LEACH, J.

¶ 1 Mary Haviland appeals a trial court's decision invalidating her deceased husband's will as the product of her undue influence. Mary1 claims that the trial court should not have applied the factors identified in Dean v. Jordan2 because they "have no meaningful application between a husband and wife." She also assigns error to the court's findings of fact and conclusions of law. Because our Supreme Court has applied Dean to analyze a claim of a spouse's undue influence, the trial court did not err by applying it here. And because substantial evidence in the record supports the trial court's written findings of fact and the conclusions that flow from them, we find no error and affirm.

FACTS

¶ 2 Haviland was born on July 18, 1911. He enjoyed a long and distinguished medical career. From the 1940s to the 1970s, he was a leader at the University of Washington School of Medicine, serving as an assistant dean, a clinical professor, and an associate dean, while maintaining a successful private medical practice. In 1962, he cofounded the Northwest Kidney Centers. Haviland and his first wife, Marion, had four children together.

¶ 3 When Marion died in 1993, the couple's assets were distributed according to the James W. Haviland and Marion B. Haviland Revised and Restated Revocable Trust, dated June 26, 1990 ("the 1990 trust agreement"). A primary purpose of this trust was "to provide common protection to the trustors against the effects of age and their increased susceptibility to the suggestions of others."3

¶ 4 Pursuant to the 1990 trust agreement, $600,000 of Marion's separate property funded a Credit Shelter Trust, Haviland's separate property and his half of the community property funded a Survivor's Trust, and the balance of Marion's separate property funded a Marital Trust. Upon Marion's death, the Credit Shelter Trust and Marital Trust became irrevocable, although Haviland could withdraw 100 percent of the principal from the Marital Trust. The Survivor's Trust remained revocable. The Credit Shelter Trust provided that upon the surviving spouse's death, the three trusts would continue for the benefit of the Haviland children and grandchildren until the death of all the children or until the youngest grandchild turned 30, whichever occurred later. Any remainder would be distributed to charitable beneficiaries.

¶ 5 In 1996, while recuperating at Providence Hospital from a leg injury, then 85-year-old Haviland met then 35-year-old Mary, a hospital nurse assistant. Haviland and Mary continued to see each other after his release from the hospital. A short time later, Haviland gave $10,000 to Mary to help pay for her education and living expenses. Three months later, Haviland agreed to pay $100,000 toward Mary's educational expenses and an additional $300,000 to $350,000 as a "nest egg." Another three months later, Haviland created the James W. Haviland Living Trust ("Living Trust"), naming himself as the beneficiary during his lifetime. Upon his death, the trust was to pay up to $500,000 to Mary for her education and living expenses and distribute the balance, if any, according to the 1990 trust agreement.

¶ 6 Haviland and Mary married in August 1997. The couple executed a prenuptial agreement that maintained the separate nature of Haviland's property. According to this agreement, Haviland had assets valued at more than $3 million, including real property on Shaw Island, Bremerton, and Canim Lake plus retirement accounts, the Survivor's Trust, charitable remainder trusts, the Living Trust, and various bank accounts. He also received substantial income from the trusts that he and Marion had established. Mary had negligible assets.

¶ 7 The following year, Haviland removed the limit on Mary's inheritance under the Living Trust, thereby eliminating the Survivor's Trust as the remainder beneficiary of that trust. Then, in 1999, Haviland transferred $765,000 from the Survivor's Trust to the Living Trust. Haviland's children were the remainder beneficiaries of the Survivor's Trust; Mary was the remainder beneficiary of the Living Trust.

¶ 8 The following year, Haviland amended the Living Trust to add Mary as a cotrustee. Haviland remained the sole beneficiary of that trust, and he continued to fund it with his separate property, the only source of funds for that trust. After this amendment, Haviland and Mary jointly approved all transactions relating to the Living Trust.

