Article, Traumatology

Obesity and seatbelt use: a fatal relationship

a b s t r a c t

Background: Seatbelts significantly reduce the risk of death in motor vehicle accidents, but a certain number of individuals from some subgroups tend not to wear their seatbelts.

Objectives: In this study, we hypothesized that obese drivers (in fatal crashes) were less likely to wear seatbelts than their normal-weight counterparts.

Methods: A retrospective study was conducted on the drivers in motor vehicle crashes entered into the Fatality Analysis Reporting System database between 2003 and 2009. A number of precrash variables were found to be significantly associated with seatbelt use. These were entered into a Multivariate logistic regression model using stepwise selection. Drivers were grouped into weight categories based on the World Health Organization definitions of obesity by body mass index. Seatbelt use was then examined by body mass index, adjusted for precrash variables that were significantly associated with seatbelt use.

Results: The odds of seatbelt use for normal-weight individuals were found to be 67% higher than the odds of seatbelt use in the morbidly obese. The relationship of seatbelt use between the different weight groups and the morbidly obese is as follows (odds ratios [ORs] for each comparison are listed with 95% confidence limits [CL]): underweight vs morbidly obese (OR, 1.62; CL, 1.47-1.79), normal weight vs morbidly obese (OR, 1.67; CL, 1.54-1.81), overweight vs morbidly obese (OR, 1.60; CL, 1.48-1.74), slightly obese vs morbidly obese (OR, 1.40; CL, 1.29-1.52), and moderately obese vs morbidly obese (OR, 1.24; CL, 1.13-1.36).

Conclusion: Seatbelt use is significantly less likely in obese individuals compared with their normal-weight counterparts.

(C) 2014

Introduction

According to the US Census Bureau, 33 808 people were killed in motor vehicle accidents in the United States in 2009 [1]. Although this number represents an overall decrease in the number of fatalities resulting from motor vehicle accidents, many more can be prevented by the use of seat belts. Research indicates that use of combined lap and shoulder belts is 45% more effective in reducing crash fatalities and 50% more effective in reducing moderate to critical injuries for front passengers in motor vehicle crashes [2].

Obesity is a known risk factor for many medical illnesses, but it has also been increasingly identified as a risk factor for morbidity and mortality after trauma. Studies have shown that there is a protective effect to being overweight; however, there is an increased risk of death for moderately and morbidly obese drivers [3,4]. Past research has largely attributed the relationship between body mass index (BMI) and increased mortality in the moderately and morbidly obese

? Funding sources/disclosures: None.

?? Prior presentations: Annual SAEM meeting; May 10, 2012; Chicago, IL, USA.

* Corresponding author. Chirag Doshi, University at Buffalo School of Medicine and Sciences, 45 Biomedical Education Building, Buffalo, NY 14214. Tel.: +0017164640319.

E-mail address: [email protected] (C. Doshi).

to greater difficulties in providing medical care (eg, surgery) and the larger number of comorbididities [5]. However, newer studies also show different injury patterns in the obese population, largely due to differences in seat belt use [6].

Body weight and seat belt use in adults have been intermittently examined over the past 30 years. The initial data came from 1981 to 1983 Behavioral Risk Factor Surveillance Surveys, which found that obesity was associated with decreased reported use of seat belts [7].A 1995 self-reporting study of internal medicine patients found that obesity put patients at risk for not wearing seat belts [8]. By 2002, the Behavioral Risk Factor Surveillance Surveys had further described the phenomenon, showing that those who reported always using seat belts decreased linearly with increasing body mass [9]. However, there are inaccuracies involved in self-reported studies of seatbelt use that can be partially overcome by the Fatality Analysis Reporting System (FARS) database. Seatbelt use is documented in FARS from direct observations of Prehospital providers, police, and other vehicle occupants; augmented by observed belt marks in crash victims wearing seatbelts.

This study seeks to retrospectively analyze the rates of seatbelt use for obese individuals as compared with normal-weight individuals, when looking at a group of drivers involved in fatal Motor vehicle collisions and controlling for confounders. We hypothesize that obese

http://dx.doi.org/10.1016/j.ajem.2014.01.010

0735-6757/(C) 2014

drivers (in fatal crashes) were less likely to wear seatbelts than normal-weight drivers.

Methods

Study design

This study was a retrospective analysis of data collected through FARS, a database run by the National Highway Traffic Safety Administration (NHTSA). To be included in FARS, a motor vehicle accident had to occur on a public roadway and had to result in the death of an individual within 30 days of the accident. The study included all such accidents occurring between the years 2003 and 2009. This study was approved by our Health Sciences Institutional Review Board.

