INTRODUCTION

¶ 1 Jesse Ogden sexually abused Victim several times before her fifth birthday. Several years later, a friend of Victim's mother repeatedly abused Victim sexually. Ogden pled guilty to two counts of aggravated sexual abuse of a child. Victim intervened in Ogden's restitution hearing to seek restitution for, among other things, the anticipated cost of mental health treatment for the remainder of her life. The district court entered orders for complete and court-ordered restitution. Ogden challenges the restitution orders, and levels a variety of arguments against them. Ogden contends that Victim's damages were caused, at least in part, by her subsequent sexual abuse, and that the district court based its complete restitution award on speculation about expenses she would incur in the future.

¶ 2 We clarify the standard the Crime Victims Restitution Act requires the district court to employ to determine whether a defendant caused the loss for which a victim seeks restitution. And we remand to permit the district court to apply that standard. Because the issue will likely arise again on remand, we also take the opportunity to address Ogden's argument that the district court did not have an evidentiary basis to *1135 support its conclusions about some of Victim's future expenses.

BACKGROUND

¶ 3 Victim, her two brothers, her sister, and Victim's mother (Mother) lived with Ogden. Ogden sexually abused Victim several times during the first few years of her life. When Victim was five years old, she disclosed the abuse to Mother. Ogden subsequently moved out of the family home. Victim's brothers blamed and ostracized Victim for Ogden's separation from the family. To protect her from her brothers' harassment, Mother sent Victim to live with several other families.

¶ 4 After Victim disclosed that Ogden had abused her, she attended therapy for approximately four years. During this time, Victim demonstrated "angry and anxious behavior along with self-injurious behavior of scratching herself when upset." Toward the end of treatment, Victim's therapist reported that Victim "had made considerable progress in treatment." Victim concluded treatment when she was ten years old.

¶ 5 At this point, Victim resided with one of Mother's friends, S.G. Mother was aware that S.G. was a registered sex offender. S.G. sexually abused Victim on multiple occasions while she stayed with him. In 2012, the Division of Child and Family Services (DCFS) removed Victim from Mother's custody and arranged therapy for Victim. Victim's therapist diagnosed her with adjustment disorder and post-traumatic stress disorder. Victim also displayed symptoms of anxiety and depression. Victim continued therapy for a year. Her therapist noted that Victim "achieved [her] treatment goals of decreasing anxiety and increasing social skills" but discharged Victim from treatment because she was "unwilling to talk about the sexual abuse." Her therapist concluded that "[t]herapy might be warranted in the future when [Victim] is ready to deal with [the sexual abuse]."

¶ 6 Ogden confessed to the abuse and the State charged him with one count of sodomy upon a child and one count of aggravated sexual abuse of a child. The State later amended the charges to two counts of attempted aggravated sexual abuse of a child. Ogden pleaded guilty to both counts and the district court sentenced him to two concurrent terms of three years to life. Victim filed a motion for restitution, arguing that she "suffered significant pecuniary damage as a result of the sexual abuse committed against her" by Ogden. 1 Victim supported her motion with a forensic evaluation from Dr. David Corwin and a life care plan from Sheryl Wainwright. Both Dr. Corwin and Ms. Wainwright testified at the restitution hearing.

The Forensic Evaluation

¶ 7 Dr. Corwin, a psychiatrist, reviewed Victim's medical, psychological, and DCFS records. Dr. Corwin also interviewed Victim and Mother on two occasions. He observed that Victim demonstrated clinically significant symptoms of post-traumatic stress disorder and moderate sexual distress. He concluded that Victim

suffers significant psychological trauma caused by [Ogden's] sexual abuse of her, the family disturbances associated with that sexual abuse including harass[ment] by her older brothers and living away from her family with various other families culminating in her spending considerable time with [S.G.] ... where she was probably sexually abused again.... The impacts of the sexual abuse are many and can last throughout life. Separate contributions of the sexual abuse by [Ogden and S.G.] cannot be reliably parsed.... The full extent of [Victim's] injuries is likely not yet fully apparent. She may well develop more serious problems in the future that may require treatment and interventions beyond what I recommend at this time.

¶ 8 Dr. Corwin also opined that Victim "is at an increased risk for a variety of behavioral, psychiatric and physical problems because of the sexual abuse, exposure to domestic violence and the other stresses and losses *1136 associated with it." He explained that these problems include "[d]epression, suicidal thoughts, drug and alcohol problems, ... school performance and vocational problems among many others including long-term health problems like increased risks of heart disease not associated with other risk factors like diabetes, smoking and high blood pressure."

¶ 9 Dr. Corwin recommended that Victim "receive a course of Trauma Focused Cognitive Behavior Treatment (TF-CBT) within the near future as soon as a therapist trained in TF-CBT, or some similarly evidence-based treatment, can be found for her." He added that Victim "should be provided the resources for five additional trauma focused treatment courses, totaling six, of evidence-based individual psychotherapy over the course of her life. Each course may last for up to two years of weekly individual therapy."

