The Importance of Non Struck Side Occupants in Side Collisions

Abstract

In a representative sample of tow-away side collisions from the UK Midlands, one third of front seat occupants were alone, on the struck side of the car. The other two thirds were either a non struck side occupant alone or two occupants sitting together. Occupant restraint, especially in perpendicular side impacts, was a notable factor in determining injury outcome for belted non struck side occupants. With both front seats occupied, there was a reduction in AIS 2+ injury to belted non struck side occupants due to a reduction in chest and lower limb injuries. Struck side occupants sustained increased injury rates to the extremities when accompanied by a belted non struck side occupant but no notable increases in moderate to serious injury to the head, chest, abdomen or pelvis.

THE RISK OF INJURY in side impacts is generally higher than in frontal impacts because there is little vehicle structure with which to control crash forces and for an occupant sitting adjacent to the struck side of the vehicle there is little room for sideways motion before striking the interior.

Regulations governing vehicle crash performance in frontal crashes appeared long before those for side crashes but in 1994, regulations were introduced in the US governing vehicle performance in a side collision and side impact regulation will soon be introduced in Europe.

In addition, the European New Car Assessment Program (Euro NCAP) also includes a side impact crash test. All these tests assess crash loadings to dummies seated alone on the struck side of the vehicle and are aimed at further reducing the injury risk for struck side occupants.

To date, much information has been analysed from real collisions and from laboratory studies concerning struck side occupants in side impacts. But there has been comparatively little work carried out to determine the limits of protection for occupants seated opposite the impacted side - the non struck side occupants.

Crash injury studies by Mackay, Parkin and Hill (1991) and Dalmotas (1983) have both suggested that measures to increase the degree of lateral restraint afforded non struck side occupants need to be explored. By contrast, Otte, Suren and Appel (1984) emphasised how well seat belts protected non struck side occupants “In a 90° impact in the region of the passenger compartment, the belt wearing passenger seated opposite the impact side is saved from any interior impacts by the restraint of the belt”. All of these studies emphasised the importance of head injuries for non struck side occupants.

A sled impact study of 3-point belt effectiveness by Horsch (1980) found that the torso of a Part 572 dummy escaped from the belt diagonal section at an impact angle of 60° from straight ahead but concluded that significant kinetic energy was removed from the upper body at all angles of impact, even at 90° from straight ahead. One consequence of this was shown by Mackay et al (1991) who emphasised the importance of chest injuries caused by seat belt loads. Faerber (1982) addressed the issue of “interaction” or a non struck side occupant moving across to strike a struck side occupant. He found that in side impact crash tests the struck side dummy showed high peak thorax accelerations caused by interaction with a belted non struck side dummy in impacts with a direction of 90° from straight ahead.

This present study set out to further examine the limits of protection for non struck side occupants in real side impacts. In particular, their exposure in the accident population and the crash conditions affecting seat belt effectiveness. It was also considered important to examine the part played by occupant interaction in the generation of injuries to both non struck side and struck side occupants.

METHODS

FIELD DATA

This study used data collected on vehicle crash performance and occupant injury from the East and West Midlands of England between June 1992 and April 1996. It forms part of an ongoing UK crash injury study and represents crashes occurring on both urban and rural roads. Cars were examined at garages and vehicle dismantlers within a few days of the crash occurring and injury details were obtained from medical notes, HM coroners and questionnaires sent to vehicle occupants.

A crash was eligible for investigation if it involved injury in a car towed from the scene of the accident. A stratified sampling procedure was employed with the aim of investigating 100 % of fatal and serious crashes, 15 % of slight crashes and 10% of non injury crashes. Crash injury severity was defined by the British government system of classification. The resulting sample was therefore biased toward the more serious injury cases. That sample was statistically weighted to reflect the population of tow away crashes in the Midlands of England. All distributions and rates illustrated and quoted in this study are based on the weighted sample but unweighted totals (N) are shown for reader information. The overall study methodology is described elsewhere [Mackay, Ashton and Galer, 1985].