Study setting and population

The FARS, which became operational in 1975, contains data on a census of fatal traffic crashes within the 50 states, the District of Columbia, and Puerto Rico. The NHTSA has a cooperative agreement with an agency in each state’s government to provide information on all qualifying fatal crashes to FARS. These agreements are managed by regional contracting officer’s technical representatives located in the

10 NHTSA Regional Offices. Trained state employees, called FARS Analysts, are responsible for gathering, translating, and transmitting their state’s data to the National Center for Statistics and Analysis in a standard format. The FARS data are obtained solely from the state’s existing documents: Police Accident Reports, State Vehicle Registra- tion Files, State Driver Licensing Files, State Highway Department Data, Vital Statistics, Death Certificates, Coroner/Medical Examiner reports, Hospital Medical Reports, Emergency Medical Service Re- ports, and other state records. From these documents, the analysts code more than 100 FARS data elements. The specific data elements may be modified slightly each year to conform to changing user needs, vehicle characteristics, and highway safety-emphasis areas. The FARS data do not include any personal information and therefore fully conform to the Privacy Act. A more extensive description of the FARS data elements can be found in a number of articles [3,15].

Selection of participants

The information obtained from the FARS database was confined to crashes that involved passenger vehicles. Large vehicles (buses, large vans, and trucks), pedestrian crashes, and motorcycle crashes were excluded from the study to minimize confounding factors. Only drivers were considered, as the data collected by FARS on drivers included height and weight, allowing for calculations of BMI. Any cases with missing or unknown information regarding factors that were found to be statistically significant by univariate analysis were excluded in the primary multivariate analysis. A secondary multivar- iate analysis was performed limiting the number of excluded cases.

Weight and BMI

slightly obese (BMI 30 to b 35), moderately obese (BMI 35 to b 40), and morbidly obese (BMI >=40).

Statistical analysis

The software used for data analysis was SAS (version 9.3; SAS, Cary, NC). After testing univariate models, a number of precrash variables were found to be significantly associated with seatbelt use. These were entered into a multivariate logistic regression model using stepwise selection. Seatbelt use was then examined by BMI, adjusted for the precrash variables that were significantly associated with seatbelt use on preliminary univariate analysis. The full model was then analyzed with backward selection, at an ? of .05.

Race

In the FARS database, race is recorded only for fatal victims. An analysis of driver deaths, controlling for confounders plus race, was performed and is presented separately.

Results

Observations

There were 336 913 drivers in the FARS database who met the preliminary inclusion criteria for the study. The characteristics of the sample population and crashes are shown in Table 1. Missing values resulted in the elimination of 142 793 drivers, yielding a sample set of 194 120 drivers that made up the study population, as shown in Table 2.

The confounding variables that were found to be statistically significant by univariate analysis in relationship to seatbelt use included: age of driver, number of occupants and fatalities, wheelbase of vehicle, Weather conditions, lighting conditions, owner of the vehicle, license status of the driver, license restrictions, alcohol use, State in which the crash occurred, and the vehicle body type. Any cases with missing or unknown information regarding age of the driver, seat belt use, vehicle body type and wheelbase, number of occupants, number of fatalities, weather conditions, light conditions, car ownership, license status of the driver, license restrictions, alcohol use, and driver height or driver weight were excluded from the primary analysis of the study. In addition, we found a significant univariate association between driver’s race and seatbelt use.

Outcomes

The adjusted rate of seatbelt use for the total study population, reported as the odds ratio (OR) of seatbelt use comparing each BMI group with morbidly obese drivers with 95% confidence intervals (CIs), is reported in Table 3. The table also highlights the comparison by breaking down results of the 3 logistic models used for analysis. Each group of drivers with a lower BMI than the morbidly obese has a statistically significant higher rate of seatbelt use. In addition, there appears to be a linear correlation between BMI class and seatbelt use: the OR for seatbelt use (compared with the morbidly obese) is

Drivers were grouped into weight categories based on the World Health Organization definitions of obesity as stratified by BMI. Body mass index is calculated as body weight (in kilograms)/height2 (in meters2). The distinction between weight and obesity is made on the

Table 1

Characteristics of the sample population

Mean SD

basis of the BMI. The BMI is calculated as follows: BMI = body weight (in kg) / height2 (in meters2). Obesity was stratified by BMI using the recommended classifications for BMI adopted by the National Institute of Health and the World Health Organization. This classification uses the BMI ranges: underweight (BMI b 18.5 kg/m2), normal weight (BMI 18.5 to b 25), overweight (BMI 25 to b 30),