The Life Care Plan

¶ 10 Ms. Wainwright, a registered nurse and certified life care planner, created a life care plan for Victim (the Plan). 2 To craft the Plan, Ms. Wainwright reviewed Victim's medical, psychological, and DCFS records. She also reviewed Dr. Corwin's forensic evaluation and included his recommendations in the Plan. She did not examine or interview Victim or Mother.

¶ 11 Ms. Wainwright projected the lifetime costs of treatment and care for Victim in several categories, two of which are relevant to this appeal: inpatient and outpatient psychiatric services, and medications.

1. Inpatient and Outpatient Psychiatric Services

¶ 12 The Plan estimated the lifetime cost of Victim's psychological counseling, which included the cost of inpatient hospitalizations, intensive outpatient programs, a psychologist for Mother, a psychologist for Victim, eye movement desensitization and reprocessing (EMDR) therapy, and a neuropsychological evaluation. The Plan explained that "[d]epression, anxiety, self-destructive behaviors, dissociative behaviors, substance abuse, borderline personality disorder, promiscuity or sexual dysfunction, issues with relationships, and physical sequelae are common among[ ] victims of childhood sexual abuse." The Plan also explained that "clients with PTSD who have suffered from childhood sexual abuse often have episodes in their lifetime that require an inpatient admission to manage their symptoms and stabilize them."

¶ 13 The Plan noted that Victim had engaged in self-harm which "required emergency room visits for deep cuts that were not repairable at the time of her admission." The Plan recommended that because Victim "was not in therapy [at the time of the Plan's preparation], and her behavior is evidence that she does not have the strategies in place to manage her symptoms, she will likely require multiple admissions throughout her lifetime." Ms. Wainwright explained that she had clients who were hospitalized multiple times a year. And the Plan included ten inpatient hospitalizations for Victim's lifetime, lasting seven to twelve days per admission. It also included the cost of ten intensive outpatient programs. The Plan recommended EMDR because it is "a recommended and accepted treatment protocol for clients with PTSD." Finally, the Plan recommended a neuropsychological evaluation "to evaluate [Victim's] functional abilities for work and her personal life."

2. Medications

¶ 14 The Plan also forecasted the lifetime cost of medications to combat anxiety, sleeping disorders, and pain. The Plan explained that even though Victim was not taking any medications, "medication management is a vital part of treatment for anxiety, depression, and PTSD." The Plan included the cost of various medications for depression because "[s]tudies indicate [they] are the most effective medications in the treatment of PTSD associated with childhood sexual abuse, with associated anxiety and depression." The Plan also explained that "clients who are the victims of childhood sexual abuse tend to develop physical symptoms including *1137 chronic fatigue, fibromyalgia, chronic pain, irritable bowel, difficulty breathing, and cardiac issues," but it is unclear whether the Plan included the cost of medications to treat these issues in its recommendation.

The Restitution Hearings

¶ 15 Victim intervened in the restitution hearing and was represented by the same counsel who was pursuing her claim for civil damages against Ogden in a separate proceeding. The State initially did not appear at the restitution hearing. The district court explained that it felt "uncomfortable without at least somebody sitting by from the State" because the court "believe[d] the State does need to be present." The State then made a brief appearance but simply requested to be excused from the hearing. The State explained that it wanted to "leave [the hearing] in the hands of [Victim's counsel]" because "[t]hey have ... all of the documents, witnesses, et cetera." Defense counsel did not object and the court excused the State. The State left the hearing, never to return. From that point on, Victim's counsel presented evidence concerning the losses she alleged Ogden had caused (or would cause) her to suffer.

¶ 16 At the hearing, Dr. Corwin spoke generally about the impacts of sexual abuse on a child and the treatment that typically benefits abuse victims. He explained that most frequently, the impacts include "increased anxiety, depression, suicidal ideation, suicidal acts, self-harm acts, ... diminished learning ability and later decreased vocational attainment, increased difficulty in interpersonal relationships and in intimate relationship[s], [and] problems with sexual functioning." Dr. Corwin also explained that "sexually abused children are more likely to get sexually abused later in life." He explained that his "view of future care of people who have experienced severe trauma is that they need to have the availability of periods of treatment throughout their life."

¶ 17 Dr. Corwin reiterated the conclusions and treatment recommendations contained in his forensic evaluation. He explained that at the time he assessed Victim "she was very symptomatic," "failing in school," and engaging in self harm behaviors. He concluded that although Victim is "at least [of] normal intelligence," she was failing school because "she is so adversely impacted from the severe stress that she has experienced." Dr. Corwin concluded that Victim "is a severely affected sexually abused child ... who shows a wide variety of psychological and behavioral problems that are known to be related to child sexual abuse and other early childhood adversity which she has also experienced, and that she is likely to have lifelong continued increased vulnerability to a wide variety of health, mental health, behavioral problems, and that the prognosis for her is guarded."

