Mercer v. Pittway Corp. , 2000 Iowa Sup. LEXIS 169 ( 2000 )

616 N.W.2d 602 (2000)

Nathan R. MERCER and Jennifer Mercer, as parents of Travis Mercer; Northwest Bank & Trust Company Corporation, Conservator for Travis Mercer, a Minor Child; Nathan R. Mercer and Jennifer Mercer, Administrators of the Estate of Bradley Mercer, Deceased; Nathan R. Mercer and Jennifer Mercer as Parents of Bradley Mercer; and Nathan R. Mercer and Jennifer Mercer, Individually and as Husband and Wife, Appellees,
v.
PITTWAY CORPORATION; BRK Brands, Inc., Appellants, and
Nicholas Bakeris, Defendant,
State of Iowa ex rel. Civil Reparations Trust Fund, Intervenor-appellee.

No. 98-1144.

Supreme Court of Iowa.

September 7, 2000.

Roger T. Stetson, Margaret C. Callahan and Mark McCormick of Belin Lamson McCormick Zumbach Flynn, P.C., Des Moines, and Stephen A. Cozen and Robert W. Hayes of Cozen & O'Connor, P.C., Philadelphia, Pennsylvania, for appellants.

W. Don Brittin, Jr. of Nymaster, Goode, McLaughlin, Voigts, West, Hansell & O'Brien, P.C., Des Moines, John O. Moeller, Davenport, and James L. Fetterly, Stephen G. Lickteig, and Diane B. Bratvold of Fetterly & Gordon, P.A., Minneapolis, Minnesota, for appellees.

Thomas J. Miller, Attorney General, Craig Kelinson, Special Assistant Attorney General, and Richard E. Mull, Assistant Attorney General, for intervenor-appellee State of Iowa ex rel. Civil Reparations Trust Fund.

Considered en banc.

McGIVERIN, Chief Justice.

Numerous issues are presented in this product liability case involving the alleged failure of a smoke detector to alarm to a fire in a home. One child died and another was severely burned as a result of the fire. In an action brought by the parents of the children against the manufacturer of the smoke detector, a jury awarded substantial compensatory and punitive damages. Defendants, the manufacturers of the smoke detector, appeal, contending the district court committed several reversible errors which include: (1) failing to grant defendants' motion for judgment notwithstanding the verdict on liability; (2) submitting the issue of punitive damages to the jury, absent substantial evidence of willful and wanton conduct by the defendants; and (3) admitting into evidence consumer complaints regarding alleged failure of defendants' smoke detectors to alarm to smoke.

Upon our review, we conclude that the district court's admission into evidence of several consumer complaints concerning the failure of defendants' smoke detectors to alarm to smoke constitutes reversible error. Also, punitive damages should not have been allowed. Accordingly, we reverse and remand for a new trial.

I. Background facts and proceedings.

A. The Mercers' fire.

Nathan and Jennifer Mercer as parents, individually, and as administrators of the estate of their son Bradley, sued defendants Pittway Corporation and BRK Brands, Inc. alleging that a smoke detector manufactured by those defendants was defective and failed to alarm to a fire in the Mercers' home.^[1] The Mercers' three-year-old son Bradley died from injuries he received during the fire and eighteen-month-old Travis was severely burned. In their petition, the Mercers alleged various theories of liability including negligence (negligent design and testing, and failure to warn), strict liability, breach of express and implied warranties, and fraudulent nondisclosure. Plaintiffs asked for both compensatory and punitive damages.

At a jury trial concerning the Mercers' claims against defendant BRK, the following evidence was presented.

In January 1993, plaintiffs Jennifer and Nathan Mercer and their sons, Bradley and Travis, were living in a duplex house in Davenport, Iowa. When his family moved into the duplex in 1991, Nathan purchased two BRK brand model 83R ionization smoke detectors. Nathan placed one smoke detector on the main floor of the duplex between the kitchen and dining room, and placed the other smoke detector on the ceiling at the top of the staircase leading to the second-floor bedrooms. Only the master bedroom opened directly onto the stairwell; the children's bedroom, in which the fire started, could be reached only by passing through the master bedroom.

In deciding where to place the smoke detectors, Nathan relied on information he learned while employed at Safety Plus, a business that sells residential fire safety equipment. As part of his employment, Nathan was informed that the ionization smoke detector "was a good universal smoke detector to have around." Nathan also learned that an ionization detector would pick up both invisible and visible smoke and that a photoelectric smoke detector would basically pick up visible smoke, which would eliminate a lot of nuisance alarms from cooking or from steam. Nathan was also informed that it was best to install photoelectric smoke detectors in the kitchen area and that the ionization detector could be installed in all other areas of the house, including in bedrooms or sleeping areas. Nathan stated he considered installing other smoke detectors on the second floor, but felt he was financially unable to do so.

Between 8:15 and 8:20 p.m. on January 18, 1993, Nathan carried Bradley to the second floor of their duplex home and put him to bed. Bradley's brother, Travis, was already sleeping in his bed in the same room. Nathan then went downstairs to the living room.

Approximately twenty minutes later, around 8:45 or 8:46 p.m., Nathan and Jennifer heard a "thump" from upstairs. The lights in the room also flickered and the furnace kicked on. Shortly thereafter, Jennifer smelled smoke and headed up the stairs to check on the children, while Nathan ran into the kitchen to get a fire extinguisher. Jennifer testified that she ran up the stairs, but stopped two or three steps from the top of the stairs, not because of smoke, but because she was unsure whether to proceed into the bedroom. Jennifer then turned and went back down the stairs, meeting Nathan halfway down. Both Jennifer and Nathan testified that they did not hear the smoke detector, installed on the ceiling of the second-floor landing, alarming at this time. Nathan told Jennifer to go call 911 and proceeded up the stairs. Jennifer then ran to the neighbor's half of the duplex and called 911 while Nathan continued up the stairs.

Nathan testified that by the time he reached the top of the stairs, the smoke extended about one foot away from the floor. Nathan testified he heard no alarm coming from the smoke detector. Nathan tried to lie down so he could slide on his stomach to reach the bedroom, but the smoke was "too choking." Nathan then tried putting a shirt around his mouth and tried to go back up the stairs, but he was not successful. He tried at least three times to get through the smoke, but was unable to enter the children's bedroom.

During this same time, the Mercers' neighbor on the other side of the duplex, Jimmy Sprout, heard a smoke detector alarming and had been halfway up his own stairs when he realized it was not his detector. He walked back down to his living room and at that point, Jennifer came into the Sprouts' side of the duplex to call 911. Jennifer's 911 call was recorded by the dispatch at 8:52 p.m.

Nathan and Jimmy Sprout then ran up the Sprout's stairs. With his bare hands, Nathan tore a hole in the common duplex wall, reached through the hole, and pulled Travis through the hole. He ran outside where he handed Travis to a firefighter. The record shows the fire department arrived at 8:57 p.m. Firefighters using self-contained breathing apparatus searched the two bedrooms and found Bradley in his bed.

Firefighters described the fire in the children's bedroom as a small, flaming fire, approximately four feet by four feet in size, that was quickly extinguished with "just a quick shot of water" from the hose. Certain firefighters also testified at trial that they had no recollection of hearing a smoke detector alarm prior to the time that they found Bradley. Fire Captain Stoelk testified he recalled hearing the smoke detector alarm at the same time he saw Lieutenant Ryan descending the stairs with Bradley in his arms and estimated that the smoke detector alarm sounded approximately two and one-half minutes after he and the other firefighters arrived at the scene.

Bradley and Travis were severely burned in the fire. Bradley died six days later. Travis survived, but sustained second and third degree burns over approximately fifty percent of his body, including his face and head.

B. Types of smoke detectors.

Defendant BRK is the leading manufacturer of smoke detectors. There are two basic types of smoke detectors: ionization detectors and photoelectric detectors, both of which are battery powered. BRK's model 83R, which was the type of detector installed in the Mercers' home, is an ionization detector. An ionization detector uses a small amount of radioactive material to create charged air particles (ions) in the detection chamber. As smoke particles enter the chamber, they combine with and essentially neutralize the charged particles. This results in a reduction in electricity which is sensed by the detector's electronic components, triggering an alarm.

A photoelectric smoke detector uses a source of light, rather than a source of radioactive material, in the detection chamber. When smoke enters the chamber, the smoke particles cause the light particles to reflect and scatter until they eventually hit a sensor, triggering an alarm.

Ionization and photoelectric smoke detectors do not react the same way in all types of fires because of the different ways in which they sense the presence of smoke particles. An ionization detector is more efficient and effective than a photoelectric detector in responding to a flaming fire because such fires generally emit smoke consisting of small smoke particles. On the other hand, a photoelectric detector may be more sensitive to smoke from a smoldering fire because such fires generally emit smoke consisting of larger sized smoke particles.

Defendant BRK manufactures both ionization and photoelectric smoke detectors. BRK is the only manufacturer that offers a combination smoke detector at the retail level. The combination detector combines both the ionization and photoelectric features in one unit. BRK's biggest selling smoke detector is the model 83R, an ionization detector. Approximately 100 million model 83R ionization smoke detectors exist in the market. A BRK ionization detector sells at retail for approximately $6-$10, a BRK photoelectric detector for approximately $20, and a combination detector for approximately $30. BRK has manufactured the model 83R since 1983 with little change in its design.

