ZAHRA, J.

Plaintiff, Kevin Krohn, who suffered an extremely severe spinal fracture that left him paraplegic, brought this suit under the no-fault act, MCL 500.3101 et seq. Plaintiff sought personal protection insurance benefits from defendant, Home-Owners Insurance Company, to cover costs incurred for a surgical procedure performed in Portugal. It is undisputed that this surgical procedure was experimental and not a generally accepted treatment for plaintiff's injury. The dispositive question presented in this case is whether this experimental procedure was a reasonably necessary service for plaintiff's care, recovery, or rehabilitation under MCL 500.3107(1)(a). We conclude that if a medical treatment is experimental and not generally accepted within the medical community, an insured seeking reimbursement for this treatment must, at a minimum, present objective and verifiable medical evidence establishing that the treatment is efficacious. A treatment or procedure that has not been shown to be efficacious cannot be reasonable or necessary under the no-fault act. An insured's subjective belief that medical treatment is efficacious, reasonable, and necessary is not enough to create a question of fact. Viewed in the light most favorable to plaintiff, the objective and verifiable medical evidence presented at trial failed to establish that the experimental surgical procedure at issue was in any way efficacious in plaintiff's care, recovery, or rehabilitation. Plaintiff's expert witnesses merely opined that plaintiff's decision to undertake the experimental surgical procedure was an "understandable" personal decision that offered plaintiff only a medically unproven "possibility," or hope, for an efficacious result. Therefore, the procedure was not an allowable expense under MCL 500.3107(1)(a). We affirm the judgment of the Court of Appeals.1

I. FACTS AND PROCEEDINGS

On December 11, 2001, plaintiff was struck head-on by a large van while riding his motorcycle. Plaintiff suffered a severe spinal fracture that left him paraplegic, without sensation or function below the mid-chest area ("injury site"). Consequently, plaintiff was unable to touch his feet, move any part of his lower body, or determine when to relieve himself. Plaintiff underwent intensive physical therapy but did not regain any sensation below the injury site and was released from the program.

While investigating treatment options, plaintiff discovered a procedure known as olfactory ensheathing glial cell transplantation, an experimental surgery performed in Portugal. The procedure involves transplanting tissue from behind the patient's sinus cavities, which contains stem cells, to the injury site. The theory behind the procedure is that, once applied to the injury site, the transplanted stem cells could develop into spinal cord nerves. The procedure is not approved by the United States Food and Drug Administration (FDA) and therefore cannot be legally performed in the United States. In addition, there is insufficient existing research to allow for clinical trials, including controlled studies, peer review, and publication for FDA evaluation. Thus far, no one has applied for FDA approval of the procedure for any purpose.

In March 2005, plaintiff visited the Rehabilitation Institute of Michigan (RIM) and discussed the procedure with Dr. Steven Hinderer. Dr. Hinderer specializes in physical medicine and rehabilitation and is the medical director of the Center for Spinal Cord Injury Recovery Program (CSCIRP). Dr. Hinderer explained to plaintiff that he could not endorse or in any way recommend the procedure because it was highly experimental, had not yet been approved by the FDA, could not be legally performed anywhere in the United States, and lacked medical evidence to establish its efficacy.2 Neither party disputes that no one had yet applied for FDA approval of the procedure for any purpose, and the existing research was insufficient to allow clinical trials to begin. Dr. Hinderer also informed plaintiff that the procedure was not part of standard clinical care and was not likely to be covered by insurance. After consulting with Dr. Hinderer, plaintiff met with a patient who had undergone the procedure. Plaintiff claimed that this individual was able to stand on a device similar to a treadmill and walk with braces after the procedure.3

After seeking advice from family members, plaintiff decided to undergo the procedure in Portugal. Plaintiff's primary health insurer denied coverage. Plaintiff then sought coverage from defendant, a no-fault auto insurance provider. Defendant's claims specialist told plaintiff that defendant would pay for testing to determine whether plaintiff medically qualified for the procedure, but would not pay for the procedure itself because it was experimental, non-FDA approved, and illegal to perform in the United States.

Plaintiff traveled to Portugal and underwent the procedure on November 10, 2005. Ten days later, plaintiff returned to the United States and began what he described as a grueling physical therapy program at the RIM, entailing four-hour therapy sessions three times a week. Defendant paid for all the postsurgical physical therapy treatment plaintiff received. Plaintiff filed suit against defendant to recover the expenses he incurred traveling to and from Portugal and undergoing the surgery. At trial, plaintiff testified that he noticed improvements immediately after the procedure. Specifically, plaintiff testified that he could sometimes move his legs, crawl forward and backward, and control bowel and bladder movements, resulting in fewer urinary tract infections.

Dr. Hinderer testified that plaintiff had experienced "some small amount of voluntary motor function" after the procedure. Dr. Hinderer also testified that it was not possible to conclude that these minor improvements were the result of the procedure. Dr. Hinderer acknowledged that the intense physical therapy program in which plaintiff engaged postprocedure could alone have accounted for plaintiff's improvements.4 Dr. Hinderer testified about the highly experimental nature of the procedure. He described plaintiff's decision to undergo the procedure as a "personal choice" and acknowledged that this procedure was not considered necessary to the treatment and care of spinal cord injuries.

Dr. Carl Lima, a neurologist and neuropathologist at a public hospital in Portugal who is not licensed in the United States, was a member of plaintiff's surgical team, but did not perform the procedure. According to Dr. Lima, experimental data showed that transplanting nasal tissue, which contains stem cells, to the injury site provides functional recovery of neurons. He testified that this research had begun 18 years earlier on guinea pigs. There was no evidence presented at trial that the procedure has been efficacious in guinea pigs. The testimony established only that the procedure could be performed on guinea pigs without the guinea pigs' developing infections or forming tumors.

