OPINION OF THE COURT
Progress begets promise and problems. Computer technology has ushered in the "information age” and helped to create a global workplace that is accessible from one’s home or office. These cases present us with a difficult question arising from the widespread use of computers in the workplace: when does a cause of action accrue against a keyboard manufacturer for repetitive stress injury (RSI) suffered by a keyboard user? We conclude that, in such cases, the cause of action accrues upon the onset of symptoms, or the last use of the keyboard, whichever is earlier.
I.
This appeal involves over 90 separate plaintiffs suing various keyboard manufacturers in a number of separate lawsuits. By order entered January 22, 1993, Administrative Judge Stanley Ostrau, in apparent recognition of the large number of RSI cases which were making their way through Supreme Court, New York County, assigned all cases involving RSIs to Justice Stephen Crane for pretrial purposes. Soon thereafter, Justice Crane established procedures for the joint briefing of various *765legal issues. Defendants’ motions raising Statute of Limitations issues were treated as dismissal motions made pursuant to CPLR 3211 (a) (5).
Most of the pleadings in the cases before us on this appeal contain common allegations with respect to the onset and manifestation of plaintiffs’ RSIs. Each plaintiff outlines his or her history of keyboard use, including, to the extent known, the particular keyboards used. The pleadings then go on to state that plaintiffs’ injuries were "insidious in their onset,” such that "it is not possible to identify [either] the precise date of the onset of symptoms,” or to say "that any initial symptoms experienced constituted the full manifestation or even partial manifestation” of the injury. Nevertheless, the plaintiffs do identify dates upon which they began experiencing some symptoms, such as numbness, tingling, pain and/or sensory motor impairments of the upper extremities, neck and torso. Plaintiffs further state the dates upon which they were diagnosed with various RSIs. Finally, the pleadings generally allege that the nature of each plaintiff’s injury is "such that there is no precise moment of injury,” but rather a "cumulative and prolonged process by which [each] plaintiff sustained injury [and] aggravated [an] existing injury.”
RSI is one of several essentially synonymous terms, all of which "connote injury to the musculo-skeletal tissues from repeated motions and exertions” (Russ, Freeman and McQuade, 5 Attorney’s Medical Advisor § 66:4, at 66-8). RSIs can be caused by activities as divergent as playing video games or working a jackhammer. Carpal tunnel syndrome, probably the most prevalent keyboard-related injury, is a subcategory of RSI involving compression of the median nerve as it passes through the wrist between the flexor tendons and the transverse carpal tunnel ligament (the area known as the carpal tunnel) (Ausman and Snyder, 3 Medical Library Lawyer’s Edition § 4:18, at 53 [1989 ed]). While carpal tunnel syndrome can have a number of causes, ranging from arthritis or benign tumors to blunt trauma, the condition is seen with increasing frequency as a result of workplace use of keyboards. Indeed, according to various surveys, the occurrence of RSIs among workers rose roughly 1,000% between 1982 and 1991, to the point that today they account for 61% of all workplace illnesses (see, Comment, Cumulative Trauma Disorders: A Hidden Downside to Technological Advancement, 11 J Contemp Health L & Pol’y 479; Juge, Stokes and Pine, Cumulative Trauma Disorders — "The Disease of the 90’s”: An Interdisciplinary Analysis, 55 La L Rev 895).
*766In deciding what accrual rule to apply to these claims, the trial court felt compelled to follow Wallen v American Tel. & Tel. Co. (Sup Ct, Bronx County, Sept. 17, 1992, Saks, J., index No. 12336/91, affd for reasons stated below 195 AD2d 417, Iv denied 82 NY2d 659). Based upon Wallen, the court ruled that accrual in RSI cases is measured from the onset of the plaintiff’s symptoms without requiring a diagnosis of their cause. The Appellate Division disagreed. Finding this Court’s exposure line of cases controlling, the Appellate Division held that a cause of action for RSI accrues upon first use of a keyboard. The Court recognized that there are some differences between the two types of claims, but held that RSIs in general "are not distinguishable from the cases where repeated, prolonged exposure to, e.g., asbestos, is necessary before the damages will develop and manifest themselves” (Blanco v American Tel. & Tel. Co., 223 AD2d 156, 164).