C. Evidence concerning nature of the fire.

The investigation of the Mercers' fire conclusively showed that the fire began in a Gerry brand baby monitor,^[2] which was sitting on the bottom shelf of a table in the children's bedroom. The major factual disputes at trial concerned when the fire started and when smoke would have reached the detector installed in the second-floor hallway of the Mercers' duplex to trigger the alarm.

As explained in plaintiffs' appellate brief, BRK's alleged liability rests on BRK's decision to market its model 83R ionization smoke detector as a "stand-alone" safety device when it knew that the model 83R did not promptly alarm to certain types of home fires, like the one that occurred in the Mercers' home. The term "stand-alone" is plaintiffs' characterization of BRK's conduct in marketing the model 83R with the impression that the model 83R will, absent any other smoke detective device, adequately alarm to a fire.

The Mercers called Dr. Joseph Zicherman, a combustion science engineer, who concluded, based upon physical evidence, testing, and research, that the "thump" the Mercers heard at approximately 8:46 p.m. occurred when the top of the table fell to the floor, after its spindle legs had burned through, due to the burning baby monitor. Dr. Zicherman determined that it would have taken eight to ten minutes for the spindle legs to burn enough to give way, causing the top shelf of the table to fall hitting the floor, which occurred at 8:46 p.m. Based on the burn time of the wood spindle legs, Dr. Zicherman estimated that the fire began between 8:30 and 8:36 p.m.

Another expert, Dr. Patrick Pagni, a professor of fire safety engineering, testified on behalf of plaintiffs that if the alarm had sounded when smoke reached the detector, the Mercers would have had about nine minutes to remove the children from the bedroom before conditions within the room became untenable. Plaintiffs' expert, Dr. B. Donald Russell, Associate Dean of the College of Engineering at Texas A & M University, also testified that had a combination or photoelectric detector been installed in the Mercer home on the day of the fire, there would have been sufficient smoke to set off the photoelectric detector and the Mercers would have had adequate warning to remove the children. The Mercers' experts also testified that burning polystyrene, such as the Gerry baby monitor and other plastics involved in the fire, produce large smoke particles, which is characteristic of a slow, smoldering fire, even though a flame is not present.

BRK, on the other hand, presented evidence showing that the fire was a short, flaming fire of intense duration that ignited between approximately 8:46 and 8:48 p.m., the approximate time that Jennifer and Nathan saw the lights flicker, heard the "thump," and just before Jennifer smelled smoke. BRK's expert, Dr. Leonard Wharton, testified that the "thump" the Mercers heard was a noise produced when the electrical components of the baby monitor ignited and that the lights flickered due to the short-circuit in the baby monitor. Wharton opined that smoke had not yet reached the detector in a sufficient amount to trigger an alarm at the time Jennifer first went upstairs because the fire was in the initial stages, but that conditions rapidly deteriorated and the smoke levels quickly increased, causing the alarm to sound as heard by Jimmy Sprout (the Mercers' neighbor) by the time that Nathan returned from the kitchen with the fire extinguisher. Dr. Wharton also commented that there would be a delay in the smoke getting from the children's bedroom, where the fire began, to the smoke detector located three rooms away from the children's bedroom.

Dr. Wharton further stated that the baby monitor was made of polystyrene, which is a fast burning plastic material that produces a dense, irritating smoke as well as a large amount of heat, which would explain why the wood spindles of the table burned so rapidly. Dr. Wharton also explained that the Mercer children were burned from the intense heat radiating from the burning plastic baby monitor. BRK's experts testified that the Mercer fire could not have lasted more than a half-hour, as suggested by plaintiffs, and that if the fire had been burning more than a half-hour, the children would have died from carbon monoxide asphyxiation. BRK's expert, Gerald Rork, also testified that if the fire had been burning for more than a half-hour, Jennifer would not have been able to get to the second-floor landing because of the intense heat from the fire.

The parties' experts seem to agree that the Mercer fire was a flaming, not a slow, smoldering fire, but that the burning polystyrene Gerry baby monitor produced large smoke particles. Additionally, all experts, including BRK's experts, agreed that smoke should have reached the detector within three minutes from ignition and that the detector should have sounded at that time.

D. Evidence concerning alleged product defect.

BRK presented evidence showing that ionization smoke detectors have been the primary residential fire detective device since the 1970s and that the number of fire deaths has decreased with the increased use of smoke detectors. BRK experts also testified that the model 83R satisfies the Underwriters Laboratory standard 217 [hereinafter UL 217 standard]. The UL 217 standard is a compilation of requirements and tests that are required for the performance and construction of smoke detectors. The UL 217 standard was created by fire department personnel, fire scientists, and members of government agencies and has been in existence for the last twenty-five years. The UL 217 standard is the standard adopted by the smoke detector industry. The UL 217 standard consists of a smoke box test whereby a smoke detector is placed in a box along with a cotton wick. The cotton wick is lighted on fire and the smoke detector's response to smoke and obscuration level of the smoke is measured. The UL 217 standard also consists of a burn test where a smoke detector is placed in a room where a fire is started, using different combustible materials, including paper, flaming wood, gasoline, plastic and smoldering wood.

Plaintiffs' experts were critical of the ability of the UL 217 standard to effectively test the type of "real world" fires representative of the conditions and types of residential fires. Plaintiffs also presented evidence suggesting that BRK knew that the model 83R ionization detector might have a delayed reaction in detecting certain types of fires, particularly in fires where large smoke particles are produced. Plaintiffs' expert, Dr. Russell, testified about various articles discussing the poor performance results of ionization detectors in detecting slow, smoldering fires. Dr. Russell stated that these publications show that it has been known in the smoke detector industry for many years that ionization detectors are not as effective as photoelectric detectors in detecting slow, smoldering fires.

Plaintiffs also presented evidence of more than 360 complaints made by consumers to BRK and presented testimony from three consumers regarding the failure of the model 83R detector to alarm to smoke. These complaints were categorized by BRK as "N/R/S" or "NRS," a shorthand reference for "no response to smoke." BRK's customer service representative, Beth Weber, testified that on average BRK received one NRS complaint per week from the time she began working at BRK in 1983. Approximately fifty percent of complaining customers returned their detectors at BRK's request. BRK would then check the battery and sensitivity of the detector and subject it to the UL 217 standard smoke box test, which measures the amount of smoke it takes to trigger the alarm. None of the returned detectors, which were submitted to the smoke box test, failed that test.

E. Jury verdict and post-trial motions.

At the close of the Mercers' evidence and of all evidence, BRK moved for a directed verdict, but the court overruled the motion.

The trial court submitted the case to the jury on each of the Mercers' theories: (1) negligence, including negligent design, negligent testing, and negligent failure to warn; (2) strict liability theory of defective condition as a "stand-alone" detector; (3) breach of implied warranty of merchantability; and (4) fraudulent nondisclosure. The court also submitted the issue of punitive damages.

The jury found in favor of plaintiffs on all theories submitted and further found Gerry (manufacturer of the baby monitor) and defendant BRK each to be fifty percent at fault for plaintiffs' compensatory damages. The jury awarded plaintiffs compensatory damages in the amount of $8.8 million and punitive damages in the amount of $12.5 million.^[3] The court, however, reduced the compensatory damages award to $4.4 million because BRK was found to be only fifty percent at fault and entered judgment in favor of plaintiffs for that amount, plus the $12.5 million punitive damages award.

In response to the verdicts and judgment, BRK filed a post-trial motion for judgment notwithstanding the verdict and new trial, asserting numerous errors by the district court. The court overruled BRK's motions, but reduced the jury's award to Travis for future medical expenses from $1 million to $397,113.76.

BRK appeals, raising several issues.

We first address whether the district court properly admitted evidence of several consumer complaints and whether the court properly submitted the punitive damages issue to the jury because our conclusions on these issues require this case to be reversed and remanded for a new trial.

II. Did the district court err in admitting evidence of all consumer complaints?

BRK contends that the district court committed reversible error by allowing the Mercers to introduce into evidence more than 300 consumer complaints to BRK regarding the alleged failure of the model 83R to alarm to smoke. BRK objected to admission of the consumer complaints on hearsay and relevancy grounds, arguing that plaintiffs failed to prove that the incidents reported in the consumer complaints were substantially similar to the circumstances of the Mercer fire.

A. General principles and scope of review.

The inquiry whether evidence is admissible under our Iowa rules of evidence 402 and 403 involves a two-step inquiry: (1) is the evidence relevant? and (2) if so, is its probative value substantially outweighed by the danger of prejudice or confusion? Evidence is relevant if it has a tendency to make a consequential fact more or less probable than it would be without the evidence. Iowa R. Evid. 401; McClure v. Walgreen Co., 613 N.W.2d 225, 235 (Iowa 2000). Even relevant evidence, however, is not admissible "if its probative value is substantially outweighed by the danger of unfair prejudice." Iowa R. Evid. 403. "Unfair prejudice arises when the evidence prompts the jury to make a decision on an improper basis...." Waits v. United Fire & Cas. Co., 572 N.W.2d 565, 569 (Iowa 1997).

