Preliminary Assessment of NASS CDS Data Related to
Rearward Seat Collapse and Occupant Injury

Introduction

Since 1988, the National Accident Sampling System (NASS) has recorded data on occupied seat performance into the data base of towaway crashes known as the Crashworthiness Data System (CDS) [1]. The types of seat failures recorded and which have been automated since 1990 are;

a) Seat Adjuster failed,

b) Seat back folding locks or "seat back" failed,

c) Seat tracks/anchors failed,

d) Deformed by impact of occupant, and

e) Deformed by passenger compartment intrusion.

A 1992 NHTSA report analyzed seat performance and occupant injuries in terms of these parameters for 1988 - 1990 data [2]. The report also addressed vehicle size and belt use as well as crash direction and severity.

A review of the NASS CDS Data Collection Manual [3] shows that parameter (a) incorporates the performance of all seat adjusting mechanisms, including the seat back recliner. A recliner failure is to be recorded only if the seat back released in a rearward direction. In an analogous, but opposite way, parameter (b) is to be recorded only if the seat lock failures causes the seat to move forward. However, it is not possible to determine from parameters (a) - (e) the initial and final position of the seat back. This could only be done by a manual search of the hard copy data.

Starting in 1995, CDS began placing the initial and final seat back position in the computerized data base. These positions are illustrated in Figure 1. The seat back position prior to impact can be coded as "upright," "slightly reclined," or "completely reclined." The dark lines shown in Figure 1 correspond to codes 14, 23 and 31, respectively, and represent the special cases where the seat position was unchanged by the impact. There are seven post-impact coding options between completely forward and completely reclined for each of the three pre-impact positions.

The seat back position parameter has, for the first time, provided an opportunity to use the automated CDS to assess occupant injuries in relation to rearward seat collapse. This report is the first attempt to make this assessment based on the limited data available to date. The analysis has been restricted to rear impacts. A method of determining the relative injury costs associated with seat backs maintaining their upright position or failing rearward is presented along with preliminary estimates of injury cost.

Probability of Rearward Seat Back Collapse in Rear Impacts

CDS data was reviewed to determine the likelihood a seat back would maintain its initial upright or slightly reclined position (code 14 or 23) or end up in a completely reclined position (code 11 or 21) subsequent to a rear impact. To maximize the data available for analysis, the preliminary 1996 data was used as well as the 1995 data. The data have been weighted to give national estimates although the 1996 data weighting factors will change after the complete data set is obtained. The results are presented in Table 1. The data were segmented into ranges of DeltaV. Crashes with an unknown DeltaV were omitted. The DeltaV value represents the change in velocity of the struck vehicle as estimated from the amount of vehicle body deformation or crush. A larger DeltaV represents a more severe crash.

†(Percent of total number of seat backs initially upright or slightly reclined)

The total number of rear impacts at the 0 - 9 kmph DeltaV range is small because only towaway crashes are admitted to the CDS data base. None of the seat backs in this DeltaV range collapsed. However, Table 1 clearly indicates that for the remaining DeltaV ranges, for more severe crashes there was a greater chance of seat back collapse. At the highest DeltaV range where a comparison can be made between the collapsed and maintained seat backs (40 - 54 kmph), approximately half the seat backs had a completely reclined post-crash position.

Injury Rates for Seat Backs Which Collapsed or Maintained Their Position

An analysis of CDS data was performed to determine the injury rate to front outboard seat occupants when the primary vehicle damage was in the rear of the vehicle and the seat back position, after the crash, was completely reclined. It was also required that the inclination of the seat back prior to impact was upright or only slightly reclined. Injury levels were segmented into four categories. One category was for moderate to maximum severity injury with an Abbreviated Injury Scale (AIS) between two and six. Another category was for whiplash injury which was defined here as an AIS 1 neck injury, excluding skin-only damage. These two categories were made completely independent of each other by separating out any cases where both whiplash and AIS 2 injury occurred into a third category. The final category was for no injury or AIS 1 non-whiplash injury. Again, to maximize the data available for analysis the preliminary 1996 data was used as well as the 1995 data. Table 2 contains the results of the data analysis.

†Parenthetical values are unweighted CDS data.

The data in Table 2 indicate that when a seat back collapsed rearward (completely reclined post-crash) in a rear impact, the occupant of that seat received an AIS 2 or greater injury an estimated 5.1% of the time. When the seat back maintained its original upright or slightly reclined position the estimated injury rate was 0.95%. Thus, the collapsed seat occupant was 5.4 times as likely to be at least moderately injured. However, this does not necessarily indicated a causal relationship between the seat back collapse and likelihood of injury. Another factor which must be considered is the severity of impact. Table 1 showed that a seat is more likely to collapse in a more severe crash. One would also expect that a more severe crash is more likely to result in an occupant injury.

In an attempt to correct for accident severity the data was segmented into ranges of DeltaV. Crashes with an unknown DeltaV were omitted. A larger DeltaV represents a more severe crash. Table 3 shows the AIS 2 or greater injury rate, without whiplash, as a function of seat back collapse for several ranges of DeltaV. The parenthetical values are the raw count of occupants injured versus the total number of involved occupants.