In the interests of brevity, Non Struck Side Occupants have been abbreviated to NSS and Struck Side Occupants to SS in the text. Only front seat occupants with known injury outcome have been considered. All were seated in bucket seats so there are no centre front occupants. Occupant injuries were described according to the Abbreviated Injury Scale 90 [Association for the Advancement of Automotive Medicine, 1990]. MAIS refers to the Maximum Abbreviated Injury Score over all body regions.

In this study, examination was made of Opposite Side Intrusion in relation to NSS occupants. This concerns the maximum residual intrusion measured at the opposite side door and is shown in Figure 1.

Figure 1.

Opposite Side Intrusion Zones

In typical UK vehicles, zone 1 encompasses intrusion up to the width of the side window horizontal to the side header. The limit of zone 2 takes side intrusion up to 1/5th or 20% of the horizontal interior width. The limit of zone 3 is crush up to half the interior width and zone 4 defines crush of more than half the interior width.

Side impact angle was also used for some of the analyses. This was derived from the direction of principal force in the collision and is expressed in degrees away from the longitudinal axis of the car. That is illustrated in Figure 2. 11 and 1 o’clock impacts were classed as 30°, 10 and 2 o’clock impacts were classed as 60°, 9 and 3 o’clock impacts were classed as 90°, 8 and 4 o’clock impacts were classed as 120° and 7 and 5 o’clock impacts were classed as 150°.

Figure 2.

Side Impact Angle

RESULTS

POPULATION ISSUES

One way to assess the importance of NSS occupants is to examine their exposure in the accident population. Figure 3 shows the kind of impact that front seat occupants were exposed to.

Fig 3.

Impact Distribution by MAIS - All Front Seat Occupants and all Levels of Belt Use

Figure 3 includes all levels of belt use but it should be remembered that the UK has had around 90% front seat occupant belt use for some time (Restraint Use by Car Occupants, 1995).

At all levels of injury severity, occupants in frontal impacts predominated followed by occupants in side impacts. Occupants in side impacts accounted for 23% of all occupants, 24% of the MAIS 2 population and 29% of the MAIS 3+ population. The side impact population did not only consist of those on the struck side however. 48% of all side impact occupants, 42% of the MAIS 2 population and 31% of the MAIS 3+ population were seated on the non struck side,

BELTED NSS OCCUPANTS

Because of the high levels of belt use in the UK, attention was focused on NSS occupants with confirmed belt use. For information, Table 1 shows the unweighted and weighted sample available for analysis together with the distributions of maximum injury severity. All the belts used were 3-point retractor systems.

Table 1.

Distribution of MAIS - Belted NSS Occupants

MAIS	UNWEIGHTED N	WEIGHTED N
0	66	461
1	150	673
2	46	137
3	13	23
4	7	7
5	9	9
6	4	4
TOTAL	295	1314

Object Struck

The struck object was known for all vehicles which contained a belted NSS occupant. Figure 4 shows the distribution of struck object by MAIS.

Figure 4.

Struck Object by MAIS - Vehicles Occupied by Belted NSS Occupants

Car to car impacts predominated for all levels of injury but as injury severity increased impacts with heavy trucks became more frequent. It should be noted that this distribution depends on the types of vehicle in the general traffic stream in the UK. Corresponding figures may be different for the US with its relatively higher population of light trucks.

Side Impact Angle

Previous laboratory studies have noted decreases in seat belt effectiveness as impact angle increases away from the longitudinal axis of the car. Table 2 shows side impact angle for the population of belted NSS occupants.

Table 2.

Side Impact Angle by MAIS - Belted NSS Occupants

	Angle of Side Impact Force°
	30°	60°	90°	120°	150°
All MAIS	31%	28%	33%	6%	2%	N=295
MAIS 2	17%	34%	42%	7%	--	N=46
MAIS 3+	9%	23%	59%	9%	--	N=33

Frontal oblique angles (30°, 60°) and straight-in impacts (90°) were the most common for all levels of injury severity, while the frequency of rear oblique angles were very small. Of particular note however is how the frequency of frontal oblique angles decreased in deference to straight in impacts as injury severity increased.