¶ 18 Dr. Corwin acknowledged that Victim had been abused by a second individual years after Ogden's abuse. Dr. Corwin testified that "[i]t's not possible" to separate the psychological trauma that Victim suffered from each perpetrator "[b]ecause they are so interwoven and sort of mixed together in her." He stated that he did not believe the damage could be "scientifically or reliably, accurately apportioned," and to do so would be "pure speculation." He opined that one could "[p]robably ... say that the subsequent stressor of being molested by [S.G.] has also contributed to the severity of her present condition," but by how much was undeterminable. And additionally, Dr. Corwin explained that the earlier trauma "builds and relates and contributes to her vulnerability to the subsequent one." He also noted that "in general it is believed that earlier traumas are more destructive because there were more developmental phases left for a person to go through. So early trauma, major trauma, is believed to be somewhat more severe than later trauma."

¶ 19 Ms. Wainwright also testified at the hearing. She explained that she is not qualified to diagnose conditions or prescribe medications. She based her estimates in the Plan on "literature, [her] experience, and ... one recommendation from Dr. Corwin." When counsel asked her how she formulated the numbers in the Plan, she responded, "I made some assumptions. And when we do a life care plan, we're kind of making assumptions and looking into a crystal ball a little bit. But ... I made assumptions based on the statistics that I read in several articles talking about the long term effects of child sex abuse and also in my experience ... case managing clients with child sex abuse."

*1138 ¶ 20 Before issuing its restitution order, the district court made several findings with respect to causation. The court acknowledged that Dr. Corwin "felt strongly that there was no way to really allocate the damage between perpetrators," because it would be "pure speculation" to do so. But the court also found that "earlier trauma was believed by [Dr. Corwin] to contribute more to the later problems." The court also found that, because Dr. Corwin testified that "sexual abuse victims are six times more likely to be sexually abused again," Victim "was more vulnerable as a sexual abuse victim because of [Ogden], because of his actions."

¶ 21 Before issuing its restitution order, the district court asked "[w]hat, if any, is the collateral effect of this Court's ruling on the civil matter or other matters that may be out there?" The court opined that "what happens here doesn't have any effect on what has to be proven in [the] civil action," because "there are no rules of evidence, very limited rules of evidence in this matter that would take place in the civil action." Despite apparently misunderstanding that the complete restitution order would become an enforceable civil judgment, the district court entered a complete restitution order that totaled $2,092,306.

¶ 22 At a supplemental restitution hearing, the district court requested information about Ogden's assets for a final determination of court-ordered restitution. Ogden held two life insurance policies-one for his wife and one for his children, including Victim. The court ordered that "the beneficiaries be changed on the policies," so that Ogden's "[c]hildren and wife ... are all equal beneficiaries on the polic[y]." Victim requested that one-half of the equity in the home be applied to the value of court-ordered restitution. The court ordered that "[h]alf of the equity based on the market value would go to [Ogden's] estate and would be used for the court ordered restitution." The court stated it could "award that one-half of the equity. That can become a lien on the home. It can be foreclosed [and] would essentially be a lien on the home, whatever that final judgment is."

The Restitution Order

¶ 23 As noted above, the court set complete restitution at $2,092,306. The court found that "[e]ach of the categories of figures from Dr. Corwin and Sheryl Wainwright are necessary for [Victim]." The court ordered that "[a]ny biological or adopted children who are already listed on the life insurance policies are to be named as equal beneficiaries along with [Ogden's] current wife. [Ogden] may not add or remove any beneficiaries from the policies." The court "award[ed] [half] of the equity of [Ogden's] home based on fair market value less [a] homestead exemption to be paid to [Victim]." Ogden's court-ordered restitution was $127,089.60, plus one-half of the equity in the home, not including Ogden's homestead exemption, plus Victim's portion of the proceeds from the life insurance policies.

ISSUES AND STANDARDS OF REVIEW

¶ 24 Ogden argues that the district court misread the statute to conclude that the Crime Victims Restitution Act required something less than proximate causation to establish that Ogden caused Victim's damages. We review questions of statutory interpretation for correctness. Nichols v. Jacobsen Constr. Co. , 2016 UT 19 , ¶ 13, 374 P.3d 3 . 3

*1139 ¶ 25 Although we vacate the district court's restitution award because it applied an incorrect causation standard, we will address one of Ogden's challenges to the restitution award because we believe it will be helpful to the district court on remand. See State v. Low , 2008 UT 58 , ¶ 61, 192 P.3d 867 (When "there are other issues presented on appeal that will likely arise [on remand]," we may "exercise our discretion to address those issues for purposes of providing guidance on remand."). Ogden contends that the district court abused its discretion in its calculation of complete restitution because Victim's damages were impermissibly speculative. We "will not disturb a district court's [restitution] determination unless the court exceeds the authority prescribed by law or abuses its discretion." State v. Laycock , 2009 UT 53 , ¶ 10, 214 P.3d 104 .