We review a district court's decision concerning the admission of relevant evidence for an abuse of discretion. See McClure, 613 N.W.2d at 235. An abuse of discretion occurs when "the court exercised [its] discretion on grounds or for reasons clearly untenable or to an extent clearly unreasonable." Waits, 572 N.W.2d at 569 (quoting State v. Maghee, 573 N.W.2d 1, 5 (Iowa 1997)). A ground or reason is untenable when it is not supported by substantial evidence or when it is based on an erroneous application of the law. See id. Not every erroneous admission of irrelevant evidence requires reversal. McClure, 613 N.W.2d at 235. Reversal is only warranted when "a substantial right of the party is affected." Id. (quoting Iowa R. Evid. 103(a)). "Although a presumption of prejudice arises when the district court has received irrelevant evidence over a proper objection, the presumption is not sufficient to require reversal if the record shows a lack of prejudice." Id.

B. Applicable law concerning admission into evidence of substantially similar incidents.

The rule is well established that evidence of prior accidents or incidents may be admissible to show the existence of a dangerous condition. Lovick v. Wil-Rich, 588 N.W.2d 688, 697 (Iowa 1999). A preliminary requirement to the admission of evidence of prior incidents, however, is a foundational showing that the prior accidents or incidents occurred under substantially the same circumstances as the incident in the present case. McClure, 613 N.W.2d at 234 (holding that evidence of thirty-four incident reports, involving claimed "error" in filling of prescriptions at defendant's pharmacy within a three-year period preceding the prescription error in plaintiff's case, were substantially similar to misfilling incident giving rise to plaintiff's claim and were properly admissible under similar incidents rule); see also Lovick, 588 N.W.2d at 697, and cases cited therein. In cases where the evidence of prior accidents or incidents is offered to show a dangerous condition, the probative value of previous accidents rests in the likelihood that the same dangerous conditions caused the accident that is the source of the present litigation. Cook v. State, 431 N.W.2d 800, 803 (Iowa 1988). We said in Cook that when the prior accidents evidence is offered to show that the alleged tortfeasor had notice of the allegedly defective condition, the proof of similarity may be more relaxed. Id. at 803 (citing E. Cleary, McCormick on Evidence § 200, at 589-90 (3d ed.1984)).

Although accident reports or other documents consisting of complaint statements made by consumers to a manufacturer constitute hearsay, and thus would be inadmissible for any purpose that assumes the truth of the allegations made in the complaints, see Iowa R. Evid. 801(c), such evidence would be admissible as tending to show the manufacturer's notice or awareness of danger, provided the incidents alleged in the complaints are substantially similar to the one at issue. See Shields v. Sturm, Ruger & Co., 864 F.2d 379, 381 (5th Cir.1989) (holding that although accident reports from consumers to gun manufacturer were hearsay and inadmissible for any purpose that assumed the truth of the allegations made in the reports, district court properly admitted 70 of the 570 consumer reports for purposes of proving notice and awareness of danger caused by misfiring of gun; reports of customers made after the date of plaintiff's accident were not relevant and not admissible).

C. The consumer complaints.

To properly address this issue, we believe it helpful to explain in detail the evidence of consumer complaints made to BRK.

At trial, BRK customer service employee Beth Weber testified concerning BRK's method of responding to consumer complaints. Weber testified that a consumer would either telephone or write to BRK concerning complaints regarding alleged failure of a smoke detector to alarm in the presence of smoke. A BRK customer service representative would document the relevant information. When a consumer complained that a smoke detector had failed to alarm in the presence of smoke, the notation "N/R/S" or "No Response to Smoke" was made on the report. Weber explained to the jury that the term No Response to Smoke is the term assigned by BRK to complaints where the consumer thought that the smoke detector should have sounded in the presence of smoke but did not. Weber was responsible for processing complaints directly from consumers.

The next step in processing a consumer complaint was that a BRK customer service representative would respond in writing to the consumer using a form-type letter, asking the consumer to complete an enclosed questionnaire. The letter also included instructions on how to return the detector to BRK for examination. BRK would send a replacement detector to the consumer prior to the consumer returning the detector that allegedly failed to alarm. From 1983 until about 1992 or 1993, BRK would send consumers a combination detector. BRK now sends the consumer the same model that is the source of the complaint.

Some of the complaining consumers completed the questionnaires and others did not. Approximately half of the customers returned their model 83R detectors. If the consumer sent the smoke detector to BRK for testing, BRK would examine the detector, check the battery connection, and submit it to the UL 217 smoke box test to determine whether the detector was working properly. According to BRK's records, of the detectors that were returned by consumers and that were subjected to the UL 217 smoke box test, none failed. After testing the detector, BRK would then send a letter to the consumer explaining what the likely problem was when the alarm did not sound.

During the course of discovery in this case, plaintiffs requested BRK to produce company documents relating to customer complaints involving the alleged failure of the BRK model 83R detector to alarm to smoke. BRK produced approximately 363 consumer complaints that had been categorized as "N/R/S" or No Response to Smoke. These complaints were received by BRK during the period from July 1989 to December 1997. Only 116 of those complaints were received by BRK prior to the Mercer fire, which occurred on January 18, 1993.

Prior to trial, BRK filed a motion in limine seeking to exclude all 363 consumer complaints from evidence at trial. BRK argued that the consumer complaints were inadmissible hearsay and were not relevant in that the Mercers had failed to show that the incidents reported in the complaints were substantially similar to the facts and circumstances of the Mercer fire.

D. The district court ruling and the parties' contentions.

After hearing extensive argument on the issue, the court ruled that all of the consumer complaints would be admissible at trial and overruled BRK's motion in limine. The court's reasoning was as follows:

THE COURT: ... the Court is going to allow into evidence the consumer complaints.... These consumer complaints... are relevant regarding punitive damages, these complaints, they are relevant to show knowledge.... The Court feels that the consumer complaints are relevant in that they are substantially similar under the Iowa law. The fact that they, perhaps, never went off or went off late, in this Court's opinion they are substantially similar.... If it doesn't go off timely, it doesn't do what it's intended to do.... Again, the Court is finding that an alarm that never goes off and one that does not go off in a timely fashion are substantially similar for the purposes of this type of product in this lawsuit. It comes in to show knowledge.

As a result, all 363 written consumer complaints were admitted into evidence. The consumer complaints were presented to the jury in three cardboard filing boxes. The only limiting instruction given by the court with respect to the consumer complaints advised the jury that in considering the conduct of BRK with regard to the issue of punitive damages, the jury could only consider the consumer complaints as to those incidents which pre-dated the Mercer fire of January 18, 1993. In addition to admission of the written consumer complaints, three of the complaining consumers, including Barbara Ware, William Watters, and Marilyn Russell, testified during trial concerning the failure of their model 83R to alarm to smoke during a fire.

Later, in overruling BRK's motion for judgment notwithstanding the verdict on the issue of the consumer complaints, the court stated, that "by BRK's own conduct and admissions, the consumer complaints at issue were substantially similar to the fire at the Mercer residence." The court specifically noted that BRK had categorized the consumer complaints under the single heading No Response to Smoke and that BRK sent out the exact same form letter regarding the identical model 83R ionization smoke detector to those consumers who complained.

On appeal, BRK argues that the district court committed reversible error in admitting into evidence all 363 complaints. Specifically, BRK argues that not all of the incidents in the 363 consumer complaints were substantially similar to the Mercer fire because BRK's records of the consumer complaints showed that the smoke detectors failed to alarm for a variety of reasons, including: (1) improper placement of the detector on wall or ceiling; (2) improper location of detector within the home; (3) cold smoke or insufficient amount of smoke to reach detector; and (4) no battery or improper battery connection in the model 83R. In other cases, the smoke detector itself had been started on fire in an attempt to test it. BRK also points out, that in some cases, the detector had not been returned so BRK could not properly test the detector for a potential defect or the consumer had not completed the questionnaire regarding the incident. BRK also points out that the Mercer smoke detector was not available for testing and BRK could therefore not examine it to determine whether it had failed to function properly.

The Mercers argue that the consumer complaints were properly admissible because they were substantially similar to the Mercer fire in that the complaints involved the exact same product and the same alleged product defect—that BRK ionization detectors have a delayed response time in responding to certain types of fires. Additionally, the Mercers argued that the consumer complaints were especially relevant to the issue of punitive damages as they showed BRK's knowledge that the 83R ionization smoke detector was not timely responding to slow smoldering fires, and BRK's knowledge that its method of testing the model 83R was not an adequate predictor of the performance of the 83R ionization detector in real world fire scenarios.