Opposite Side Intrusion

One of the important parameters associated with occupant injury severity is passenger compartment intrusion. For NSS occupants, the most relevant intrusion is that which occurs to the opposite side passenger compartment and this was known for most of the NSS occupants in the sample. Figure 5 shows how maximum injury severity rates were distributed within different levels of intrusion.

Figure 5.

Maximum Injury Severity Rates by Opposite Side Intrusion-Belted NSS Occupants

Generally, injury severity was seen to increase as opposite side intrusion increased. In particular the rate of AIS 3+ injuries showed a large increase once intrusion had passed the mid point of the vehicle. The results do show however that, in most cases intrusion was less than 25 cm (72% of the accident sample, rising with weighting to 82% in the accident population).

Crash Severity

Another important factor linked to occupant injury severity is the speed of impact. Determination of impact speed using the CRASH3 program has many limitations, especially in side crashes for which it was not designed. In this study however, the Equivalent Test Speed (ETS) was the best available information and this was known for 45% of belted NSS occupants.

Figure 6 shows how maximum injury severity rates were distributed within different levels of ETS in the population of belted NSS occupants. Although the sample was compromised by small numbers at the higher speeds, moderate to serious injury rates were seen to increase with ETS. Despite some of the reservations concerning the use of CRASH3 in side impacts, this trend indicates that ETS can at least be used as a general indicator of crash severity. Of particular interest were the rates of MAIS 3+ injury. Up to 50 km/h ETS, occupants with serious to life threatening injury did not constitute substantial proportions of the MAIS distributions within ETS bands.

Figure 6.

Maximum Injury Severity Rates by ETS- Belted NSS Occupants

Figure 7 shows the cumulative frequency distributions of ETS by MAIS for the population of belted NSS occupants. It shows that extremely high crash severity is not a pre-requisite for moderate or serious injury. Half of the occupants with MAIS 2 experienced an ETS of < 25 km/h and the median speed for occupants with MAIS 3+ was 33 km/h. 75% of all occupants experienced ETS below 24 km/h, 75% of all occupants with MAIS 2 experienced ETS below 35 km/h and the 75 %’ile value for occupants with MAIS 3+ was 48 km/h.

Figure 7.

ETS by MAIS - Belted NSS Occupants

OCCUPANCY IN SIDE IMPACTS

This study has already indicated that injury outcome for belted NSS occupants is associated with impact speed and the degree of opposite side intrusion. But other issues which have bearing on injury outcome are those concerned with seat belt effectiveness and interaction with occupants sitting on the struck side.

It has already been indicated that, in side crashes, occupants sitting on the struck side are not the only ones of concern. In addition to this, occupants sitting on their own are not the only ones of concern. Figure 8 shows the percentage of all occupants in side impacts who were alone or accompanied by another occupant in the adjacent front seat. Belt use was not considered here and there were a few cases where there was uncertainty concerning front seat occupancy. Those have been excluded.

Fig 8.

Occupancy in Side Impacts

Figure 8 shows that the situation where a SS occupant was alone, only accounted for a third of front seat occupants in side impacts, while a NSS occupant alone accounted for 29%. 38% of front seat occupants in side crashes were accompanied by another occupant.

Focussing only on NSS occupants showed that 40% were accompanied by a SS occupant. Focussing only on SS occupants showed that 37% were accompanied by a NSS occupant.

SEAT BELT EFFECTIVENESS AND OCCUPANCY

For belted NSS occupants both seat belt use and occupancy have a bearing on injury outcome. Figure 9 shows the MAIS 2+ injury risk for unbelted NSS occupants alone, belted NSS occupants alone and belted NSS occupants with a SS occupant present.

Fig 9.

MAIS 2+ Injury Rates by Occupancy and Belt Use - Belted NSS Occupants

There was an obvious decrease in injury rate for occupants wearing seat belts and it must be concluded that the seat belt had a beneficial effect for NSS occupants. The presence of a SS occupant also appeared to benefit the Belted NSS occupant but to a lesser degree than seat belt use.

