15% of individuals with less than a high school degree owned cellular phones, compared to 26% of high school graduates, 30% of those with some college education, and 40% of those with a college degree (NHTSA, 1997a).

If present trends continue, a majority of American households may be owners of cellular phones by the year 2005. A hand-held cellular phone can be used in a variety of circumstances: while walking down the street, waiting in line at a store or doctor’s office, or operating a motor vehicle. The use of a cellular phone while driving is a common application of this technology. Several surveys have found that 80-90% of cellular phone owners use these devices while driving at least some of the time (NHTSA, 1997a; PCIA, 1999; Cain and Burris, 1999).

The extent of use while driving varies substantially among owners.

One recent survey found that 15% of cellular phone owners use their cellular phone for more than 1 hour per month while driving, 15% for 30-60 minutes, 20% for 10-30 minutes, and 39% for less than 10 minutes per month while driving; 11% did not respond to the survey (PCIA, 1999).

More precise data are needed on users who frequently initiate and/or receive calls while driving. One survey found that 29% of all cellular phone users regularly use their phone in their car (Hart, 1997)3, yet publicly available data do not reveal what fraction of owners use this technology primarily or only while driving a motor vehicle.

Individuals who are particularly likely to use cellular phones regularly while driving include owners who use them for business purposes (48%), owners who are commuters (40%), and owners between the ages of 18 and 34 (35%) (Hart, 1997).

The elderly and low-income individuals are less likely to use a cellular phone regularly while driving. Men are reported to use their phones regularly while driving more frequently than women (Hart, 1997), though this disparity may decline in the future as the technology becomes further integrated into personal, family, and business life. In certain regions 12 f the country where average commuting distances are long and highways congested, we suspect that cellular phone owners may initiate the vast majority of their total number of calls while operating a motor vehicle. A richer database on characteristics of individuals who use cellular phones while driving and their purpose for use is needed. Call duration also varies substantially (Table 1).

The most frequent type of call while driving lasts between 30 seconds to 2 minutes (37%), followed by calls of less than 30 seconds (23%), calls of 2-5 minutes (18%), and calls of longer than 5 minutes (4%) (PCIA, 1999).

Table 1. Cellular Phone Call Duration While Driving. Call Duration While Driving Less than 30 seconds 30 seconds – 2 minutes 2 – 5 minutes More than 5 minutes Do not use while driving No response Source: PCIA, 1999 Percentage of Respondents (%) 23 37 18 4 10 8 3 Survey among 409 wireless users/407 non-users nationwide, May 15-17, 1997. 13 Section 3. Risks of Using a Cellular Phone While Driving motor vehicle collisions are harmful in many ways. For motorists these collisions can result in minor, serious, crippling, and even fatal injuries.

Collisions result in property damage to vehicles that is often expensive to repair. There are also related health care costs, litigation expenses, insurance administration, lost work time, and other adverse ramifications of collisions. In urban areas, each collision occurring in rush hour can induce a cascade of traffic congestion, emergency response time, and grief for those affected. Thus, if use of cellular phones while driving causes more collisions, motorists and policymakers should be concerned.

Businesses and insurers may also become concerned since these firms bear a significant fraction of the monetary costs of motor vehicle collisions. In response to growing interest in this issue, a number of recent studies have conducted rigorous reviews of the literature. The most comprehensive was NHTSA (1997a); more recent authors have drawn heavily from this initial review (e. g. , Hahn and Tetlock, 1999; Cain and Burris, 1999).

In this report, we highlight information and studies that may be of strongest interest and/or relevance to the regulatory issue.

Four types of information address the risks of using a cellular phone while driving: (1) driver performance studies, (2) case reports of crashes where use of a cellular phone appears to have played a role in the crash, (3) statistical comparisons of trends in motor vehicle crashes and cellular phone usage (so-called “ecological” studies of risk), and (4) “epidemiological” studies that use individual-level data on phone use and crash experience to determine the statistical association between use of a cellular phone and collision, injury, and/or fatality.

As we shall see, this entire body of evidence is consistent with the common-sense judgment that use of a cellular phone while driving can increase the risk of a motor vehicle collision with attendant injuries and 14 fatalities. However, it is not yet clear what the magnitude and severity of these risks are under real-world conditions. Driver Performance Studies. The use of a hand-held cellular phone while driving may entail a variety of different maneuvers: searching for a phone in the vehicle, reaching for a hone to initiate or receive a call, dialing, holding a phone near the ear while talking and driving, picking up a phone that has been dropped, and so forth. Even use of a hands-free phone can be distracting to the driver as conversation consumes mental energy while driving. The various tasks entailed in using a cellular phone each require a different amount of time, mental energy, and coordination, leading to potentially different complications of the driving task and resulting risk of collision.

A number of studies, beginning as early as the late 1960s, have examined how cellular phone use affects driver behavior (Brown, Tickner, & Simmonds, 1969; Kames, 1978; Drory, 1985; Stein, Pareghian, & Allen, 1987; Hayes, Kurokawa, & Wierwille, 1989).

Some studies observe experimental subjects operating a motor vehicle while other studies make use of a computer-generated driving simulator. The relevant studies published since 1990 are summarized in Table 2. A comprehensive review of the entire literature through 1997 is available in NHTSA (1997a).

15 Table 2. Summary of Driver Performance Studies, 1991-1999.

Reference Brookhuis, de Vries, & de Waard, 1991 Subject Characteristics 12 drivers 10 men, 2 women Equally divided into 3 age categories: 23-35, 35-50, 50-65. Methods & Measurement -Modified automobile (i. e. redundant controls for use if needed by accompanying driving instructor).

-Hand-held and hands-free phone use equally distributed. -Measures included lateral position (swerving) of vehicle, ability to follow car-in-front, steering wheel motions, checking the rearview mirror, and mental workload. -Computer-based driving simulation using hands-free device. -Measured driver response rate (decelerating, braking, turning away).

Intervention(s) Using hand-held or hands-free cellular phone while driving modified vehicle. Comparison Not using cellular phone while driving modified vehicle. Key Findings -Hands-free phone users had better control than hand-held users, as measured only by steering wheel movement. -Phone use decreased swerving and delayed reaction time to car following. -Mental workload increased when phoning. -Age effect: none. McKnight & McKnight, 1993 150 drivers 75 men, 75 women Mean age 39 yrs. -Radio tuning -Call placing (handsfree) -Casual conversation (hands-free) -Intense conversation (hands-free) Driving with no distraction.

Serafin, Wen, Paelke, & Green, 1993 12 drivers 6 men, 6 women Mean age Young group, 24 yrs. Elderly group, 70 yrs. -Driving simulator. -Manual (hand-held) phones & voice command (hands-free) used. -Measured driving performance and dialing performance. -Dialing phone while driving. -Performing tasks while driving . (involving recall and conversation).

Driving without using phone. -All interventions and comparison significantly different. Tuning radio and intense conversation most distracting overall. -Age effect: younger drivers more distracted while tuning radio; older drivers more distracted while placing a call or conversing. Dialing while driving disturbed driving performance most. Voice input led to better driving performance than manual handset. -Age effect: younger drivers performed better than elderly on driving performance. 16 Table 2. Summary of Driver Performance Studies, 1991-1999. (continued) Reference Alm & Nilsson, 1994 Subject Characteristics 40 drivers 20 men, 20 women Mean age 32. 4 yrs Methods & Measurement -Driving simulator with moving base, wide-angle view, vibration, sound, and temperature regulation. -Used hands-free cellular phone mounted on steering wheel. Measured driver reaction time and mental workload, and lateral position (swerving) of vehicle, with easy versus hard driving tasks. -Methods and measurement identical to Alm & Nilsson, 1994 (above) with the additional measurement of headway (distance b/t front of subject’s car and end of lead vehicle).

Intervention(s) Using a hands-free cellular phone while driving simulator. Comparison Not using a cellular phone while driving simulator. Key Findings -Longer reaction time (slower reaction) for phone group in easy driving task; no difference found for hard driving task. Greater deviation from lateral position (swerving) of vehicle for phone group in hard driving task. -Mental workload increased for phone group. -Contrary to predictions, strongest effects found in phone group exposed to easy driving tasks. -Longer reaction time and shorter minimum headway for phone group. -Mental workload increased for phone group. -No difference in lateral position (swerving) of vehicle. -Age effect: elderly group had longer reaction time than younger group. -Road position improved slightly with introduction of a second task. Road position most affected by dialing radio, followed by difficult & easy conversation, respectively. -Age effect: none. -Male drivers performed better -Difficult conversation may effect driving adversely. -Driver’s detection ability significantly impaired regarding BRT and TTC when either dialing or performing cognitive task. -Cognitive and phone dialing tasks roughly equally distracting. Alm & Nilsson, 1995 40 drivers 30 men, 10 women Mean age Young group, 29. 3 yrs. Elderly group, 67. 6 yrs. Using a hands-free cellular phone while driving simulator involving interaction with other road users.