¶ 9 Over the course of their marriage, millions of dollars of Haviland's separate assets were transferred from the Living Trust into the couple's joint checking account, Mary's separate checking account, or Mary's separate line of credit. In turn, bank statements document the withdrawal of millions of dollars from the joint checking account. The trial court found little evidence as to the ultimate purpose for which the money withdrawn from the joint checking account was used. Haviland also conveyed two parcels of his separate real property to Mary as her separate property. In addition, Haviland's retirement accounts were cashed in, and substantial sums of money were gifted to Mary's children from a previous marriage and to other designees. Haviland did not make comparable gifts to his own children.

¶ 10 Meanwhile, Haviland's physical health substantially deteriorated. In 2002, Haviland changed primary care physicians and indicated on the new patient registration form that he was having memory problems. Mary also filled out a new patient registration form for Haviland identical to the one Haviland completed, except she omitted any indication of his memory impairment. The new physician testified that he likely relied on the form Mary provided because that form contained that physician's signature. This physician did not evaluate Haviland's mental state at that time.

¶ 11 In 2005, Mary quit her job to care full time for Haviland. Mary explained that Haviland began refusing care from Mary's daughter, who had been hired to care for Haviland while Mary was at work. Mary reported that

Mary claimed that these symptoms abated within two weeks of her staying home to care for Haviland.

¶ 12 Before his marriage to Mary, Haviland was known as a "frugal" man, who made generous gifts to education, the arts, and charitable organizations. During his marriage to Mary, he made four revisions to his estate plan. Each change resulted in a greater portion of his estate going to Mary and less going to his children and designated charities.

¶ 13 The last major revision occurred in 2006, when Mary phoned Alan Kane, an attorney at K & L Gates LLP, to advise him that Haviland wished to change his will. After the phone call, Mary typed a letter to Kane and enclosed a copy of Haviland's 2002 will with requested revisions. Mary's handwritten revisions, initialed by Haviland, provided that (1) she would inherit the personal property in the Mercer Island residence, previously given to Haviland's children; (2) the remainder of Haviland's estate would pass to the Living Trust for Mary's benefit and not to the Credit Shelter Trust; (3) $55,000 of the previous $105,000 given to specific individuals and charities would be eliminated; and (4) certain waiting periods would be reduced from four months to two months. These changes effectively disinherited Haviland's children, leaving them with only a right of first refusal for the Shaw Island residence.

¶ 14 Three days before Haviland signed the 2006 will, Mary took him to his physician. This was his first visit with his physician since 2003. Mary informed the physician that Haviland's "mentation was good."

¶ 15 On the day of the will signing, Mary brought Haviland to Kane's office. Kane testified that he met with Haviland for five minutes before the will signing but that he did not discuss with him his family, the objects of his bounty, or ask Haviland questions about the nature and extent of his estate. Kane recalled having last met with Haviland in 2002, and Kane noticed a moderate decline in Haviland's mental functioning. According to Kane, the only thing Haviland said during this meeting was "yes" in response to questions about the signing. Nevertheless, Kane testified to his opinion that Haviland possessed testamentary capacity.

¶ 16 By 2007, Haviland's mental condition deteriorated to the point that he could not recognize Mary 75 percent of the time. On November 6, Haviland visited the emergency room for dehydration. The emergency room consultation report describes Haviland as suffering from advanced dementia. A few days later, Mary prepared a request, signed by Haviland, to pay his debts, including debts owed jointly with Mary, with principal from the Credit Shelter Trust. The trust manager denied this request. Haviland died approximately one week later.

¶ 17 After the court admitted Haviland's 2006 will to probate, three of Haviland's children commenced this will contest. They alleged that Haviland lacked testamentary capacity at the time he signed the 2006 will and that the will was a product of undue influence. A lengthy bench trial followed.

¶ 18 At trial, Dr. Elaine Peskind, a geriatric psychiatrist, testified about her review of Haviland's records and her "retrospective" analysis. In her professional opinion, Haviland began exhibiting symptoms of Alzheimer's disease as early as 2000 and had progressed to the early stages of the disease by 2002. Peskind identified witness statements indicating that by 2004, Haviland required cueing for his basic daily activities. Peskind stated that Haviland could not perform the most basic activities of daily living without assistance. Peskind also testified that Haviland lacked the ability to form independent thoughts and likely suffered from severe, advanced Alzheimer's at the time of his death.