SIDE IMPACT ANGLE AND INJURY OUTCOME

Because the presence of a SS occupant has a bearing on injury outcome for the belted NSS occupant the clearest way to quantify the effect of side impact angle is to examine the situation where belted NSS occupants are alone.

Figure 10 shows the effect of impact angle and opposite side intrusion on the MAIS 2+ injury rate for belted NSS occupants sitting alone.

Fig 10.

Effect of Side Impact Angle and Intrusion on MAIS 2+ Injury Rates - Belted NSS Occupants Sitting Alone

At the lowest level of intrusion the MAIS 2+ injury rate was much higher with oblique impact angles than with perpendicular impacts. With 10–24 cm intrusion the rate of injury in perpendicular impacts far outweighed that in oblique impacts although the rate of injury for those oblique impacts was similar to that for oblique impacts with 0–9cm of intrusion. Where 25–49 cm of intrusion occurred, the injury rate was still higher for perpendicular impact angles. The greatest risk of AIS 2+ injury was seen once intrusion passed the centre line of the car (50+ cm). There, the impact angle was of little consequence.

NSS OCCUPANTS AND INTERACTION

Interaction effects are extremely difficult to isolate from crash injury data because often there is little physical evidence available with which to correlate injuries with contacts. Examination of injury patterns between belted NSS occupants alone and those accompanied by a SS occupant provides some insight into the effects of interaction on injury outcome. Figure 11 provides the comparison. AIS 2+ injury rates have been shown per 10,000 occupants to enable a clearer comparison.

Fig 11.

AIS 2+ Injury Rates by Body Region - Belted NSS Occupants Alone and with a SS Occupant Present

For occupants alone, the head was clearly the most often injured body region to the AIS 2+ level followed by the chest and then the lower limb. With a struck side occupant present the head was still the most often injured but then followed by the upper limb and then the chest. With a SS occupant present, the AIS 2+ chest injury rate halved and lower limb injuries almost disappeared. The head injury rate also dropped slightly so there is no evidence that moderate to serious head injuries increased due to the possibility of a head to head impact.

Of interest in figure 11 were the AIS 2+ injuries occurring to the upper and lower limbs. Table 3 illustrates which parts of the limbs sustained AIS 2+ injury and how this varied with and without a struck side occupant present. For NSS occupants, the inboard limb is the one nearest to the struck side of the vehicle.

Table 3.

Rate of AIS 2+ Injury to the Lower and Upper Extremities (Rate/10,000 occupants) - Belted NSS Occupants Alone and with a SS Occupant Present

	Alone N=118		With SS Occupant Present N=160
	Inboard	Outboard	Inboard	Outboard
Thigh	207	30	--	--
Knee	15	15	--	--
Leg	89	15	--	--
Foot/ankle complex	74	30	--	21
Shoulder	30	15	104	--
Arm/hand	103	59	69	189

For occupants alone, the inboard thigh was the part of the lower extremity most often injured at the AIS 2+ level. All of those thigh injuries were femur fractures caused by contact with the intruding opposite side. In all cases except for the knee, there was a large bias toward inboard lower limb injuries. By contrast, very few of the occupants with a SS occupant present sustained any AIS 2+ lower limb injury.

Within the upper extremity, the inboard arm/hand was the most frequently injured area at AIS 2+ for occupants alone. For those occupants inboard shoulder injury occurred more often than outboard but the rates were low. The presence of a SS occupant increased the inboard shoulder injury rate threefold. Most were clavicle fractures.

SS OCCUPANTS AND INTERACTION

The frequency with which there were two occupants together in side impacts prompted an examination of the effect of a belted NSS occupant on injury outcome for the SS occupant. Figure 12 provides the comparison. Again, AIS 2+ injury rates have been shown per 10,000 occupants to enable a clearer comparison. Belt use was not considered for the SS occupant.

Fig 12.