While the Appellate Division found our toxic torts cases to be controlling, it also held that CPLR 214-c (providing a discovery rule in toxic torts cases) was inapplicable. The Court recognized the harshness of the rule it had pronounced, but held that the remedy, if there was to be one, lay with the Legislature rather than the courts.
II.
At the outset, we agree with the Appellate Division that CPLR 214-c is inapplicable in this case. CPLR 214-c was enacted in 1986 as part of a larger "tort reform” package (L 1986, ch 682), and provides that "the three year period within which an action to recover damages for personal injury or injury to property caused by the latent effects of exposure to any substance or combination of substances * * * must be commenced shall be computed from the date of discovery of the injury by the plaintiff or from the date when through the exercise of reasonable diligence such injury should have been discovered by the plaintiff, whichever is earlier.” (CPLR 214-c [2].) CPLR 214-c was enacted to abrogate the exposure rule which this Court had formulated and adhered to in a line of cases stretching from Schmidt v Merchants Desp. Transp. Co. (270 NY 287) to Consorti v Owens-Corning Fiberglas Corp. (86 NY2d 449).
While CPLR 214-c is a remedial statute and as such, should be interpreted broadly, this maxim does not allow us to stretch the statute beyond its intended coverage. "[E]ven a remedial statute must be given a meaning consistent with the words *767chosen by the Legislature” (Enright v Lilly & Co., 77 NY2d 377, 385, n 1). Plaintiffs argue that these cases fall within the plain language of CPLR 214-c, since they allege that their injuries occurred from contact with a substance — to wit, a keyboard. While CPLR 214-c does nominally cover situations where a plaintiff is injured due to "contact” with a "substance,” it is plain from reading the statute as a whole that the types of substances intended to be covered are toxic substances. To the extent that confirmation of this point is necessary, it can be found in abundance in the legislative history. Virtually every memorandum in support of the bill, including that of the Senate sponsor, the Governor and the Attorney-General, refers to the law enacting section 214-c as the "toxic torts” bill. Obviously, a keyboard is not a toxic substance and therefore, CPLR 214-c is inapplicable to plaintiffs’ claims. Thus, it is CPLR 214, rather than section 214-c, that controls the disposition of this case.
We also reject plaintiffs’ contention that the continuing duty to warn doctrine may be applied to extend the limitations period. The gravamen of plaintiffs’ claims is that defendants’ keyboards were defectively designed. In the analogous case of liability predicated upon a negligent act, this Court in Schwartz v Heyden Newport Chem. Corp. (12 NY2d 212) said that "the bar of the Statute of Limitations cannot be avoided through an allegation that defendant had a continuing duty to warn plaintiff of the possible damaging consequences resulting from the act alleged to be negligent” (id., at 216). As a rule, the date of injury is the benchmark for determining the accrual of a cause of action, and "the mere continuation of [the duty to warn] into the limitations period is not enough to resurrect a cause of action premised upon an injury occurring earlier” (Coughlin v International Bus. Machs. Corp., 225 AD2d 256, 260).
III.
CPLR 214 provides that an action for damages for personal injuries "must be commenced within three years.” A cause of action accrues for purposes of CPLR 214 "when all of the facts necessary to the cause of action have occurred so that the party would be entitled to obtain relief in court” (Aetna Life & Cas. Co. v Nelson, 67 NY2d 169, 175). Prior to the enactment of CPLR 214-c, this Court had held that, in the case of latent injuries arising out of exposure to toxic substances, the cause of action accrued upon plaintiffs initial exposure (see, Consorti *768v Owens-Corning Fiberglas Corp., supra; Schwartz v Heyden Newport Chem. Corp., supra; Schmidt v Merchants Desp. Transp. Co., supra). Both parties assume that this case is analogous to our exposure rule cases. We agree that there are similarities between the two types of cases, but there are significant differences as well.
In Martin v Edwards Labs. (60 NY2d 417), this Court identified three separate categories of latent injury cases. The first, typified by Schmidt v Merchants Desp. Transp. Co. (supra) and Schwartz v Heyden Newport Chem. Corp. (supra), were characterized as injuries arising from the inhalation, ingestion or injection of harmful substances which are assimilated into the body. As indicated above, in toxic tort exposure cases, we had long applied a first exposure rule, under which the time to sue for resultant injuries ran from plaintiff’s first exposure to the substance, rather than from the discovery of the injury. This rule will be discussed in greater detail later.
