Opinion
LEVY, Acting P. J.In May 2003, Gerardo Ceja and Simon Olivarez were killed in a cross-median accident on State Route 99 (SR 99). Appellants, the decedents’ surviving family members, sued respondent, the Department of Transportation (Department), for wrongful death damages. Appellants alleged that the lack of a median barrier created a dangerous condition.
In 1992, the Department conducted an investigation of the stretch of SR 99 that included the accident location. A Department engineer recommended that a median barrier be installed. However, because this section of highway was to be reconfigured, it was decided that a median barrier would not be installed. This reconfiguration was completed in 1994. Four lanes of expressway and freeway with an 84-foot median became six lanes of freeway with a 60-foot median.
Before trial, the Department moved in limine to exclude evidence of four accidents that occurred before 1994 on the ground that the physical condition at the accident location had substantially changed. The trial court granted the motion.
The jury found that the accident location was not in a dangerous condition and judgment was entered in the Department’s favor. Appellants contend this *1478judgment must be reversed because the trial court excluded relevant evidence of pre-1994 cross-median accidents when it granted the Department’s in limine motion.
As discussed below, the trial court did not abuse its discretion when it excluded the pre-1994 accident evidence. The physical conditions existing before 1994 were substantially different from those existing in 2003. Accordingly, the judgment will be affirmed.
BACKGROUND
On May 25, 2003, Ceja was driving his car northbound on SR 99. Olivarez was a passenger. Ceja’s car crossed the median and struck two southbound vehicles. Ceja and Olivarez died at the scene. This accident occurred 661 feet north of Keyes Road. The roadway consisted of six lanes of freeway with opposing traffic separated by a 60-foot median.
1. Median barrier warrants.
A median barrier is not necessarily a beneficial addition to a freeway. While a median barrier prevents nearly all cross-median accidents, it also halves the recovery room for an out-of-control vehicle and is a fixed object that can cause serious injury either by direct impact or by deflecting vehicles back into traffic. (Alvarez v. State of California (1999) 79 Cal.App.4th 720, 724 [95 Cal.Rptr.2d 719] (Alvarez), abrogated on other grounds in Cornette v. Department of Transportation (2001) 26 Cal.4th 63, 74, fn. 3 [109 Cal.Rptr.2d 1, 26 P.3d 332] (Cornette).) Thus, a median barrier usually results in an increase in accidents and injuries. However, cross-median accidents tend to be more severe than other types of collisions. Accordingly, the Department’s challenge is to balance the reduction of accident severity against a greater frequency of accidents and injuries.
In an effort to strike this balance, the Department has developed criteria, referred to as warrants, to identify those locations that should be evaluated to decide whether a median barrier is appropriate. A warrant is not a determination that a median barrier is required at the location, only that conditions there merit further study.
There are two types of warrants, traffic volume/median width warrants and accident warrants. Traffic volume/median width warrants index traffic volume to median width. This warrant reflects the fact that as traffic volumes rise, the chance that an errant vehicle will cross the median increases. At the same time, the possibility that such an errant vehicle will reach the opposing lanes and collide with another vehicle decreases as the median becomes wider. *1479(Alvarez, supra, 79 Cal.App.4th at pp. 724-725.) Between 1964 and 1997 a median width greater than 45 feet did not trigger an investigation into a median barrier regardless of traffic volume. In 1997 the Department changed that policy to trigger an investigation for median widths up to 75 feet.
The second independent basis for considering the placement of a median barrier is the accident warrant. This warrant reflects the actual operating history of the location. It is based on the premise that, for whatever reason, some locations experience a higher number of cross-median accidents than would be anticipated for a highway with the same median width and traffic volume. Since 1978, the threshold rate has been met by either 0.50 cross-median accidents of all types per mile per year or 0.12 fatal cross-median accidents per mile per year. A minimum of three accidents in a five-year period is required.
2. The 1992 accident warrant.
In June 1992, the accident warrant was met for a 0.39-mile stretch of SR 99 that included the subject accident location. There had been three nonfatal cross-median accidents from 1987 to 1991, creating an accident rate of 1.54 cross-median accidents per mile per year. Accordingly, an investigation was conducted.
A state civil engineer initially recommended that a median barrier be constructed. However, upon review and discussions with that state engineer, the lead engineer for the median barrier monitoring program concluded that a median barrier should not be installed. The state engineers based this decision on the following factors: (1) the number of cross-median accidents just met the minimum of three accidents; (2) only one of the cross-median accidents was an injury accident; (3) the road was going to be reconfigured to add two additional lanes that would result in a reduction of traffic congestion; and (4) the resulting median width would be 60 feet, which was more than the 45-foot width specified in the median barrier guidelines at the time.