E. Application of law to facts.

1. We first conclude that only those consumer complaints received by BRK prior to the date of the Mercer fire, January 18, 1993, (which the record shows to be 116) are relevant to plaintiffs' theories of recovery in this case. Therefore, only those complaints received by BRK prior to January 18, 1993, were properly subject to consideration for admission into evidence. Our recent cases also suggest that the rule allowing evidence of similar incidents is generally limited to incidents occurring prior to the one in question. See Lovick, 588 N.W.2d at 697 (making reference to prior incidents); Fell v. Kewanee Farm Equip. Co., 457 N.W.2d 911, 920 (Iowa 1990) (same); Cook, 431 N.W.2d at 803 (same); Oberreuter v. Orion Indus., Inc., 398 N.W.2d 206, 211-12 (Iowa App. 1986) (same); cf. Shields, 864 F.2d at 381 (holding that district court properly excluded reports from consumers made after the date of plaintiff's accident on relevancy grounds). But see Rattenborg by Rattenborg v. Montgomery Elevator Co., 438 N.W.2d 602, 606 (Iowa App.1989) (holding evidence of two subsequent accidents was admissible); Caterpillar Tractor Co. v. Beck, 624 P.2d 790, 794 (Alaska 1981) (holding that evidence of post-injury accidents is admissible in strict liability cases); Bass v. Cincinnati, Inc., 180 Ill.App.3d 1076, 129 Ill.Dec. 781, 536 N.E.2d 831, 833-34 (1989) (stating that rationale of why evidence of prior accidents is admissible—to show a product is dangerous or defective—applies to evidence of accidents which occur after the accident involving a particular plaintiff; but holding that evidence of similar post-accident occurrences is not admissible on issue of punitive damages).

2. We next consider which of the 116 consumer complaints received by BRK before the Mercer fire meet the substantially similar requirement. Plaintiffs' theory at trial was that their model 83R detector was fully operable and properly installed, that smoke reached the detector, but the detector did not sound an alarm due to its defective condition in that ionization detectors do not timely alarm to certain types of fires. The 116 consumer complaints received prior to the Mercer fire all involved the same product—the BRK model 83R ionization detector—and also involved the same general allegation—that the model 83R failed to alarm to smoke.

At first glance, it may seem that the 116 consumer complaints were properly admitted. However, the consumer complaints show that there are a number of variables that affect the performance of a smoke detector. Thus, the specific reason why the model 83R failed to alarm to the presence of smoke in the Mercer fire therefore became an issue in the case. Plaintiffs had the burden of showing that the facts and circumstances of each consumer complaint were substantially similar to the Mercer fire to be admissible into evidence. Accordingly, this means that plaintiffs had to do more than just show that each consumer complaint involved a model 83R that did not alarm to smoke. Rather, plaintiffs had to show that there was a sufficient similarity between the reason why the particular consumer's detector failed to alarm and why the Mercers' detector failed to alarm; it was not BRK's burden to disprove any common factors between the Mercer fire and the incidents reported in the consumer complaints. Thus, the fact that BRK categorized the complaints as NRS or that BRK sent out the same form letter to the consumer does not prove that the reason the particular consumer's detector allegedly failed to alarm is sufficiently similar to the reason the Mercers' detector allegedly did not alarm.

Upon our review of the evidence related to the consumer complaints, we conclude that plaintiffs failed in their burden of proving that all the consumer complaints received by BRK prior to January 18, 1993, are substantially similar to the facts and circumstances of the Mercer fire. The customer complaints reveal that some of the smoke detectors referenced in the complaints failed to alarm for reasons unrelated to those advanced by plaintiffs in this case. For example, in some of the consumer complaints, there was no battery in the detector or the battery or the smoke detector itself was started on fire in an attempt to test it. The failure of a smoke detector to alarm for these reasons is not relevant because it was unlikely to make a finding that the Mercers' detector was defective more or less probable.

We also believe that any probative value of the complaints was outweighed by the danger of unfair prejudice when the jury observed the three boxes of 363 consumer complaints being admitted into evidence with the implication that each consumer complaint involved an incident similar to the Mercer fire where the model 83R allegedly failed to alarm to smoke because it was defective. See Iowa R. Evid. 403 (relevant evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice).

We realize that the record in this case was voluminous. However, with respect to the 116 consumer complaints received by BRK prior to the fire, it would be necessary for the court to examine each prior incident to determine if it truly is substantially similar to the incident in the subject case. The court simply could not rely on BRK's classification system of processing consumer complaints for its conclusion that the complaints were substantially similar to the Mercer fire. Rather, it was plaintiffs' burden to establish a threshold foundation that in each consumer complaint the smoke detector was fully operable, that the detector was installed properly as to location in the home and on the wall/ceiling, and that a sufficient amount of smoke reached the detector to trigger the alarm. Those incidents where the consumer reported that the detector was otherwise operable, properly installed, and smoke reached the detector, but was informed by BRK that the detector did not alarm due to cold smoke or improper location of detector or some other subjective reason, would seem to relate more to weight of the evidence rather than its admissibility.

We also point out that the district court would have discretion under Iowa rule of evidence 403 to exclude evidence of consumer complaints due to considerations of undue trial delay in analyzing whether each particular consumer complaint was sufficiently similar to the facts and circumstances of the Mercer fire. See Wilson v. Bicycle South, Inc., 915 F.2d 1503, 1510 n. 10 (11th Cir.1990) (upholding trial court's exclusion of evidence on grounds that evidence was not probative because of the necessity for considerable amount of extrinsic evidence to determine whether incidents were sufficiently similar to meet the standards of Fed.R.Evid. 403; even if trial court had reached the issue of whether the two incidents were substantially similar, admissibility of evidence would have required a trial within a trial); Brooks v. Chrysler Corp., 786 F.2d 1191, 1198-99 (D.C.Cir.1986) (holding that district court did not abuse its discretion in excluding evidence of 330 consumer complaints and other exhibits under Fed.R.Evid. 403 where minimal probative value of exhibits was substantially outweighed by risk of unfair prejudice, trial delay, and jury confusion; noting that trial would have been delayed by defendant attempting to rebut substance of each of the 330 complaints or to distinguish nature of complaints from alleged defect in plaintiff's case). The court might also consider giving a cautionary instruction advising the jury that the evidence of the consumer complaints may only be considered on the issue of whether BRK has notice or knowledge of the defective condition of the model 83R.

We therefore conclude that the court's admission of all 300 plus consumer complaints constitutes reversible error.

III. Was the evidence sufficient to support submission of punitive damages to the jury?

The jury awarded $12.5 million punitive damages to plaintiffs and the court entered judgment against BRK thereon. The parties stipulated that defendant's conduct was not specifically directed at plaintiffs.

BRK contends there was insufficient evidence to support the submission of plaintiffs' punitive damage claim to the jury and that BRK's motion for directed verdict and for judgment notwithstanding the verdict in that regard should have been sustained.

A. Applicable law.

Iowa Code section 668A.1(1)(a) sets the standard for an award of punitive damages. Under this section, an award of punitive damages will stand when there is proof of conduct that establishes a "willful and wanton disregard for the rights or safety of another." Wilson v. IBP, Inc., 558 N.W.2d 132, 142 (Iowa 1996) (citing Iowa Code § 668A.1(1)(a)). We have approved the following definition of "willful and wanton" conduct for section 668A.1(1) purposes:

[T]he actor has intentionally done an act of an unreasonable character in disregard of a known or obvious risk that was so great as to make it highly probable that harm would follow, and which thus is usually accompanied by a conscious indifference to the consequences.

Fell, 457 N.W.2d at 919 (quoting W. Page Keeton, et al., Prosser & Keeton on the Law of Torts § 34, at 213 (5th ed.1984)). The evidence to support an award must be clear, convincing and satisfactory. Iowa Code § 668A.1(1)(a).

We have noted that punitive damages serve "``as a form of punishment and to deter others from conduct which is sufficiently egregious to call for the remedy.'" McClure, 613 N.W.2d at 230 (quoting Coster v. Crookham, 468 N.W.2d 802, 810 (Iowa 1991)). Consequently, punitive damages are appropriate only when actual or legal malice is shown. Schultz v. Security Nat'l Bank, 583 N.W.2d 886, 888 (Iowa 1998). Mere negligent conduct is therefore not sufficient to support a claim for punitive damages. Beeman v. Manville Corp. Asbestos Disease Compensation Fund, 496 N.W.2d 247, 256 (Iowa 1993).

B. Application of law to facts.

The question is whether a rational jury could find by "a preponderance of clear, convincing, and satisfactory evidence," Iowa Code § 668A.1(1)(a), that BRK's conduct constituted willful and wanton disregard for the rights or safety of another. See Schultz, 583 N.W.2d at 888.

As will be explained in more detail later in this opinion, plaintiffs' theory at trial was that BRK had knowledge that the model 83R was not as effective as a photoelectric smoke detector at alarming to slow, smoldering fires, but yet continued to market the model 83R ionization detector as a sufficient fire detective device.

We conclude later that plaintiffs generated a jury question on their theories of strict liability and negligence as to the effectiveness of ionization smoke detectors, and how the Mercers' smoke detector did, or should have, responded to the fire in their home. However, we reject plaintiffs' contention that BRK's knowledge of the 83R's failure to alarm in certain types of fires, coupled with BRK's failure to test ionization detectors in real world fires and its failure to place a specific warning on the box concerning the 83R's delayed response to certain types of fires, creates a jury question on punitive damages.

Both parties presented evidence concerning the effectiveness of the tests incorporated by the UL 217 standard to test "real world" fires. Plaintiffs' witnesses were especially critical of the smoke box test, which BRK uses to test returned smoke detectors from complaining consumers. On the other hand, BRK's experts testified that the criticisms of the smoke box test and of the UL 217 standard in general, as expressed by plaintiffs' experts, were unfounded.