¶ 19 The trial court concluded that respondents failed to meet their burden of proof on the issue of testamentary capacity. However, the court determined that clear, cogent, and convincing evidence established that the 2006 will was the product of undue influence by Mary. The court set that will aside, admitted the 2002 will into probate, removed Mary as personal representative, and appointed a new administrator of Haviland's estate.

¶ 20 Mary appeals.

ANALYSIS

Undue Influence

¶ 21 Mary asserts that the trial court used the wrong legal standard to determine whether the 2006 will was a product of undue influence. We review this legal question de novo.4

¶ 22 "The law presumes the validity of a rational will."5 However, a will executed by a person who otherwise possesses testamentary capacity may be set aside if the beneficiary exercised undue influence over the testator.6 To invalidate a will on this basis, the court must find that the beneficiary exerted sufficient influence at the time of the testamentary act to have "`controlled the volition of the testator, interfered with his free will, and prevented an exercise of his judgment and choice.'"7 The party challenging the will must prove undue influence by clear, cogent, and convincing evidence.8

¶ 23 The clear, cogent, and convincing standard requires evidence that convinces the trier of fact that the fact in issue is "`highly probable.'"9 "In determining whether the evidence meets the `clear, cogent, and convincing' standard of persuasion, the trial court must make credibility determinations and weigh and evaluate the evidence."10 We do not review the trial court's credibility determinations.11

¶ 24 In Dean, our Supreme Court announced that a combination of suspicious facts and circumstances may give rise to a rebuttable presumption of undue influence.12 The "most important" suspicion-raising facts include

Once the evidence raises this presumption, the burden shifts to the will proponent to rebut it with "evidence sufficient at least to balance the scales and restore the equilibrium of evidence touching the validity of the will."14 In the absence of rebuttal evidence, the evidence raising the presumption may be sufficient to invalidate the will.15

¶ 25 Here, the trial court applied Dean and invalidated the will as the product of Mary's undue influence. It found that Mary was Haviland's fiduciary, participated in the creation of the 2006 will, and received an unnaturally large share of Haviland's estate in comparison to his earlier estate plans. It also found that Haviland was extremely vulnerable to undue influence due to physical disabilities, some cognitive impairment, and Mary's position as his primary caregiver. Finally, it found that Mary depleted his estate through a systematic, persistent, and largely unexplained pattern of transferring assets from Haviland's estate for her benefit and that of her children and other designees.

¶ 26 Mary argues that the Dean factors should not apply when a will contestant asserts undue influence against a spouse. Mary bases her claim upon two observations: that a common trait of marriage is that each spouse occupies a fiduciary relationship with respect to the other and that spouses routinely participate in the preparation and procurement of each other's wills.

¶ 27 Mary's argument does not persuade us to depart from Dean for two reasons. First, Mary's argument cannot be reconciled with In re Estate of Lint.16 In Lint, our Supreme Court described Dean as a scholarly opinion and affirmed the trial court's use of the Dean factors to determine whether the decedent's spouse exerted undue influence to procure her will. The Lint court specifically discussed the suspicion-raising factors Dean identified as most important.17 Because we are bound by Lint and Mary provides no persuasive basis for distinguishing it, we conclude that the Dean factors apply here.

¶ 28 Second, precedent contradicts Mary's suggestion that spouses are at greater risk than other beneficiaries of having a will invalidated under Dean. In the case of In re Estate of Beck,18 which Dean relies upon, our Supreme Court explained, "`No presumption of undue influence invariably arises from the fact that a [beneficiary actively participated in the preparation or procurement of the will].'" Rather, "`it is a fact to be considered with other facts. It is undoubtedly a suspicious fact, but its weight depends, not solely upon its character, but upon the facts and circumstances with which it is connected.'"19 Dean, and the case law interpreting it, makes clear that the same analysis applies to each Dean factor. Consequently, no single Dean factor is determinative. Instead, Dean requires that each case be decided based upon a "combination of facts shown by the evidence in a particular case [to] be of such [a] suspicious nature as to raise a presumption of . . . undue influence."20 Without such evidence, an undue influence challenge will fail.21 Accordingly, Dean does not put spouses at a substantially greater risk of a will being set aside than nonspouse beneficiaries.