Dr. Lima testified that he started conducting human trials of the procedure in the government-operated hospital where he works, which sanctions the procedure for research purposes. No testimony was offered to suggest that the hospital had sanctioned the procedure because of its efficaciousness. Dr. Lima testified that, since 2001, 110 patients have undergone the procedure; however, Dr. Lima did not offer testimony regarding individual patients, the severity or location of their injuries, the outcomes following their procedures, or their prognoses. Dr. Lima published a paper in 2006 that summarized the outcome for seven patients who had undergone the procedure. All seven patients engaged in physical therapy following the procedure, but only two of the seven showed improvements in bladder and bowel control. Although there had been no controlled clinical studies regarding this procedure, Dr. Lima testified, "I would say the majority of the patients have some kind of improvement."5

Dr. Lima found plaintiff's spinal cord injury to be one of the most severe injuries that he had ever treated. Dr. Lima testified that he was very surprised by the "quite unexpected" results of plaintiff's procedure. Dr. Lima acknowledged that plaintiff would never fully recover from such a severe injury. Nonetheless, Dr. Lima testified that the procedure was necessary to allow plaintiff a chance at some recovery. He added that any degree of recovery requires physical therapy. Although Dr. Lima conceded that the procedure was experimental, he opined that it was reasonably necessary because a person with a chronic spinal cord injury has no other available option. The lack of FDA approval did not change Dr. Lima's opinion.

Defendant moved for a directed verdict, arguing that as a matter of law, experimental surgery is not "reasonably necessary" under the no-fault act. The trial court denied defendant's motion, ruling that whether the procedure was "reasonably necessary" was a question of fact. The jury rendered a verdict in favor of plaintiff, concluding that the procedure was reasonably necessary. Judgment was entered, awarding plaintiff $51,412.85 in allowable expenses, plus interest, case-evaluation sanctions, and taxable costs.

The Court of Appeals reversed. The Court of Appeals observed that because the dispositive issue required a review of medical judgment, plaintiff was required to present expert testimony.6 Citing SPECT Imaging, Inc. v. Allstate Ins. Co.,7 the Court of Appeals concluded that plaintiff was required to demonstrate that the procedure had gained general acceptance in the medical community.8 Because plaintiff lacked such proof, the Court of Appeals concluded that a directed verdict in favor of defendant was required.9 The dissent criticized the majority for sua sponte raising the issue of admissibility of scientific evidence because the issue was not preserved for appellate review.10 The dissent also concluded that the question whether the procedure was "reasonably necessary" was properly submitted to the jury.11

Plaintiff applied for leave to appeal in this Court. We granted the application to consider, among other issues, whether the experimental surgical procedure plaintiff underwent in Portugal was an allowable expense under MCL 500.3107(1)(a) of the no-fault act.12

II. STANDARD OF REVIEW

We review de novo a trial court's decision to direct a verdict.13 In doing so, we "review the evidence and all legitimate inferences in the light most favorable to the nonmoving party."14 Only if the evidence, when viewed in this light, fails to establish a claim as a matter of law should a motion for a directed verdict be granted.15

Issues of statutory interpretation are questions of law that this Court reviews de novo.16

III. ANALYSIS

A. BACKGROUND

The Michigan no-fault act requires that owners and registrants of automobiles carry personal protection insurance to cover an insured's medical care arising from injuries sustained in an automobile accident.17 This case requires us to determine whether the experimental surgical procedure undergone by plaintiff constituted a compensable expense under the personal protection insurance requirements of MCL 500.3107(1)(a). MCL 500.3107(1) provides in pertinent part:

B. PRINCIPLES OF STATUTORY INTERPRETATION

The primary goal of statutory interpretation is to "ascertain the legislative intent that may reasonably be inferred from the statutory language."18 "The first step in that determination is to review the language of the statute itself."19 Unless statutorily defined, every word or phrase of a statute should be accorded its plain and ordinary meaning,20 taking into account the context in which the words are used.21 We may consult dictionary definitions to give words their common and ordinary meaning.22 When given their common and ordinary meaning,23 "[t]he words of a statute provide `the most reliable evidence of its intent....'"24

C. PRECEDENT

This is not the first time this Court has been called upon to interpret MCL 500.3107(1)(a). In Nasser v. Auto Club Ins. Ass'n,25 this Court held that under MCL 500.3107(1)(a), "an insurer is not liable for any medical expense ... if the product or service itself is not reasonably necessary."26 This Court further observed that "[t]he plain and unambiguous language of [MCL 500.3107] makes both reasonableness and necessity explicit and necessary elements of a claimant's recovery, and thus renders their absence a defense to the insurer's liability."27 This Court rejected the notion that public-policy concerns would require the payment of expenses for medical care not shown to be reasonable and necessary to the care of an insured. Justice BOYLE, writing for the majority, observed that "[w]hile policy considerations may indeed cause some reluctance on the part of courts to allow insureds to be `stuck' with unnecessary expenses" that they incurred, "that determination was made by the Legislature when it drafted [MCL 500.3107] and restricted [personal protection insurance] benefits under a rule of reasonableness."28 Finally, this Court recognized that while the question of reasonable necessity under this provision is generally one for a jury, "it may in some cases be possible for the court to decide the question of the reasonableness or necessity of particular expenses as a matter of law...."29

While Nasser made clear that the language of MCL 500.3107 only permits an insured to recover expenses that are reasonable and necessary to the care, recovery, or rehabilitation of the insured, Nasser provided little guidance on how properly to determine what is a reasonably necessary expense or when such a determination may be made as a matter of law. To provide guidance along these lines, we observe that the no-fault act does not require coverage for all treatments. Obviously, treatments such as apricot pit therapy, coning (ear candling), homeopathy, magnet therapy and psychic surgery are patently unreasonable. Even if administered by licensed health-care providers, these so-called treatments not only lack a scientific basis to conclude that they are generally accepted by the medical community, but there is simply no basis to conclude that they are at all efficacious. On the other hand, we presume, subject to rebuttal, that services generally accepted by the medical community for treatment or care of a specific and diagnosed injury are reasonably necessary under MCL 500.3107(1)(a). Less clear is the case presented here, in which an insured has undergone a surgical procedure that is not generally accepted by the medical community. Defendant maintains that experimental procedures, by their nature, cannot, as a matter of law, be reasonably necessary under the no-fault act. We reject defendant's position and conclude that experimental treatments are not necessarily barred from being compensable under the no-fault act. The ultimate question whether the surgical procedure at issue here is a covered expense under the no-fault act does not turn on its status as experimental. Rather, like all claims for allowable expenses, the question turns on whether the procedure was reasonably necessary for plaintiff's care, recovery, or rehabilitation.