ANALYSIS

I. The Crime Victims Restitution Act Permits Recovery of Costs the Defendant Has Proximately Caused the Victim to Suffer

¶ 26 The Crime Victims Restitution Act (CVRA) 4 requires courts to order restitution "[w]hen a defendant enters into a plea disposition or is convicted of criminal activity that has resulted in pecuniary damages...." UTAH CODE § 77-38a-302(1). The CVRA requires the district court to calculate two types of restitution: complete restitution and court-ordered restitution. See id. § 77-38a-302(2).

¶ 27 "Complete restitution" means the "restitution necessary to compensate a victim for all losses caused by the defendant." Id. § 77-38a-302(2)(a). The court determines complete restitution based solely on the losses the victim has suffered, without regard to the defendant's ability to pay. See id. § 77-38a-302(5)(b). Once the district court determines "that a defendant owes restitution, the clerk of the court ... enter[s] an order of complete restitution ... on the civil judgment docket...." 5 Id. § 77-38a-401(1).

¶ 28 "Court-ordered restitution," on the other hand, "means the restitution the court having criminal jurisdiction orders the defendant to pay as a part of the criminal sentence." Id. § 77-38a-302(2)(b). Unlike complete restitution, courts consider the effect on the defendant to set the amount of court-ordered restitution, including the defendant's "financial resources," "other obligations,"

*1140 "the rehabilitative effect," and "other circumstances...." Id. § 77-38a-302(5)(c). The district court then orders the defendant to pay the restitution as "part of the criminal sentence." Id. § 77-38a-302(2)(b).

¶ 29 Ogden argues that the CVRA requires a defendant to only pay for losses that he proximately caused and that the district court applied the wrong causation standard. As an initial matter, it is not entirely clear what causation standard the district court used. It may have been but-for causation, or the "modified but for" test that our court of appeals has applied on a number of occasions. See State v. Poulsen , 2012 UT App 292 , ¶ 11, 288 P.3d 601 ("Utah has adopted a modified 'but for' test to determine whether pecuniary damages actually arise out of criminal activities." (quoting State v. Brown , 2009 UT App 285 , ¶ 11, 221 P.3d 273 ) ). See also State v. Ruiz , 2016 UT App 18 , ¶ 12, 366 P.3d 1230 ("In order to determine complete restitution, the trial court was required to employ '[a] modified "but for" test.' " (alteration in original) (quoting State v. Ruiz , 2013 UT App 166 , ¶ 8, 305 P.3d 223 ) ); State v. Birkeland , 2011 UT App 227 , ¶ 11, 258 P.3d 662 ("The circumstances ... readily support an award of restitution for the loss ... under our modified but for test."). Notwithstanding that ambiguity, it is clear that the district court employed something other than proximate cause to assess which of Victim's losses Ogden caused. And this matters because the causation standard applied could well affect the outcome of this case.

¶ 30 The State argues that but-for causation is the appropriate standard, but also contends that the causation standard is irrelevant because any standard-including proximate cause-would support the restitution award the court entered.

¶ 31 To determine what causal connection the Legislature intended to apply to the CVRA, we start with the statute's language. " 'When interpreting a statute, it is axiomatic that this court's primary goal "is to give effect to the legislature's intent in light of the purpose that the statute was meant to achieve." ' " Garfield Cty. v. United States , 2017 UT 41 , ¶ 15, --- P.3d ---- (quoting Biddle v. Wash. Terrace City , 1999 UT 110 , ¶ 14, 993 P.2d 875 ). "The best evidence of the legislature's intent is 'the plain language of the statute itself.' " State v. Miller , 2008 UT 61 , ¶ 18, 193 P.3d 92 (citation omitted).

¶ 32 The CVRA addresses causation in two provisions. The statute allows the district court to enter an order "[w]hen a defendant enters into a plea disposition or is convicted of criminal activity that has resulted in pecuniary damages...." UTAH CODE § 77-38a-302(1) (emphasis added). And the district court awards " '[c]omplete restitution' ... to compensate a victim for all losses caused by the defendant." 6 Id. § 77-38a-302(2)(a) (emphasis added). Thus, the relevant language for our analysis is "resulted in" and "caused by." And, unfortunately, neither speaks explicitly to the causation standard the Legislature intended. 7

¶ 33 First, the phrase "resulted in" does not, by itself, tell us what type of causation the statute requires. To "result" is "to proceed or arise as a consequence, effect, or conclusion." Result , MERRIAM-WEBSTER , https://www.merriam-webster.com/dictionary/result (last visited Feb. 7, 2018). However, something that would not have happened but for someone's action (a but-for cause) could be described as "a consequence, effect, or conclusion," just as easily as an action that had a more direct causal connection (such as proximate cause). See id.

¶ 34 "Caused by" reveals just as little about what the Legislature intended. "Caused by" could refer to a but-for cause or a proximate cause. UTAH CODE § 77-38a-302(2)(a). As such, the plain language by itself does not answer the question.