Not using a cellular phone while driving simulator involving interaction with other road users. Driving condition (compared with multiple distraction).

Briem & Hedman, 1995 20 drivers 10 men, 10 women Mean age Young group, 21. 0 yrs. Elderly group, 45. 5 yrs. Lamble, Kauranen, Laakso, & Summala, 1999 19 drivers 10 men, 9 women Mean age 22. 7 yrs. -Radio use -Simulated driving on slippery or -Easy conversation firm road w/ or w/o secondary -Difficult task (communication or conversation instrument manipulation).

-Used hands-free cellular phone mounted on console to the right of steering wheel. Measured road position, collisions with obstacles, and speed of driving. -Phone dialing task -Modified automobile. -Mounted (hands-free) phone used. -Cognitive tasks -Measured time-to-collision (TTC) (involving memory and addition) and brake reaction time (BRT).

Driving without tasks. 17 Virtually all of these studies document some decrement in driver performance that is associated with use of a cellular phone while driving. The studies vary tremendously in the methods employed, the type of phone models (hand-held or hands-free), the type of phone use (initiating vs. receiving a call), and the nature of the conversation (casual vs. ntense).

Which task contributes most to poor driving performance: searching for the phone in the vehicle, dialing, conversing, hanging up, picking up a dropped phone, or receiving a call? The answer to this question is not entirely clear because studies to date are of small sample size and usually unreplicated. According to Serafin et al. , (1993), the predominant disturbance in driving performance was dialing the phone. Crash data from Japan support this finding. On the other hand, U. S. crash reports suggest that the distraction of conversation may contribute more to risk than the acts of dialing or receiving a call (see below).

However, Briem & Hedman (1995) suggested that conversing alone – as long as the conversation was not emotional or involved – did not contribute to driver distraction. There is insufficient literature on whether or not hands-free models offer any safety advantage over hand-held models. Most studies have investigated only one model type (generally, hands-free).

One exception is Serafin et al. (1993), who compared hand-held phones to hands-free phones with voice activation. The authors concluded that, when driving and dialing, voice input through a hands-free phone led to better driving performance compared to a hand-held phone (Serafin et al. 1993).

Several studies have found that older-aged individuals demonstrated diminished driving performance while using a cellular phone (Serafin et al. , 1993; Alm & Nilsson, 1995; McKnight 18 & McKnight, 1993).

In contrast, Briem & Hedman (1995) found no difference in effect on driving proficiency between age groups. 4 The real-world implications of driver performance studies are uncertain. Drivers may become aware of the risks entailed in using a cellular phone while driving, either through their own intuition and/or through what they learn from family members, friends, or the mass media.

Such awareness would be predicted, at least by some behavioral scientists, to induce a change in driving behavior that is aimed at restoring the pre-existing level of driver safety (Peltzman, 1975; Evans, 1991).

These theories are rooted in the assumption that the driver cares about his or her own safety. They are not rooted in any altruistic concerns for the safety of other road users, although altruism could also motivate drivers to take some kinds of risk compensation measures.

These compensatory behaviors might include maintenance of a constant speed, reduction in passing maneuvers, initiation of calls only when the vehicle is at a stop or is being operated in a remote area, and participation in lengthy and/or intense calls only on certain kinds of trips or in certain kinds of road conditions where there is a perception of relative safety. Compensatory behaviors are not necessarily exact in their offsetting influences and thus could reduce risk to a lesser or even greater extent than the additional risk induced by use of the cellular phone while driving.

Moreover, compensatory behaviors could be greater among some drivers than others. Although driver performance studies may implicitly include some compensatory effects in their research design, it is very difficult for small-scale experimental studies to account for the full range of potential compensatory effects. In summary, driver performance studies have provided insight into the mechanisms whereby use of a cellular phone while driving could cause or contribute to motor vehicle crashes. They therefore provide a solid basis for believing that use of cellular phone while driving will 4 See Table 2 for mean sample ages of referenced studies. 9 cause collisions in the real world. However, the actual risks of real world collisions cannot be accurately inferred from these experimental studies because compensatory behaviors may not be accounted for in the study design. Case Reports of Crashes Involving Cellular Phones. Information about the influence of cellular phones in particular crashes is difficult to obtain. With the exception of the states of Oklahoma, Minnesota, and Tennessee, police-accident reports typically do not elicit information about whether a cellular phone was present in the vehicle and, if so, whether it was in use immediately before the crash.

Even in these states, where such information is supposed to be included on police-accident reports, there are serious questions about the completeness and accuracy of the coded information about cellular phone presence and use (NHTSA, 1997a).

5 Drivers may be reluctant to report their use of a cellular phone to police for fear that they may be admitting culpability for the crash, which might in turn create liability or insurance problems for the driver.

Witnesses may be the most reliable source of information about cellular phone use prior to a crash but they are often unavailable or their information is not recorded on a policeaccident report. Even if it is known that a cellular phone was in use by the driver immediately prior to a crash, it is not always clear whether use of the cellular phone was the single causative factor in the crash, a contributing factor, or simply an unrelated fact. In a strictly scientific sense, it may be impossible for a police officer or witness to discern whether a particular crash was caused by a driver’s use of a cellular phone. The olice officer or witness has no basis for judging how the driver would have behaved behind the wheel if he or she had not been using a cellular phone. Thus, determining whether a particular crash was “phone-related” is a very subjective decision. 5 Though NHTSA, 1997a refers only to Oklahoma and Minnesota, Vuong, 2000 reports that Tennessee also requires the documentation of cellular-phone related accidents on police reports. 20 NHTSA’s Fatal Analysis Reporting System (FARS) is a census of all fatal motor vehicle crashes that occur in the United States each year, with information on each crash derived primarily from police-accident reports.

FARS does include some information on the role of driver distractions in crash causation, but information on cellular phone use is incomplete. In calendar years 1994 and 1995, a total of 36 and 40 FARS fatal crashes, respectively (of approximately 40,000 fatalities in each year), were identified that included cellular phone use as a “possible distraction inside the vehicle” (NHTSA, 1997a, Section 3. 2, p. 3).

Over half of these cases occurred in Oklahoma each year. When NHTSA’s project staff attempted to verify the phone-related crashes in Oklahoma, problems emerged.

Only 2 of the 21 Oklahoma cases in 1994 FARS could be verified as phone-related; only 1 of the 26 Oklahoma cases in 1995 FARS could be verified as phone-related. The national number of FARS cases where fatal crashes were verified as phone-related (17 in 1994 and 15 in 1995) are likely a substantial undercount since FARS depends on a police report. Yet some suggestive patterns emerge even from this incomplete data. In only 3 of the 32 cases was the operator of the cellular phone a driver of the struck vehicle. The drivers using the cellular phone were typically operating the striking vehicle.

In one third of the cases, the striking vehicle ran off the road in single-vehicle collisions. The other cases involved pedestrians, bicyclists, or multi-vehicle collisions. NHTSA’s National Accident Sampling System (NASS) is a different database that includes a stratified, random sample of 5,000 police-reported crashes per year in the U. S. Information on cellular phone use in NASS is present only if drivers and/or occupants of crashinvolved vehicles were willing to state such involvement in an interview with NASS researchers.

Thus, the reported data on cellular phone use in NASS accidents are also likely to be incomplete. 21 In 1995, for example, there were 8 relevant cases out of 4,555 (one involving a driver dialing a phone; the other seven involving a driver conversing on the cellular phone).

In a special one-time investigation, a team of crash analysts at NHTSA reviewed 60,233 police-crash report narratives—only ten were cellular-phone related. Again, these reports were likely subject to a vast undercount of the true number of phone-related crashes.

The following are short descriptions of these 10 crashes, as reported by NHTSA (1997a, Table 3-10): § Case 1: The driver of the vehicle was talking on his cellular phone to get directions when his vehicle hit a concrete island on the left, and veered through the right lane down an embankment into a tree. Case 2: The driver of the vehicle took his eyes off the road while attempting to “use” a cellular phone and the vehicle veered to the right, striking a curb. Case 3: The driver was talking on a cellular phone and proceeded to turn left at an intersection as the light was turning yellow/red, and turned into the path of an on-coming vehicle.

