Feature Cases
Multiple cause birth injury analysis
This case involved a birth with a multitude of risk factors for hypoxic ischemic encephalopathy (HIE), a type of brain injury that results from prolonged fetal oxygen deprivation. The actionable event that the healthcare provider failed to diagnose was an obvious placental abruption (separation of the placenta from its maternal blood source), which resulted in hours of continued bleeding during the delivery. What the plaintiff alleged should have happened, and did not, was that an emergency c-section should have been performed.
The defense in the case was that there was evidence of an infection in the placenta (called chorioamnionitis), and that the effect of this condition along with the placental abruption could not be quantified, so that no one could tell what the cause of the HIE was. The alternative explanation from the defense was that the HIE was unpreventable because of all of the risk factors, and thus rapid diagnosis would not have changed the outcome.
In order to analyze the contribution of risk factors for the HIE injury, an analysis was undertaken by FR+A on the largest database of births (700K) in which the mother's and baby's records are linked, so that maternal and neonatal risk factors could be considered together for the analysis.
The known HIE risk factors that were present in the birth that was the basis for the medical negligence action were 1) maternal obesity, 2) an abnormal fetal heart rate during delivery, 3) chorioamnionitis, and 4) placental abruption. When the risks associated with each factor were analyzed, taking into account the effect of the other risk factors, they were as follows:
| HIE Risk Factor | risk ratio | 95% confidence limit | |
|---|---|---|---|
| Placental Abruption | 10.2 | 7.4 | 14.2 |
| Chorioamnionitis | 4.8 | 3.5 | 6.5 |
| Abnormal fetal heart rate | 8.6 | 6.7 | 11.0 |
| Maternal obesity | 2.8 | 1.9 | 3.9 |
The fact that the lower bound of the 95% confidence limit did not drop below 1.0 for any of the risk ratios means that they all were "statistically significant."
The table illustrates the fact that the largest HIE risk factor was the placental abruption, which carries more than 2 times the risk of the chorioamnionitis. These values don't tell quantify what would have happened if the abruption had not occurred (or been diagnosed timely, as alleged by the plaintiff), however. In order to do this, the absolute risk of HIE with various combinations of the risk factors had to be calculated. The following was the result:
A newborn who is not exposed to any of the risk factors has a 0.02% chance of HIE (1 in 5,000). A newborn exposed to 2 of the risk factors (abnormal fetal heart rate and maternal obesity) has a 0.36% chance of HIE (1 in 278). The newborn exposed to 3 risk factors (chorioamnionitis, abnormal fetal heart rate and maternal obesity, but no abruption) has an HIE risk of 1.7% (1 in 59).
In comparison, in a newborn exposed to all 4 risk factors, including placental abruption, there is a 14.9% chance of HIE (1 in 6.7).
Based on these findings, of the baby's total risk of 14.9% with all factors present, 13.2% is attributable to the abruption (this number is calculated by subtracting 1.7% [3 risk factors] from 14.9% [all 4 risk factors]. Because 13.2 is 89% of 14.9, we can say that the attributable risk of the placental abruption to the HIE, also known as the probability of causation, is 89%.*
Further analysis indicated that when all 4 risk factors were associated with an emergency C-section, 85.1% of the risk of injury was eliminated.
* The concept of attributable risk as a measure of causation is part of US case law and well described in the Reference Manual on Scientific Evidence, the guide for Federal judges.
Crash reconstruction issue: airbag non-deployment as an indicator of crash-related delta V
In this case, the defendant’s engineering/ crash reconstruction expert performed a download of the event data recorder (EDR) after a minimal damage crash, and used the fact that no recording of the crash was found as evidence that the crash-related delta V must have been below 5 mph. The basis for the claim was the federal standard that indicates that 5 mph delta V is the “wake up” threshold for the EDR to record the crash, and thus, if the crash were over 5 mph it would have been recorded.
The defense expert’s assumption was based on the logically fallacious conclusion that a mandated minimum delta V threshold at which an EDR should begin to record equates to a mandate that the EDR should never record below the threshold, as well as the assumption that the EDR always records above the threshold.
To evaluate the rate at which non-EDR-recorded crashes are distributed for rear impact collisions by delta V, an FR+A analysis was performed in which data from the US National Automotive Sampling System-Crashworthiness Data System (NASS-CDS) database for 2000-2015 were pulled. Crashes in which a vehicle sustained a rear impact only (no secondary crashes), and had a delta V that was reconstructed using a typical momentum, energy, and restitution (MER) approach, AND there had been an EDR download (regardless of whether there was a recorded event) were then evaluated and tabulated. The results of the analysis were as follows:
There were an estimated total of 261,910 rear impacted vehicles with a reconstructed delta V, with 17,838 impacts reconstructed at under 5 mph, 122,833 at 5-10 mph, and the remaining 121,238 vehicle impacts reconstructed at over 10 mph delta V. The crashes were then matched to their EDR data, and dichotomized into the crashes for which there was a recording and the ones which did not have a recorded non-deployment event (Note:all rear impacts are “non-deployment events, as airbags are designed to not deploy in such crashes). The results of this analysis are depicted in the chart below:
The chart above demonstrates that the majority of rear impacts with missing EDR data are, in actuality, greater than 5 mph delta V.
The chart below demonstrates the percent of crashes, by speed category, with missing EDR data (the orange sections from the chart above). It is easily seen that it is far more likely that missing EDR data is related to a crash of more than, rather than less than, 5 mph delta V. Indeed, if the 15.4% of crashes with missing EDR data at less than 5 mph delta V are combined with the 27.2% at 5-10 mph delta V (42.6% total), these data indicate a 66% probability the crash will be at the higher, rather than lower speed category.