¶ 35 We confronted a similar linguistic puzzler in Barneck v. Utah Department of Transportation . 2015 UT 50 , 353 P.3d 140 . There, we needed to determine what causation standard the Legislature intended to apply to a provision of the Governmental *1141 Immunity Act (GIA). Id. ¶ 2. The GIA waives immunity for "any injury caused by ... a defective, unsafe, or dangerous condition of any highway [or] ... culvert," but "reinstated [immunity] for injuries that 'arise [ ] out of, in connection with, or result[ ] from' a latent defective condition of a culvert; from the 'management of flood waters'; or from the 'construction, repair, or operation of flood or storm systems.' " Id. ¶¶ 2, 36 (alterations in original) (omissions in original) (citations omitted).

¶ 36 We ultimately concluded that the GIA reinstates immunity only when proximate cause exists between the injury and the defect, management, or construction. Id. ¶¶ 36, 44. There, we recognized that "the [']results from' formulation may properly be understood as the invocation of a but-for test," but that "[s]ometimes 'results from' is understood to convey the principle of proximate cause." Id. ¶¶ 39, 41.

¶ 37 We also reasoned that "courts 'read phrases like "results from" to require but-for causality' only '[w]here there is no textual or contextual indication to the contrary.' " Id. ¶ 41 (alteration in original) (citation omitted). We then looked at the statutory framework for a contrary indication-an indication we found, in part, by acknowledging that a but-for standard for the reinstatement of immunity would, in practice, swallow the waiver. 8 Id. ¶¶ 42-44.

¶ 38 Examining the CVRA's structure, we conclude that the Legislature intended that the same causation standard apply in a restitution hearing that would apply in a parallel civil action. See, e.g. , Retherford v. AT&T Commc'ns of Mountain States, Inc. , 844 P.2d 949 , 970-71 (Utah 1992) (requiring a showing that "conduct proximately caused [the plaintiff's] emotional distress" to sustain an intentional infliction of emotional distress claim); Reynolds v. MacFarlane , 2014 UT App 57 , ¶ 16, 322 P.3d 755 (" 'Damages ... may ... be recovered only to the extent that [the plaintiff] proves they were a proximate result' of the nonconsensual touching." (alteration in original) (second omission in original) (citation omitted) ). The most compelling evidence springs from the provision that converts a complete restitution order into an enforceable civil judgment. See UTAH CODE § 77-38a-401. This means that even though the restitution proceeding takes place as part of a criminal proceeding, the end result is a civil judgment that a victim is entitled to enforce.

¶ 39 To accept the State's reading of the CVRA, we would need to conclude that the Legislature designed a system with dual tracks to a civil judgment with each track offering the victim a different causation standard. It is unlikely that the Legislature would have created a framework where the burden to prove causation varied depending on whether the victim proceeded in a criminal restitution hearing or a civil trial. At the very least, it is unlikely that the Legislature would have implemented such a system without explicitly expressing that intent in the statute. In the absence of such an expression of legislative intent, we opt for the reading that harmonizes with the causation standard that would apply in an analogous civil action: proximate cause.

¶ 40 We find further support for this reading in the CVRA's definition of "pecuniary damages." "Pecuniary damages" are "all demonstrable economic injury, whether or not yet incurred, which a person could recover in a civil action...." Id. § 77-38a-102(6) (2015). 9

*1142 This also suggests that the Legislature did not intend to create a separate track that would allow a victim to obtain a subset 10 of the relief available in a civil case under a different causation standard.

¶ 41 The State offers several arguments to combat this conclusion. First, the State argues that the Legislature's use of the term "all" in the phrase "all losses caused by defendant" requires the imposition of but-for causation. (Quoting Id. § 77-38a-302(2)(a) ). We agree that the CVRA requires the court to consider all losses the defendant causes, but that simply raises the question of what losses the defendant caused. In other words, "all" does not tell us whether the Legislature intended proximate or but-for causation.

¶ 42 Second, the State points to our statement in State v. Laycock that "proximate cause ... [is] best left to civil litigation." 2009 UT 53 , ¶ 29, 214 P.3d 104 . We were not asked in Laycock to consider the CVRA's causation standard. Rather, we were asked to examine if the CVRA required a district court to order both complete and court-ordered restitution. Id. ¶ 19. We concluded that the CVRA requires a district court to "determine" complete restitution, but gives it discretion with regard to the imposition of court-ordered restitution. Id. ¶ 23. Along the analytical path, we remarked on a number of the difficulties a sentencing court would face in trying to resolve issues of causation and losses, most notably a lack of opportunity for discovery that would allow a defendant to "raise issues of proximate cause and comparative negligence by using depositions and interrogatories to gather relevant information." Id. ¶ 22. This caused us to quote favorably the court of appeals' observation that "[m]atters of negligence, proximate cause and the amount of resulting damages are best left to civil litigation." Id. ¶ 29 (quoting State v. Robinson , 860 P.2d 979 , 983 (Utah Ct. App. 1993) ). This statement was dicta then and is dicta now. And although we continue to agree that the civil arena appears to be a superior forum to adjudicate those issues, that statement should not be read to suggest that we had concluded the CVRA requires something other than proximate causation. 11

¶ 43 Third, the State argues that the application of proximate cause does not comport with the purposes of the restitution statute that we identified in Laycock : compensation, rehabilitation, and deterrence. 2009 UT 53 , ¶ 18, 214 P.3d 104 . As an initial matter, reading statutory language in light of perceived legislative purpose is a very tricky species of argument. The plain language of the text provides the best evidence of legislative intent. Attempts to read statutory language in the light of a perceived purpose (or purposes) can distort our view of the plain language. This is, in part, because the statutory text already reflects the balance the Legislature *1143 has struck between competing policies and purposes.