Case 4: The driver attempted to answer the cellular phone and ran off the road into a tree. Case 5: The driver, reaching for a cellular phone, ran a red light and struck another vehicle. Case 6: The driver was answering his cellular phone, when he looked up and saw a vehicle stopped in front of him and was not able to stop in time. Case 7: The driver “was distracted by a cellular phone” and skidded into an intersection against a red signal and struck another vehicle. Case 8: The driver was talking on a cellular phone and “not paying attention” when she rear-ended another vehicle stopped for a crossing pedestrian. § § § § § § 22 § Case 9: The driver was talking on a cellular phone and did not see a red light until it was too late to stop for an on-coming vehicle. Case 10: The driver was talking on a cellular phone and made a left turn into another turning vehicle in an adjacent turning lane. § These case reports, while they do not provide definitive cause-effect evidence, suggest that cellular phone use while driving has proven to be hazardous in certain real-world circumstances. In a pilot program in the Baltimore-Washington area, Dynamic Science Inc. DSI) was contacted by police agencies whenever a cellular phone-related crash came to the attention of the police (NHTSA, 1997a).

A special DSI crash investigation team was then commissioned to perform an in-depth investigation of the crash. During the 6-month period of the pilot, seven crashes were reported to DSI. Each of these crashes had notable circumstances or consequences: one involved a driver denying reports of witnesses that he was using a cellular phone prior to the collision; another involved a state trooper who refused o release information about the crash to DSI; another involved a driver talking and writing down notes (directions) prior to the collision; two cases involved drivers receiving calls prior to their respective collisions; one involving the death of the cellular phone user’s child; one involved a cellular phone user involved in a head-on collision where a police officer was killed; and one involved a driver who struck a school bus from behind while picking up her cellular phone (NHTSA, 1997a, Section 3. 7).

The National Police Agency of Japan also studied the frequency of cellular phone use antecedent to motor vehicle crashes (NHTSA, 1997a).

In the month of June 1996, 129 crashes involving cellular phones were identified: 76% involved rear-end impacts, 2. 3% were singlevehicle crashes, and 19% were categorized as “others” (presumably frontal or side impacts with other vehicles at intersections or during lane changes).

At the time of the crash, 42% of drivers 23 were responding to a call, 32% were dialing, 16% were conversing on the phone, and 5% were hanging up the phone (remaining cases unknown).

There are also news reports that the Japanese ban on the use of hand-held cellular phones while driving has been followed by a reduction in phone-related crashes, though a scholarly evaluation has not yet been published (Yomiuri, 1999).

It is not known whether the determinations in the Japanese data have greater or lesser validity than the determinations found in U. S. investigations of crashes. NHTSA (1997a) has noted that the percentage of Japanese crashes involving dialing appears to be larger than observed in the U. S. , where distraction from conversation appears to be a more common mechanism of crash causation.

Crash circumstances in the U. S. may also be somewhat different. In order of frequency, the crash circumstances involving cellular phones in the U. S. include drivers moving out of designated lanes, drivers striking a stopped vehicle in the same lane, and drivers failing to stop for a red traffic signal. In summary, despite the uneven quality of case reports and the inability to draw rigorous cause-effect inferences from post-crash investigations, these case reports suggest that use of a cellular phone while driving has caused or contributed to motor vehicle crashes, njuries and fatalities in the U. S. and abroad. The types of driver distractions noted in police reports of phone-related crashes are consistent with those predicted by driver performance studies. What cannot be determined from any individual crash report is the magnitude of risk (elevation in crash probability) that is associated with use of a cellular phone while driving. Overall Trends in Crashes and Fatalities.

If use of a cellular phone while driving is an important causal factor in motor vehicle collisions, injuries, and fatalities, one might expect that the recent growth in cellular phone use would produce an increase in the overall number of motor vehicle collisions, injuries, and fatalities. The growth in cellular phone use has been 24 explosive and concentrated in a relatively short time period in the U. S. ; therefore, any substantial adverse effect should be immediately observable, since there would not be any expected latency period to detectable effect that is typical of technologies or behaviors that cause chronic diseases.

Analysts have searched for such a correlation in national, state, and local data. We have chosen to report selected data of this type because they are often cited in the cellular-phone policy debate. Yet, as we shall see, these data are unlikely to provide any persuasive evidence, of either an incriminating or reassuring nature. When national traffic fatality counts are used as the dependent variable, there does not appear to be any simple association between fatalities and national cellular phone subscriptions (a surrogate for the extent of cellular phone use while driving).

Figure 2, for example, plots data for the U.

S. mileage fatality rate and U. S. cellular phone subscriptions (in millions) for the 19701999 period. Similarly, figure 3 presents traffic fatalities and cellular phone subscriptions in the U. S. There is no indication in Figures 2 or 3 of an explosive growth in the number of traffic fatalities or in the mileage fatality rate that is attributable to cellular phone use. Traffic safety researchers do not find much reassurance in the data presented in Figures 2 and 3 because there are many powerful variables (beneficial and adverse) that influence overall fatal crash statistics.

As an example, if cellular phones were in fact causing 500 additional fatalities each year in the U. S. , the problem – even though large in absolute magnitude – might be masked in the aggregate data by recent reductions in accident fatalities from campaigns against drunk driving and for safety belt use. Alternatively, if cellular phone use were to increase the risk of motor vehicle collisions but primarily in less severe crashes (i. e. , those collisions least likely to cause a fatality, such as rear-end impacts), then one would not expect to see a simple correlation between traffic fatalities 5 and cellular phone use. For example, in rush-hour traffic where cellular phone use is common, fatal crashes account for a disproportionately small share of crashes because congestion produces low-speed collisions in which vehicles may be damaged but occupants receive little or no injury. The times of the day, week, and year when drivers are particularly susceptible to severe, fatal crashes (e. g. , weekend nights) are not necessarily periods when use of a cellular phone while driving is particularly likely.

Although fatal crashes are of obvious human significance, they may not be the most important outcome when scientists study the risks of using a cellular phone while driving. Figure 2. Traffic Fatalities per Billion Vehicle Miles Traveled and U. S. Cellular Phone Subscribers in Millions, 1970-1999. 80 70 60 50 40 30 20 10 0 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 Traffic Fatalities per Billion Vehicle Miles Traveled Number of Cellular Phone Subscribers in Millions 26 Figure 3. Traffic Fatalities and U. S.

Cellular Phone Subscribers, 1970-1999. 100,000,000 80,000,000 60,000,000 40,000,000 20,000,000 0 Number of Traffic Fatalities (x 1,000) Number of Cellular Phone Subscribers Researchers in the state of North Carolina made an effort to determine whether increases in cellular phone use from 1989 to 1995 were associated with an increase in the frequency of police-reported motor vehicle crashes. The circumstances of each accident were detailed in the reporting police officer’s “narrative”, a description of how the crash occurred and what the contributing factors were.

Researchers then used a key-word search of narratives to identify crashes involving cellular phone use. This exploratory study, cited in NHTSA (1997a), did report a positive correlation between cellular phone subscriptions and collisions. The absolute number of phone-related crashes did not increase as sharply as the number of cellular phone subscriptions. Nonetheless, useful information was extracted from these narratives regarding the nature of cellular phone use that may have contributed to crashes. Reaching for cellular phones and picking up dropped cellular phones were considered major contributors to crashes.

The major limitations of this study are that errors in police reporting of crashes are likely to be significant, with both false-negative errors (missing cases) and false-positive errors (crashes unrelated to cellular phone use).

27 A more recent study in metropolitan Toronto, Canada also used police-reported crashes as the dependent variable in a time-trend analysis (Min and Redelmeier, 1998).

The objective was to determine whether the change in collision rates from 1984 to 1993, at a street-by-street level, was correlated with the growth of cellular phone usage.

Cellular phone usage at each of the study’s locations was estimated by the density of cellular towers (or antennae) in the surrounding area. The police officer’s judgment as to whether cellular phone use played a role in the crash was not used in this study. The authors controlled statistically for two potentially confounding variables: pedestrian flow and traffic flow. The average street location in Toronto had 9 more collisions in 1993 than 1984. However, locations in Toronto with 6 or more towers had a smaller average increase in collision rates than locations with 5 or fewer towers.

The authors caution that there may have been biases in their research design that could have concealed a real, positive association between cellular phone use and collision frequency. Their conclusion was that “. . . the effects of cellular phones on collision risk are smaller than the biases of ecological analysis” (Min and Redelmeier, 1998, p. 160).

In summary, aggregate statistical analyses of crash data (so-called “ecological” studies) do not provide convincing evidence that the growth of cellular phone usage while driving is associated with an increase in the number of traffic fatalities or collisions.