Age (y) 39.8 19.1

Height (in) 67.7 4

Weight (lb) 168.4 40.4

No. of occupants 1.7 1.1

No. of fatalities 1.2 0.5

BMI (kg/m2) 25.7 5.3

Vehicle’s wheelbase (in) 1069.3 74.8

Table 2

Reasons for sample exclusion

Variable Missing observations (% of total exclusion)

Seatbelt use

0 (0)

BMI

0 (0)

Age

124 (0.09)

Occupants

0 (0)

Fatalities

0 (0)

Weather

1025 (0.72)

Lighting

0 (0)

Owner

2199 (1.55)

Vehicle license status

762 (0.64)

License restrictions

1560 (1.10)

Alcohol use

0 (0)

State of crash

0 (0)

Wheelbase

56 763 (39.99)

Vehicle body type

135 629 (95.55)

Percentages sum to greater than 100 because a single record can have multiple missing variables.

estimated to increase by 13% with each successive decrease in BMI category, as shown in Fig.

The adjusted rate of seatbelt use for the total study population, reported as the OR of seatbelt use comparing each BMI group with normal-weight drivers with 95% CIs, is reported in Table 4. Each group of drivers with a higher BMI than normal has statistically smaller odds of seatbelt use. Odds of seatbelt use in underweight drivers are not statistically different from odds of seatbelt use in drivers of normal BMI classification.

The secondary multivariate analysis was performed without specific vehicle type and wheelbase (responsible for approximately 99% of the exclusions). There were 332 102 drivers in the FARS database who met these secondary inclusion criteria reducing the number of excluded drivers to 4811. The trends were similar to what was found in the primary analysis. The ORs of seatbelt use of drivers in severe motor vehicle crashes, stratified by BMI and compared with morbidly obese for the population that met these secondary inclusion criteria, were as follows: underweight vs morbidly obese (OR, 1.38), normal weight vs morbidly obese (OR, 1.39), overweight vs morbidly obese (OR, 1.32), slightly obese vs morbidly obese (OR, 1.19), and moderately obese vs morbidly obese (OR, 1.15). All results were statistically significant with P b .02.

Restricting the sample to instances where the driver’s race was recorded, there were 88 291 drivers in the FARS database. After adjusting for confounders, we found similar trends in seatbelt use controlling for BMI. The ORs of seatbelt use of drivers in severe motor

Table 3

Odds ratio of seatbelt use of drivers in severe motor vehicle crashes, stratified by BMI and compared with morbidly obese for total population

Model

BMI

OR (95% CI; P value)

A

B

Underweight vs morbidly obese Normal weight vs morbidly obese Overweight vs morbidly obese Slightly Obese vs morbidly obese Moderately obese vs morbidly obese Underweight vs morbidly obese Normal weight vs morbidly obese

1.62 (1.47-1.79; b.0001)

1.67 (1.54-1.81; b.0001)

1.60 (1.48-1.74; b.0001)

1.40 (1.29-1.52; b.0001)

1.24 (1.13-1.36; b.0001)

1.38 (1.27-1.49; b.0001)

1.39 (1.31-1.47; b.0001)

C

Overweight vs morbidly obese Slightly obese vs morbidly obese Moderately obese vs morbidly obese Underweight vs morbidly obese Normal weight vs morbidly obese Overweight vs morbidly obese Slightly Obese vs morbidly obese Moderately obese vs morbidly obese

1.32 (1.24-1.40; b.0001)

1.19 (1.12-1.27; .0174)

1.15 (1.07-1.23; .0001)

1.50 (1.31-1.71; b.0001)

1.51 (1.36-1.68; b.0001)

1.41 (1.27-1.56; b.0001)

1.27 (1.13-1.41; b.0001)

1.19 (1.05-1.35; .0063)

A: model including all covariates significant on univariate analysis. B: model removing specific vehicle type and wheelbase. C: model including race of driver.

vehicle crashes, stratified by BMI and compared with morbidly obese for the total population (after restricting samples to race and adjusting for confounders), were as follows: underweight vs morbidly obese (OR, 1.50), normal weight vs morbidly obese (OR, 1.51), overweight vs morbidly obese (OR, 1.41), slightly obese vs morbidly obese (OR, 1.27), and moderately obese vs morbidly obese (OR, 1.19). All results were statistically significant with P b .01.