D. MCL 500.3107(1)(a) MUST BE ASSESSED USING AN OBJECTIVE STANDARD

In order to give meaning to this statutory provision, we start by examining the perspective from which reasonable necessity is determined. Stated more precisely, when the Legislature provided that allowable expenses consist of "all reasonable charges incurred for reasonably necessary products, services and accommodations for an injured person's care, recovery, or rehabilitation," did it intend for reasonable necessity to be determined under a subjective or objective standard?

The term "reasonable" commonly refers to that which is "agreeable to or in accord with reason; logical," or "not exceeding the limit prescribed by reason; not excessive[.]"30 The term "reasonable" has also been defined to mean "fair, proper, or moderate under the circumstances"31 and "[f]it and appropriate to the end in view."32 These definitions evidence an absence of the personal sentiment, prejudice, and bias associated with a subjective point of view, which is "based on an individual's perceptions, feelings, or intentions," rather than the "externally verifiable phenomena" associated with an objective viewpoint.33 Accordingly, we conclude that reasonableness is not based merely on the subjective perception that a service is necessary for an injured person's care, recovery, or rehabilitation. Rather, the term "reasonably" must be determined under an objective perspective.

This conclusion is entirely consistent with this Court's precedent interpreting MCL 500.3107. In Nasser, the plaintiff was involved in a minor accident. Complaining of pain in his "head, neck, chest, shoulder, and both upper and lower back, as well as blurred vision and nausea," he initially sought medical treatment from an internist, who then admitted him to a hospital.34 Over the following three months, he spent 50 days in the hospital and underwent a battery of medical tests.35

The plaintiff's no-fault insurer refused to pay for the plaintiff's hospitalization, and the plaintiff sued to recover allowable expenses under the no-fault act.36 The trial court granted the plaintiff summary disposition on the issue of liability.37 The Court of Appeals affirmed, citing policy considerations to justify allowing the plaintiff to rely on his subjective beliefs that his hospital expenses were "reasonably necessary" when he accepted treatment.38 In doing so, the Courts of Appeals also agreed with the plaintiff that "[t]he reasonableness of medical expenses cannot be used as a defense to liability in a no-fault accident case."39 This Court rejected both the plaintiff's claims, clearly stating that the defendant insurer could challenge the reasonableness of the plaintiff's expenses and impliedly rejecting the plaintiff's reliance on his subjective belief of reasonableness.40

This Court has also held when interpreting insurance contracts that the use of the term "reasonably" requires the application an objective standard unless it is used in reference to a particular person's point of view or expectation under certain circumstances.41 In the companion cases of Allstate Ins. Co. v. Freeman and Metro. Prop. & Liability Ins. Co. v. DiCicco,42 this Court distinguished between language identifying objective and subjective standards in exclusionary insurance clauses. In Freeman, this Court unanimously held that the phrase "reasonably be expected" unambiguously directed the use of an objective standard of expectation.43 In Di-Cicco, a majority of this Court applied a subjective standard to an insurance policy that excluded "bodily injury or property damage which is either expected or intended from the standpoint of the insured."44

More recently, in Allstate Ins. Co. v. McCarn (After Remand), we addressed an insurance policy that excluded coverage for damage that "may reasonably be expected to result" from an insured's intentional or criminal acts.45 Because the contract used the phrase "reasonably expected," six members of this Court agreed that the contract required the application of an objective standard.46

The statutory provision at issue in this case uses the term "reasonably," and there is no statutory language suggesting that "reasonably" should be determined on a subjective basis. Most indicative that an objective standard applies is the absence of language providing for any particular point of view, such as "from the standpoint of the insured" or "by an insured person." Thus, although "reasonably necessary" is a broadly worded phrase, we conclude that this phrase must be assessed by using an objective standard.47

E. AN EXPERIMENTAL SURGICAL PROCEDURE CANNOT BE REASONABLY NECESSARY IF IT IS NOT EFFICACIOUS

Having determined that the term "reasonably necessary" must be assessed from an objective perspective, we next consider what it is that must be reasonably necessary under MCL 500.3107(1)(a): "products, services and accommodations" that are provided "for an injured person's care, recovery, or rehabilitation."48 Thus, a service, product, or accommodation must be (1) objectively reasonable and (2) necessary for an insured's care, recovery, or rehabilitation.49 If, as in this case, the service under consideration is an experimental surgical procedure, the insured must present evidence that the surgery may result in care, recovery, or rehabilitation. In other words, there must be evidence that the surgery is efficacious. Further, because a surgery involves the exercise of medical judgment,50 the efficacy determination must be based on objective and verifiable medical evidence. Experimental surgical procedures lacking objective and verifiable medical evidence of their efficacy cannot be "reasonably necessary" simply because it cannot be shown to effect the insured's care, recovery, or rehabilitation. To interpret MCL 500.3107(1)(a) as allowing reimbursement for nonefficacious experimental treatments "for an injured person's care, recovery, or rehabilitation" would be to read the phrase "reasonably necessary" out of this provision.51

Requiring the minimum threshold of efficacy in the context of experimental surgical procedures is consistent with our precedent regarding nonmedical allowable expenses. In Griffith, for example, we rejected the proposition that insurers were "obligated to pay for any expenses that an injured person would otherwise be provided in an institutional setting as long as they are remotely related to the person's general care."52 Rather, we concluded that coverage "requires that allowable expenses be causally connected to a person's injury."53 We also emphasized that "the statute specifically limits compensation to charges for products or services that are reasonably necessary `for an injured person's care, recovery, or rehabilitation[,]'... suggest[ing] that `care' must be related to the insured's injuries."54 Just as Griffith required that expenses for food actually be related to a person's injury, so also do we require here that expenses for experimental medical treatment actually be for an injured person's care, recovery, or rehabilitation. This requires, at a minimum, that services be efficacious in an injured person's care, recovery, or rehabilitation.