While this lack of evidence suggests that use of cellular phones while driving has not caused an explosive growth of traffic collisions, injuries, and fatalities, these kinds of “ecological” studies are too crude (imprecise) to detect substantial risks that would be of interest to motorists and policymakers (Min and Redelmeier, 1998).

Epidemiological Studies. Four epidemiological studies in the peer-reviewed scientific literature have assessed the association between use of a cellular phone and involvement in 28 motor vehicle collisions.

By “epidemiological”, we refer to studies that employ individual-level data on cellular phone use and traffic crashes in conjunction with modern statistical tools. Violanti and Marshall (1996) were the first to study the association between use of a cellular phone and motor vehicle collisions. Their “case-control” study of New York State residents compared the experience of 100 randomly selected drivers (the “cases”) who had been in an accident within the past two years to the experience of 100 different, randomly selected drivers (the “controls”) who had not been in an accident for the previous 10 years.

The cases were drivers in crashes severe enough to cause at least $1,000 in property damage or personal injury (so-called “reportable” accidents).

A mail survey was conducted to obtain demographic and risk factor information on each case and control (response rates were 60% for the cases and 77% for controls), including information on cellular phone use while driving. Of those subjects reporting use of a cellular phone while driving (11%), the majority reported use of a hand-held design. Both cases (13%) and controls (9%) reported at least some use of a cellular phone while driving.

The authors, using logistic regression analysis, reported that the amount of cellular phone use, categorized as median time per month on the phone while in the vehicle, was strongly associated with the risk of a traffic collision (adjusted odds ratio (OR) = 5. 59; 95% confidence interval (CI) [1. 19-37. 33]).

An important feature of this study was control for other risk factors for accidents including average miles driven per year, years of driving experience, driver age, and 18 self-reported behavioral variables ranging from smoking while driving to dozing off while driving.

This study has several important limitations. No information is provided about crashes involving serious injuries or fatalities. The sample sizes in both groups are small and the data are 29 too sparse to compare hand-held and hands-free designs. The reliability and validity of the selfreported information in this study – the information on cellular phone use and other risk-taking behaviors (e. g. , miles driven) – is not assessed by the authors. No information was collected about whether a cellular phone was actually in use by cases when the accident of interest occurred.

A larger and better designed study was performed by Redelmeier and Tibshirani (1997) in metropolitan Toronto, Canada. Among the subject pool of 5,890 drivers, of whom 1064 acknowledged having a cellular phone, 699 met study inclusion criteria of having been in a car collision resulting in substantial property damage but no personal injury. In addition, each driver’s cellular phone calls on the day of the collision and during the previous week were analyzed through detailed billing records. A total of 26,798 cellular phone calls were made during the 14-month study period.

This study employs a “case-crossover design”, wherein each driver serves as his or her own control. Cellular phone usage 10 minutes prior to the collision is compared to usage during a control period on the day before the collision. Case-crossover analysis would identify an increase in risk from cellular phone use if there were more phone calls immediately before the collision than during the time period on the previous day, allowing for chance fluctuations in the frequency of cellular phone calls. Overall, 170 ubjects (24%) had used a cellular phone during the 10 minutes immediately before the collision, 37 (5%) had used the cellular phone during the same period on the day before the collision, and 13 (2%) had used the cellular phone during both periods. An adjusted statistical analysis revealed that cellular phone activity was associated with a relative risk of motor vehicle collision of 4. 3 (95% CI, 3. 0 – 6. 5).

The relative risk was 4. 8 for calls within 5 30 minutes before the collision as compared to 1. for calls more than 15 minutes before the collision. A similar result was found when the sample was restricted to those crashes whose time of occurrence was known exactly. The authors detected no statistical safety advantage of handheld over hands-free cellular phones, and therefore postulated that the distraction of conversation or attention, rather than manual dexterity involved in activities such as dialing, is responsible for cellular-phone related collisions. Despite its many strengths, this study also has important limitations.

As the authors note, the reported association may not be causal since underlying road conditions (e. g. , congestion or poor weather) might cause both cellular phone use and crashes. No information is provided about crashes involving serious injuries or fatalities. The refusal of many drivers (30% of 1064 drivers) to participate in the study may have induced a downward bias in the relative-risk estimates. The validity of the reported information regarding time of collision – even for the socalled “exact times” – is unknown.

Although most of the calls were less than 2 minutes in length, the authors focused their analyses on 5-minute and 10-minute time intervals prior to the collision (though shorter intervals were also analyzed).

On the other hand, it is possible that driver distraction (due to the cellular phone) can cause a crash before or after a call takes place. In order for the relative-risk estimates to be interpreted as causal, the collision day and the day preceding the collision must be assumed to be comparable with regard to multiple risk factors.

Violanti (1998) used a case-control design in the state of Oklahoma to determine whether or not use of a cellular phone was associated with fatal crash involvement. Since 1992, police in Oklahoma have been provided, on standardized accident report forms, a “check box” for officers to indicate the presence and/or use of a cellular phone. Over a four-year period (1992-1995), 223,137 cellular phone-related accident reports were obtained from the Oklahoma State 31 Department of Public Safety. Of 1,548 driver fatalities, 65 (4. %) had a phone present in the vehicle, and 5 (7. 7%) were reported to have been using their phones at the time of the collision. In order to determine whether or not cellular phone use increased the risk of fatal crash involvement, the author compared “cases” (fatally injured drivers) to “controls” (drivers who survived).

The author hypothesized that if cellular phones were to increase fatality risk, the presence of phones and/or the reported use of phones should be more frequent among cases than the controls.

To justify this analysis, it was presumed that phone presence indicated an elevated probability of dialing and/or use, since use may be underreported to or by police. A limited number of potential confounding variables were controlled in statistical analyses (age, gender, type of collision, driver actions, reported cause of collision).

The author calculated an adjusted odds ratio of 9. 29 (95% CI, 3. 70-23. 14) for cellular phones in use by the driver and a corresponding 2. 11 (95% CI, 1. 64-2. 71) for cellular phone presence in the vehicle.

This study suffers from several serious limitations. First, the author acknowledged that cases may differ from controls by a variety of unmeasured dimensions associated with fatal crash involvement (e. g. , total miles driven, number of miles driven on high-speed roads, number of miles driven while fatigued, and other sources of driver distraction such as eating or smoking while driving).

Second, the author expressed confidence that police reporting of cellular phone presence and use is accurate and complete.

This confidence is not buttressed by any reported validity data. In particular, the study’s key findings are biased against the safety of cellular phones if police are more successful in capturing cellular phone presence/use for cases (fatallyinjured drivers) than for controls (drivers who survive).

It is certainly plausible that police are more diligent in investigating and reporting information about fatal than nonfatal crashes. It is also plausible that surviving drivers are more successful in concealing the presence and/or use of 32 ellular phones. No phone records are typically used by police to establish phone use and thus the author reports “phone use was likely the result of expert judgment on the part of police or reported by witnesses” (p. 523).

In a separate report, NHTSA (1997a) analyzed the process by which police in Oklahoma code information on cellular phone presence and use. The data are collected by the investigating officer, generally at the scene of the crash. Officers are trained to look into the crash-involved vehicle to see if a cellular phone is present.

Installed phones and large portable units are likely to be visible but more popular hand-held models may not be readily visible. If a phone is observed, the driver (or any other witness) is asked if (s)he was using the phone at the time of the crash. If a positive response is received, then the “in use” box is checked. NHTSA (1997a) believes that underreporting of presence is likely for hand-held phones and for phones that are not visible. Moreover, potentially culpable or simply nervous drivers may be less inclined to acknowledge that they were using their cellular phones at the time of the crash.

Witness testimony may be more reliable in some cases but it is often unavailable since many crashes that kill drivers involve no other passengers. NHTSA (1997a) authors also concluded, based on discussions with instructors at the Oklahoma State Police Training Academy, that “there are no strict guidelines for collecting this (cellular phone) information, and it cannot be determined from the data whether a cellular phone was being used at the time of the crash or was being used to report the crash” (Ch. 6, p. 5).

In addition to these problems, Cher et al. 1999) have expressed concern about the sensitivity of Violanti’s findings to the small number of driver fatalities where cellular phones were reported in use (N = 5).

They also note several anomalies in how the data are reported and raise serious questions about whether the data were analyzed properly. 33 Dreyer et al. (1999), extending previous work by Rothman et al. (1996), have reported the mortality experiences of a large cohort of users of cellular phones. The cohort was compiled from two large U. S. cellular phone carriers serving several metropolitan areas.