While we cannot say that BRK's adoption of the UL 217 standard proves its state of the art defense, we conclude that this evidence shows a reasonable disagreement over the relative risks and utilities of BRK's conduct in the manufacture and production of the model 83R ionization detector. See Hillrichs v. Avco Corp., 514 N.W.2d 94, 100 (Iowa 1994) (stating "that an award of punitive damages is inappropriate when room exists for reasonable disagreement over the relative risks and utilities of the conduct and the device at issue"). BRK's experts also testified that the sensitivity level of the ionization and photoelectric components in a combination smoke detector are not as sensitive as a stand-alone device to avoid nuisance alarms. Thus, although a jury could reasonably conclude that BRK was at fault, we do not believe that a rational fact finder could find by clear, convincing, and satisfactory proof a willful and wanton disregard by BRK for the rights of another. Consequently, the district court erred in submitting plaintiffs' claim of punitive damages to the jury and in overruling BRK's post-trial motion in this regard. We reverse on this issue.

Although our decision on the admissibility of the consumer complaints into evidence and the punitive damages issue requires reversal, we address other issues raised on appeal that will likely arise again upon retrial.

IV. Submitted theories of liability.

Plaintiffs sought recovery on and the district court submitted to the jury, four theories of liability: negligence (negligent failure to warn, negligent design, negligent testing), strict liability (defective design), breach of warranty, and fraudulent nondisclosure. The jury found in plaintiffs' favor on each theory. Defendant first contends that plaintiffs failed to present sufficient evidence to justify submitting the case to the jury on any of plaintiffs' theories of recovery and the court therefore should have sustained its motion for judgment notwithstanding the verdict concerning liability. We examine the theories submitted.

V. Strict liability claim.

The trial court's Instruction no. 21 to the jury sets forth the law on plaintiffs' claim of strict liability. It states in relevant part:

In order to recover on the claim of strict liability against Defendants Pittway and BRK Brands, Plaintiffs must prove all of the following propositions:

....

3. The Model 83R battery-operated ionization smoke detector was in a defective condition at the time it left defendants' control in that it was not adequate as a "stand-alone" smoke detector because it would not timely detect certain types of common household fires.

(Emphasis added.) Based on the highlighted language in the instruction, it seems that plaintiffs' strict liability claim is a design defect claim, i.e., that the model 83R ionization detector has a design defect in that it fails to alarm to certain types of fires and that this defect causes the product to be inherently dangerous.

A. Applicable law—design defect.

We have adopted section 402A of the Restatement (Second) of Torts (1965) regarding strict liability claims. See Leaf v. Goodyear Tire & Rubber Co., 590 N.W.2d 525, 528 (Iowa 1999). Section 402A provides in relevant part:

(1) [o]ne who sells any product in a defective condition unreasonably dangerous to the user or consumer or to his property is subject to liability for physical harm thereby caused to the ultimate user or consumer, or to his property, if

(a) the seller is engaged in the business of selling such a product, and

(b) it is expected to and does reach the user or consumer without substantial change in the condition in which it is sold.

(2) The rule stated in Subsection (1) applies although

(a) the seller has exercised all possible care in the preparation and sale of his product, and

(b) the user or consumer has not bought the product from or entered into any contractual relation with the seller.

Leaf, 590 N.W.2d at 528 (quoting Restatement (Second) of Torts § 402A, at 347-48 (1965)) (emphasis added). Under this rule, the seller is subject to liability to the user or consumer even though the seller has exercised all possible care in the preparation of the product. Aller v. Rodgers Mach. Mfg. Co., 268 N.W.2d 830, 834 (Iowa 1978).

Thus, to prove a "defective condition unreasonably dangerous," the plaintiff must show that the defect in the product was not one contemplated by the consumer, which would be unreasonably dangerous to the plaintiff in the normal and intended use of the product.^[4]Id.; accord Fell, 457 N.W.2d at 916 (quoting Restatement (Second) of Torts § 402A cmt. i). The unreasonableness element of a strict liability claim is determined by balancing the utility of the product against the risk of its use. Fell, 457 N.W.2d at 918; Chown v. USM Corp., 297 N.W.2d 218, 220 (Iowa 1980). In analyzing whether the risks of the product outweigh its utility, the jury may consider the availability of "a safer alternative design." Chown, 297 N.W.2d at 220-21. In other words, evidence of an alternative design is helpful to a plaintiff in proving that the design in question is unreasonably dangerous. Leaf, 590 N.W.2d at 535.

Strict liability claims generally involve factual issues to be resolved by a jury or other trier of fact. Bredberg v. Pepsico, Inc., 551 N.W.2d 321, 326 (Iowa 1996). Proof that a product is unreasonably dangerous may be proven by circumstantial evidence. Fell, 457 N.W.2d at 918.

B. Application of law to facts.

As we explained above, plaintiffs' strict liability claim is based on the contention that the model 83R is defective in design. Thus, it was plaintiffs' burden to prove by a preponderance of the evidence that the model 83R ionization detector was in a defective condition and was unreasonably dangerous at the time the product left defendant's possession. See Bredberg, 551 N.W.2d at 328. We need not conclude that the model 83R is a defective product, but rather need only determine whether the evidence, when viewed in the light most favorable to plaintiffs, was such that reasonable minds could differ on the issue.

The recitation of the facts stated above shows that both parties presented detailed, thorough and well-documented evidence from qualified experts concerning the various theories as to when the Mercer fire started, the burning pattern of the fire, and the time at which smoke reached the detector. The parties also presented thorough and well-documented evidence concerning the manufacturing design of an 83R ionization smoke detector and how it detects smoke. Not surprisingly, most of the evidence was conflicting. Viewing the evidence presented in a light most favorable to plaintiffs, we believe plaintiffs presented substantial evidence to justify submission of their strict liability claim to the jury. It was for the jury to determine whether a sufficient amount of smoke reached the detector, whether the detector alarmed or should have alarmed, and whether the detector failed to alarm because of a defect. See id. at 326 (strict liability claims generally involve factual issues to be resolved by the jury or other trier of fact). Evidence of properly admitted consumer complaints also puts BRK on notice that its product was not performing its intended purpose.

We likewise conclude that plaintiffs presented substantial evidence concerning whether the allegedly defective model 83R creates an unreasonably dangerous risk which is not outweighed by the benefits of the product.^[5] The most relevant evidence on this issue is the fact that an alternative design to the model 83R—the combination ionization/photoelectric detector—not only exists, but is manufactured by BRK. This evidence is relevant because it arguably shows that the model 83R is an unreasonably dangerous product, that a safer product exists, and that perhaps the model 83R should have been been removed from the market. Any properly admitted consumer complaints would also put BRK on notice about the potential danger associated with the failure of its smoke detector to alarm to smoke.

We conclude it was for the jury to consider all of the competent evidence presented and determine whether BRK's model 83R was defective and whether that defective condition was unreasonably dangerous. Consequently, the district court properly overruled BRK's motion for directed verdict and denied its post-trial motion for judgment notwithstanding the verdict concerning submission of plaintiffs' strict liability claim.

C. Was it proper to submit both a strict liability and a breach of warranty theory to the jury?

Plaintiffs also sought recovery under a theory of breach of implied warranty of merchantability. BRK contends that the district court erred in submitting to the jury both the strict liability and breach of warranty theories, relying on Nelson v. Todd's Ltd., 426 N.W.2d 120 (Iowa 1988).

We said in Nelson that "[f]ew cases will sufficiently involve the theories of breach of warranty and strict liability to warrant presenting both theories to a jury." 426 N.W.2d at 125 (citing Hawkeye Sec. Ins. Co. v. Ford Motor Co., 199 N.W.2d 373, 381 (Iowa 1972)). Nelson, however, involved the issue of whether a strict liability theory, in addition to a warranty claim, should be submitted to the jury when the plaintiff has only suffered economic loss. We stated the well established general rule that "purely economic injuries without accompanying physical injury to the user or consumer or to the user or consumer's property is not recoverable under strict liability." 426 N.W.2d at 123. We said in Kemin Industries v. KPMG Peat Marwick, 578 N.W.2d 212, 220 (Iowa 1998), that Nelson was "limited to deciding whether purely economic injuries without accompanying physical injury are recoverable under a theory of strict liability in tort."

In their case against BRK, the Mercers do not seek property damages sustained as a result of the fire, but rather seek damages for personal injuries under a strict liability claim based on the theory that the defect in the model 83R, which allegedly resulted in the failure of the detector to alarm to the Mercers' fire, created an unreasonable risk of danger. Therefore, nothing we said in Nelson supports BRK's contention that the district court was somehow precluded from submitting both strict liability and warranty claims to the jury. Thus, the damages sustained by plaintiffs caused by an alleged safety hazard in BRK's product relate to more than just their disappointed expectations of the product to perform as expected. See American Fire & Cas. Co. v. Ford Motor, 588 N.W.2d 437, 439 (Iowa 1999) (stating that "defects of suitability and quality are redressed through contract actions and safety hazards through tort actions") (cited sources omitted).

