The opinion of the Court was delivered by
COLEMAN, J.This appeal requires us to revisit the standard for determining the appropriateness of consecutive sentences imposed during a single sentencing proceeding following a conviction on two counts of vehicular'homicide that involved multiple victims.
On the night of December 26, 1995, defendant, Joseph Carey, got behind the wheel of his motor vehicle after a long evening of drinking beer. That ill-fated decision to drive while intoxicated *419resulted in a head-on collision that caused two deaths. Defendant was convicted of, among other things, two counts of vehicular homicide. The trial court sentenced him to consecutive seven-year terms of imprisonment on each count with a three-year parole disqualifier on each count, as mandated by N.J.S.A. 2C:11-5b(1). The Appellate Division vacated those sentences and remanded to the trial court for resentencing to concurrent terms, noting that the only factor militating in favor of consecutive sentences was the presence of multiple victims. We granted certification in this case and in State v. Molina, 168 N.J. 436, 775 A.2d 509, which we also decide today, to determine whether the New Jersey Code of Criminal Justice, N.J.S.A. 2C:1-1 to 98-4 (Code), and the sentencing guidelines established in State v. Yarbough, 100 N.J. 627, 498 A.2d 1239 (1985), cert. denied, 475 U.S. 1014, 106 S.Ct. 1193, 89 L.Ed.2d 308 (1986), permit consecutive sentences in cases where, as here, a person under the influence of alcohol causes an accident that results in more than one death. Although we do not adopt a per se rule, we conclude that it was not an abuse of the trial court’s discretion to impose consecutive sentences. Accordingly, we reverse the Appellate Division’s judgment and reinstate the trial court’s sentences.
I.
On December 26, 1995, defendant, who was twenty-two years old at the time, spent the evening playing darts and drinking beer with friends and family at his father’s house in Franklin Township, New Jersey. While there, he made the acquaintance of Joyce and Melissa Snook, two sisters who lived in an apartment above defendant’s father. Shortly after midnight, Melissa and Joyce accepted defendant’s invitation to accompany him back to his house to meet his roommate. The three of them left together in defendant’s pick-up truck and headed north on North Church Road.
Joyce Snook testified that defendant drove very fast and she became scared. She asked defendant to slow down, but he did not *420respond. There was also testimony that defendant nearly struck a volunteer firefighter who was on his way home from a firehouse. Minutes later, while speeding around a bend in the road, defendant’s pick-up truck veered into the oncoming lane of traffic and collided head-on with a Ford Mustang. William Ferguson, a twenty-one-year-old passenger in the Mustang, did not survive the impact. Melissa Snook was airlifted from the scene by helicopter, but she died at a hospital later that morning. Michael DiGangi, who had been driving the Mustang, and Joyce Snook both sustained severe injuries, but survived the accident.
The first police officer at the accident scene testified that he smelled alcohol on defendant’s breath. Based on a blood sample taken later that morning, the State’s expert opined that defendant had a .169 blood alcohol level at the time of the accident. Another expert retained by the State estimated that defendant was traveling between sixty-five and eighty-one miles per hour at the time of impact. A traffic sign warned that the maximum safe speed at the curve in the road was twenty-five miles per hour.
A jury convicted defendant of two counts of vehicular homicide, N.J.S.A. 2C:11-5, and two counts of assault by auto, N.J.S.A. 2C:12-1c. Twelve days before defendant’s accident, the Legislature amended the vehicular homicide statute. Among other things, the Legislature elevated the crime from third degree to second degree, punishable by a term of imprisonment of five to ten years. L. 1995, c. 285, § 1. The amendments took effect December 15, 1995, thus applying “to offenses committed on or after the effective date.” Id. § 2.
At the sentencing hearing, the trial court found three aggravating factors: the serious nature of the harm inflicted on the victims, the risk that defendant would re-offend, and the need to deter defendant and others from violating the law. N.J.S.A. 2C:44-1a(2), (3), (9). The court did not find any mitigating factors. After determining that the aggravating factors outweighed any potential mitigating factors, the court sentenced defendant to the presumptive term of seven years imprisonment with a three-year parole disqualifier on each of the two counts of vehicular homicide.