†(Unweighted CDS injury count/Unweighted total number of involved occupants)

From Table 3 the data from three DeltaV ranges can be compared. The AIS 2 or greater injury rate when the initial seat back positions are maintained is greater than when seat backs collapse, for two of three DeltaV ranges. The largest variation was for the 40 - 54 kmph range, with a 8.3% injury rate for the collapsed seat backs and a 45.8 % injury rate for the maintained seat backs. However, these estimates are based on very limited data and further analysis would be necessary to determine if the differences are statistically significant.

A frequent injury with long term health effects occurring in rear impacts is whiplash. It is well established that the relative position of the occupant and head restraint have an effect on the occurrence of whiplash and its severity. However, the CDS data base doesn't contain this head restraint information. Table 2 shows that the whiplash rate for seat backs that remained in their initial upright or slightly reclined position is 2.1 times as great as for seat backs which collapsed. This may indicate some benefit for reduced whiplash when the seat back collapses. However, there may be some other factors contributing to this result.

In an attempt to correct for accident severity, Table 4 shows the data segmented into ranges of DeltaV. This is a re-creation of Table 3 using whiplash injuries. As was the case for the data of AIS 2 and greater injury, the whiplash data become sparse when segmented by DeltaV. Therefore, any patterns indicated by the data may not be statistically significant. For the 25 - 39 kmph DeltaV range, the maintained seat backs have 27 times the whiplash rate of collapsed seat backs. For the highest DeltaV range (40 - 54 kmph) where comparison can be made between the collapsed and maintained seat backs, the collapsed seat backs have a whiplash rate of 2.7% and the maintained seat backs have a whiplash rate of 26.5%.

†(Unweighted CDS injury count/Unweighted total number of involved occupants)

Table 5 shows the injury category where both a whiplash and AIS 2 and greater injury occur. Clearly this data is extremely limited with only seven unweighted cases in the data base.

†(Unweighted CDS injury count/Unweighted total number of involved occupants)

Occupant Injury Cost Estimates

The injury rate data presented above seem to indicate a trade-off in the type and quantity of injuries depending on whether a seat back maintained its initial position or was in a reclined position after the rear crash event. The relative injury costs associated with seat back collapse must be considered when determining preferable seat performance. The following is an attempt to calculate the relative injury costs.

As was the case for the injury rate estimates, to compensate for crash severity, the injury cost estimates are made within the DeltaV ranges previously defined. Equation (1) defines the injury cost per seat when a seat back maintains its upright or slightly reclined position for a rear impact in the 10 - 24 kmph DeltaV range. Equation (2) is the per seat cost when a seat collapses after impact. The subscripts "M" and "C" indicate the maintained or collapsed case. Similar pairs of equations can be written for the 25 - 39 kmph and 40- 54 kmph DeltaV ranges. The 0 - 9 kmph and 55+ are not used because no data exist for the collapsed seat situation.

Where,

mc = Injury cost per seat if a seat back maintains its original position

cc = Injury cost per seat if a seat back collapses

AR = Injury Rate for AIS 2

WR = Injury Rate for Whiplash

AWR = Injury Rate when Whiplash and AIS 2 are present

AC = Estimated Cost of AIS 2 Injury

WC = Estimated Cost of Whiplash

AwC = Estimated Cost of AIS 2 Injury when a whiplash is also present

According to 1994 NHTSA estimates, when the most severe injury to an occupant is an AIS 1 injury to the face, head or neck the cost is $5,893 per person [4, pp.66]. This is assumed to be the cost of a whiplash injury (WC). This estimate excludes property damage and travel costs, but includes medical, legal, insurance, productivity, and work costs. The same NHTSA report estimates the average cost of AIS 2 - AIS 5 injuries and fatality excluding property damage and travel costs [4, pp.8]. The cost of AIS 2 injury (AC) was estimated by determining the frequency of occurrence of each injury type for each DeltaV range and seat performance. The frequency was divided by the total number of injuries and fatalities to calculate a weighting factor to apply to each cost. The costs were then summed to arrive at the values of AC given in Table 6.

For the cases where both an AIS 2 injury and whiplash were present, the same process was followed to get the cost estimate (AwC). However, all injuries were at the AIS = 2 level, so AwC was set equal to $31,164. Substitution of the injury cost estimates, along with the injury rates of Tables 3 - 5, into equations for "mc" and "cc" yield the values in Table 7.

At the 10 - 24 kmph DeltaV range the injury cost per collapsed seat is nearly 3.8 times that of a maintained seat. For the 25 - 39 kmph DeltaV ranges the cost per collapsed seat is 1.2 times as large as that of a maintained seat. At the 40 - 54 kmph DeltaV range the injury cost per maintained seat is nearly 3.1 times that of a collapsed seat.