The analysis resulted in the conclusion that the single principle upon which the defendant’s reconstruction depended was demonstrably false.
Analysis of the effect of seat belt non-use on the risk of death in a high-speed crash
The decedent, a 22 year-old male, was the unrestrained driver of a 2019 4 door Chevrolet Silverado 3500 traveling at highway speed on a rural 2 lane road. The vehicle was struck on the left front/side by the left front of a 2019 White International truck, that had been traveling the opposite direction and crossed the center line and entered the decedent’s lane of travel. The decedent was killed outright by the crash, and could not be resuscitated after he was extricated from the vehicle. The figures below depict a police diagram and photos of the subject vehicles post-collision:
Police diagram (left, the decedent’s Chevrolet is “Unit 1”), scene photo (right)
The decedent’s Chevrolet post-collision (photos taken at a wrecking yard)
The International truck, post-collision (photos taken at a wrecking yard)
The photographs indicate that the Chevrolet Silverado sustained more than 3 or 4 feet of crush damage extending from the left front to the left rear side. The PDOF (Principle direction of force) of the crash was from the front left, or approximately the 10:30 to 11 o’clock direction. The driver’s side front airbag had deployed, and the occupant space on the driver’s side was severely compromised.
The crush damage to the International truck essentially matched the damage to the Chevrolet, and included >4-5 feet of crush to the left front extending beyond the left front wheel well, with induced damage from the front and into the cabin.
The autopsy demonstrated massive and unsurvivable head and chest trauma.
FR+A reconstruction of the crash
All of the reviewed evidence indicated that the eastbound International truck crossed the oncoming lane at highway speeds and struck the westbound Chevrolet Silverado on the westbound shoulder of the highway. The collision was essentially head-on, with a sustained left side impact to the Chevrolet Silverado.
An inspection of the vehicles indicated that the reason the International truck crossed into the westbound lanes and caused the crash was due to the failure of the left front tire of the International, which was found to be excessively worn. An event data recorder (EDR) download was performed on the decedent’s Chevrolet Silverado, resulting in a retrieved airbag deployment event. The report indicated the Chevrolet Silverado was traveling 58 mph at one half second prior to the crash, with the brakes activated. The download data indicates that the Chevrolet Silverado was subjected to a total delta V of approximately 55.8 mph with a peak total acceleration of approximately 71 g, putting the crash in the upper 1% of frontal crashes in the US for severity.
Seatbelt non-use attributable risk analysis*
As the decedent was killed in the subject collision, the outcome of interest for the analysis was death. Data were accessed from the US national crash injury database, the NASS-CDS, specific to the following parameters: included were all drivers that were exposed to a single frontal collision, with a direction of force ranging from 9 to 11 o’clock (to account for the fact that the vehicle damage was primarily focused on the driver’s side of the vehicle and at the decedent’s position). Only cases in which seatbelt use or non-use was known, and where it was known if there was a driver’s side frontal airbag deployment were included. Crashes that resulted in a rollover were excluded. The outcome of interest was death within 30 days of the crash. The analysis utilized the years 2001-2015 (15 years, inclusive).
A logistic regression model was constructed from the results in the analysis. The model accommodated the effect of covariables on the outcome of interest (death) aside from the 2 main variables of interest, which were delta V and seatbelt use. Other factors that needed to be accounted for, specific to the circumstances of the decedent’s crash, were the fact that he was driving a truck, the airbag in the vehicle deployed, the decedent was not ejected, and the year of the Chevrolet was 2019. The results of the analysis were as follows:
There were a weighted total of 2,272,756 belted and 669,097 unbelted drivers used for the logistic regression model. Among the belted drivers there were 592,311 with an airbag deployment, and 99,579 unbelted with an airbag deployment.
The results of the logistic model are presented in the chart below, with the dashed straight line demonstrating the risk of death for a belted (red solid line) and unbelted (blue solid line) driver at delta Vs ranging from 30 to 80 mph, with 56 mph (the delta V for the subject crash) indicated with a vertical black dashed line.
At 56 mph it can be seen that the solid red curved line representing the death risk for belted drivers intersects with the dashed black line at 0.78 (78%) and the solid blue curved line, representing the death risk for unbelted drivers intersects at 0.66 (66%). These results can be interpreted as follows:
First, belted drivers are actually at higher risk of death versus unbelted occupants in very high-speed frontal crashes like the subject collision. This result is likely due to the fact that at such high speeds a seatbelt is ineffective in preventing the types of serious head and chest injuries that are most likely to result in death, and unbelted drivers are more likely to be ejected into the passenger side of the vehicle and away from the driver’s side where the greatest crush is occurring.
Regardless of seatbelt use and airbag deployment, there is a more than 50% risk of death in a 56 mph delta V frontal crash. This means that the decedent would have died in the subject crash, on a more likely than not basis, regardless of seatbelt use.
Because there are so few very high-speed crashes like the subject collision that occur in the US and which are in the NASS-CDS data, the 95% confidence intervals of the belted and unbelted drivers overlap. This means that the above-described death risk differences between belted and unbelted drivers are not statistically significant (i.e. they may be explained by random scatter in the data). Thus, after reviewing all of the results of the analysis, the final conclusion that must be drawn from the analysis of differential risk by restraint status is that there is no statistically significant difference in the risk of death for a driver who is belted versus unbelted.
Based on the analysis, it was concluded that the decedent’s seatbelt non-use did not contribute to his risk of death in the subject high speed left frontal crash.
* The concept of attributable risk as a measure of causation is part of US case law and well described in the Reference Manual on Scientific Evidence, the guide for Federal judges.