*421In deciding whether those sentences should run concurrently or consecutively, the trial court posited that the “one factor that has become more prevalent as a consideration in successive sentence and consecutive sentencing rather than concurrent is multiple victims and the nature of the charges.” Thus, the trial court determined that defendant’s sentences should run consecutively because defendant’s crime had resulted in multiple victims. The court also sentenced defendant to a one-year term for each count of fourth-degree assault by auto, to run concurrent with each other and with the vehicular-homicide sentences. Hence, defendant received an aggregate sentence of fourteen years imprisonment with six years of parole ineligibility.
In an unpublished opinion, the Appellate Division affirmed defendant’s convictions, but vacated the sentences on the vehicular homicides. The Appellate Division concluded that the trial court abused its discretion in imposing consecutive sentences because, “at least in the context of the death-by-auto crime, the multiple-victim factor ... has not itself, as the only factor, ordinarily resulted in consecutive sentencing.” The panel also determined that the trial court had erred in identifying and weighing the aggravating and mitigating factors. Specifically, the panel found that the trial court had improperly double-counted the two deaths, given that death was an element of each of the vehicular-homicide offenses. The panel also found no support in the record for the trial court’s finding that defendant was likely to reoffend or that there was a need for deterrence. Therefore, the panel vacated the consecutive sentences on the vehicular homicide convictions and remanded to the trial court for the purpose of imposing concurrent sentences on all offenses. We granted certification, 165 N.J. 674, 762 A.2d 655 (2000), and now reverse.
II.
A.
Prior to the enactment of the Code, the focus of our criminal laws was on rehabilitating and reforming offenders. Yar*422bough, supra, 100 N.J. at 637, 498 A.2d 1239. Consistent with that goal, there was a presumption in favor of concurrent sentences in cases involving multiple convictions. Ibid. By contrast, the Code, which is ‘“based on notions of proportionality and desert/ ” focuses on the gravity of the offense. State v. Roth, 95 N.J. 334, 355, 471 A.2d 370 (1984) (quoting Andrew von Hirsch, Utilitarian Sentencing Resuscitated: The American Bar Association’s Second Report on Criminal Sentencing, 33 Rutgers L.Rev. 772, 773 (1981)). Consequently, under the Code’s “just deserts” sentencing model “[cjoncurrent sentences frustrate this objective, and consecutive sentences thus should be the rule.” Yarbough, supra, 100 N.J. at 637, 498 A.2d 1239 (quoting Perlman and Stebbins, “Implementing an Equitable Sentencing System: The Uniform Law Commissioners’ Model Sentencing and Corrections Act,” 65 Va. L.Rev. 1175, 1220 (1979)).
Generally, the Code does not specify when multiple sentences should run concurrently and when they should run consecutively. The Code simply states that “multiple sentences shall run concurrently or consecutively as the court determines at the time of sentence.” N.J.S.A. 2C:44-5a. We recognized early on that investing unbridled discretion in sentencing judges would inevitably lead to a lack of sentencing uniformity, so in Yarbough, supra, 100 N.J. 627, 498 A.2d 1239, we set forth six guidelines to assist trial courts in deciding whether to impose concurrent or consecutive sentences:
(1) there can be no free crimes in a system for which the punishment shall fit the crime;
(2) the reasons for imposing either a consecutive or concurrent sentence should be separately stated in the sentencing decision;
(3) some reasons to be considered by the sentencing court should include facts relating to the crimes, including whether or not:
(a) the crimes and their objectives were predominantly independent of each other;
(b) the crimes involved separate acts of violence or threats of violence;
(c) the crimes were committed at different times or separate places, rather than being committed so closely in time and place as to indicate a single period of aberrant behavior;
*423(d) any of the crimes involved multiple victims;
(e) the convictions for which the sentences are to be imposed are numerous;
(4) there should be no double counting of aggravating factors;
(5) successive terms for the same offense should not ordinarily be equal to the punishment for the first offense; and
(6) there should be an overall outer limit on the cumulation of consecutive sentences for multiple offenses not to exceed the sum of the longest terms (including an extended term, if eligible) that could be imposed for the two most serious offenses.1
[Id. at 643-44, 498 A.2d 1239 (footnote omitted).]