¶ 29 We therefore hold that the factors and presumption identified in Dean apply when assessing the conduct of a spouse in a will contest.

Findings of Fact and Conclusions of Law

¶ 30 Mary assigns error to 5 of the trial court's 135 findings of fact, arguing that substantial evidence does not support them. She also assigns error to 3 conclusions of law.

¶ 31 Where, as here, some findings are actually conclusions of law or mixed findings of fact and conclusions of law, we review the factual components under the substantial evidence standard and the conclusions of law, including those mistakenly characterized as findings of fact, de novo.22 Substantial evidence is that "quantum of evidence sufficient to persuade a rational fairminded person the premise is true."23 "[W]here there is conflicting evidence, the court needs only to determine whether the evidence viewed most favorable to respondent supports the challenged finding."24

¶ 32 Mary challenges the following findings of fact:

To the extent that these challenged findings contain factual determinations, substantial evidence in the record supports them.

¶ 33 Here, Haviland was admitted to the emergency room for dehydration on November 6, 2007. The emergency room consultation states, "[Haviland has] advanced dementia. He is essentially bed ridden [sic] at home." It also indicates that Mary was his primary and complete caregiver and that she provided Haviland's medical history. Unchallenged finding of fact 124 states,

Unchallenged findings are verities on appeal.25 Thus, these findings, together with the testimony of Dr. Peskind described earlier, when viewed most favorably to the Haviland children, provide ample support for a finding that Haviland suffered from advanced dementia before his death. And because the request to pay debt with the principal from the Credit Shelter Trust came two days after Haviland's admission to the emergency room, it also sufficiently supports a finding that Haviland "was suffering from dementia prior to th[at] request" and that Mary made the request "at a time when Dr. Haviland clearly lacked capacity to make that decision."

¶ 34 Abundant evidence in the record also supports the court's finding that Mary depleted Haviland's assets. The parties do not dispute that Haviland's separate property provided the only funding for the Living Trust, and numerous unchallenged findings of fact detail the large sums repeatedly transferred to Mary from the Living Trust and the joint checking account with no credible explanation as to how she used the money.

¶ 35 Similarly, the 2006 will reflected a steadily increasing share of Haviland's estate being gifted to Mary and away from his children and charities. In Haviland's 1997 will, he devised the Shaw Island property to Mary, provided that she did not predecease him. If she did, the property passed to his children. The residue of his estate passed to the Survivor's Trust in accordance with the 1990 trust agreement. Haviland's children and various charities were the lifetime beneficiaries of that trust. But in 1998, Haviland executed a new will. That will devised the Bremerton home and Shaw Island property to Mary, stipulated that Mary's children would inherit the Shaw Island property even if she predeceased Haviland, and eliminated certain charitable bequests. Then, in 2002, Haviland withdrew his assets from the Marital Trust, revoked the Survivor's Trust, and transferred those assets to the Living Trust for Mary's benefit. Haviland also executed a will that devised the residue of his estate to the Credit Shelter Trust. Haviland's children were the beneficiaries of that trust. But in 2006, Haviland executed a will that devised the residue of his estate to the Living Trust, thereby effectively disinheriting his children, and cut bequests to charitable organizations by nearly half, increasing the remainder bequest to Mary.

¶ 36 For the foregoing reasons, we are satisfied that the findings of fact are supported by substantial evidence. Because the remaining challenged findings are not necessary to the undue influence determination, we do not address them.