If a surgical procedure is experimental, an insured cannot establish its reasonable necessity under MCL 500.3107 unless expert testimony indicates that the surgery presents a reasonable chance that it will be efficacious in the injured person's care, recovery, or rehabilitation. Contrary to the Court of Appeals' holding in this case, an insured is not required to prove that an experimental surgical procedure gained general acceptance in the medical community before its reasonable necessity becomes a question for consideration by the trier of fact.55 MCL 500.3107(1)(a) does not require that medical treatment be shown to have gained general acceptance within the medical community. Rather, an insured must present objective and verifiable medical evidence that medical treatment is efficacious in an injured person's care, recovery, or rehabilitation.56 If there is objective and verifiable evidence that an experimental surgical procedure is efficacious, the finder of fact can begin to make an informed decision in regard to whether the treatment was reasonably necessary by considering whatever factors were relevant in that case, which may include but are not limited to the severity and chronicity of the condition, the outcomes of any previous treatments, the likelihood that alternative treatments would be efficacious, a personal physician's recommendation in conjunction with the a patient's preference, and both the short-term and long-term risks and benefits.57 Absent objective evidence to establish that the experimental surgical procedure is at least efficacious, there would not exist a material question of fact about whether the medical treatment was reasonably necessary to the care recovery or rehabilitation of an insured.58

We also observe that MRE 702 imposes an obligation on the trial court to ensure that any expert testimony or scientific evidence admitted at all stages of a proceeding is reliable.59 "While the exercise of this gatekeeper role is within a court's discretion, a trial judge may neither `abandon' this obligation nor `perform the function inadequately.'"60 The trial court must specifically ensure that expert testimony is based on sufficient facts or data, the product of reliable principles and methods, and that the witness has applied the principles and methods reliably to the facts of the case.

F. THE PROCEDURE WAS NOT REASONABLY NECESSARY FOR THE CARE, RECOVERY, OR REHABILITATION OF PLAINTIFF

In this case, plaintiff failed to present evidence to establish that the experimental surgical procedure at issue presented him with an objectively verifiable chance that it would be efficacious in his care, recovery, or rehabilitation. Therefore, defendant was entitled to judgment as a matter of law because plaintiff did not meet the minimum threshold for recovery.

Plaintiff relied on the testimony of two expert witnesses, Dr. Hinderer of the RIM and Dr. Lima,61 to establish that the procedure was "reasonably necessary." Dr. Hinderer's testimony cast doubt on whether the procedure was efficacious in plaintiff's care, recovery, or rehabilitation.62 In particular, Dr. Hinderer did not endorse, recommend, or prescribe the procedure to plaintiff.63 Dr. Hinderer testified that the procedure is not regarded as necessary in his field of medicine and that "[i]t's certainly not standard of practice given its experimental nature." More importantly, when asked whether the surgical procedure increased the chances of an injured person's potential for recovery, Dr. Hinderer agreed with defense counsel's statement that "we don't know the outcomes yet because this is such a new procedure."

Further, Dr. Hinderer's testimony actually confirmed that the decision to undergo the procedure was purely subjective. He candidly testified that

Dr. Lima's testimony does not save plaintiff's claim.64 Plaintiff maintains that the procedure afforded him the possibility or opportunity to recover as much function as possible below his injury site. Dr. Lima did indeed claim that the procedure was reasonably necessary if plaintiff wanted the opportunity to recover some function below the injury site. And plaintiff has clearly relied on Dr. Lima's affirmative answer to counsel's question whether, "as a result of the procedure, the possibility exists that [plaintiff] may regain some level of function below the injury site[.]"

This possibility, however, cannot be measured without objective evidence establishing efficacy in the first place. Further, as with the legal standard for establishing causation, the mere possibility of efficacy is not enough, and "when the matter remains one of pure speculation or conjecture, ... it becomes the duty of the court to direct a verdict for the defendant."65 While Dr. Lima articulated his theory, he failed to present medical evidence to support it. Whatever research he may have conducted, it was unsupported by any controlled studies, it had not been subjected to peer review, and the medical evidence had not been not debated in scholarly publications. Dr. Lima did not base his testimony on any verifiable evidence that undergoing the procedure would be efficacious. The record reflects that his testimony, at best, reflects his personal belief, or hope, that many of the patients who undergo the procedure improve. This is clearly established by the fact that Dr. Lima was very "surprised" by the "quite unexpected" results of plaintiff's procedure. In sum, Dr. Lima's testimony also fails to provide an objective basis by which a jury could conclude that the experimental surgical procedure was reasonably necessary for plaintiff's care, recovery, or rehabilitation.66

G. RESPONSE TO THE DISSENT

The dissent's declaration that "[t]oday's decision rewrites [MCL 500.3107] to require that a procedure be `medically necessary' or `medically appropriate' in order for an insured to be reimbursed by his or her insurer"67 is patently false. Nowhere in this opinion, except in response to the dissent, will you find the phrases "medically necessary" or "medically appropriate." After falsely ascribing these standards to us, the dissent uses them to set up the straw-man argument that we are thwarting the will of the people by enacting standards that were rejected when 1993 PA 143 was rejected by referendum. This is also patently false.

1. WE ARE INTERPRETING MCL 500.3107

The dissent claims that our opinion adds language to MCL 500.3107. We obviously disagree with this characterization. We believe that the dissent fails to give meaning to the portion of the provision that states "for an injured person's care, recovery, or rehabilitation" by concluding that evidence of a treatment's efficaciousness is not required to prove that it is reasonably necessary. A treatment or procedure that has not been shown to be efficacious can be neither "reasonable" nor "necessary" under the no-fault act.