Additionally, we conclude the evidence with respect to plaintiffs' negligence theory of liability is likewise sufficient to create a jury question on plaintiffs' breach of implied warranty claim. See 63A Am. Jur.2d Products Liability § 1110, at 265 (1997) (stating that factual questions regarding duty to warn and causation underlying both a negligence and warranty claim are essentially identical; although the standards applied under the two claims differ slightly a defendant cannot be found to have been negligent without having breached the warranty of merchantability); see also Lovick, 588 N.W.2d at 698 (stating that although claims for negligence, strict liability, and breach of warranty are separate and distinct theories of liability under products liability law, the same facts often give rise to those three claims).

Accordingly, we conclude that the district court did not err in submitting both strict liability and breach of warranty theories to the jury.

D. State of the art defense.

The jury found that the model 83R detector was not state of the art. On appeal, BRK contends that it presented uncontested evidence that the model 83R ionization smoke detector was state of the art and that the district court therefore erred in failing to sustain its post-trial motion on that basis as a complete defense.^[6] BRK therefore contends it was entitled to judgment as a matter of law concerning plaintiffs' strict liability claim.

1. Applicable law.

The state of the art defense in product liability cases is codified in Iowa Code section 668.12. It provides that in any action brought against a manufacturer for damages arising from alleged defects in the design of a product, "a percentage of fault shall not be assigned to such persons if they plead and prove that the product conformed to the state of the art in existence at the time the product was designed, tested, manufactured, formulated, packaged, provided with a warning, or labeled." Iowa Code § 668.12. Section 668.12 further states that,

[n]othing contained in this section shall diminish the duty of an assembler, designer, supplier of specifications, distributor, manufacturer or seller to warn concerning subsequently acquired knowledge of a defect or dangerous condition that would render the product unreasonably dangerous for its foreseeable use or diminish the liability for failure to so warn.

In short, the state of the art defense may be a complete defense in product defect cases, but only as to a claim other than a negligent failure to warn. Olson v. Prosoco, 522 N.W.2d 284, 291 (Iowa 1994) (concluding that the state of the art defense is not applicable to claims based on negligent failure to warn). As we explained in Fell:

a duty and corresponding liability still exists in a case of a defect that is unreasonably dangerous and that a defendant learns of later. For example, even though a product is state of the art, if a manufacturer later learns of a defect that is unreasonably dangerous, the manufacturer then has a duty to warn. And a failure to do so could render the manufacturer liable.

457 N.W.2d at 920.

2. Standards in the industry.

BRK contends that it proved the state of the art defense as a matter of law because the model 83R complied with the UL 217 standard, which is the standard used in the industry to test ionization smoke detectors.

In Chown, we distinguished between what is considered "custom in the industry" and what is "state of the art." 297 N.W.2d at 221. We said that "[c]ustom refers to what was being done in the industry; state of the art refers to what feasibly could have been done." Id. (emphasis added). We said the relevant question "is not whether anyone else was doing more, although that may be considered, but whether the evidence disclosed that anything more could reasonably and economically be done." Id. (quoting Hancock v. Paccar, Inc., 204 Neb. 468, 283 N.W.2d 25, 35 (1979)); accord Hughes v. Massey-Ferguson, Inc., 522 N.W.2d 294, 295-96 (Iowa 1994). We noted in Hughes that evidence of a product design that is practically, as well as technologically sound, or evidence regarding government regulations, is relevant and material on the question of whether a state of the art defense has been sufficiently established to warrant submission to the jury. 522 N.W.2d at 296.

3. Application of law to facts.

At trial, Mark Devine, BRK vice president of engineering, testified that the model 83R conforms to the UL 217 standard, the current manufacturing and technology standard available in the industry. Plaintiffs' experts, however, testified in detail about the ineffectiveness of the UL 217 standard to test ionization detectors. For instance, Dr. Torero testified that the UL 217 smoke box test was only testing for the type of smoke (small smoke particles) that the ionization detector was capable of detecting (small smoke particles) and did not really test the ionization detector's ability to detect large smoke particles common in slow, smoldering fires.

We believe that the evidence presented created a jury question concerning BRK's state of the art defense and it could not be decided as a matter of law. BRK was therefore not entitled to a directed verdict concerning plaintiffs' strict liability claim and the court therefore properly submitted the state of the art instruction to the jury.

VI. Negligence theories.

Three theories of negligence by BRK were submitted by the trial court to the jury: (1) negligent failure to warn, (2) negligent design; and (3) negligent testing. BRK contends this was error.

A. Failure to warn.

The Mercers' theory at trial was that BRK had a duty to warn consumers or otherwise give consumers notice that the model 83R ionization smoke detector might fail to alarm in certain types of fires. Stated another way, plaintiffs say that defendant had a duty to warn that a photoelectric smoke detector might be more suitable for certain types of fires. Based on the arguments of the parties during trial and the court's instructions to the jury, we will assume that the Mercers' theory of negligent failure to warn is based on a point of sale failure to warn versus a post-sale failure to warn. See Lovick, 588 N.W.2d at 695 (discussing distinction between point of sale and post-sale failure to warn).

1. Law regarding negligent failure to warn.

We have adopted section 388 of the Restatement (Second) of Torts for determining whether a manufacturer of goods has fulfilled its duty to warn of a product's dangerous propensities. Lamb v. Manitowoc Co., 570 N.W.2d 65, 68 (Iowa 1997). That section states as follows:

One who supplies directly or through a third person a chattel for another to use is subject to liability to those whom the supplier should expect to use the chattel with the consent of the other or to be endangered by its probable use, for physical harm caused by the use of the chattel in the manner for which and by a person for whose use it is supplied, if the supplier

(a) knows or has reason to know that the chattel is or is likely to be dangerous for the use for which it is supplied, and

(b) has no reason to believe that those for whose use the chattel is supplied will realize its dangerous condition, and

(c) fails to exercise reasonable care to inform them of its dangerous condition or of the facts which make it likely to be dangerous.

Restatement (Second) of Torts § 388 (1965). While section 388 deals specifically with suppliers of chattels, section 394 of the Restatement (Second) of Torts subjects manufacturers to the same liability. Lamb, 570 N.W.2d at 68. A claim alleging a manufacturer failed to warn of the dangers involved in using a product is properly based on a theory of negligence, not strict liability. Id.

We have said that the proper analysis concerning a manufacturer's duty to warn at the point of sale essentially focuses on the foreseeability of a defective product. Lovick, 588 N.W.2d at 695. In other words, the inquiry is whether a reasonable manufacturer knew or should have known of the danger, in light of the generally recognized and prevailing best scientific knowledge, yet failed to provide adequate warning to users or customers. Olson, 522 N.W.2d at 289-90. A manufacturer therefore has a duty to warn "when a party ``reasonably foresee[s] a danger of injury or damage to one less knowledgeable unless an adequate warning is given.'" Lamb, 570 N.W.2d at 68 (quoting Beeman, 496 N.W.2d at 252). Whether a warning should have been given is a question for the jury. Beeman, 496 N.W.2d at 252. "In the context of a failure to warn claim, proximate cause can be established by showing a warning would have altered the plaintiff's conduct so as to avoid injury." Lovick, 588 N.W.2d at 700.

2. The model 83R warning and the parties' contentions.

The following information appears on the product insert pamphlet of the model 83R box.

WARNING

LIMITATIONS OF SMOKE DETECTORS:

WHAT SMOKE DETECTORS CANNOT DO

Smoke detectors have played a key role in reducing home fire deaths in the United States. However, according to the Federal Emergency Management Agency (an agency of the U.S. Government), they may not go off or give early enough warning in as many as 35% of all fires. What are some reasons smoke detectors may not work?

Smoke detectors will not work without power. Battery operated smoke detectors will not work without batteries, if the batteries are dead, if the wrong kind of batteries are used, or if the batteries are put in wrong. AC powered smoke detectors will not work if the power supply is cut off for any reason....

Smoke detectors will not sense fires which start where smoke does not reach the detectors. Smoke from fires in chimneys, in walls, on roofs, or on the other side of closed doors may not reach your smoke detector and alarm it. If bedroom doors are usually closed at night, detectors should be put in each bedroom and in the hallway between them.

....

Smoke detectors are not fool-proof. Like all other electronic devices smoke detectors have limitations. Smoke detectors may not sense every kind of fire every time. They cannot be expected to sense dangerous fires caused by carelessness or safety hazards. They may not give early warning of fast-growing fires caused by smoking in bed, violent explosions, escaping gas, poor storage of flammable liquids, overloaded electrical circuits, children playing with matches or lighters, or persons who set fires on purpose.....

(Emphasis added.)

The box containing a model 83R like the one Nathan Mercer purchased does not mention that an ionization detector may react slower to smoldering fires than a photoelectric detector or otherwise explain the difference between an ionization and photoelectric smoke detector. The user's manual for the model 83R that Nathan purchased generally describes how to install the unit, and as noted above, warns that "[s]moke detectors may not sense every kind of fire every time." A BRK photoelectric smoke detector box states that "[p]hotoelectric sensors are generally more effective at detecting slow, smoldering fires: fires which smolder for hours before bursting into flame." A BRK combination smoke detector box states that it is "[e]ffective at detecting both slow, smoldering fires and fast-flaming fires."

Gary Lederer, BRK's acting chief executive officer, testified at trial that the reason BRK does not explain in the product information for the model 83R (like the one Nathan purchased) the difference between photoelectric, ionization and combination detectors is that "if we try to educate the consumer to that degree, we are going to confuse them."