The second line of cases is typified by Flanagan v Mount Eden Gen. Hosp. (24 NY2d 427). Flanagan involved a medical malpractice claim against a doctor for leaving surgical clamps in the plaintiff’s abdomen following gall bladder surgery. The surgery was performed in 1958, and it was not until the spring of 1966 that the plaintiff actually experienced problems resulting from the unremoved clamps. She consulted a doctor at that time, and in June of 1966, X rays revealed the clamps. After surgery to remove the clamps, plaintiff commenced an action in October 1966. This Court applied a discovery rule and held that the claim was timely.
Martin characterized the type of claim at issue in Flanagan as one based upon objects being implanted, but not assimilated, into the body. We reasoned that a discovery rule was justified in such cases based upon a balancing of typical Statute of Limitations policies. Specifically, we noted that there was little danger of feigned claims since the objects retained their identity within the body; that for the same reason the defendant was not unduly handicapped in defending such claims; that there was little or no possible causal break between the negligence and the injury; and that the claim did not rest to a significant degree on questions of credibility or professional diagnostic judgment (Martin v Edwards Labs., supra, 60 NY2d, at 425).
The third line of cases identified in Martin is typified by Victorson v Bock Laundry Mach. Co. (37 NY2d 395). Victorson *769involved three product liability claims against manufacturers of defective products by remote users who were injured when the product defect manifested itself several years after manufacture. Because plaintiff suffered no actual injury until the defect manifested itself, we held that the cause of action did not accrue until the defect actually caused injury.
Martin itself involved claims against the manufacturer of a defective heart valve, and a defective Daikon Shield. Even though these products were intended to be inserted into the body, and thus would seem to have fit into the second line of cases, we held that the third line of cases was controlling (Martin v Edwards Labs., supra, 60 NY2d, at 428). Thus, the claims accrued upon causation of injury, which we noted in most cases would be correlated with the date the product malfunctioned. In the case of an artificial heart valve, the injury was the release of teflon particles into the blood stream; in the case of an IUD, the injury was the unintended introduction of infectious bacteria into the uterus.
Our conclusion that the implantation cases in Martin fell within the third category was reached after a careful evaluation of the various policy considerations underlying Statutes of Limitation. The Court noted that a discovery rule was not necessary in implantation cases, since "the implantation or insertion is with the recipient’s knowledge and consent, knowledge which he or she can pass on to a physician seeking to diagnose the cause of later developing bodily problems” (Martin v Edwards Labs., supra, 60 NY2d, at 427). The Martin Court also reasoned that an exposure rule was inappropriate, because, unlike the case where a toxic substance is ingested, and "the forces of harm are inexorably set in motion when the substance enters and is assimilated into the body,” implants are intended to perform a continuing function, and "cause[ ] no injury until the product malfunctions” (id., at 427). Up until that time, the plaintiff has no reason to complain.
Martin also noted that the age of the claim would not present a great problem, because unlike assimilated substances, "[i]f through malfunction the product is thought to have caused harm, it can in most cases be removed and examined to ascertain whether in fact it malfunctioned and, if so, whether that was the cause of the harm” (Martin v Edwards Labs., supra, 60 NY2d, at 427). Thus, credibility determinations would not play a major role in defining when the cause of action accrued. Because the product could be made available upon malfunction, there was not a great danger of feigned or frivo*770lous claims. Finally, the Court noted that, while "complicated medical questions may be involved and professional diagnostic judgment implicated,” these would concern primarily recent, rather than remote, events, and thus there was no significant staleness problem (id., at 427).
A computer keyboard is not a toxic substance which is ingested into the body, nor is it an object implanted, but not assimilated, into the body. However, unlike a typical case involving products remaining outside of the body, there is no one readily discernible first date of injury in RSI cases. Thus, these claims would not appéar to fit neatly within any of the three categories identified in Martin. It follows that this case requires us to strike a new balance with respect to the competing policy interests at stake in order to correctly determine the accrual question. However, given the analogies which may be drawn between this type of case and an exposure case, as identified by both the parties and the Court below, any principled balancing necessarily must take into consideration our holdings in the exposure cases.