Our interpretation of MCL 500.3107 gives meaning to the phrase "reasonably necessary ... for an injured person's care recovery or rehabilitation," and, in doing so, we define the minimum amount of evidence that must be presented on the question before the matter becomes a genuine and material question of fact sufficient to be submitted to a jury for its determination.68 We merely conclude that the reasonably necessary standard cannot be met when there is no evidence that medical treatment will have any beneficial effect on the "injured person's care, recovery, or rehabilitation."

It is a bedrock legal principle that "[i]t is, emphatically, the province and duty of the judicial department, to say what the law is. Those who apply the rule to particular cases, must of necessity expound and interpret that rule."69 This Court stated that "it is necessary ... that the law shall be known and certain, and shall not depend on each jury that tries a cause."70 It is axiomatic that courts decide questions of law and juries apply the law given them to the facts as they have found them. This principle is reflected in our model civil jury instructions, and the trial court instructed the jury consistently with those instructions.71 We do not add language to MCL 500.3107(1)(a), but expound upon the phrase "reasonably necessary ... for an injured person's care recovery or rehabilitation" to provide essential legal guidance.72

We find overwrought the dissent's protestations regarding the so-called "stringent" standard that the dissent claims this opinion articulates. Again, we merely hold that an insured must establish that medical treatment is efficacious in his or her care, recovery, or rehabilitation. We conclude that this standard is entirely consistent with the common meaning of the phrase "reasonably necessary ... for an injured person's care, recovery, or rehabilitation." On the other hand, the dissent's position that a reasonably necessary treatment is any treatment that a person hopes could possibly work falls far short of any commonly accepted meaning of "reasonably necessary ... for an injured person's care, recovery, or rehabilitation." The dissent's standard would allow a nonefficacious treatment—which is worthless—to be considered "reasonably necessary" for the sole reason that an expert witness offered an opinion that the medical treatment is reasonably necessary. For the same reasons that we caution trial courts not to "admit opinion evidence which is connected to existing data only by the ipse dixit of the expert,"73 we believe that what constitutes a reasonably necessary medical treatment cannot solely be based on the ipse dixit of a physician. And while we recognize that our standard, which requires evidence of efficacy, is more imposing than the dissent's standard of let the jury figure it out, it bears repeating that medical treatments, like any other personal protection insurance benefit, must be efficacious to be reasonably necessary. Ignoring this basic principle sanctions the dissent's anything-goes approach, whereby every insured's demand for payment would inexplicably become a question of fact and no-fault benefits would be paid for treatments not even shown to be reliable or effective, let alone reasonable or necessary.

2. 1993 PA 143

The dissent maintains that our interpretation of MCL 500.3107(1)(a) thwarts the will of the voters as expressed in the 1994 general election, in which the voters rejected 1993 PA 143. But, as explained earlier, in order to advance this argument, the dissent must ascribe to us legal standards not adopted in this opinion. Further, we are unfamiliar with a method of statutory interpretation that commences interpretation of an existing statute not by reviewing the words of that statute, but instead by examining the language of one rejected by referendum. Employing this method of interpretation, the dissent maintains that an experimental surgical procedure may never be deemed unreasonable as a matter of law and that a jury must always determine what is reasonable and necessary, regardless of the evidence presented at trial.

1993 PA 143 would indeed have amended MCL 500.3107 to state in subsection (4)(c) that "[e]xpenses within personal protection insurance coverage shall not include experimental treatment or participation in research projects." But it defies logic to presume that because a total bar on experimental treatments was rejected by the voters, the reasonableness and necessity of all experimental treatments must be resolved by a jury. Following the dissent's reasoning, if a medical doctor opined that treatments such as apricot pit theory, ear candling, homeopathy, magnet therapy, and psychic surgery could possibly give an insured a chance to recover, a jury would have to resolve whether those treatments were reasonable and necessary to the care, recovery, or rehabilitation of the insured.74 This could not possibly be the effect of the rejection of 1993 PA 143.

Further, [a]t least one Michigan court has declined to adopt the method of statutory construction adopted here by the dissent. In Michigan Chiropractic Council v. Office of Fin. & Ins. Servs. Comm'r,75 the Court of Appeals was asked to assume that the rejection by ballot referendum of 1993 PA 143 amounted to a rejection by the voters of every single aspect of the act. The Court of Appeals did not accept this argument. The Court of Appeals observed that "1993 PA 143 made comprehensive changes to Michigan's no-fault insurance scheme. Because the referendum rejected the act in its entirety, it has little bearing on the [specific issue presented in] this case."76 We agree with the Chiropractic Council panel that 1993 PA 143 was a comprehensive insurance reform bill, and one can only speculate whether the rejection of 1993 PA 143 signified that voters expected every type of experimental treatment to be covered under the no-fault act. Indeed, the language on the ballot proposal did not even mention experimental treatment.77 The dissent engages in guesswork and, on this basis, believes that Michigan voters intended its courts to rubber-stamp all determinations under the no-fault act, regardless of the paucity of evidence supporting a jury's finding and regardless of how unreasonable and unnecessary the expense may be. We base our opinion on the current law. We do not base it on what the law once was or could have been.