In its brief on appeal, BRK argues that the Mercers failed to show causation because Nathan never completely read the product warnings that came with the detectors he purchased and that the Mercers therefore cannot challenge the adequacy of the content of the information provided in connection with the smoke detector's performance limitations.^[7] The Mercers respond by arguing that BRK did not adequately preserve error on this contention in district court.

Upon our review, we conclude that BRK failed to preserve error on the issue of whether Nathan read the product warnings accompanying the model 83R detector that he purchased and installed in his home. The record shows, that in BRK's directed verdict motion, BRK made the statement that Nathan never completely read the product information of the model 83R that he purchased. BRK, however, did not explain how this fact established a lack of proximate cause concerning plaintiffs' failure to warn claim and did not otherwise alert the court to the issue. The district court also did not address this specific issue in ruling on BRK's directed verdict motion. We believe that BRK's mention of the alleged fact that Nathan did not read the model 83R product warnings was not sufficient to preserve error on this issue for purposes of our appellate review. Cf. Soo Line R.R. v. Dep't of Transp., 521 N.W.2d 685, 691 (Iowa 1994) (concluding that random mention of an issue in party's appellate brief "without elaboration of supportive authority, is insufficient to raise the issue for [appellate] consideration"). Accordingly, we proceed to examine whether plaintiffs presented sufficient evidence to create a jury question on their failure to warn claim.

3. Application of law to facts.

Our inquiry is whether plaintiffs presented sufficient evidence to raise a jury issue regarding: (1) whether BRK had knowledge that the model 83R might not be as effective as a photoelectric detector at detecting certain types of household fires; (2) whether BRK knew of the possible danger caused by the 83R's ineffectiveness with respect to certain types of fires; and (3) whether BRK was negligent in failing to tell consumers of its knowledge. Upon our review of the record, we conclude that plaintiffs presented sufficient evidence on all of these points to generate a jury question concerning their failure to warn theory.

As noted above, BRK was aware that the model 83R ionization detector might have a delayed response time in detecting certain types of fires. BRK's receipt of certain consumer complaints^[8] concerning the alleged failure of the model 83R to alarm to smoke also put BRK on notice as to potential performance problems of its product. This evidence also establishes the reasonable foreseeability on the part of BRK of danger to consumers due to a failure of the model 83R to sound a timely alarm in the presence of smoke. The fact that BRK could reasonably foresee such danger triggered a duty on the part of BRK to warn consumers of the possible limitations of the model 83R to detect certain types of fires.

Additionally, the fact that the product insert pamphlet for the model 83R states that the detector "may not go off or give early enough warning in as many as 35% of all fires" is not necessarily determinative on plaintiffs' failure to warn claim. This is because the evidence showed that plaintiffs' detector was otherwise operable, and there were no closed doors to block the smoke from reaching the detector, which are main factors listed in the product information pamphlet as reasons a smoke detector may not alarm. Additionally, Nathan testified that smoke conditions prevented him from entering the children's room, but that he did not recall hearing the smoke detector alarm.

We conclude that plaintiffs therefore presented sufficient evidence to justify submitting their negligent failure to warn claim to the jury. Cf. Laaperi v. Sears, Roebuck & Co., 787 F.2d 726, 730 (1st Cir.1986) (holding that whether manufacturer was negligent in failing to warn purchasers that a short circuit which causes an electrical fire may also render smoke detector useless was a question for jury); Carruth v. Pittway Corp., 643 So.2d 1340, 1346 (Ala.1994) (holding that jury question was presented whether pamphlet provided sufficient warning about danger of installing smoke detector in dead airspace).

B. Negligent design.

To prove a negligent design claim, a plaintiff must show that the product was unreasonably dangerous because of defendant's failure to use reasonable care in its design. See Chown, 297 N.W.2d at 220 (citing Restatement (Second) of Torts § 397).

Upon our review, we conclude that plaintiffs presented sufficient evidence to create a jury question as to whether BRK was negligent in designing the model 83R ionization detector. As noted above, the parties presented conflicting evidence concerning the effectiveness of a model 83R to sense certain types of fires like the one that occurred in the Mercer home and whether the smoke detector in the Mercer home sounded a timely alarm or should have so sounded. The district court therefore properly submitted the negligent design theory to the jury and properly overruled BRK's post-trial motion on this issue.

C. Negligent testing.

1. Applicable law.

Instruction no. 19 instructed that BRK was negligent if it "[f]ail[ed] to conduct field investigations or appropriate testing after receiving consumer complaints regarding the failure of the model 83R detector to function properly in real-world fire events."

The duty of a manufacturer with regard to testing is reasonable care. West v. Broderick & Bascom Rope Co., 197 N.W.2d 202, 212 (Iowa 1972).

In West, we discussed when a negligent testing claim may, in fact, be a negligent failure to warn claim. West involved a claim by an ironworker against the manufacturer of a wire sling. The plaintiff was injured when the wire sling broke. 197 N.W.2d at 208. Plaintiff alleged negligent failure to warn and negligent testing of the product. Id. Both claims were submitted to the jury, which returned a substantial verdict in plaintiff's favor on both claims. Id. On appeal, we found no error in submission of the failure to warn claim. However, we concluded that plaintiff's negligent testing claim was really a failure to warn claim and that the trial court therefore erred in submitting the negligent testing claim to the jury. Id. at 214. In doing so, we made the following observations:

Failure to test arises in two main factual situations: where adequate tests for defects are not conducted in production and an article containing a defect is marketed, and where the particular article is not defective but the properties of the product in general are not adequately tested before it is released to the public.

Id. at 212.

We said that plaintiff's negligent testing claim in West fell into the latter category, more thoroughly described as follows:

As to the second situation—testing to ascertain the properties of a product— the commonest case is one in which a new product such as a compound is launched on the market without adequate testing to ascertain harmful effects when the product is used in various ways.

Id. at 213. We concluded that the evidence revealed that the defendants had conducted tests and experiments that showed the safety capacities of its wire slings, which was the source of plaintiff's negligent testing claim. Id. We then said that the defendant's "neglect, if there was neglect, was not in failing to ascertain information on capacities by testing, but in failing adequately to bring that information home to users." Id. We also noted in West that the evidence showed that the ironworkers would not have used the sling had they been warned of its rated capacity. Id. Thus, the relevant negligence on defendant's part was failure to give such warning, not failure to test for such dangers. Id. at 213-14. We therefore concluded that the specification of negligence based on a failure to test should not have been submitted to the jury.

2. Application of law to facts.

We examine the evidence to determine whether plaintiffs actually submitted proof of a material failure to test on the part of BRK, or whether plaintiffs merely proved a failure to warn claim.

Plaintiffs' expert, Dr. Russell, was critical of the ability of the UL 217 standard to test the effectiveness of the model 83R to sense certain types of household fires. Plaintiffs therefore argued that had BRK submitted the model 83R to tests in addition to the UL 217 standard, it would have learned that the model 83R and all ionization smoke detectors are not effective in responding to certain types of fires. The record shows, however, that BRK was aware of the limitations of ionization detectors and that this information was widely known in the industry. In fact, BRK had used the knowledge it had gained from testing to develop and market a combination detector. Thus, BRK has performed the sort of testing that plaintiffs complain should have been performed and has also developed a product to remedy the alleged defect present in an ionization detector.

We therefore conclude that the evidence presented by plaintiffs in support of their negligent testing claim is in reality a negligent failure to warn claim. BRK had knowledge of the performance limitations of ionization smoke detectors in detecting certain types of fires. Thus, any amount of further testing by BRK would not have increased BRK's knowledge of the effectiveness of ionization detectors to alarm to certain types of fires. Cf. Ackerman, 586 N.W.2d at 223 (Ternus, J., dissenting) (noting that defendant "conducted tests that showed the very information material to the plaintiff's harm" and that plaintiff's negligent testing claim was really a challenge to the information contained on the product label). Therefore, BRK's "neglect, if there was neglect, was not in failing to ascertain information on [the limitations of an ionization detector], but in failing adequately to bring that information home to users." West, 197 N.W.2d at 213. Based on this analysis, plaintiffs' negligent testing claim should not have been submitted to the jury.

VII. Fraudulent nondisclosure.

The jury was instructed on plaintiffs' claim of fraudulent nondisclosure. BRK, however, did not raise any issue in its initial brief concerning plaintiffs' fraudulent nondisclosure claim. We therefore need not consider any error with respect to this claim. See Iowa R.App. P. 14(a)(3) (failure in brief to state or argue an issue may be deemed waiver of the issue); Hollingsworth v. Schminkey, 553 N.W.2d 591, 596 (Iowa 1996).

VIII. Admission of expert witness testimony.

During trial, plaintiffs offered evidence of expert witness testimony concerning how smoke from the Mercer fire flowed from the point of origin to the smoke detector and concerning the response time of ionization smoke detectors. An expert witness for plaintiffs also testified regarding BRK's pricing of its smoke detectors. BRK objected to admission of the conclusions reached by the expert witnesses on grounds that their testimony was incompetent and unreliable and contends on appeal that the district court erred in allowing the testimony of plaintiffs' expert witnesses.