AIS 2+ Injury Rates by Body Region - SS Occupants Alone and with a Belted NSS Occupant Present

For occupants alone, the head was the body region most often injured to AIS 2+, followed by the chest. With a belted NSS occupant present this was still the case. With a belted NSS occupant present the AIS 2+ head injury rate was lower than for occupants alone but chest injury rates were similar. The presence of belted NSS occupants did not therefore result in any increase of moderate to serious injuries to the heads or chests of SS occupants. Neither were there any notable increases in AIS 2+ injury rates to the abdomen or pelvis There was however a substantial increase in AIS 2+ injury rates to the upper and lower limbs when a belted NSS occupant was present.

Table 4 illustrates which parts of the limbs sustained AIS 2+ injury and how this varied with and without a belted NSS occupant present. For SS occupants, the inboard limb is the one furthest away from the struck side of the vehicle.

Table 4.

Rate of AIS 2+ Injury to the Lower and Upper Extremities (Rate/10,000 occupants) - SS Occupants Alone and with a Belted NSS Occupant Present

	Alone N=243		With Belted NSS Occupant Present N=160
	Inboard	Outboard	Inboard	Outboard
Thigh	25	109	16	143
Knee	42	25	--	16
Leg	17	67	79	127
Foot/ankle complex	--	25	48	63
Shoulder	25	125	63	269
Arm/hand	33	159	79	190

For occupants alone, the outboard thigh was the most often injured part of the lower extremity at the AIS 2+ level. All of those thigh injuries were femur fractures caused by contact with the intruding adjacent side. In all cases except for the knee, there was a large bias toward outboard lower limb injuries. Injuries to the knee were biased toward the inboard limb. With a belted NSS occupant present there was a large increase in both inboard and outboard leg and foot/ankle complex injury rates. There was a smaller increase in the injury rate to the outboard thigh.

Within the upper extremity, the outboard shoulder and outboard arm/hand were the most frequently injured areas at AIS 2+ for occupants alone. With a belted NSS occupant present there were substantial increases in injury rates to both the inboard and outboard shoulders and to the inboard arm/hand. There was also an increase in rate for the outboard arm/hand.

DISCUSSION

This study has presented information on the exposure of non struck side occupants in the crash population. The crash conditions where belted non struck side occupants were injured have also been explored. Some aspects of non struck side occupant restraint were examined and the issue of interaction between front seat occupants in side impacts was considered.

In side impacts it was found that about one half of all occupants were seated on the non struck side. When occupants were seriously injured in side impacts about a third were on the non struck side. With the current introduction of new side impact tests and the emphasis on new technology for occupant protection in side crashes, these results suggest there may be additional opportunities for injury reduction in side impacts by also considering non struck side occupants.

Opposite side intrusion and crash severity were found to correlate with injury severity for belted non struck side occupants. Generally, injury severity increased with increasing intrusion and crash severity. However, median crash severities were not exceptionally high (33 km/h for the most seriously injured occupants).

Similarly, for the majority of belted non struck side occupants, opposite side intrusion was less than expected (in 82% of cases this did not pass 20% of the vehicle width). That suggested there were additional factors which affected injury outcome.

Previous research by Horsch (1980) showed how belt effectiveness could be compromised in a non struck side situation as impact angle reached 60° but that work also showed that the seat belt shoulder section was able to reduce the kinetic energy of a belted dummy to varying degrees at all impact angles. This study also indicates that the seat belt has an overall benefit for non struck side occupants because the MAIS 2+ injury rate for unbelted occupants was twice that for those who were belted.

Horsch reported that even though the test dummy torso escaped from the shoulder belt at a 60° impact angle it was found that most of the upper body kinetic energy was dissipated, resulting in little motion of the upper body after escape. His measurements of maximum shoulder belt tension at varying impact angles showed the 90° or perpendicular impact angle to be the one which most severely compromised shoulder belt restraint.