Detailed billing information and demographic data were obtained on each noncorporate customer, with these records linked to the National Death Index to ascertain mortality experience. The authors found that only one cause of mortality (of several dozen examined) was positively associated with cellular phone use (average minutes of daily use): death from a motor vehicle collision. To evaluate the association between phone use and fatal motor vehicle collisions, data for hand-held cellular phone users and car phone users were pooled and examined for trends of fatality risk with increasing phone use.

The reference group for this comparison was phone users with the least amount of use; thus, the association relates to more use compared with less use. For categories of use of less than 1 minute per day, 1 to 3 minutes per day, and more than 3 minutes per day, the authors found motor vehicle mortality rates of 5, 10, and 12 per 100,000, respectively. A similar association was reported with number of calls per day, though an inverse trend was found with history of cellular phone use (in years).

The authors acknowledge that they do not know which calls were placed from a motor vehicle or whether a cellular phone was in use immediately prior to a crash. This promising study reveals an important statistical association that needs to be analyzed in more detail. The association may reflect a cause-effect relationship, or it may be that drivers who use cellular phones more frequently, compared to less frequent users, happen to have more exposure to fatal crashes (e. g. , because they drive more frequently or in a more risk-prone manner).

34

In summary, there are four epidemiological studies published in the peer-reviewed literature which examine whether use of a cellular phone while driving is associated with risk of traffic accident involvement. The two studies of police-reported collisions suggest that use of a cellular phone is associated with collisions that involve property damage, but these data do not address serious or fatal injuries. The two studies that address risk of fatal collisions are positive but problems with data quality and potential confounding variables preclude a confident causeeffect inference.

Nonetheless, the epidemiological evidence is qualitatively consistent with the suggestions of risk found in the driver performance studies and the case reports of crashes. The strongest epidemiological evidence to date (Redelmeier and Tibshirani, 1997) is used below in a quantitative risk comparison. Risk Comparisons. The purpose of risk comparisons is to provide readers an intuitive grasp of risk magnitude (Crouch and Wilson, 1982).

Risk comparisons are a helpful risk communication device but they do not, by themselves, establish whether or not a risk is acceptable (Roth, 1990).

Judgments about risk acceptability require risk-benefit comparisons as well as ethical considerations. Although the precise risks of using a cellular phone while driving are unknown, it is feasible to perform a risk assessment using the best available data and plausible (yet unverifiable) assumptions. We estimated two types of risks: (1) the risk of fatality to the user of the cellular phone, which we compared to other voluntary risks that drivers accept; and (2) the risk of cellular phone use to other road users (occupants of other vehicles, pedestrians, bicyclists), which we compare to other involuntary risks in daily life.

We recognize that there is subjectivity in deciding which risks are voluntary and which are involuntary. 35 We omitted from our calculations risks incurred by passengers traveling with a driver who uses a cellular phone because it is not clear whether these individuals incur the associated risks voluntarily or involuntarily. It is likely that many passengers choose to ride with a driver even though they know the driver may use a cellular phone.

On the other hand, some passengers may not be aware beforehand that the driver will use a cellular phone, or may have no choice about riding with the driver (e. g. , children).

The “voluntary” incremental fatality risks as a driver were calculated by multiplying the annual risk of being in a collision while driving and using a cellular phone (using estimates reported by Redelmeier and Tibshirani, 1997) and the conditional probability of being fatally injured in a collision.

The “involuntary” incremental fatality risk to third parties is the product of three quantities: the number of individuals using a cellular phone while driving, the annual probability of being in a collision while driving and using a cellular phone, and the average number of fatalities per collision to individuals not riding with the cellular phone user. More detail on these calculations, including input data and assumptions, is reported in Appendix A. Table 3 reports the magnitude of voluntary risks assumed by drivers.

The annual probability of driver fatality due to cellular phone use is smaller than the risk of not wearing a safety belt, yet larger than the risk of taking a short trip on a rural road instead of a safer interstate highway. In absolute terms, the average annual probability of death to a driver using a cellular phone is about 6 chances in 1,000,000 (assuming 318 minutes of use per year, or about 26 minutes per month – the estimated average usage rate among cellular phone owners while driving during 1999 [PCIA, 1999]).

36 Table 3. Voluntary Risk Factors Affecting Driver Fatality Rates.

Risk Factor Relative Risk While Risk Factor is Applicable Annual Time or Miles During Which Risk Factor is Applicable 318 minutes 360 minutes Annual Fatalities per Million Drivers Driving while using a cellular phone Driving with a blood alcohol concentration at the legal limit of 0. 10% for one-half hour, 12 times per year (hypothetical) Driving without wearing a lap and shoulder belt (assumes vehicle has airbags) Driving in a small car instead of a large car (1,000 pound difference in weight) Driving 60 miles once per year on a noninterstate rural roadway rather than on a rural interstate highway 4. 3 15 6. 30. 9 1. 41 Always 49. 3 1. 11 Always 14. 5 2. 78 60 miles 1. 5 Note that the estimated risk associated with using a cellular phone is uncertain. Redelmeier and Weinstein (1999) point out that the confidence interval for the relative risk, which reflects uncertainty introduced by the limited sample size in Redelmeier and Tibshirani (1997), ranges from 3. 0 to 6. 5. That source of uncertainty alone indicates that the true incremental risk to a driver associated with the use of cellular phone for 318 minutes per year while operating a motor vehicle ranges from 61% to 167% of the central value reported in Table 3.

There are other potentially important sources of uncertainty that are discussed in Section 1. 1 of Appendix A, including: • The possibility that fatalities are underrepresented in accidents at times when cellular phones are typically used (e. g. , rush hour), a factor not reflected in the Redelmeier and Tibshirani relative risk estimate; The possibility that collision rates are higher at times when cellular phones are typically used (e. g. , because traffic density is greater during rush hour) or lower at these times (e. g. because typical road speeds are lower); The possibility that the collision and fatality risks remain elevated after a cellular phone call ends. 37 • • The magnitude of the uncertainty induced by these factors is difficult to quantify. The overall uncertainty associated with the cellular-phone risk estimate is likely to be larger than the uncertainties associated with the other risk factors reported in Table 3. Involuntary risk estimates are reported in Table 4. Given the current rate of cellular phone use in the U. S. , the average annual probability of being killed in a crash by a cellular phone user is about 1 in 1,000,000.

This probability will increase as cellular phone use increases, unless new users are more cautious than current users, or unless motorists take new precautionary measures to protect themselves. A person is less likely to be killed in a crash caused by a cellular phone user than to be killed as a pedestrian, to be killed by a drunk driver, or to be killed in a crash involving a heavy truck. The average risks imposed on others by cellular phone users are larger than the risks of being struck and killed on the ground by a crashing airplane. Table 4. Involuntary Risk Factors Affecting Fatality Risks.

Risk Factor Motorist struck and killed by driver using cellular phone Sober driver struck and killed by driver with a non-zero blood alcohol concentration Motorist struck and killed in crash with large truck Person struck and killed on ground by crashing airplane Pedestrian struck and killed in motor vehicle crash Annual Fatalities per Million Individuals in the U. S. Population 1. 5 17. 6 16. 8 0. 013 22. 2 Note that the same sources of uncertainty discussed in the case of voluntary risks associated with cellular phone use also apply to the first risk estimate listed in Table 4. 8 Section 4. Benefits of Using a Cellular Phone While Driving In this section, we define “benefits” as any positive consequences – whether tangible or intangible – of using a cellular phone while driving that may accrue to the user of the phone, the user’s family or household, the user’s social network of friends and acquaintances, the user’s business, or the community as a whole. 6 We define “disbenefits” as adverse consequences of cellular phone use while driving other than those related to the risk of motor vehicle collisions.

In any discussion of the benefits of cellular phone use, we recognize that it is difficult to determine which of these benefits relate to use of the phone while driving and which could be maintained if calls were made at other times. We address this complicated issue in Section 5, reserving Section 4 for a qualitative identification of benefits and disbenefits that should be considered. Focus Groups Our review of the scientific literature uncovered few studies that identified the benefits of using a cellular phone while driving.

Given the lack of previous research to identify benefits, we conducted a series of exploratory focus groups to identify the types of perceived benefits of using cellular phones while driving. The authors commissioned Axiom Research Company, LLC of Cambridge, MA to conduct a total of three focus groups in each of two markets of varying size. Springfield, MA and Los Angeles, CA were selected to represent mid-size and large-size markets, respectively. In each market, two focus groups were conducted among consumers and one was conducted among a specialized group of emergency medical personnel and dispatchers (e. . , police dispatchers, AAA, and tow-truck personnel).