IV.
The "first exposure” rule has its roots in Schmidt v Merchants Desp. Transp. Co. (270 NY 287, supra). The plaintiff in Schmidt sued his employer when he developed pneumoconiosis as a result of exposure to toxic silicone dust in the workplace. This Court pointed out that " 'in actions of negligence damage is of the very gist and essence of the plaintiff’s cause’ ” (Schmidt v Merchants Desp. Transp. Co., supra, 270 NY, at 300, quoting Comstock v Wilson, 257 NY 231, 235). However, in answering the question of what constitutes injury, the Court pointed not to the physical manifestation of a disease such that it became identifiable to plaintiff, but instead held that "[t]he injury occurs when there is a wrongful invasion of personal or property rights * * * [w]hen substantial damage may result from any wrong affecting the person or property of another, a cause of action for such wrong immediately accrues.” (Schmidt v Merchants Desp. Transp. Co., supra, 270 NY, at 300-301.) Applying this principle, the Court held that "[t]he injury to the plaintiff was complete when the alleged negligence of the defendant caused the plaintiff to inhale the deleterious dust * * * [t]he disease of the lungs was a consequence of that injury” (id., at 301).
In Schwartz v Heyden Newport Chem. Corp. (12 NY2d 212, supra), a. case arising from a plaintiff’s exposure to a substance *771used to make the sinus cavity perceptible to X rays, this Court applied Schmidt and held that the action accrued upon the plaintiffs exposure to the toxic substance. We stated that "the action accrues only when there is some actual deterioration of a plaintiffs bodily structure,” but went on to hold that "we must assume that the dust immediately acted upon Schmidt’s lung tissue,” thus producing injury (id., at 217). Schwartz also acknowledged certain practical policy justifications for the first exposure rule. Specifically, we noted that, notwithstanding the "insidious and 'inherently unknowable’ nature of cancer * * * a potential defendant’s equities are the same whether the plaintiff knows of his condition or not. Repose is as beneficial to society in the one case as in the other” (Schwartz v Heyden Newport Chem. Corp., supra, 12 NY2d, at 218). Schwartz thus recognized that knowledge of injury was not solely determinative in defining the moment a cause of action for exposure to a toxic substance accrues. The seeds planted in Schwartz came to full fruition in Consorti v Owens-Corning Fiberglas Corp. (supra, 86 NY2d, at 451-452), in which we recognized that the original premise of the first exposure rule was tenuous, but held that continued adherence to the rule was justified based upon various policy considerations.
Admittedly, there are similarities between toxic tort and repetitive stress injuries which would militate in favor of applying the same accrual rule in both types of cases. In each case, the plaintiffs appreciable injury does not manifest itself until some time after (and in many cases significantly after) initial exposure to the product. Thus, in both types of cases there is a gap between the manufacturer’s breach of duty and the resulting ultimate injury to the plaintiff. As we recognized in Consorti v Owens-Corning Fiberglas Corp. (supra), this injects a substantial degree of uncertainty into a manufacturer’s risk assessment calculation, and may give rise to significant problems of proof. Thus, in Consorti, we made clear that our rejection of a fact-based date of manifestation of injury test
"was made for practical and policy reasons articulated in the developing case law, that is, the need to provide manufacturers, employers and other economic actors who are potential defendants with a degree of certainty or predictability in assessing the risk of liability and to avoid stale claims which often turn on questions of credibility or disputed medical judgments. Therefore, a bright line, readily *772verifiable rule was adopted in which, as a matter of law, the tortious injury is deemed to have occurred upon the introduction of the toxic substance into the body.” (Id., at 451-452.)
These same policy considerations would appear to point to the adoption of a first exposure rule in this case. Like a toxic tort, RSI may not develop until years after a person first uses a keyboard; indeed, according to plaintiffs’ expert, the condition may not develop at all, or not until after one keyboard is discarded and a person subsequently uses different keyboards.