A. Standard of review and rules regarding admission of expert testimony.

We ordinarily will not reverse a district court's decision concerning admission of evidence absent an abuse of discretion to the prejudice of the complaining party. Leaf, 590 N.W.2d at 531. "Only in a clear case of abuse will the admission of such evidence be found to be prejudicial." State v. Atwood, 602 N.W.2d 775, 783 (Iowa 1999), cert. denied, ___ U.S. ___, 120 S.Ct. 1729, 146 L.Ed.2d 649 (2000).

We stated in Mensink v. American Grain, Related Industries, 564 N.W.2d 376, 380 (Iowa 1997), that "we are committed to a liberal view on the admissibility of expert testimony, and we have been quite deferential to the district court in the exercise of its discretion in that area." We also explained in Leaf that trial courts are not required to apply the analysis set forth in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), concerning the admission of expert testimony, but that trial courts may find it helpful in complex cases to use one or more of the relevant Daubert "considerations" in assessing the reliability of expert testimony. Leaf, 590 N.W.2d at 533. We then articulated the proper test for admission of expert testimony in Iowa as set forth in Iowa rule of evidence 702. Id. First, the evidence, of course, must be relevant. Iowa R. Evid. 402. Second, the evidence must be in the form of "scientific, technical, or other specialized knowledge [that] will assist the trier of fact to understand the evidence or to determine a fact in issue." Iowa R. Evid. 702. Third, the witness must be qualified as an expert by knowledge, skill, experience, training, or education. Id.; accord Leaf, 590 N.W.2d at 533. Other evidence might be so novel or complex that the court will require proof of acceptance of the theory or technique in the scientific community before the evidence is admissible. Leaf, 590 N.W.2d at 534.

B. Evidence of expert witnesses regarding Mercer fire.

Dr. Patrick Pagni testified on behalf of plaintiffs concerning results of an experiment he performed to demonstrate the flow of smoke in the Mercer fire from the point of origin to the smoke detector. Dr. Jose Torero also presented testimony of behalf of plaintiffs concerning results of an experiment he supervised involving the response time of ionization smoke detectors in detecting smoke from a fire.

Upon our review, we find no abuse of discretion by the district court in allowing into evidence the testimony of Drs. Pagni and Torero. The record shows that both witnesses were qualified experts in the area of fire safety engineering. Plaintiffs also presented an adequate foundation for the subject matter of each expert witness. Additionally, BRK's own expert witnesses criticized the results and reasoning of plaintiffs' experts and BRK's experts also presented their own opinions concerning the smoke pattern of the Mercer fire and the response time of ionization detectors to certain types of fires. The jury could evaluate the testimony of the expert witnesses and make its own conclusions, based on all evidence presented, concerning the Mercer fire and the effectiveness of ionization detectors in responding to certain types of fires. No error appears here.

C. Testimony regarding pricing of BRK smoke detectors.

Over BRK's objection based on relevancy, Dr. Russell testified on behalf of plaintiffs regarding a report prepared by BRK showing the evaluation of the internal costs of production of BRK's smoke detectors as part of a cost-benefit analysis.

Based upon our review of the record, it appears that BRK did not make a timely objection to Dr. Russell's testimony on the pricing of BRK's detectors. The objection was made after Dr. Russell's comments were made. BRK therefore failed to preserve error concerning this issue. In any event, evidence of BRK's pricing of smoke detectors was relevant to the risk-utility analysis under plaintiffs' strict liability claim. No error exists here.

IX. Future medical expenses of Travis.

BRK contends that the district court erred when it simply substituted the amount of $397,113.76 for the jury's $1 million award for Travis's future medical expenses. We express no opinion as to an appropriate amount for Travis's future medical expenses because that issue will be decided anew upon remand and retrial.

X. Viability of recovery under tort.

BRK contends that the district court should have sustained its motion for directed verdict because the Mercers failed to state a claim in tort. As outlined in BRK's brief, this contention seems to be based on the argument that plaintiffs are limited to a contract theory of recovery because their complaint about the model 83R smoke detector is that it does not perform as well as a combination smoke detector, which BRK says is a contract-based theory and for which recovery under tort is not available. Upon our review, we conclude that BRK did not raise this specific contention in its motion for directed verdict. We therefore conclude that BRK failed to preserve error on whether plaintiffs are limited to a contract-based theory of recovery and we need not consider the merits of that issue on appeal. See Lamb, 570 N.W.2d at 67 (stating that a motion for judgment notwithstanding the verdict pursuant to Iowa rule of civil procedure 243 must stand on grounds raised in the movant's motion for directed verdict and review on appeal is limited to grounds raised in the directed verdict motion).

XI. Disposition.

We conclude the district court committed reversible error by admitting all 363 consumer complaints into evidence. Only those consumer complaints received by BRK prior to the Mercer fire are admissible, assuming plaintiffs meet their burden of showing that each of the complaints is substantially similar to the facts and circumstances of the Mercer fire situation.

We further conclude that plaintiffs failed to prove that BRK's conduct was willful and wanton. Consequently, the district court erred in submitting the issue of punitive damages to the jury.

We conclude the Mercers presented substantial evidence concerning the elements of a strict liability claim to justify submitting that claim to the jury. Based on the evidence presented, a fact question to be answered by the jury was created concerning whether defendant BRK proved that its model 83R is state of the art. The district court also properly submitted plaintiffs' breach of implied warranty claim as well as plaintiffs' strict liability claim.

No reversible error is found in submission to the jury of the theories of negligent failure to warn and negligent design. The Mercers' negligent testing claim should not be submitted to the jury upon remand.

We affirm in part and reverse in part the district court rulings. The judgment is reversed and the case is remanded for a new trial consistent with this opinion.

Costs on appeal are taxed three-fourths to the Mercers and one-fourth to BRK.

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED.

All justices concur except NEUMAN, J., who takes no part.

NOTES

[1] Northwest Bank & Trust Company, the conservator for Travis Mercer, was also named as a plaintiff. For convenience, we will refer to the Mercers as the primary plaintiffs and to BRK as the primary defendant.

[2] The Mercers brought their claims against defendant BRK after settling their claims against Gerry, the manufacturer of the baby monitor.

[3] The State of Iowa ex rel. Civil Reparations Trust Fund intervened to claim a share of any possible punitive damages award. See Iowa Code § 668A.1(2)(b) (1995).

[4] Although some authorities suggest that the negligence/strict liability distinction in design defect cases should be abandoned, see Keith Miller, Design Defect Litigation in Iowa: The Myths of Strict Liability, 40 Drake L.Rev. 465 (1991), we continue to recognize that distinction. See Lovick, 588 N.W.2d at 698-99. The argument that the distinction between design defect cases in strict liability and negligence should be abandoned finds support in the fact that the Restatement (Third) of Torts: Product Liability (1997), dropped the strict liability/negligence distinction. Lovick, 588 N.W.2d at 698 (citing Restatement (Third) of Torts: Products Liability § 2 cmt. n). The Restatement (Third) also eliminated the consumer expectation test traditionally used in the strict liability analysis. Id. The Restatement now follows the risk-utility approach traditionally found in negligence analysis, which in most instances will require proof of a reasonable alternative design to establish a design defect. Id.

The parties do not argue that we should abandon the distinction between strict liability and negligence in design defect cases or that we should otherwise adopt the rules of the Restatement (Third) of Torts concerning product liability claims. Therefore, we will apply the principles of section 402A of the Restatement (Second) of Torts for purposes of this case.

[5] Cf. Butler v. Pittway Corp., 770 F.2d 7, 11 (2d Cir.1985) (holding that plaintiffs "are entitled to an opportunity to show that the alleged failure of the smoke detector to sound a timely alarm exposed them to an unreasonably dangerous condition, and that their damages were attributable to this alleged defect"; stating that a "malfunctioning smoke detector can create an unreasonable risk of harm in that inhabitants of a structure who rely on alarm may be lulled into an unjustified sense of safety and fail to be forewarned of the existence of a fire"). But see Werner v. Pittway Corp., 90 F.Supp.2d 1018, 1032 (W.D.Wis.2000) (holding that plaintiff's claim that battery-operated ionization smoke detectors are defective because design problems reduce a smoke detector's ability to detect slow, smoldering fires fails because battery-operated ionization smoke detector installed in plaintiff's bedroom hallway sounded an alarm during the fire in plaintiff's home and provided plaintiff's family adequate time to escape).

[6] The district court submitted the state of the art defense in a separate verdict form. See Hillrichs v. Avco Corp., 478 N.W.2d 70, 76 (Iowa 1991) (Hillrichs I) (recommending that state of the art defense be submitted by way of special verdict and that defendant must establish the defense with respect to the specific claims made by plaintiff), abrogated on other grounds by Reed v. Chrysler Corp., 494 N.W.2d 224, 230 (Iowa 1992).

[7] In its motion for directed verdict and in its post-trial motion, BRK argued that plaintiffs failed to prove causation on their failure to warn claim because according to Jennifer's testimony, concerning the smoke conditions when she first went upstairs, a sufficient amount of smoke had not reached the smoke detector to trigger the alarm. BRK does not raise this issue on appeal, and we therefore need not address it.

[8] We indicated above which type of consumer complaints should have been admitted.