A total of 52 participants attended the groups, with an average of 7 in each of the emergency personnel groups and 9. 5 in each of the consumer 39 groups. The two consumer groups were divided so that one consisted primarily of “commuters” (defined as those with at least a 20-minute drive to work in Springfield or at least a 30-minute drive to work in Los Angeles) and the other “non-commuters” (including those who commute, but less than the durations required for commuters).

All participants in the consumer groups were required to own cellular phones and report that they use them while driving a motor vehicle. Each focus group lasted approximately 90 minutes; all were conducted in early March, 2000. Protocols for the focus groups were reviewed and approved by the Human Subjects review committee at Harvard University. More information on the recruitment and analysis methods, in addition to participant characteristics, is included as Appendix B. A full report of the results of the focus groups (Axiom Research Company, 2000) is available from the authors upon request.

Our focus groups did not include people who own cellular phones but do not use them in the car. We also did not include people who do not own cellular phones. Thus, drivers in general are likely to have different perceptions of risks and benefits and thus are worthy of study in the future. Several limitations inherent in focus group research are important to keep in mind while reviewing this section. The purpose of focus groups is to provide context and depth to a research topic (Barbour and Kitzinger, 1999).

The results of focus group research are neither representative of nor generalizable to the population of cellular phone users in a particular geographic area or across the United States. Focus groups can be subject to a “social desirability” bias whereby participants may provide responses that are perceived as being pleasing to the moderator or to other participants. In addition, results can be biased if they reflect the opinions of participants who are naturally more vocal than other participants. While 6 The economics of regulating the use of cellular phones while driving is addressed in Section 5. 40 killed focus group moderators can mitigate these biases, they cannot be completely controlled for during the group discussion or in the analysis. Below we discuss the major types of benefits that were identified either in the focus groups or the existing literature. It should be noted that while this section describes the range of benefits perceived by consumers and emergency personnel, it does not provide context as to the frequency with which they occur, nor their magnitude or value. We also do not address which types of benefits are most likely to be lost by a prohibition on cellular phone use while driving.

Personal Benefits The following are categories of benefits to the user of a cellular phone that are salient to cellular phone owners and users: § Preventing Unnecessary Trips. Use of a cellular phone while driving can reduce unnecessary trips, minimize the length of trips, and diminish overall time on the road by allowing more effective communication with household members, friends/acquaintances, and other parties whose schedules and transit plans need to be coordinated with the schedule of the cellular phone owner/user.

Curtailing unnecessary trips and travel time is not just a time-saving matter; it also curtails exposure to traffic crashes, injuries, and fatalities, cuts fuel expenses, reduces pollution, and reduces wear and tear on a motor vehicle. The ability to make calls while driving may also stimulate some additional travel such as making a stop at the grocery store. Typically this travel has additional value to the owner; otherwise, (s)he would not undertake the additional travel. § Diminishing the Tendency to Speed. When a driver is running late, (s)he may have a endency to speed in order to reach the destination quickly, thereby reducing the imposition on others at his/her destination. Yet, a call from a cellular phone by the driver can notify colleagues 41 of late arrival and diminish the driver’s urge to speed. A possible disbenefit would be excessive speed by a driver who is distracted by use of the cellular phone and does not pay attention as his/her speed exceeds the posted limit. § Contributing to Security and Peace of Mind. Knowing that it is permissible to use a cellular hone while driving reduces worry and stress while contributing to peace of mind. The ability to achieve virtually instant communication, whether from or to a driver, provides the driver a psychological reassurance that unexpected events of the day can be managed effectively and that worries related to uncertainties can be resolved by instant news conveyed through a phone call. The added sense of security may arise when traveling alone, at night, in poor weather, in a crimeridden part of town, in an unfamiliar ocation, or at any moment when the driver is feeling vulnerable. Cellular phone use may also contribute to actual (as well as perceived) security by decreasing stress and time on the road when a driver is lost. Women in particular mention using cellular phones while driving to obtain directions when trying to find an unfamiliar location or when disoriented by environmental conditions (e. g. , inclement weather or nightfall).

A disbenefit of cellular phones is the potential to increase stress while driving in the case of “road rage. For example, a motorist may call emergency services to report another motorist who cut him off (a minor traffic offense) rather than forgetting the incident. This rage can now be acted on, with the driver weaving through traffic to follow the offending car and reporting its position to police. Being able to be contacted while on the road may not always be beneficial. Another disbenefit of cellular phones is that the increased accessibility, including repeated incoming calls from others, may lead to increased stress and ultimately decreased personal peace of mind. 2 § Improving Mental Alertness. Cellular phones may also increase alertness on what might otherwise become a long, monotonous drive. There is some evidence of increased awareness in the literature: Drory (1985) found that voice communication while driving long distances significantly decreased driver fatigue. § Facilitating Privacy in Communication. Making a phone call while driving alone in one’s vehicle can provide a measure of privacy in communications that is not always available when phone calls are initiated from one’s home, business, or a public place.

Moreover, emergency services personnel report that they use cellular phones if they want to talk freely about sensitive information that they do not want broadcast across police radio. § Expanding Productivity for Commuters. Among consumers with long commutes to and from work, conducting personal and business matters by phone while driving is sometimes considered a necessity since these people spend a substantial percentage of their workday in their motor vehicles.

Yet, consumers also report using their cellular phones to call radio stations for promotions, giveaways, or to join talk shows. For the cellular phone user, being accessible while driving (and being able to reach others while driving) is not always perceived as being beneficial. There are some occasions where the driver might prefer not to be reached, at least by particular people or on particular matters, but even in these situations new technology (such as caller ID) is allowing drivers to be more selective about who can reach the driver and when they can be reached.

The consumers in the focus groups were of the general view that the personal benefits of using a cellular phone while driving far outweighed any disbenefits and risks, even though they recognized a measure of risk from this particular use. 43 Family/Household Benefits As the hand-held cellular phone has transitioned from primarily business use to a wide range of uses within households, there are clear day-to-day benefits that have accrued to households. Here are some categories of household benefits and examples that are salient to cellular phone owners/users: § More Efficient Execution of Household Responsibilities.

Communication among household members, while at least one member is in transit, facilitates the accomplishment of household chores, whether it is picking up a gallon of milk on the way home from work, providing notice of an early or late arrival to the dinner table, or making a last-minute adjustment as to who can more readily pick up younger members of the household from day care, soccer practice, or other activities. § Parental and Familial Peace of Mind. When teenagers drive, parents worry. Similar concern is expressed for other members of the household (spouses or parents), articularly if they are driving long distances or driving late at night. Knowing that these individuals can, if necessary, call home while driving or can be contacted while in transit – even if this capability is rarely exercised – contributes to peace of mind for all concerned family members. Yet the risks of teens using a cellular phone while driving are also a source of worry. There may also be a tangible aspect of this benefit. A teacher or caregiver can reach the parent at any time (even in the car), and parents can check on their child at any time.

Use of these capabilities may result in safer and healthier children. § More Time at Home. Use of a cellular phone while driving permits a parent to leave the office earlier, making calls on the commute home, and thereby spend more time with his/her children. 44 Social Network Benefits Some researchers have suggested that America is losing its “social connectedness,” sometimes referred to as social capital, as networks of friends and acquaintances weaken in the face of growing time pressures and the commercialization of life (Putnam, 1995).

Declines in social capital have also been suggested to have detrimental effects on health-related outcomes (Kawachi, 1999).

The use of a cellular phone while driving may help counteract the deterioration of social networks by allowing motorists to capitalize on otherwise idle time to stay in touch. § Increased Social Connectedness. Caught unexpectedly in a 90-minute traffic jam, the driver places a call to a friend he or she has been remiss about contacting, using this time window as a way to reconnect with friends and associates. Coordinating Social Engagements. A social engagement that is jeopardized by scheduling conflicts is rescued rather than cancelled as a driver in transit makes one or more calls at a critical time to hold the engagement together or to reschedule it. Coordinating such events reduces stress on both parties, which might otherwise strain interpersonal relationships. Focus group participants also noted the social disbenefits of widespread cellular phone use. Cellular phones ring during meetings or at the theatre.

People talk on cellular phones while they are at a party or business function. Inappropriate or rude use of cellular phones has a negative effect on social connectedness. Business Benefits Increased use of automated answering or voice messaging systems – often a result of workers on the move – has contributed to frustration among workers and clients trying to get in touch with employees. Possession of mobile technology, such as cellular phones, by workers 45 enhances relationships with clients by improving response time to their calls.