The reconfiguration of that stretch of SR 99 was completed in July 1994. Two lanes were added for a total of six, the transition on both ends from six lanes to four was eliminated, the entire stretch became freeway, and the median width was decreased from 84 feet to 60 feet in the area where the Ceja accident would later occur.
3. The 1998 decision to install a median barrier.
There were no cross-median accidents along the reconfigured section of SR 99 until April 17, 1999. Thus, over four years and nine months passed *1480without a cross-median accident. Nevertheless, due to the change in the traffic volume/median width criteria to include median widths up to 75 feet, this location met the traffic volume/median width warrant in 1998. This area was investigated and a median barrier was recommended. In 1999 the median barrier project was approved with a completion date of October 2004. The median barrier was completed in July 2004.
However, after the Department decided to install a median barrier but before the project was completed, a new accident history developed for the reconfigured highway. Between April 17, 1999, and February 22, 2003, there were 11 cross-median accidents, six of them fatal, along a 2.75-mile stretch of roadway. As noted above, the Ceja accident occurred on May 25, 2003.
4. The Department’s motion in limine.
The Department moved in limine for “an order instructing all parties and counsel not to refer to, question any witness concerning, comment on, or suggest to the jury in any way, anything about the number of cross-median accidents at the subject location prior to July 1994 when the conditions of the freeway were substantially different from those conditions that existed after July 1994.” The trial court granted this motion.
5. The trial and special verdict.
At trial, appellants argued that the accident location constituted a dangerous condition of public property. Appellants presented evidence of the 11 cross-median accidents that occurred between 1994 and 2003 and expert testimony that SR 99 was unsafe.
The jury found that SR 99 at the accident location was not in a dangerous condition at the time of the collision. Accordingly, judgment was entered in favor of the Department. The jury did not reach the issues of causation or the Department’s reasonableness.
DISCUSSION
Appellants contend the trial court erred in granting the motion in limine and in excluding both the pre-1994 accidents and the state engineers’ 1992 recommendation that a median barrier be installed. According to appellants, this evidence was relevant to prove a dangerous condition and its exclusion was prejudicial.
1. Standard of review.
A. motion in limine is made to exclude evidence before it is offered at trial on the ground that the evidence is either irrelevant or subject to *1481discretionary exclusion as unduly prejudicial. (Ulloa v. McMillin Real Estate & Mortgage, Inc. (2007) 149 Cal.App.4th 333, 337-338 [57 Cal.Rptr.3d 1]; Condon-Johnson & Associates, Inc. v. Sacramento Municipal Utility Dist. (2007) 149 Cal.App.4th 1384, 1392 [57 Cal.Rptr.3d 849].)
Generally, a trial court’s ruling on the admissibility of evidence is reviewed for an abuse of discretion. (People v. Waidla (2000) 22 Cal.4th 690, 717 [94 Cal.Rptr.2d 396, 996 P.2d 46].) Accordingly, an in limine ruling to keep particular items of evidence from the jury is subject to reversal only where the trial court exceeded the bounds of reason. (Shamblin v. Brattain (1988) 44 Cal.3d 474, 478 [243 Cal.Rptr. 902, 749 P.2d 339].) In other words, the appellate court will not disturb the trial court’s decision unless the trial court exceeded the limits of legal discretion by making an arbitrary, capricious or patently absurd determination. (In re Raymundo B. (1988) 203 Cal.App.3d 1447, 1456 [250 Cal.Rptr. 812].) Moreover, when two or more inferences can reasonably be deduced from the facts, the appellate court cannot substitute its decision for that of the trial court. (Shamblin v. Brattain, supra, 44 Cal.3d at pp. 478-479.)
2. The trial court did not abuse its discretion in excluding the pre-1994 accident history.
Under Government Code section 835, a public entity may be liable for injury proximately caused by a dangerous condition of its property if the entity either creates a dangerous condition on its property or fails to remedy a dangerous condition when it has actual or constructive notice of the condition and had sufficient time to take preventive measures before the injury. (Cornette, supra, 26 Cal.4th at p. 66; Brenner v. City of El Cajon (2003) 113 Cal.App.4th 434, 438-439 [6 Cal.Rptr.3d 316].) “The state’s failure to erect median barriers to prevent cross-median accidents may result in such liability.” (Cornette, supra, 26 Cal.4th at p. 68.) However, to qualify as a “dangerous condition,” the condition must create a substantial, as distinguished from a minor, trivial or insignificant, risk of injury when such property is used with due care in a manner in which it is reasonably foreseeable that it will be used. (Gov. Code, § 830, subd. (a).)