The above injury cost comparison doesn't consider the population of rear impacts in each DeltaV range. One way of doing this is to compare total injury cost in two theoretical seat performance cases. In the first case all seat backs in the population of rear impacts are designed such that they maintain their position. In the second, all seat backs in the population are designed to collapse rearward. The calculation is done for each DeltaV range and summed together. The total injury cost in any DeltaV range is the injury cost per seat times the number of seats in this population. Equation (3) shows the total injury cost if all seat backs maintained their position. Analogously, equation (4) represents the injury cost if all seat backs collapsed.

Where,

MC = Injury cost if all seat backs where maintained in original position.

CC = Injury cost if all seat backs collapsed.

TP = Total population of rear impacts in each DeltaV range.

The ratio of the total cost of injury if all seat backs collapsed over the total cost of injury if all seat backs were maintained is represented by equation (5).

Using the per seat cost values from Table 7 and the population values from Table 2,

Which means that if all seat backs collapsed the injury cost would be 2.83 times as much as if all seat backs maintained their initial position.

Another way of assessing the cost of injury while considering the population of rear impacts in each DeltaV range is the following. First, the real or actual cost of injury over the entire available data set is calculated by equation (7).

RC = Real total injury cost.

PM = Population of maintained seats in each DeltaV range.

PC = Population of collapsed seats in each DeltaV range.

Next, another theoretical injury cost estimate is made assuming that at each DeltaV range all seats either maintain their position or collapse depending on which produces the minimum cost. This will produce the minimum or optimum injury cost estimate for the current data set and is represented by equation (8).

OC = Optimum injury cost estimate.

By substitution of the values of Tables 2 and 7, RC = $457 million and OC = $428 million.

This indicates that the optimum cost of injury is $29 million less than the real injury cost. This is a difference of 6.3 percent or $130 per seat.

Discussion and Conclusions

The 1992 NHTSA analysis of CDS seat performance data did not determine the relationship between seat back collapse and occupant injury because the post-crash seat back position was not coded until 1995. The coding of seat back position has, for the first time, allowed for preliminary analysis of seat back collapse and injury potential. However, the limited amount of data does not allow the exploration of the effect of parameters such as belt use, vehicle weight, and occupant descriptive information.

The data seem to indicate a trade-off in the type and quantity of injuries depending on whether a seat back maintained its initial position or was in a reclined position after the rear crash event. The AIS 2 level injury rate increased when the seat back collapsed and the whiplash injury rate increased if the seat back position was maintained. This pattern held true for whiplash injury even when accidents were segmented by DeltaV. The ratio of whiplash to AIS 2 level injury was 2:1 and 21:1 for collapsed and maintained seat backs, respectively, for all rear crashes.

It was estimated that the injury cost per collapsed seat was higher than for a maintained seat at the 10- 24 kmph and 25 - 39 kmph DeltaV ranges. At the 40 - 54 kmph DeltaV range the maintained seat injury cost was higher. It was estimated that, over a 10 - 54 kmph DeltaV range, if all seat backs collapsed the injury cost would be at least 2.83 times as much as if all seat backs maintained their initial position. Therefore, it is clearly not preferable to design all seats to collapse upon rear impact. This injury cost ratio does not necessarily mean that if all seats maintained their initial position there would be a cost reduction from the real total cost. In fact, the data available suggest that at the 40 - 54 kmph DeltaV range a collapsed seat minimizes the occupant injury cost. With optimally performing seats a 6.3% or $130 per seat injury cost reduction might be realized.

The paucity of data collected since 1995 reduces the level of confidence in this injury cost analysis. An evaluation of the statistical significance of the results was not performed because of the limited amount of data. Future additions to the data will allow more definitive conclusions to be reached. However, the trends reflected in the current analysis may hold as new crash data become available. A factor which may have an influence on any future analysis is an upgrade of the head restraint requirement of FMVSS 202. This may reduce whiplash rates, especially for seat backs which don't collapse. Also, the introduction of seats with integrated shoulder belts may affect injury rates. These seats may be stronger and, therefore, less likely to collapse. When they collapse, they may be able to retain the occupant in the seat more effectively, leading to a reduced injury rate.

It is important to reiterate that the analysis presented here only considers injuries to the front outboard seat occupants in rear impacts. In a more detailed analysis it may be important to assess the injuries caused by the interaction of rear and front seat occupants due to the front seat back collapse. This is especially true with the increased emphasis of placing children in the rear seats. However, it may not be possible to obtain this information from the CDS automated data base, so a manual search of the accident files may be required.

[1] Occupant Assessment Form, National Accident Sampling System Crashworthiness Data System, USDOT, NHTSA, HS Form 433A, 1996.

[2] Partyka, S.C; Seat Damage and Occupant Injury in Passenger Car Towaway Crashes, NHTSA, June 8, 1992.

[3] Data Collection, Coding, and Editing Manual, National Accident Sampling System Crashworthiness Data System, USDOT, NHTSA, NCSA, 1996.

[4] Blincoe, L.H. (1994): The Economic Cost of Motor Vehicle Crashes, 1994. US DOT, NHTSA, HS 808 425, Washington, D.C.