This technology also increases business productivity by making employees more efficient and accessible while traveling. § Increased Productivity and Efficiency. Mentioned earlier as a benefit to the user, cellular phone use can also help businesses by creating productive time out of idle time otherwise spent commuting to the office or traveling between job sites. In business, time is money; conducting business over the phone while driving can generate revenue and allow employees to seize opportunities in a fast-paced, competitive environment. Increased Responsiveness to Clients and Co-workers. Making it possible to contact workers while driving improves an individual’s responsiveness to issues that arise with his/her clients or co-workers. For example, an anxious co-worker is trying to find an answer to a question before a client meeting and is able to do so on his cellular phone while he drives between other client sites. This increased responsiveness likely improves relationships with clients. Clients know that they can reach the consultant in charge of their account almost immediately in a time-sensitive situation.

This ability is especially valuable in industries such as financial services that require quick action. Immediate contact or quick response contributes to clients’ feeling more secure about the organization, and about the value of their business to the company. Community Benefits Communities as a whole benefit from the use of cellular phones while driving. These benefits range from providing emergency services personnel with more accurate and complete information as they approach an accident or crime scene, increased ability to apprehend 6 criminals or thwart a crime in progress, and decreased response time to arrive at roadside emergencies. § Improved Knowledge of Emergencies. Emergency personnel report that, because motorists use cellular phones from their cars to report emergencies, emergency personnel are better able to anticipate the emergency situation and what types of equipment might be needed. Several emergency workers reported that they receive information from multiple callers with different views of the same scene.

This information may help emergency services to better distribute resources (vehicles, manpower), and it has the potential to save the lives of emergency workers (especially police).

For example, a motorist reporting a roadside emergency who notes that a gun is being waved warns police to approach the scene with greater caution. § Apprehending Criminals. Emergency services personnel are receiving more help in identifying and apprehending criminals from citizen drivers. Drivers report more suspicious activity (speeding, road rage, reckless driving) or offenses (e. g. carpool lane violators) on the road than ever before and can provide detailed information about the route the car is taking, car make and model, license plate, and driver characteristics – largely because they have the ability to call while driving. Emergency services personnel overwhelmingly noted that motorists report drunk drivers, which previously occurred only rarely. Getting drunk drivers off the road improves road safety for the drunk driver himself, for any passengers in his/her vehicle, and for all other road users (pedestrians, bicyclists, drivers, passengers).

More generally, the growing proportion cellular phone users in motor vehicles may have an unmeasured deterrent effect on reckless driving and other criminal behavior. Would-be offenders are now aware of the ease with which citizens can report a crime through prompt use of a cellular phone while driving. 47 § Decreased Accident Response Time. The hour immediately following a traumatic injury is referred to as the “golden hour” in medicine: the time during which an individual who has suffered severe injury has an increased likelihood of survival if medical attention is received.

Motorists with cellular phones report roadside accidents immediately after they occur, which improves the overall response time of emergency personnel. Reporting roadside emergencies may benefit the user if he or she were involved in the accident. However, this benefit also extends beyond the user to the community at-large, in a “Good Samaritan” effect. A survey of over 700 Australian cellular phone users confirmed the widespread use of cellular phones for this purpose: 12% of respondents reported using their cellular phones to contact authorities to report roadside accidents involving others (Chapman and Schofield, 1998).

The ability to use cellular phones while driving to report accidents has a downside. Increased numbers of calls to 911 have led to emergency response systems being inundated with calls reporting the same event. It has been suggested that this increased volume can delay other emergencies from being reported and responded to quickly. In addition, emergency personnel have reported an increased number of calls for non-emergency situations. For example, motorists will call to report that their vehicle has broken down along the highway, an event that does not require an immediate emergency response.

Summary There are numerous benefits to being able to use cellular phones while driving. However, the frequency of these benefits is unknown, as researchers have yet to quantify them. Moreover, the value of each benefit has not been assessed: some benefits may be worth more than others as contributions to the common good. Many of these benefits also have tradeoffs. For instance, the increased accessibility of individuals by cellular phone may contribute to a 48 feeling of lost privacy, and the increased reporting of accidents can overwhelm emergency response systems.

Perhaps most importantly, it is not clear which of these benefits would be lost if cellular phone use while driving were banned, since use of the technology at other times would still be permitted (see Section 5).

49 Section 5. Benefits Foregone if Cellular Phone Use While Driving is Restricted A critical question is whether a ban or restriction on the use of cellular phones while driving would cause any loss of the benefits of cellular phone use. It is often argued that users can make whatever calls they need to make at times other than when they are operating a motor vehicle.

For example, a driver may pull over to the side of the road, stop the vehicle, make a call, and re-enter traffic after the call is completed. There are several ways that a regulation of cellular phone use while driving might result in a loss of benefits from the use of cellular phones. None of these possibilities has yet been investigated rigorously in the scientific or policy literature. First, a prohibition on the use of cellular phones while driving might curtail cellular-phone ownership rates among users who find the device useful only or predominantly during motor vehicle travel.

We know of no information as to whether such a user group exists (e. g. , long-distance commuters) or the size of the group and the benefits of the calls they make. In light of the many users who also find this technology useful in non-transportation settings, a decline in overall ownership after a prohibition on use while driving does not seem plausible. Second, a prohibition on the use of cellular phones while driving might reduce the rate at which cellular phones are carried (transported) during motor vehicle trips.

Although a decline in carrying might occur, it seems that many of the security (“peace of mind”) advantages of carrying a cellular phone would remain, even if use while driving were prohibited, thus causing continued possession of phones by drivers. Third, the location/accessibility of the phone in the vehicle might be diminished by a prohibition. Since the benefits and safety of phone use would certainly be diminished if the phone is 50 not readily accessible to the driver, the effects of regulation on phone accessibility need to be investigated.

Finally, a prohibition on the use of a cellular phone while driving could jeopardize the convenience and time-saving feature of communications while driving, perhaps resulting in an overall decline (and/or delay) in the amount of communications that occur between people compared to what would occur if cellular phone use while driving were permitted. Permitting the use of cellular phones while driving does facilitate the productive use of relatively idle time, with potentially important benefits to users, families, social networks, businesses, and communities.

It has been suggested that the benefits of cellular phone use while driving could be preserved under a restrictive regulation since a driver could simply pull over on the side of the road and place (or return) a call, re-entering traffic when the call is completed. We believe that this argument is naive and potentially encouraging of hazardous behavior. If drivers retain a cellular phone in their car and pull over to the side of the road to make their calls, there is an incremental risk of collision associated with stopping on the shoulder of a road/highway and re-entering traffic after the call has been made.

These risks are of unknown magnitude and are not included in the epidemiological comparisons reviewed above. Moreover, many roads and highways are not designed with an adequate shoulder for drivers to pull over safely and make phone calls. In these settings especially, which are common on congested urban freeways, it is not realistic to expect drivers to simply stop for a brief moment at the roadside to make a phone call. If it is suggested that drivers should not make calls until they reach an appropriate stopping point (e. g. a rest area, an exit ramp, or refueling station), then there may be incremental risks from exiting and re-entering traffic streams, as well as additional time on the road. Moreover, the longer the delay between the point at which a call could be made or received (while driving) and the point when it 51 can legally be made or received (under a regulatory policy), the less likely it is that call will occur. Delays in communication may also reduce the supply of accurate and up-to-date information. Delays in making and returning calls can clearly cause diminished benefits due to the time sensitivity of many calls.

There may also be extreme situations where continuous contact between the driver and another party is critical to beneficial communication (e. g. , following another vehicle with a suspected criminal).

If prohibitions apply only to a subset of wireless technologies (e. g. , hand-held designs), then policymakers need to consider the risks and benefits of drivers switching to other communications devices (hands-free phones, pagers, fax machines, and so forth).

Drivers can already convert a handheld phone into a hands-free phone with use of an earpiece or head set.

If a prohibition is more comprehensive with respect to wireless devices or technological distractions, then driver compliance with the prohibition is likely to be lower, especially since these kinds of prohibitions are difficult and tedious for police to enforce. A risk-benefit analysis of a prohibition with partial compliance is even more complicated than an analysis of complete prohibition, as the risks and benefits of discouraged calls may be different from the risks and benefits of calls permitted by law.

If a weakly enforced prohibition curtails only the high-risk and low-benefit calls, the prohibition might be attractive. Yet partial compliance also raises the cost of police enforcement and suggests possible diversion of police time from more cost-effective safety activities (e. g. , primary enforcement of child seat and safety belt laws or drunk driving laws).