This case does not turn on any aspect of 1993 PA 143. As already stated, we reject defendant's claim that plaintiff cannot prevail merely because the procedure was experimental. We also reject the Court of Appeals' holding that an insured is required to prove that an experimental surgical procedure has gained general acceptance in the medical community before consideration by the trier of fact. The question whether an experimental treatment is reasonably necessary for an insured's care, recovery, or rehabilitation must be resolved by a fact-finder if the insured can present objective and verifiable medical evidence to support the conclusion that the treatment is efficacious. The dissent maintains that this is an illusory standard because, if objective and verifiable medical evidence of efficacy exists, "it is unclear how the procedure would still be termed `experimental' or in the `research' phase."78 But the practices of the FDA establish that an efficacious treatment may nonetheless be experimental. The FDA has three phases of testing before a medical procedure or product receives full FDA approval. Phases II and III of the FDA process, in which the treatment is still experimental or in the research phase, focus on efficacy. This is but one example. And contrary to the dissent's representation, we do not hold that the objective and verifiable medical evidence must include "controlled studies subject to peer review or scholarly publications" supporting the science behind the surgery.79 Likewise, we are in no way suggesting that data from the FDA are required. Rather, these are additional examples of objective and verifiable evidence that can establish efficacy. Presentation by an expert witness of any of this objective and verifiable evidence, even if opposed by several witnesses claiming the proposed medical treatment is not efficacious, is sufficient to sustain plaintiff's burden. Since we are not medical experts, we are not going to artificially limit the types of objective and verifiable evidence that a party may present to support its claim; instead, we simply note that there must be some evidence from the medical community that a particular procedure would have some beneficial effect on a person's "care, recovery, or rehabilitation" in accordance with MCL 500.3107(1)(a).

IV. CONCLUSION

We conclude that the question whether a product, service or accommodation is reasonably necessary for an injured person's care, recovery, or rehabilitation must be determined under an objective standard. We further conclude that when medical treatment is experimental, an insured seeking reimbursement for this treatment must present objective and verifiable medical evidence establishing that the treatment is efficacious. A treatment or procedure that has not been shown to be efficacious cannot be reasonable or necessary under the no-fault act. An insured's subjective belief that medical treatment is efficacious, reasonable, and necessary is not sufficient to create a question of fact. Viewed in the light most favorable to plaintiff, the objective and verifiable medical evidence presented at trial failed to establish that the experimental surgical procedure at issue in this case was any way efficacious in the care, recovery, or rehabilitation of plaintiff's injury.80 Accordingly, we affirm the judgment of the Court of Appeals.

ROBERT P. YOUNG, STEPHEN J. MARKMAN, MARY BETH KELLY, JJ., concurs.

HATHAWAY, J. (dissenting).

This case addresses whether a medical procedure performed on plaintiff to treat his severe accident-related spinal-cord injuries was "reasonably necessary" under MCL 500.3107(1)(a) of the no-fault act, MCL 500.3101 et seq. The majority holds that the procedure was not "reasonably necessary" and, in doing so, adds language to the no-fault act that was rejected by ballot referendum in 1994. The majority reaches its result by erroneously removing the determination of which expenses are "reasonably necessary" from the jury. Additionally, the majority's new judicially crafted definition of "reasonably necessary" elevates the standard for proving that treatment is "reasonably necessary" to one that is more stringent than MCL 500.3107(1)(a) requires. I respectfully dissent because today's decision erroneously changes the mandates of the no-fault act and replaces them with standards that are inconsistent with the language and history of that act. I would apply the statute as written and uphold the jury's finding that the procedure performed on plaintiff was "reasonably necessary." Therefore, I would reverse the Court of Appeals' judgment and hold that plaintiff is entitled to reimbursement of the costs associated with the procedure.

This case involves plaintiff's request that his no-fault insurer, defendant Home-Owners Insurance Company, reimburse him for the expenses surrounding an experimental procedure that he underwent in Portugal. The procedure was performed to treat the serious spinal-cord injuries plaintiff had sustained in a motorcycle accident. The accident left plaintiff a paraplegic with no sensation in or control of his lower body, leaving him confined to a wheelchair and in need of assistance in urinating and defecating. His condition showed no improvement during the four years between his accident and the procedure. According to plaintiff's expert witness, plaintiff's condition improved following the procedure. However, defendant refused to pay for the procedure, arguing that it was not "reasonably necessary" for plaintiff's "care, recovery, or rehabilitation"1 because it was experimental in nature.

During his jury trial, plaintiff presented testimony from Dr. Carlos Lima, a neurologist on the surgical team that performed the procedure. Dr. Lima testified that the procedure involved harvesting tissue containing stem cells from plaintiff's own sinus cavities and transplanting the tissue into the injured area of the spinal cord. Dr. Lima testified that this procedure fosters growth of new cells in the injured spinal cord, while avoiding the ethical and technical issues surrounding the use of embryonic stem cells. Although the procedure had not been presented for approval by the federal Food and Drug Administration (FDA), Dr. Lima testified that it was conducted within the standards of the European Commission's guidelines regarding clinical procedures. The procedure was performed in a governmental hospital in Lisbon, Portugal, after the presiding physician had obtained approval from the hospital board. Dr. Lima testified that of the 110 patients who had undergone the treatment in his program, a majority of the patients showed improvement. Dr. Lima's testimony describing the success of the procedure included the following:

Plaintiff's treating doctor in the United States, Dr. Steven Hinderer, also testified concerning the reasonableness of the procedure, and responded to questioning as follows:

The jury found that the expenses related to the surgery in Portugal were reasonable charges for reasonably necessary products, services, and accommodations for plaintiff's care, recovery, and rehabilitation under the no-fault act. The Court of Appeals, however, reversed the jury's finding and ordered the trial court to enter a judgment in defendant's favor.2 Plaintiff now appeals that decision.

The issue before this Court is whether an experimental medical procedure can be "reasonably necessary" for an injured person's care, recovery, or rehabilitation.3 Deciding this issue requires application of MCL 500.3107(1)(a).