This Court has refused to waver from the first exposure rule where the policy considerations articulated in Consorti have dictated its application. "While recognizing that the rule may seem unjust, we have generally left ameliorative efforts to the Legislature. Indeed, where the Legislature has deemed it necessary, it has abrogated the exposure rule (see, e.g., CPLR 214-b [agent orange exposure]; CPLR 214-c [toxic torts]). Thus, defendants argue, given that the Legislature has, in fact, modified our accrual jurisprudence where it has deemed fit, this Court should leave the exposure rule in place, and allow the Legislature to modify it further if it so chooses. We reject this argument as applied in this case.
The exposure rule is premised upon the proposition that, while injury is necessary for the accrual of a cause of action, injury in a toxic tort case accrues upon exposure to the toxic substance, because it is at that point that there has been "a wrongful invasion of personal or property rights” (Schmidt v Merchants Desp. Transp. Co., supra, 270 NY, at 300). This was based upon the assumption that a toxic substance acts immediately upon the body to produce injury (see, Schwartz v Heyden Newport Chem. Corp., 12 NY2d 212, supra; cf., Consorti v Owens-Corning Fiberglas Corp., supra, 86 NY2d, at 451-452). A keyboard is obviously not an inherently toxic or dangerous substance, and exposure to a keyboard is not a "wrongful invasion” in the same sense as is exposure to a toxic substance. Indeed, if we were to apply a first exposure rule literally, then we would have to justify it on the legal fiction that an injury occurs the first moment a person presses a button on a keyboard. It would be absurd to argue that a person who simply touches a keyboard sustains the type of trauma that will eventually lead to RSI. As plaintiffs’ doctor noted in his affidavit, this type of injury has "no single initiating event, but rather an accumulation of events.” Thus, because a keyboard is not an inherently toxic or dangerous substance, and a person *773who initially touches one does not necessarily sustain an injury, the justifications that gave rise to the exposure rule do not apply here.
Defendants are correct that we have long given deference to the Legislature to recast the accrual methodology of toxic torts if it saw fit and as noted earlier, the Legislature has on several occasions done so. However, in cases such as this, where we are presented with new categories of tort claims and injuries, we have always done a careful and balanced analysis of the nature of the claim and its intricate interplay with the policy considerations at the heart of our Statute of Limitations jurisprudence (Martin v Edwards Labs., supra, 60 NY2d, at 427-428). Our analysis here is not a rejection of our long line of cases from Schmidt to Consorti; it is a recognition that RSI cases present new and different challenges in defining the accrual dates for plaintiffs’ causes of action. To accept defendants’ argument that we should rely upon the analysis we found appropriate in toxic tort cases is to ignore the significantly different nature of both plaintiffs’ injuries and the alleged causes. With this in mind, we turn to the task of evaluating the policy interests at stake in RSI cases.
V.
Statutes of Limitation were "designed to promote justice by preventing surprises through the revival of claims that have been allowed to slumber until evidence has been lost, memories have faded, and witnesses have disappeared” (Telegraphers v Railway Express Agency, 321 US 342, 348-349). Other considerations include "promot[ing] repose by giving security and stability to human affairs” (Wood v Carpenter, 101 US 135, 139 [1879]), judicial economy, discouraging courts from reaching dubious results, recognition of self-reformation by defendants, and the perceived unfairness to defendants of having to defend claims long past (see, Fischer, The Limits of Statutes of Limitation, 16 SW U L Rev 1, 1-2; see also, Harshaw, Not Enough Time?: The Constitutionality of Short Statutes of Limitations for Civil Child Sexual Abuse Litigation, 50 Ohio St LJ 753).
In examining the policies underlying Statutes of Limitation, this Court has emphasized both a defendant’s interest in repose and "in defending a claim before his ability to do so has deteriorated through passage of time” (Martin v Edwards Labs., supra, 60 NY2d, at 425). We have also taken note of considerations of judicial economy and possible plaintiff fraud where *774excessive factual inquiries would be necessary (supra, 60 NY2d, at 425-426). Balanced against these considerations is the injured person’s interest in having a reasonable opportunity to assert a claim.
Weighing these competing policy considerations, we conclude that the proper rule in RSI cases is that the cause of action accrues against a given manufacturer upon the onset of symptoms or the last use of the injury-producing device, whichever is earlier (see, Piper v International Bus. Machs. Corp., 219 AD2d 56, 60-62). Clearly in this case fairness to the plaintiff militates against a rule based on first use. Plaintiffs suffering from RSI have no reason to complain of their condition until a symptom actually manifests itself, and cannot be faulted for failing to exercise their legal rights prior to that time.