Discussion

This study shows that obesity is a significant risk factor for not wearing a seat belt. Not buckling up is a deadly decision. Obese drivers are far less likely to wear seatbelts than are drivers of normal weight, which puts them at a greater risk of being subjected to higher impact forces and being ejected from the vehicle, both of which lead to more severe injury and/or death. Our study shows that normal-weight drivers are 67% more likely to wear a seatbelt than morbidly obese drivers and that the relationship between the degree of obesity and seatbelt use was linear. This confirms with observational data that had previously been based on self-reporting surveys: the more obese the driver, the less likely that seatbelts are used [9]. This is by far the largest study to date to look at the association of obesity with seatbelt use in motor vehicle crashes. The strength of the results comes from the large size of the database, the mandatory entry into the database for any crash with a death, and the relatively low dropout rate due to the required completion of the data as part of a legal document.

Obesity is second only to cigarette smoking as a cause of Preventable deaths in the United States [16]. Obesity has been shown to be a significant risk factor for mortality in motor vehicle crashes, and morbidly obese individuals are 56% more likely to die in a crash than individuals of normal weight [3]. An explanation for the higher mortality rates of obese individuals in crashes has previously focused on comorbid conditions associated with obesity and the closer proximity to the steering column for obese drivers.

It is well known that obesity puts patients at risk for a number of other health conditions, including hypertension, dyslipidemia, car- diovascular disease, stroke, Gallbladder disease, osteoarthritis, Sleep apnea, and Non-Insulin-dependent diabetes mellitus [10]. Obesity has also been associated with mortality after surgical procedures and postoperative complications such as myocardial infarction, Wound infection, nerve injury, and urinary tract infection [11]. Even without surgery, patients with blunt trauma having higher BMIs have been found to have an increased likelihood of developing acute respiratory failure, acute respiratory distress syndrome, acute renal failure, pneumonia, deep vein thrombosis, decubitus ulcers, and multiorgan system failure [12].

Our study hypothesizes that the reason obese drivers are less likely to wear seatbelts than their normal-weight counterparts is that obese drivers may find it more cumbersome to buckle up a standard seatbelt. Our study, therefore, has implications for the design of car interiors and the design of seat belts. These findings raise questions about current vehicle crash testing and safety protocols. Crash testing and occupant crash protection designs, which include seat belt assemblies, follow guidelines from the Federal Motor Vehicle Safety Standards. According to the Federal Motor Vehicle Safety Standards, the standard dummy used in the driver position for the frontal crash test and other designs is the H3CD, which represents a 50th percentile male subject (5 ft 10 in, 170 lb [1.78 m, 77.11 kg]; BMI 24 kg/m2)[13]. The dummies that are used in crash tests should be representative of the population as a whole; however, there are no obese dummies.

According to the Centers for Disease Control and Prevention, as of 2009 to 2010, one-third of the US population is overweight (not obese) and one-third is considered obese (BMI N 30 kg/m2). Thus, current crash testing and car designs may not be optimal for most users. Obese drivers may find it more difficult to buckle up the current

Fig. Odds ratio for seatbelt use (compared with morbidly obese).

standard seatbelt. Current testing protocols do not represent the ever- growing obese population, and the study by Forman [17] shows that using current restraint mechanisms, the kinematics of morbidly obese restrained occupants results in less protective effect and higher risk of injury in motor vehicle crashes.

Improved vehicular design may lead to more seat belt use, making cars safer for the obese. How can the automobile industry make it more likely for people, including the overweight or obese, to wear seatbelts? Is there something the automobile industry can do to make cars safer for the obese? In the interim, an online search of seatbelt extenders reveals that there are extension pieces that can be attached to the existing seatbelt to make them longer. If purchased, the one selected needs to match up correctly with the most common seatbelt attachment designs (of which there are at least 4). Some may find this process burdening and abandon trying to find a solution to their seatbelt issue altogether. In the short term, perhaps these kinds of devices could work to make seatbelts more comfortable, whereas the automotive industry improves their designs. This is an important consideration given the growing number of overweight and obese members of the population who are putting themselves at risk on a daily basis as they get behind the wheel.

Limitations

There are several limitations to our study. First, stratifying patients by BMI may not completely reflect percentage of body fat. For example, if an individual has a large muscle mass, he or she will have a BMI that reflects obesity when, in fact, he or she has increased lean mass rather than fat. It has been postulated that waist circumference

Table 4 Odds ratio of seatbelt use of drivers in severe motor vehicle crashes, stratified by BMI and compared with normal weight for total population (after adjusting for confounders)

BMI

OR (95% CI)

Underweight vs normal

0.97 (0.91-1.03)

Overweight vs normal

0.96 (0.93-0.98)

Slightly obese vs normal

0.84 (0.81-0.87)

Moderately obese vs normal

0.74 (0.70-0.79)

Morbidly obese vs normal

0.60 (0.55-0.65)

size may actually be more accurate than BMI in defining obesity [14]; however, this is not available in the FARS database.