As discussed above, the Department uses the warrant system to identify those locations that should be evaluated to decide whether a median barrier is appropriate. In 1992 the accident warrant was met for the subject location. However, the Department decided not to install a median barrier because the highway was to be reconfigured in 1994. Appellants acknowledge that the highway changed but nevertheless argue that the accident history for the road when it was four lanes is relevant to prove that the now six-lane road was in a dangerous condition at the time of the accident.
*1482The trial court has discretion to admit evidence of prior accidents where the conditions existing at the time of the respective accidents are shown to be similar. (Johnston v. County of Yolo (1969) 274 Cal.App.2d 46, 59 [79 Cal.Rptr. 33]; Martindale v. Atchison, T. & S. F. Ry. Co. (1948) 89 Cal.App.2d 400, 411 [201 P.2d 48].) “Where a dangerous condition of property is involved, there must be proof that there was no substantial change during the interval between the accidents under consideration.” (Alwood v. City of Los Angeles (1956) 139 Cal.App.2d 49, 58 [293 P.2d 69].)
Here, there was a substantial change in the physical conditions existing at the time of the pre-1994 accidents and the Ceja accident. The highway went from four lanes of freeway and expressway to six lanes of freeway and the transitions from six lanes to four lanes at each end of the section were eliminated. Thus, traffic congestion was relieved at this location. Under these circumstances, the trial court did not exceed the bounds of reason when it excluded the pre-1994 accident history. This evidence was not relevant. The accidents that occurred on the four-lane highway would not have a tendency to prove that the six-lane highway was in a dangerous condition. Rather, one would expect a six-lane highway to be safer than a four-lane highway.
Appellants argue that, because the number of lanes is not a factor in the median barrier warrant program, the addition of the two interior lanes in 1994 was immaterial to the accident history. It was a physical difference, but made no difference as to how the accident warrant is applied. Thus, appellants contend, the six-lane highway in 1994 was not substantially dissimilar from the previous four-lane highway and the pre-1994 accident history should have been admitted.
Contrary to appellants’ position, the accident warrant parameters do not dictate what constitutes a substantial change in the location’s physical condition. A warrant merely indicates that the location should be further evaluated, not that the location is in a dangerous condition.
Appellants further point out that in 1994 the median width narrowed from 84 feet to 60 feet. Therefore, appellants assert, the reconfigured highway was more dangerous than it had been when it was four lanes. Nevertheless, at that time, the median width warrant was not met unless the width was 45 feet or less.
Before ruling on the Department’s in limine motion, the trial court mentioned that “[a]t some point it becomes cumulative as to the number of accidents.” Based on this statement, appellants contend that the trial court granted the motion on the ground that the evidence was cumulative and that this reason did not justify the exclusion.
*1483However, the Department did not make this argument. Rather, the Department’s motion was based on the pre-1994 highway being substantially different from the reconfigured highway and the concomitant irrelevance of the pre-1994 accidents. Moreover, if the exclusion of evidence is proper on any theory, the exclusion must be sustained. (Philip Chang & Sons Associates v. La Casa Novato (1986) 177 Cal.App.3d 159, 173 [222 Cal.Rptr. 800].) A lower court order is presumed correct. (Denham v. Superior Court (1970) 2 Cal.3d 557, 564 [86 Cal.Rptr. 65, 468 P.2d 193].) Accordingly, a correct ruling will not be reversed simply because it may have been based on an incorrect reason. (Nmsbpcsldhb v. County of Fresno (2007) 152 Cal.App.4th 954, 966 [61 Cal.Rptr.3d 425].)
In sum, the trial court’s exclusion of the accidents that occurred before the highway was substantially changed by the addition of two lanes was not arbitrary, capricious or patently absurd. Therefore, the trial court did not abuse its discretion in granting the Department’s motion in limine.
DISPOSITION
The judgment is affirmed. Costs on appeal are awarded to respondent.
Cornell, J., and Franson, J., concurred.
A petition for a rehearing was denied January 13, 2012, and appellants’ petition for review by the Supreme Court was denied March 14, 2012, S199773.