We reject plaintiffs’ argument that accrual should be measured not from the time they experienced RSI symptoms, but from the date they actually became aware of the nature of their injuries. A date of diagnosis test would give a plaintiff "the power to put off the running of the Statute of Limitations indefinitely” (Snyder v Town Insulation, 81 NY2d 429, 435). Moreover, this approach would only further ensnare trial courts in a fact-based application of the rule (see, Consorti v Owens-Corning Fiberglas Corp., supra, 86 NY2d, at 452-453).
Judicial economy and the possibility of feigned claims militate in favor of some type of first use rule. However, it is likely that a plaintiff will seek medical treatment when the injury manifests itself in the form of RSI or pre-RSI symptomatology. This significantly reduces the possibility of a plaintiff indefinitely postponing or feigning a claim, since in the majority of cases there will be objective evidence of the onset of symptoms in the form of medical or work records. Although plaintiffs contend they are not able to identify a date for the onset of RSI symptoms, they have alleged a date on which the signs /symptoms of RSI were manifest. Requiring a symptom-based benchmark does not impose an inappropriate or unfair burden on plaintiffs.
Moreover, while an accrual period which bears some relationship to the onset of symptoms will certainly involve lower courts in limited factual inquiries, this was also true of the rule we adopted in Martin v Edwards Labs. (supra, 60 NY2d, at 428-429). As in Martin, the products in these cases retain their identity. Thus, the evidence would not appear to be stale except in the case of remote products long since discarded. By *775tying the accrual date to the date of last use, as well as the onset of symptoms, the rule we formulate today strikes a proper balance between these competing considerations. * In addition, the last use condition will serve as an additional filter for claims where the date of injury remains in doubt.
Finally, the possibility of causal breaks between the claimed negligence and the injury is also a legitimate concern. However, the last use or onset of symptoms rule solves the causal breakdown problem, which is most pronounced in the case of keyboards used years prior to the onset of symptoms. For example, in one of the cases before us, Sharib, Heholt & DeForest v Zenith Data Sys. (Sup Ct, NY County, index No. 125155/93), plaintiff Heholt sues two different manufacturers of four different keyboards which she used from 1979 through 1992. She commenced her action on October 5,1993, and identifies April 1992 as the date upon which she began to experience some symptoms of RSI. Nevertheless, by suing the manufacturer of every keyboard she ever used, plaintiff implicitly alleges that each use of the defendants’ keyboards contributed to her condition. This includes the keyboard which she used from 1979 to 1984 with no symptomatology. Certainly there comes a point in time when the causal link between use and injury becomes too remote as a matter of law.
The rule we announce today will bar direct actions against such remote manufacturers. It strikes a proper balance between giving a plaintiff an opportunity to commence an action after becoming aware of a symptom of injury and providing certainty and predictability to manufacturers, employers and other economic actors in their risk assessment, while also avoiding stale claims.
In adopting this rule, we are also cognizant of the fact that problems of proof will be inherent in these cases whether we define accrual by reference to the first use of the device or by the onset of symptoms. Unless first use is defined as any touching of a keyboard whatsoever (which, as previously noted, bears no relationship to actual injury), then the problem in defining the moment of injury would become determining when a plaintiff began using a keyboard with any substantial regularity. In an increasingly technological world, exposure to com*776puter keyboards and other devices with the potential to cause RSI has become commonplace. Given this reality, defining "first use” would pose equally serious, if not greater, difficulty. Having carefully considered the various competing concerns presented by this information age injury, we conclude that the balance is best struck by defining accrual of a cause of action for RSI as the earlier of the onset of symptoms or last use.
Accordingly, the order of the Appellate Division should be modified, without costs, the actions remitted for action in accordance with this opinion, and the certified question answered in the negative.
Chief Judge Kaye and Judges Titone, Smith and Ciparick concur; Judges Bellacosa and Levine taking no part.
Order modified, without costs, actions remitted to Supreme Court, New York County, for further proceedings in accordance with the opinion herein, and certified question answered in the negative.