Another limitation is the exclusion of subjects from the study due to missing data where 42.1% of the original study population was excluded. As a secondary analysis, we performed a similar analysis limiting the number of excluded drivers to approximately 1.4% of the study population. When specific vehicle type and wheelbase were eliminated from the analysis (approximately 99% of exclusions), we found results similar to the primary analysis.

The rate of seatbelt use in a database that requires death will be lower than population at large. However, the ratios between different BMI categories should be relatively unchanged.

Conclusion

Obese drivers are far less likely to wear seatbelts than are drivers of normal weight, a behavior that puts them at greater risk for severe injury or death during motor vehicle crashes. These findings have important implications for traffic safety and motor vehicle design.

References

  1. U.S. Census Bureau. The 2012 Statistical Abstract. “1105–Fatal Motor Vehicle Accidents–National Summary”. http://www.census.gov/compendia/statab/ cats/transportation/motor_vehicle_accidents_and_fatalities.html Date accessed: November 15, 2012.
  2. Dellinger HM, Sleet DA, Shults RA, Rinehart CF. Interventions to prevent motor vehicle injuries. In: Doll LS, Bonzo SE, Mercy JA, Sleet DA, editors. Handbook of injury and Violence prevention. New York, NY: Springer; 2007. p. 55-79.
  3. Jehle D, Gemme S, Jehle C. Influence of obesity on mortality of drivers in severe motor vehicle crashes. Am J Emerg Med 2012;30(1):191-5.
  4. Neville AL, Brown CV, Weng J, Demetriades D, Velmahos GC. Obesity is an independent risk factor of mortality in severely injured blunt trauma patients. Arch Surg 2004;139(9):983-7.
  5. Arbabi S, Wahl WL, Hemmila MR, Kohoyda-Inglis C, Taheri PA, Wang SC. The cushion effect. J Trauma 2003;54:1090-3.
  6. Tagliaferri F, Compagnone C, Yoganandan N, Gennarelli TA. Traumatic brain injury after frontal crashes: relationship with body mass index. J Trauma 2009;66(3): 727-9.
  7. Goldbaum GM, Remington PL, Powell KE, Hogelin GC, Gentry EM. Failure to use seat belts in the United States. The 1981-1983 behavior risk factor surveys. JAMA. 1986 May 9; 255(18):2459-62.
  8. Hunt DK, Lowenstein SR, Badgett RG, Steiner JF. Seat belt nonuse by internal medicine patients: a missed opportunity in clinical preventive medicine. Am J Med 1995;98(4):343-8.
  9. Schlundt DG, Briggs NC, Miller ST, Arthur CM, Goldzweig IA. BMI and seatbelt use. Obesity 2007;15(11):2541-5.
  10. Significant advances in the treatment of obesity. Secaucus, N.J.: Network for Continuing Medical Education; 1997.
  11. Bamgbade OA, Rutter TW, Nafiu OO, et al. postoperative complications in obese and nonobese patients. World J Surg 2007;31(3):556-60 [discussion 561].
  12. Newell MA, Bard MR, Goettler CE, et al. Body mass index and outcomes in critically injured blunt trauma patients: weighing the impact. J Am Coll Surg 2007;204(5): 1056-61 [discussion1062-1054].
  13. Moran SG, McGwin Jr G, Metsger JS, Windham ST, Reiff DA, Rue III LW. injury rates among restrained drivers in motor vehicle collisions: the role of body habitus. J Trauma 2002;52:1116-20.
  14. Wang J, Wang J. Waist circumference: a simple, inexpensive, and reliable tool that should be included as part of physical examinations in the doctor’s office. Am J Clin Nutr 2003;78(5):902-3.
  15. NHTSA Fatality Analysis Reporting System Fatal Crash Data Overview (Brochure) [Publication number 809726, Year 2005]. http://www-nrd. nhtsa.dot.gov/CATS/listpublications.aspx?Id=I&ShowBy=DocType Last accessed: 11/15/2012
  16. Mokdad AH, Marks JS, Stroup DF, et al. Actual causes of death in the United States, 2000. [Erratum appears in JAMA. 2005 Jan 19;293(3):293-4; PMID: 15657315].

    Jama 2004;291(10): 1238-45.

    Forman J. The effect of obesity on the restraint of automobile occupants. Annals of Advances in Automotive Medicine 2009;53.

Leave a Reply

Your email address will not be published. Required fields are marked *