Admittedly, the second, fourth, fifth, and sixth guidelines do not assist a sentencing court in making the threshold decision whether to impose concurrent or consecutive sentences; rather, they establish certain procedural requirements. The first guideline — “no free crimes” — tilts in the direction of consecutive sentences because the Code focuses on the crime, not the criminal. Id. at 630, 498 A.2d 1239. Obviously, the “no free crimes” guideline is not easily implemented. State v. Rogers, 124 N.J. 113, 119, 590 A.2d 234 (1991).
The Yarbough guideline that provides the clearest guidance to sentencing courts faced with a choice between concurrent and consecutive sentences is the third guideline, which focuses on the “facts relating to the crimes.” See, e.g., id. at 121, 590 A.2d 234 (instructing trial court to focus on “the five ‘facts relating to the crimes’ referred to in [Yarbough] guideline 3”). In State v. Baylass, 114 N.J. 169, 180, 553 A.2d 326 (1989), we summarized those five factors as follows: “[I]t suffices to note that they generally concentrate on such considerations as the nature and number of offenses for which the defendant is being sentenced, whether the offenses occurred at different times or places, and whether they involve numerous or separate victims.”
*424In the present case, three of the five factors militate in favor of concurrent sentences: (i) defendant’s crimes (two counts of vehicular homicide and two counts of fourth-degree assault by auto based on causing serious bodily injuries to the victims) and their objectives were not predominantly independent of each other; (ii) defendant’s crimes involved a single act of violence (one vehicular accident); and (iii) the crimes involved a single episode of aberrant behavior. The fourth factor weighs in favor of consecutive sentences because the accident resulted in multiple deaths. The fifth factor — “whether there are numerous convictions” — provides some support for consecutive sentences because defendant’s four convictions approach the “numerous” range. In sum, three factors support concurrent sentences and two factors support consecutive sentences. Our task is to determine whether the multiple-victims and numerous-convictions factors, by themselves, support consecutive sentences. That decision is influenced to a large degree by the aggravating factors, which are disputed in this case. It is undisputed, however, that the trial court found no mitigating factors.
When a trial court fails to give proper reasons for imposing consecutive sentences at a single sentencing proceeding, ordinarily a remand should be required for resentencing. State v. Miller, 108 N.J. 112, 122, 527 A.2d 1362 (1987). The same is true when the trial court double counts or considers an improper aggravating factor as is alleged in this case. The reason for such a remand is that appellate courts should exercise original sentencing jurisdiction sparingly. State v. Kromphold, 162 N.J. 345, 355, 744 A.2d 640 (2000); State v. Ghertler, 114 N.J. 383, 387-88, 555 A.2d 553 (1989). An additional reason why a remand is the preferred procedure is that the trial court may restructure the sentence on remand without violating the defendant’s due process or double jeopardy rights, so long as the defendant’s aggregate sentence is not increased. State v. Rodriguez, 97 N.J. 263, 277, 478 A.2d 408 (1984).
*425B.
Before deciding whether the vehicular-homicide sentences should be consecutive or concurrent, we address the trial court’s handling of the aggravating factors. As we noted before, the Appellate Division concluded that the trial court erred in identifying and weighing the aggravating factors. The panel determined that the trial court had impermissibly double counted “the two deaths as an aggravating factor despite the fact that the deaths were an element of the offense.” It is well-settled that where the death of an individual is an element of the offense, that fact cannot be used as an aggravating factor for sentencing purposes. State v. Pineda, 119 N.J. 621, 627, 575 A.2d 855 (1990). Hence, it was improper for the trial court to rely on the victims’ deaths as part of an aggravating factor in this case.