When interpreting a statute, we follow the established rules of statutory construction. The purpose of statutory construction is to discern and give effect to the intent of the Legislature.4 In doing so, we first look to the actual language of the statute.5 If a statute is clear and unambiguous, it must be enforced as written and no further judicial construction is allowed.6 Simply stated, we must avoid a construction that would render any part of the statute nugatory,7 and similarly, we are "not free to add language to a statute or to interpret a statute on the basis of this Court's own sense of how the statute should have been written."8 Further, a statute must be read as a whole,9 and while individual words and phrases are important, the words and phrases should be read in the context of the entire legislative scheme.10 And "when courts interpret the no-fault act in particular, they are to remember that the act is remedial in nature and must be liberally construed in favor of the persons intended to benefit from it."11

The statute at issue, MCL 500.3107(1), provides in pertinent part:

The majority holds that in order for an expense related to an experimental surgical procedure to be "reasonably necessary," a court must first determine as a matter of law that there is "objective and verifiable medical evidence establishing that [the experimental surgical procedure] is efficacious."12 Further, the majority holds that plaintiff did not meet the "objective and verifiable medical evidence" standard because Dr. Lima's research "was unsupported by any controlled studies, it was not subject to peer review, and the medical evidence was not debated in scholarly publications."13 Thus, the majority's new standards add language to the statute that is simply not there.

In this case, there was testimony from two doctors who assessed plaintiff's condition before the procedure was performed. Dr. Lima testified that it would be necessary for plaintiff to undergo the procedure in order to have a chance at recovery. Dr. Hinderer did state that he was not able to recommend the procedure to plaintiff because the procedure was not an authorized procedure in the United States, but he also testified that the procedure was a reasonable form of treatment for plaintiff. The majority characterizes Dr. Hinderer's testimony as casting doubt on the efficacy of the procedure because "Dr. Hinderer did not endorse, recommend, or prescribe the procedure to plaintiff."14 This characterization is erroneous.

The majority dismisses the fact that the facility where plaintiff was treated by Dr. Hinderer, the Rehabilitation Institute of Michigan, has a professional relationship with Dr. Lima's program in Portugal pursuant to which the Rehabilitation Institute screens patients to determine whether they meet the criteria to be eligible for the procedure. Dr. Lima's program has performed the procedure on 110 patients from around the world. According to Dr. Lima, the Rehabilitation Institute has screened nearly 60 patients for the procedure. Of the 60 patients from the Rehabilitation Institute, 40 were Dr. Hinderer's patients. Thus, I disagree with the majority's assertion that Dr. Hinderer cast doubt on the efficacy of the procedure. More than a third of the patients in the worldwide program were patients of Dr. Hinderer, which, when viewed in a light most favorable to the plaintiff,15 suggests that Dr. Hinderer does not doubt the effectiveness of the procedure. However, even if doubt was cast by one of the two assessing physicians, an issue of fact still existed for the jury to resolve under Owens v. Auto Club Ins. Ass'n, 444 Mich. 314, 326, 506 N.W.2d 850 (1993).16 Today's decision erroneously holds that the jury should not have decided this genuine issue of material fact.

The majority holds that the jury incorrectly concluded that the procedure was "reasonably necessary." In reaching this result, the majority disregards much of the actual testimony presented.17 For instance, Dr. Lima testified that without the procedure, "there's no possibility for [plaintiff] to have any recovery with such [an injury] which is not just functionally complete, but it was anatomically very destructive and complete also." Dr. Lima's statement shows that a doctor who assessed plaintiff's condition found that there was no possibility of recovery before the procedure.

The majority argues that Dr. Lima's testimony suggests merely "the possibility or opportunity to recover" and that a "possibility. . . cannot be measured without objective evidence establishing efficacy in the first place."18 This argument is at odds with the actual language of MCL 500.3107(1)(a) because it contains standards not found in the language of the statute. The statute only requires a procedure to be "reasonably necessary" to qualify as an allowable expense. Therefore, the analysis should be limited to whether a procedure was "reasonably necessary" under the commonly understood meaning of those words.19 The statute does not contain any language limiting the basis of a "reasonably necessary" determination to objective and verifiable medical data, as is required by today's decision.

The majority also errs because it misconstrues the meaning of the term "reasonably necessary." Without any statutory support, it interprets the word "reasonably" to mean "objective and verifiable." The majority then declares that the term "necessary" creates a strict standard requiring "evidence" of "efficacy."20 The evidence that the majority refers to can only be satisfied with thorough evidence from the "medical community."21 However, the statute before us does not contain the terms "objective," "verifiable," "evidence," "efficacy," or "medical community."

Under the majority's own stated principle, words cannot be read into this statute. In order to provide support for the majority's new standard, the statute would have to contain, at a minimum, language indicating that expenses are only allowable if they relate to procedures that are "proven to be efficacious by the medical community or the FDA." However, such language is not in the no-fault act. For all practical purposes, this is a "medically necessary" or "medically appropriate" standard, despite the majority's statements to the contrary. Thus, today's decision rewrites the statute to require that a procedure be "medically necessary" or "medically appropriate" in order for an insured to be reimbursed by his or her insurer.

In response to this criticism, the majority proclaims that nowhere in its opinion does it use the phrases "medically necessary" or "medically appropriate," except in its response to this dissent. But the majority need not invoke those magic words for it to be obvious to all that this is precisely what the majority's new standard requires. A standard that requires the presentation of objective and verifiable medical evidence establishing that a treatment is generally efficacious, based on controlled studies subject to peer review or scholarly publications, is a "medically necessary" standard. The majority's statements to the contrary do not change the practical reality of its new standard.

In reviewing the actual language of the statute, it is clear that the determination of whether a procedure is "reasonably necessary" involves analyzing whether the decision to undergo the procedure was within reason, in light of the testimony that plaintiff would not recover if he did nothing.22 Moreover, it must not be forgotten that a jury of plaintiff's peers found that the procedure was "reasonably necessary" for plaintiff's "care, recovery, and rehabilitation." By making this broad decision today, the majority has turned a procedure that was found to be "reasonably necessary" for plaintiff's "care, recovery, or rehabilitation" into an unreasonable choice. In this case, the majority effectively asserts that it was unreasonable as a matter of law for this plaintiff to have pursued the only procedure that could possibly prevent him from being a paraplegic for the rest of his life.

Further, the majority states that "[t]he ultimate question whether the surgical procedure at issue here is a covered expense under the no-fault act does not turn on its status as experimental."23 However, despite this statement, experimental procedures and participation in research projects are effectively excluded from coverage as a matter of law under the majority's new standard. The majority's standard requires objective and verifiable medical evidence proving a procedure's efficacy, but if such data were to exist, it is unclear how the procedure would still be termed "experimental" or in the "research" phase.