If such laws are adopted, it is important to consider whether primary police enforcement should be permitted or whether only secondary police enforcement (where cellular phone violations are enforced only in the context of other offenses) should be permitted. 2 In jurisdictions where restrictions or prohibitions have been adopted, there are often exceptions for calls made while driving that concern emergencies or other public purposes. Although such exemptions are sensible, guidance to motorists and police is required as to what constitutes a compelling emergency or public purpose. If laws were passed restricting cellular phone use while driving, a variety of crucial questions may be treated differently in various jurisdictions: the specific devices covered (e. g. hand-held versus any wireless phone), the nature of calls exempted (if any) from a ban, the type of police enforcement permitted, and the magnitude of fines for noncompliance. It will be difficult for drivers who travel across jurisdictions to know how the traffic laws concerning use of cellular phones vary from locality to locality, from state to state, and from country to country. Given the potential confusion created by non-uniform traffic laws, policymakers should have a clear rationale before introducing new traffic laws in the U. S. 53

Section 6. Cost-Effectiveness of Alternative Safety Measures When considering restrictions on the use of cellular phones while driving, it may be instructive to compare the cost-effectiveness of such restrictions to the cost-effectiveness of other measures to improve traffic safety. The key summary statistic in a cost-effectiveness analysis is the costeffectiveness ratio: the net monetary costs of the measure divided by the number of years of life saved, compared to a well-defined policy alternative. A recent review article by Graham et al. 1998) summarizes cost-effectiveness ratios for a wide range of public health and medical interventions, including selected highway safety programs. The cost effectiveness of a ban on the use of cellular phones while driving is estimated using cost information in Hahn and Tetlock (1999), who quantify the monetary cost of such a ban, and effectiveness information in Redelmeier and Weinstein (1999), who estimate the health benefits, expressed in Quality Adjusted Life Years (QALYs) saved. The QALY is a measure that reflects the value of both reduced morbidity and extended life.

Hahn and Tetlock (1999) offer the most comprehensive estimate of the monetary costs associated with such a proposal. They define costs of a ban as the lost welfare estimated from the willingness of consumers to pay money for cellular phone calls. Aggregated over the U. S. population in 1999, Hahn and Tetlock estimate those costs to amount to $25 billion (p. 11).

From this figure, the direct costs associated with accidents caused by cellular phones (medical expenses, lost productivity, property damage) must be subtracted.

Redelmeier and Weinstein (1999) offer the most comprehensive calculation of this quantity, concluding that it amounts to around $3,000 per collision (p. 3), or approximately $2 billion per year. Hence, the net cost of banning cellular phone use while driving is $25 billion – $2 billion, or $23 billion per year. Weinstein and Redelmeier (1999) also offer the most comprehensive calculation of a ban’s effectiveness, defined as the number of QALYs saved by 54 reventing collisions caused by the use of cellular phones while operating a motor vehicle. They estimate that each year in the United States the collisions caused by use of cellular phones while driving result in a loss of approximately 33,000 QALYs. Hence, the cost-effectiveness ratio for a ban on the use of cellular phones while driving is $23 billion ? 33,000 QALYs, or approximately $700,000 per QALY saved. Both the monetary cost and the effectiveness (QALYs saved) estimates just described are uncertain.

In order to estimate the value consumers place on being able to use their cellular phones while driving, Hahn and Tetlock (1999) first estimate the total value of all cellular phone calls and then estimate the fraction of that value associated with calls made while driving. Estimating the value of all cellular phone calls involves determining how much consumers would insist on being compensated if they were no longer allowed to make those calls, or if they had to substitute for those calls other means of communication (pagers, land-based phone links, etc. ).

The authors conclude that cellular phone calls made while driving have a particularly high value, as these calls are typically made by people who place a high value on their time. For many such people, the next best communication alternatives are inferior to the use of a cellular phone, especially when an individual (such as business person or long-distance commuter) is in his or her own car. Nonetheless, the authors acknowledge that the price elasticity of demand for cellular phone use does have a range of plausible values. They also state that the fraction of all cellular phone calls made while driving may range from 40% to 70%.

Based on these two sources of uncertainty, Hahn and Tetlock (1999) calculate that the net benefits associated with cellular phone use while driving range from $14 to $73 billion. Replacing the direct costs associated with accidents caused by cellular phones estimated by Hahn and Tetlock ($1. 2 billion) with the corresponding figure estimated by Redelmeier and Weinstein ($1. 9 billion) indicates that a ban on cellular phones would result in a net cost to society of $13 billion to $72 billion, with a central estimate 55 of $23 billion.

This source of uncertainty alone indicates that the cost effectiveness ratio may range from approximately one-half ($13 billion ? $23 billion) to more than three times ($72 billion ? $23 billion) its central estimate. The estimates of effectiveness (QALYs saved) are uncertain for two reasons. First, as discussed in Section 3, the incremental collision and fatality risks associated with the use of a cellular phone while driving are uncertain. Taking into account only the stochastic (statistical) uncertainty about the relative risks estimated by Redelmeier and Tibshirani (confidence interval of 3. 0 to 6. ) produces cost-effectiveness ratios ranging from 60% to 165% of the central value7. Second, the average number of QALYs saved per averted collision is uncertain because the value of each averted injury or fatality is uncertain. In the case of the cost-effectiveness ratio calculated by Redelmeier and Weinstein (1999), uncertainty associated with the QALY value of each averted injury is nearly as important as the stochastic uncertainty in the relative-risk estimate. Table 5 reports cost-effectiveness ratios for eight highway safety measures that have been analyzed using comparable analytic methods and assumptions.

From this comparative information, it appears that a prohibition on the use of cellular phones while driving is a relatively inefficient way for policymakers to save lives and reduce injuries from traffic crashes. Efforts to promote safety belt use, for example, are far more cost-effective. 7 Effectiveness (QALYs saved) appears in the denominator of the cost effectiveness ratio. Hence, use of the upper bound on the confidence interval (CI) for the relative risk (6. 5) reduces the cost effectiveness ratio to 60% of its value (6. 5-1. 0 ? 4. 3-1. )-1. Similarly, use of the lower bound on the CI for the relative risk (3. 0) increases the cost effectiveness ratio (3. 01. 0 ? 4. 3-1. 0)-1. 56 Table 5. Cost-Effectiveness Ratios for Selected Highway Safety Investments. 8 Intervention Lap/shoulder belts (assuming 50% use) Daytime running lights Front-crash airbags Side door beams Frontal-crash airbags 55 MPH speed limit Add shoulder belts to lap belts (assuming 9% use) for rear outboard seats Cellular phone restrictions Target Population Front-seat occupants Net Cost Per Life-Year* Saved < $0

All motor vehicles Drivers only Light trucks Front-right passengers Rural interstate travelers Passengers using rear outboard seats < $0 $24,000 $53,000 $61,000 $82,000 $160,000 All drivers $700,000 Passengers using rear center seats > $2,400,000 Add shoulder belts to lap belts (assuming 9% use) for rear center seats * Life-years saved have been adjusted to account for both enhanced life expectancy and improvements in quality of life due to reductions in morbidity and functional impairment due to trauma.

The adjustments are based on the quality-adjusted life year (QALY), a preference-based system that accounts for trauma severity and the health preferences of consumers for quality of life. The information presented on speed limits is particularly instructive. Many states have recently raised speed limits on rural interstate highways from 55 to 65 miles per hour (MPH) (or even higher), despite the fact that the cost-effectiveness ratio for the 55 MPH limit ($82,000) is close to ratios published for passenger airbags, a mandated safety feature in all new motor vehicles.

The “costs” of a lower speed limit are primarily the time/productivity costs to motorists and truckers, the sort of “inconvenience” cost that is central to the policy debate about using cellular phones while driving. Yet a speed limit of 55 MPH on rural interstates would appear to be far more cost-effective than a prohibition on the use of cellular phones while driving. This conclusion should be tempered with the 8 Note: For technical details on assumptions, input data, and primary references regarding specific interventions, see Graham et al. 1998) and Redelmeier and Weinstein (1999).

57 qualification that the cost-effectiveness ratio for the cellular-phone prohibition is highly uncertain. The estimates from the sensitivity analysis reported in Redelmeier and Weinstein (1999) range from $50,000 to $700,000 per QALY saved. There is, moreover, no public consensus about when a cost-effectiveness ratio is small enough to be acceptable. Some public health programs that cost more than $1 million per life-year saved have been adopted while others that cost less than $100,000 per life-year saved have been rejected (Tengs et al. 1995; Tengs and Graham, 1996).

Recent research suggests that public preferences for lifesaving investments may be influenced by a variety of qualitative and ethical considerations (Cookson, 2000).

One of the limitations of cost-effectiveness analysis is that effectiveness is not expressed in the same units as costs, and thus it is impossible to determine whether the effectiveness of a policy is worth the costs. A cost-benefit analysis can provide such a result, at least from the