¶ 44 The State's argument illustrates the challenge. The State posits that "the more stringent a test of legal causation adopted, the more these statutory purposes [of compensation and deterrence] are inhibited." That may be true, but the Legislature may have weighed the purposes differently and concluded that proximate causation best balanced compensation and deterrence with the ability for rehabilitation. Or the Legislature may have been motivated by a policy we did not manage to divine when we penned Laycock . And, in this case, resorting to the CVRA's perceived purposes does not help us understand what the Legislature intended.

¶ 45 Fourth, the State argues that proximate cause would be inappropriate because we have observed that "the foreseeability aspect of proximate causation is frequently relaxed in the case of intentional torts." Graves v. N. E. Servs., Inc., 2015 UT 28 , ¶ 51 n.3, 345 P.3d 619 (citation omitted). But this does not tell us what the statutory language means. The fact that the "foreseeability aspect ... is frequently relaxed" should inform the district court's application of proximate cause principles on remand, not our decision on whether to apply a proximate cause standard at the outset. Id. (citation omitted).

¶ 46 Fifth, the State argues for a standard other than proximate cause based on our interpretation of two insurance contracts. In National Farmers Union Property & Casualty Co. v. Western Casualty & Surety Co. , we considered a policy that excluded coverage for "bodily injury or property damages arising out of any premises, other than an insured premises, owned, rented, or controlled by any insured." 577 P.2d 961 , 962 (Utah 1978). We reasoned that "[a]s used in a liability insurance policy, the words 'arising out of' are very broad, general and comprehensive. They are commonly understood to mean originating from, growing out of, or flowing from, and require only that there be some causal relationship between the injury and the risk for which coverage is provided." Id. at 963 ; See also Viking Ins. Co. of Wis. v. Coleman , 927 P.2d 661 , 665 (Utah Ct. App. 1996) (applying a causal nexus test similar to the modified but-for test the court of appeals applied in restitution cases). Because we interpret insurance policies broadly and in favor of the insured, our interpretation of insurance policies carries little persuasive value to help us understand this statutory language. See U.S. Fid. & Guar. Co. v. Sandt , 854 P.2d 519 , 522 (Utah 1993) ("[I]nsurance policies should be strictly construed against the insurer and in favor of the insured....").

¶ 47 Finally, the State argues that we need not definitively resolve what the CVRA requires because the district court would reach the same result regardless of what causation standard is applied. We are not convinced that the outcome would have been the same had the district court required a showing of proximate cause. "For proximate cause to exist, the relationship between the negligent act and the injury must be foreseeable." Fordham v. Oldroyd , 2007 UT 74 , ¶ 30, 171 P.3d 411 . Here, Ogden argued that Victim's losses were not foreseeable because "Ogden would not reasonably foresee when he abused [Victim] that she would be again sexually abused eight years later by a third party...." We do not opine on the foreseeability of Victim's subsequent abuse, other than to acknowledge that Ogden has a good faith argument that it was not foreseeable. Thus, we rest our remand on the premise that application of an erroneous standard may have had a material effect and that Ogden is entitled to a proceeding conducted with the proper causative test in clear focus.

¶ 48 Because we find that proximate cause is required to find that a "criminal activity ... has resulted in pecuniary damages" and had the district court applied this causation standard, the outcome may have changed, we vacate and remand for further proceedings consistent with this opinion. 12 See *1144 UTAH CODE § 77-38a-302(1).

II. On Remand, the District Court Should Ensure That the Restitution Award Is Not Based on Speculative Evidence

¶ 49 Although we vacate the district court's restitution award and remand, Ogden presents another question that will likely arise again before the district court. Ogden argues that the district court erred by including speculative damages in its restitution calculation. 13 Although it is unnecessary to our decision, we retain the authority to reach issues when we believe our analysis could prove helpful on remand. 14 See State v. Low , 2008 UT 58 , ¶ 61, 192 P.3d 867 (When "there are other issues presented on appeal that will likely arise [on remand]," we may "exercise our discretion to address those issues for purposes of providing guidance on remand."). The issue we examine is whether the damages included in the complete restitution award were speculative and therefore lacked a sufficient evidentiary basis.