By comparing injury rates for belted non struck side occupants sitting alone in oblique and perpendicular side impacts, Horsch’s results have been vindicated to some degree in this study. With opposite side intrusion of 10–24 cm and 25–49 cm, there was a markedly higher MAIS 2+ injury rate in perpendicular as opposed to oblique impacts. It is highly likely that occupant slippage out of the belt diagonal section (with subsequent impact on opposite side interior structures) contributed to this result. This may also have occurred in some of the more oblique impacts but more of the occupant’s kinetic energy will have been removed first, resulting in lower injury rates in oblique crashes overall.

With low levels of opposite side intrusion (0–9cm) the MAIS 2+ injury rate was found to be much higher for occupants in oblique compared to perpendicular impacts. In those low intrusion conditions, belted occupants did not strike the opposite side interior. The explanation for the increased injury rate in oblique impacts may lie in the fact that the higher restraining force of the seat belt resulted in more AIS 2+ injuries being sustained by the chest.

Once intrusion had passed the half way line of the car, AIS 2+ injury risk increased dramatically and restraint effectiveness appeared of little consequence.

These results suggest that belted non struck side occupants would benefit from improved restraint in perpendicular crashes with between 10 and 49 cm of opposite side intrusion. Overall, non struck side occupants with moderate to serious injury were frequently exposed to perpendicular impacts (42% of the MAIS 2 population and 59% of the MAIS 3+ population).

Current occupant protection measures for side impact are focused on struck side occupants sitting alone. The side impact population studied show that situation occurs for only one third of occupants in side crashes and that 38% of front seat occupants in side impacts are accompanied by an adjacent occupant. As seat belt performance may be compromised in certain circumstances, there is the possibility that occupants could impact each other. This was an important point to consider, especially as economical and environmental pressures are encouraging multiple occupancy of vehicles.

For belted non struck side occupants, the presence of a struck side occupant reduced the overall rate of moderate to serious injury. That concurs with the findings of Mackay et al (1991). This study was able to look further at this finding by examining where injury reduction was occurring. Reductions in AIS 2+ chest and lower limb injuries were the major factors, likely due to the struck side occupant preventing contact on the opposite side structures. If this is the case, it would certainly suggest that some AIS 2+ chest injuries are caused by contacting intruding structures and are not solely due to seat belt loads.

There was however, an increase in upper extremity injuries, largely due to increases in inboard clavicle fractures from contact on the struck side occupant. Outboard arm and hand injuries increased to a lesser degree but it is difficult to see why this should occur with a struck side occupant present. Predicting how the arms can be injured is fraught with difficulty because of the wide variety of positions they can flail into at impact.

For both occupants alone and accompanied, the head was the most frequently injured body region at moderate to serious levels. That has been pointed out by previous research but notably there was no increase in head injuries with a struck side occupant present, suggesting that head to head contact was not responsible for AIS 2+ injury.

For occupants alone, the lower limb was the third most frequently injured body region at the AIS 2+ level and for accompanied occupants the upper limb was the second most frequently injured body region at the AIS 2+ level. This finding contrasts with those of Mackay et al (1991) who found that injuries to the upper and lower limbs were infrequent and almost always minor.

When struck side occupants were accompanied by a belted non struck side occupant, no notable increases in AIS 2+ injuries to the struck side occupant’s head, chest, abdomen or pelvis occurred. Crash test work by Faerber (1982) showed that the presence of a belted non struck side dummy substantially increased peak accelerations of the struck side dummy’s chest. The crash injury data suggests that there is no interaction which increases the rate of moderate to serious chest injuries for struck side occupants.

The presence of a belted non struck side occupant was associated with increases in AIS 2+ injury to the upper and lower limbs of struck side occupants. Injury rates increased substantially for the outboard thigh, the inboard and outboard leg and foot ankle complex. Injury rates also increased for both the outboard and inboard shoulder and arm/hand. The increases in inboard and outboard shoulder injury correlate well with increases in inboard shoulder injury to accompanied belted non struck side occupants and are likely due to shoulder to shoulder contact. Increases in injury rates to the other parts of the struck side occupants limbs may be associated with interaction but the nature of the interaction is not as easily defined as that for shoulder injuries. What is likely however, is that improved restraint for non struck side occupants could have beneficial effects for struck side occupants as well.