Despite the majority's protestations to the contrary, its decision today also abandons well-established precedent. In Nasser v. Auto Club Ins. Ass'n, 435 Mich. 33, 54, 457 N.W.2d 637 (1990), this Court stated that "the question of whether expenses are reasonable and reasonably necessary is generally one of fact for the jury" and that summary disposition should only be granted when the reasonableness and necessity of a procedure can be determined with "certainty" when the evidence is viewed in the light most favorable to the nonmoving party.24 In Owens, 444 Mich. at 326, 506 N.W.2d 850, this Court stated that the presentation of competing professional opinions from doctors who assessed the plaintiff was enough to create a question of fact regarding whether the procedure was "reasonably necessary." Rather than following existing precedent and holding that the determination of whether a procedure is "reasonably necessary" is one for the jury, the majority transforms this question into a question of law.

Today's decision is particularly troubling given that, as recently as December 2010, a majority of this Court clarified a Court of Appeals remand order that had stated "`[w]hether a cost constitutes an allowable expense is a question of law and so it is to be determined by the court, not the jury.'" Wilcox v. State Farm Mut. Auto. Ins. Co., 488 Mich. 1011, 791 N.W.2d 723 (2010). This Court's clarifying order instructed that

In light of this Court's recent decision in Wilcox, it is unclear why it is suddenly necessary to change the way that "reasonably necessary" is decided. How has this Court's precedent become so unclear in such a short time? Why must this Court now effectively reverse the instructions in Wilcox and disregard Nasser?

Finally, perhaps the most significant evidence that the majority errs is that the Legislature enacted a bill inserting language similar to that which the majority adds to the statute today, and the voters of this state rejected it by referendum. In 1993 PA 143, the Legislature amended the no-fault act, creating a standard to determine allowable expenses similar to the standard that the majority has adopted today.25 MCL 500.3107(1), as amended by 1993 PA 143, stated:

Additionally, MCL 500.3107(4), as added by 1993 PA 143, stated in pertinent part:

As used in this section:

In November 1994, Proposal C asked the voters of this state to consider whether the amended requirements imposed by 1993 PA 143 embodied what the law of this state ought to be. In the referendum, Michigan voters overwhelmingly answered "No."26 Thus, the citizens of Michigan expressly rejected a "medically necessary or medically appropriate" standard, a requirement that disputes be decided by courts as a question of law, and, most significantly, a prohibition against coverage for "experimental treatment or participation in research projects."27

Despite the voters' rejections of these three elements, today's decision inserts them into the no-fault act. The majority argues that, because 1993 PA 143 attempted to broadly reform the no-fault system with numerous changes to MCL 500.3107, it is somehow unclear whether the voters actually rejected the specific reforms that the majority judicially enacts today.28 This reasoning is misguided and illogical.

First, it is improper for this Court to insert elements of a rejected law into a statute because such action amounts to judicial engineering of a statute. The voters spoke on 1993 PA 143 in Proposal C, the Legislature has not chosen to subsequently add these three rejected elements into the no-fault act, and it is wrong for the majority to do so today. Most importantly, Const. 1963, art. 2, § 9 states that "[n]o law as to which the power of referendum properly has been invoked shall be effective thereafter unless approved by a majority of the electors voting thereon at the next general election."

Second, it is disingenuous to argue that there is no way to determine which specific element of the law the voters rejected. The voters rejected the entire law. Plain and simple, the voters said "No." Thus, it borders on nonsensical for this Court to argue that the voters only disagreed with specific elements of the act and that we do not know which elements. A referendum vote, such as that taken on Proposal C, is an all-or-nothing vote, and, with respect to what voters wanted added to the no-fault act, the voters chose nothing.29

Moreover, the majority fails to recognize the unique importance of referenda. As Justice RILEY stated in In re Executive Message from the Governor, 444 Mich. 1214, 514 N.W.2d 465 (1994):

Thus, the majority's decision today is in direct conflict with the will of the voters of this state.

The pertinent part of the statute only uses the phrase "reasonably necessary" and specifies that the procedure must be for the "injured person's care, recovery, or rehabilitation." As noted earlier, if there is any factual dispute about whether a treatment is "reasonably necessary," that dispute must properly be decided by a jury. Rather than focusing on one factor, such as objective and verifiable medical evidence establishing the efficacy of the procedure, a determination by the jury could include an analysis of any number of factors. Such factors could include medical professionals' conclusions regarding the reasonable necessity of a procedure, lay persons' conclusions regarding the reasonable necessity of a procedure, scientific support for the effectiveness of a procedure, or possibly even the subjective belief of the plaintiff. The point, however, is that it is up to the jury, on a case-by-case basis, to decide what is reasonable or unreasonable. Michigan's Constitution affords parties "[t]he right of trial by jury. . . ."30 This Court should not disregard the important fact-finding role of the jury. This Court must respect the no-fault act as it is currently written.

CONCLUSION

Today's decision rewrites the requirements for an insurer to pay allowable expenses under MCL 500.3107 of Michigan's no-fault act. The majority holds that the procedure in this case was not "reasonably necessary" and, in doing so, adds language to the no-fault act that was rejected by referendum in 1994. The majority reaches its result by erroneously removing the determination of which expenses are "reasonably necessary" from the jury. Additionally, the majority's new judicially crafted definition of "reasonably necessary" elevates the standard for proving that treatment is "reasonably necessary" to one that is more stringent than MCL 500.3107(1)(a) requires. I would apply the no-fault act as written, I would uphold the jury's finding in this case that the procedure performed on plaintiff was "reasonably necessary," and I would hold that plaintiff is entitled to reimbursement of the costs associated with the procedure. Accordingly, I dissent.

MICHAEL F. CAVANAGH (except footnote 20), MARILYN KELLY, JJ., agrees.