The trial court, however, did not rely on the two deaths alone to establish the gravity and seriousness of harm aggravating factor. It also took into account the serious bodily injuries sustained by the two surviving victims, which formed the basis for the two fourth-degree convictions. Michael DiGangi suffered compound fractures to the femur and tibia in his right leg, injuries to his back, and the loss of a tooth. Joyce Snook broke both of her femurs and her ankles and suffered a collapsed lung. Although the two separate convictions for assault by auto required the jury to find that defendant had caused “serious bodily injury” to Snook and DeGangi, N.J.S.A. 2C:12-1c,2 we recently held that the statutory definition of “serious bodily injury” and the aggravating factor focusing on the “seriousness of harm inflicted on the victim” *426address different concepts. Kromphold, supra, 162 N.J. at 358, 744 A.2d 640. There, we explained:
When a sentencing court considers the harm a defendant caused to a victim for purposes of determining whether [the gravity and seriousness of harm] aggravating factor is implicated, it should engage in a pragmatic assessment of the totality of harm inflicted by the offender on the victim, to the end that defendants who purposely or recklessly inflict substantial harm receive more severe sentences than other defendants. Although the definition of “serious bodily injury” in N.J.S.A. 2C:11-1(b) clearly contemplates a level of injury severe enough to trigger the [gravity and seriousness of harm] aggravating factor, “the gravity and seriousness of the harm” encompassed by that aggravating factor is a broader and less precise concept that pennits the exercise of sound discretion by the sentencing court in determining whether the extent of the harm to the victim warrants application of that aggravating factor.
[Ibid.]
Accordingly, we are persuaded that the extensive injuries sustained by Joyce Snook and Michael DiGangi warranted the trial court’s reliance on the gravity and seriousness of harm aggravating factor independent of the deaths of the two other victims.
The Appellate Division also concluded that the trial court erred in finding two other aggravating factors: deterrence and risk of re-offense. With respect to the deterrence factor, the Appellate Division reasoned that the vehicular-homicide statute was deterrence enough because the Legislature substantially increased the penalties for that crime just days before defendant’s accident. We do not agree. The need for public safety and deterrence increase proportionally with the degree of the offense. State v. Megargel, 143 N.J. 484, 500, 673 A.2d 259 (1996). Here, the Legislature increased vehicular homicide from a third-degree offense to a second-degree offense. It follows that the need for deterrence increased as part of a legislative plan to reduce the slaughter and mayhem that occurs on our roads at the hands of drunken drivers. Thus, the trial court was correct in finding a need for deterrence.
Regarding whether defendant was a risk to commit another offense, we have stated that “an appellate court should not second-guess a trial court’s finding of sufficient facts to support an aggravating or mitigating factor if that finding is supported by *427substantial evidence in the record.” State v. O’Donnell, 117 N.J. 210, 216, 564 A.2d 1202 (1989). The trial court found that defendant denied responsibility for the crash and did not acknowledge that he had an alcohol problem. That conclusion is supported by a letter sent to the judge by defendant. In that letter, defendant expresses remorse, but does not directly accept responsibility for the crash or admit that he has a problem of drinking and driving. In addition, William Ferguson’s mother testified at the sentencing hearing that she recently had encountered defendant at a grocery store and that he had acted in a manner indicative of denial. That evidence does not irrefutably prove that defendant is likely to re-offend, but it does provide support for the trial court’s conclusion. Considering also that the trial court was in a far better position to develop a “feel of the case” than was the Appellate Division, we conclude that the Appellate Division should not have disturbed the trial court’s finding that defendant was likely to re-offend. We conclude, therefore, that the three aggravating factors found by the trial court are supported by the record and should not have been disturbed by the Appellate Division.
C.