A. Speculative Damages

¶ 50 The district court's complete restitution award totaled $2,092,306-the projection of expected future life care plan costs the Plan predicted, accounting for inflation over Victim's lifetime. In the restitution order, the district court found that "[e]ach of the categories of figures from Dr. Corwin and Sheryl Wainwright are necessary" for Victim. Ogden argues that several of these figures were *1145 speculative and therefore lacked a sufficient evidentiary basis.

¶ 51 Restitution shall be awarded for pecuniary damages arising out of a criminal defendant's activity. UTAH CODE § 77-38a-302(1). Pecuniary damages are

all demonstrable economic injury, whether or not yet incurred, including those which a person could recover in a civil action arising out of the facts or events constituting the defendant's criminal activities and includes the fair market value of property taken, destroyed, broken, or otherwise harmed, and losses, including lost earnings, ... and medical and other expenses, but excludes punitive or exemplary damages and pain and suffering.

Id. § 77-38a-102(6).

¶ 52 A trial court's restitution award must rely on a sufficient evidentiary basis. See State v. Weeks , 2002 UT 98 , ¶ 26, 61 P.3d 1000 . "Although an award of damages based only on speculation cannot be upheld, it is generally recognized that some degree of uncertainty in the evidence of damages will not suffice to relieve a defendant from recompensing a wronged plaintiff." Bastian v. King , 661 P.2d 953 , 956 (Utah 1983). "The amount of damages may be based upon approximations," however the approximations must be "based upon reasonable assumptions or projections." Atkin Wright & Miles v. Mountain States Tel. & Tel. Co. , 709 P.2d 330 , 336 (Utah 1985).

¶ 53 "[T]he determination of [a] restitution amount is by nature an inexact science," and consequently "a reasonable estimate of the loss" may be the only evidence available. United States v. Osman , 853 F.3d 1184 , 1189 (11th Cir. 2017) (citation omitted) (internal quotation marks omitted). However, evidence of the loss should bear "sufficient indicia of reliability to support its probable accuracy." Id. (citation omitted) (internal quotation marks omitted). "[B]allpark figures ... and purely speculative calculations" constitute "insufficient information" for a district court to rely on in awarding restitution. State v. Passwater , 379 Mont. 372 , 350 P.3d 382 , 385 (2015) (citation omitted). Rather, anticipated future expenses should be "firmly established" in the record. See Lawrence v. State , 764 P.2d 318 , 322 (Alaska Ct. App. 1988).

¶ 54 We acknowledge that it is impossible to avoid some measure of uncertainty when a district court is asked to base an award on damages a victim will suffer in the future. However, to establish a sufficient evidentiary basis for a restitution award, a district court should require the State (or the victim, in those cases where the State decides not to appear) to demonstrate that the expenses are necessary and that the amount needed to cover those expenses is firmly established in the record. In light of these clarifications, we express concern with aspects of the district court's restitution award.

¶ 55 The Plan allocated a portion of the lifetime care costs for depression, anxiety, sleeping, and pain medications. Ogden argues that this allocation was speculative because Ms. Wainwright "was not a psychiatrist, had no medical diagnostic or prescribing ability, and had no underlying data or 'careful assessment' that [Victim] would need any of the specific medications recommended on the chart." We agree with Ogden that her opinion lacked "underlying data or 'careful assessment' " that Victim would actually need all the medication the Plan contemplated.

¶ 56 The expert best positioned to opine on the medications Victim would need-Dr. Corwin-neither recommended nor prescribed any medication for her. Dr. Corwin testified at the restitution hearing that to his knowledge, Victim was not taking any type of medication. Ms. Wainwright was not qualified to diagnose psychological disorders or prescribe medication to treat them. She based her conclusion in part on her experience "in case managing [sexual abuse victims]" because "they need medication as they get older." She explained that to inform her recommendation, she "went to the studies and the literature and there's a couple of articles that [she] cited in [her] report that talk about ... therapy for people who have suffered post-traumatic stress and child sex abuse. And they recommend[ ] ... Zoloft, ... Paxil or ... Prozac as the most effective psychiatric treatment for clients who have suffered child sex abuse." Additionally, Ms. Wainwright recommended anxiety medication because *1146 anxiety is "another common side effect for clients who have suffered child sex abuse," and because Victim already exhibited "symptoms of depression, as she's cutting herself ... they're all signs of some sort of anxiety."

¶ 57 Speaking again of her experience with her clients, Ms. Wainwright explained that:

They all have sleep disorders. They have trouble sleeping. So I put in Ambien. And then the pain medications, because a lot of these clients, ... probably over 80 percent of the clients that I manage, are on some sort of pain medications for the physical symptoms that have manifested as a result of the psychiatric condition.

The medications the Plan recommended "reflect [Ms. Wainwright's] experience of the types and costs of medications that frequently help victims of sexual abuse," and "the majority of [her clients] are on these medications."