Next, we apply the Yarbough guidelines to this case to determine whether the vehicular-homicide sentences should be consecutive or concurrent. We begin our analysis by stressing that the Yarbough guidelines are just that — guidelines. They were intended to promote uniformity in sentencing while retaining a fair degree of discretion in the sentencing courts. As such, the five “facts relating to the crimes” contained in Yarbough’s third guideline should be applied qualitatively, not quantitatively. Cf. State v. Kruse, 105 N.J. 354, 363, 521 A.2d 836 (1987) (observing that “[m]erely enumerating [aggravating and mitigating] factors does not provide any insight into the sentencing decision, which follows not from a quantitative, but from a qualitative, analysis”). It follows that a sentencing court may impose consecutive sentences even though a majority of the Yarbough factors support *428concurrent sentences. See, e.g., State v. Brown, 138 N.J. 481, 560, 651 A.2d 19 (1994) (declaring that “consecutive sentencing for multiple murders that occur in close sequence is not improper”), overruled on other grounds, State v. Cooper, 151 N.J. 326, 377, 700 A.2d 306 (1997); State v. Perry, 124 N.J. 128, 177, 590 A.2d 624 (1991) (concluding that trial court properly imposed consecutive sentences even though four of Yarbough’s five factors militated in favor of concurrent sentences).
When a trial court is faced with the decision whether to impose consecutive or concurrent sentences, the court must determine whether the Yarbough factor under consideration “renders the collective group of offenses distinctively worse than the group of offenses would be were that circumstance not present.” People v. Leung, 5 Cal.App.4th. 482, 7 Cal.Rptr.2d 290, 303 (1992). Indeed, there are some cases that are so extreme and so extraordinary that deviation from the guidelines may be warranted. Yarbough, supra, 100 N.J. at 647, 498 A.2d 1239; State v. Louis, 117 N.J. 250, 252, 566 A.2d 511 (1989). Crimes involving multiple deaths or victims who have sustained serious bodily injuries represent especially suitable circumstances for the imposition of consecutive sentences. See, e.g., State v. J.G., 261 N.J.Super. 409, 426, 619 A.2d 232 (App.Div.), certif. denied, 133 N.J. 436, 627 A.2d 1142 (1993); State v. Russo, 243 N.J.Super. 383, 413, 579 A.2d 834 (App.Div.1990), certif. denied, 126 N.J. 322, 598 A.2d 882 (1991). However, “[t]hat does not mean that all consecutive sentencing criteria are to be disregarded in favor of fashioning the longest sentence possible.” Louis, supra, 117 N.J. at 258, 566 A.2d 511.
We have little difficulty concluding that a drunk-driving accident that results in two deaths and serious bodily injuries to two others is distinctively worse than a drunk-driving accident that results in death or serious bodily injuries to a single individual. As a result of the accident, Melissa Snook’s two young children will be forced to grow up without a mother. William Ferguson’s parents will never see their son graduate from college. Michael DeGangi and Joyce Snook, who suffered serious bodily injuries, have undergone *429multiple surgeries and spent countless hours in physical therapy. We make those observations simply to illustrate that “[t]he total impact of singular offenses against different victims will generally exceed the total impact on a single individual who is victimized multiple times.” Leung, supra, 7 Cal.Rptr.2d at 303-04. Accordingly, defendant’s culpability exceeds the culpability of someone who commits the same group of offenses against a single victim because culpability is influenced not only by the single act of driving while drunk, but also by the number of victims killed or caused to sustain serious bodily injuries the singular criminal event generates.
Although that principle resonates most clearly in eases in which a perpetrator intentionally targets multiple victims (e.g., a double murder or robbery), it also applies to cases in which, as here, the defendant does not intend to harm multiple victims but it is foreseeable that his or her reckless conduct will result in multiple victims. Defendant has argued that punishing him for both deaths unjustly bases his punishment on the misfortune that he happened to be involved in an accident with multiple victims. The District of Columbia Court of Appeals cogently explained the fallacy in defendant’s argument:
That appellant’s conduct would have resulted in the tragedy which occurred was not fortuitous but, unhappily, was almost inevitable. The combination of an undue ingestion of alcohol and the resultant mishandling of automobiles causes awesome carnage on our highways daily. In fairness it can be said that appellant could hardly have chosen a means which would have been more likely to result in injury to many persons.
[Murray v. United States, 358 A.2d 314, 321-22 (D.C.1976).]
In the present ease, it was especially foreseeable that driving while intoxicated would result in multiple victims because defendant had two passengers in his vehicle.