¶ 58 Leaving aside the question of whether Ms. Wainwright possessed the expertise to offer these opinions, her recommendations were based exclusively on her assessment of her other clients, not on an individual assessment of Victim's specific needs. Although Victim demonstrated clinically significant symptoms of depression and post-traumatic stress disorder, she was not diagnosed with anxiety, nor did she report trouble sleeping. Without a diagnosis and recommendation for a particular medication from an individual qualified to make that assessment, the inclusion of costs for those medications crosses the line into impermissible speculation. To establish a sufficient evidentiary basis for this portion of the restitution award, Victim would need to present credible evidence demonstrating her need for the medications.

¶ 59 The Plan also recommended inpatient and outpatient counseling. This included the cost of ten inpatient hospitalizations, ten intensive outpatient counseling programs, six courses of weekly therapy sessions, one course of Eye Movement Desensitization and Reprocessing (EMDR) treatment, and one neuropsychological evaluation. Ogden argues that this recommendation was "not based on [Victim's] actual needs and amounts to nothing more than an unfounded and speculative assertion about what [Victim's] needs might be." (Emphasis omitted).

¶ 60 The Plan based part of its conclusion on Dr. Corwin's recommendation that Victim receive six courses of weekly therapy sessions, each course lasting up to two years. In his forensic evaluation, Dr. Corwin concluded that Victim would benefit from therapy because she "suffered significant psychological trauma caused by her father's sexual abuse of her, the family disturbances associated with that sexual abuse including being harassed by her older brothers and living away from her family with various other families," which culminated in the subsequent sexual abuse by S.G. Dr. Corwin explained that "[e]ffective treatment helps diminish symptoms associated with lifelong increased emotional burdens and memories." The district court did not abuse its discretion by including these costs in the complete restitution award.

¶ 61 But the Plan recommended more than what Dr. Corwin opined would be necessary. Instead, Ms. Wainwright relied on literature and her experience with other clients as a basis to include other categories of psychological care in the Plan. Ms. Wainwright explained:

[W]hen we do a life care plan, we're kind of making assumptions and looking into a crystal ball a little bit. ... I made assumptions based on the statistics that I read in several articles talking about the long term effects of child sex abuse and also in my experience ... case managing clients with [a history of] child sex abuse.

¶ 62 Ms. Wainwright concluded it was "more likely than not" that Victim would need inpatient hospitalizations because her clients that have suffered from sexual abuse "end up in the hospital not just for psychiatric issues but for medical issues." In the Plan, she stated that because Victim was not "in therapy at [the] time, and her behavior is evidence that she does not have the strategies in place to manage her symptoms, she will likely require multiple admissions throughout her lifetime." At the hearing she explained that many of her clients, "especially women, tend to have eating disorders" and "over fifty percent" of Ms. Wainwright's clients "end up in the hospital because they've got metabolic dysfunction because *1147 they're either bulimic or anorexic." She explained that some of her clients are hospitalized once every four to six weeks, and other clients are hospitalized once or twice a year, and ultimately concluded that Victim would likely need ten inpatient hospitalizations.

¶ 63 In other words, Ms. Wainwright recommended that Victim participate in intensive outpatient programs because her other clients "do well with intensive outpatient programs." She conceded there was "no guarantee" that Victim would need to participate in an intensive outpatient program, but concluded it was "more likely than not" that Victim would. However, when asked whether this projection was "somewhat speculative," Ms. Wainwright responded, "Yes." Ms. Wainwright allocated the cost of EMDR therapy because it is "a recommended and accepted treatment protocol for clients with PTSD." Finally, she included the cost of a neuropsychological evaluation "to determine the effects and any improvements that are coming from treatment."

¶ 64 Again leaving aside the question of Ms. Wainwright's qualifications to offer that opinion, Ms. Wainwright's recommendations appear to be based on speculation about what treatments Victim may benefit from based on her other clients' treatment plans. See Bastian , 661 P.2d at 956 . Leaving aside the therapy Dr. Corwin recommended, Ms. Wainwright provided no evidence that Victim needed the treatment outlined in her recommendations. In other words, record evidence supported the inclusion of costs related to the treatment Dr. Corwin recommended. But there is more than "some degree of uncertainty in the evidence" with respect to those costs flowing from Ms. Wainwright's generalizations about what sexual abuse victims typically need. See id.

¶ 65 On remand, Victim may be able to justify an award of the items in Ms. Wainwright's report. But to meet that burden, the State or Victim will need to present evidence of her individual need for inpatient hospitalizations, intensive outpatient programs, EMDR therapy, and a neuropsychological evaluation. And although we recognize that there will be some degree of uncertainty in predicting needs decades into the future, to combat some of this uncertainty, the district court should have before it an adequate foundation that Victim will likely need these treatments.

CONCLUSION

¶ 66 The CVRA requires that a district court include the losses that a defendant proximately causes in its complete restitution determination. On remand, we remind the district court to ensure that it rests its restitution calculation on non-speculative evidence of losses that Victim has incurred or will likely incur. We vacate the district court's restitution orders and remand for further proceedings consistent with this opinion.