In light of the foregoing principles, we hold that, in order to facilitate sentencing under Yarbough in vehicular homicide cases, the multiple-victims factor is entitled to great weight and should ordinarily result in the imposition of at least two consecutive terms when multiple deaths or serious bodily injuries have *430been inflicted upon multiple victims by the defendant. Kromphold, supra, 162 N.J. at 347-48, 359, 744 A.2d 640 (affirming consecutive sentences of drunk driver who caused single accident resulting in multiple victims); State v. Serrone, 95 N.J. 23, 27-28, 468 A.2d 1050 (1983) (suggesting that multiple-victims factor especially supports imposition of consecutive sentences). The fact that two consecutive terms of imprisonment should ordinarily be imposed in multiple-victims cases does not prevent the sentencing court from setting the base term of each sentence below the maximum provided by the Code. That is precisely what occurred here.
D.
Finally, we address whether it was an abuse of the trial court’s discretion to impose consecutive sentences. The Appellate Division concluded that the trial court abused its discretion, and remanded for the imposition of concurrent sentences. We disagree.
“Although appellate courts possess original jurisdiction over sentencing, the exercise of that jurisdiction ‘should not occur regularly or routinely; in the face of a deficient sentence, a remand to the trial court for resentencing is strongly to be preferred.’ ” Kromphold supra, 162 N.J. at 355, 744 A.2d 640 (quoting State v. Jarbath, 114 N.J. 394, 411, 555 A.2d 559 (1989)). An appellate court may disturb a sentence imposed by the trial court in only three situations: (1) the trial court failed to follow the sentencing guidelines, (2) the aggravating and mitigating factors found by the trial court are not supported by the record, or (3) application of the guidelines renders a specific sentence clearly unreasonable. Roth, supra, 95 N.J. at 365-66, 471 A.2d 370.
The Appellate Division asserted that in vehicular-homicide cases, the multiple-victims factor does not, by itself, “ordinarily” result in consecutive sentences. The decisional law has gone both ways. Compare State v. Travers, 229 N.J.Super. 144, 147, 550 A.2d 1281 (App.Div.1988) (affirming concurrent sentences of de*431fendant who caused multiple deaths in alcohol-related automobile accident); with State v. Pindale, 279 N.J.Super. 123, 128, 652 A.2d 237 (App.Div.) (approving consecutive sentences of defendant who caused multiple deaths in alcohol-related automobile accident), certif. denied, 142 N.J. 449, 663 A.2d 1357 (1995). Regardless, we are convinced that, as a general matter, it is appropriate to impose consecutive sentences on defendants who drive while they are drunk and cause accidents that result in multiple deaths or multiple persons sustaining serious personal injuries. Although the record could support a more lenient sentence in this case, we do not find that the judgment of the trial court was “so wide of the mark as to require its modification by this Court.” State v. Peace, 63 N.J. 127, 129, 305 A.2d 410 (1973).
III.
We reverse the judgment of the Appellate Division and reinstate the trial court’s original sentence. Under the guidelines established by Yarbough, supra, 100 N.J. 627, 498 A.2d 1239, and Jarbath, supra, 114 N.J. 394, 555 A.2d 559, the Appellate Division should not have vacated the sentences and exercised its original sentencing jurisdiction.
Following our decision in Yarbough, the Legislature amended N.J.S.A. 2C:44-5a to provide that "[t]here shall be no overall outer limit on the cumulation of consecutive sentences,” thereby eliminating guideline number six. L. 1993, c. 223, § 1; see also State v. Eisenman, 153 N.J. 462, 478, 710 A.2d 441 (1998) (recognizing supersedence).
Today, the State can obtain a fourth-degree assault-by-auto conviction in one of two ways. N.J.S.A. 2C:12-1c. The State can either prove that the defendant drove recklessly and caused serious bodily injury, or that the defendant was legally intoxicated (or refused to submit to a breathalyzer test) and caused bodily injury. Ibid. However, when defendant's accident occurred in 1995, the State could only obtain a fourth-degree assault-by-auto conviction by proving that the defendant drove recklessly and caused serious bodily injury.