¶ 37 Next, we examine the findings in view of the factors set forth in Dean to determine whether the court erred in entering the challenged conclusions of law. The challenged conclusions state,

After reviewing the findings, we are satisfied that they support the conclusions of law.

¶ 38 Notably, conclusion of law 9 succinctly addresses the Dean factors. The first Dean factor requires that Mary be Haviland's fiduciary. The fiduciary duty of loyalty prohibits the use of the principal's property for the benefit of the trustee.26 A spouse is not relieved of fiduciary duties when she serves as a trustee for her husband.27 Finding of fact 29 states that Mary was a cotrustee on the Living Trust and that Haviland was the only named beneficiary of that account. Therefore, clear, cogent, and convincing evidence establishes that Mary occupied a fiduciary relationship to Haviland.

¶ 39 The second inquiry under Dean is whether Mary participated in the procurement of the 2006 will. As indicated before, the weight of this factor "`depends, not solely upon its character, but upon the facts and circumstances with which it is connected.'"28 And although there may be circumstances in which the beneficiary's mere participation in the procurement of the will is insufficient to cast suspicion upon it,29 such is not the case here. Finding of fact 72 states that Mary called Kane to inform him that Haviland wanted to make revisions to his 2002 estate plan. Finding of fact 73 indicates that Mary typed a letter to Kane and enclosed a copy of the previous will upon which Mary interlineated draft changes. Mary does not challenge these findings of fact. Further, the 2006 will revision comes on the heels of nearly a decade-long campaign of draining Haviland's estate and at a time when Haviland's mental and physical faculties were clearly declining. Thus, Mary's participation in the drafting of the 2006 will supports the conclusion that there is clear, cogent, and convincing evidence of her undue influence.

¶ 40 The third Dean factor, examines whether the 2006 will was unnatural in comparison to earlier estate plans. "A will is unnatural when it is contrary to what the testator, from his known views, feelings, and intentions would have been expected to make."30 Courts determine naturalness by looking to a testator's prior wills.31 In this case, it is clear that each change to Haviland's estate plan resulted in Mary receiving more of his estate while less went to his children and designated charities. The 2006 will also reduced bequests to charitable organizations for Mary's obvious benefit, conflicting with Haviland's reputation for charitable giving. We therefore conclude that there is clear, cogent, and convincing evidence that the 2006 will was unnatural.

¶ 41 The remaining Dean factor, whether additional justification exists for setting aside a will as product of undue influence, is also present. Here, the court relied on two considerations: (1) that Haviland was extremely vulnerable to Mary's undue influence and (2) that Mary engaged in a systematic and persistent pattern of draining assets from Haviland's estate for her own benefit. Both are supported by clear, cogent, and convincing evidence. Haviland experienced substantial physical disabilities, exhibited symptoms of dementia as early as 2000, required Mary's full-time care in 2005, and had advanced dementia at the time of his death in 2007. Unchallenged findings describe in detail the many large asset transfers without explanation as to their purpose.

¶ 42 In summary, the petitioners presented sufficient evidence to raise a presumption of undue influence. The burden then shifted to Mary to produce credible evidence sufficient to "balance the scales and restore the equilibrium of evidence touching the validity of the will."32 Mary challenges the trial court's determination that she failed to do so. She submits she presented testimony from legal professionals and numerous disinterested witnesses sufficient to restore the equilibrium.

¶ 43 Specifically, Mary argues she presented evidence that Haviland was alert and in good mental health at the time of the will signing. But this argument does not survive scrutiny. "[E]vidence of testamentary capacity is not inconsistent with the conclusion that undue influence had overmastered free agency."33 Thus, discrepancies in the record regarding testamentary capacity do not necessarily rebut the presumption of undue influence, especially where, as here, overwhelming evidence demonstrates that the testator suffered at least some decline in his mental faculties and depended on the beneficiary for his care.

¶ 44 Next, Mary alleges that her participation in the procurement of the 2006 will was no more than would be expected of any spouse under similar circumstances. The relevant inquiry, however, is not whether Mary involved herself to the same degree as any other spouse, but whether her participation in the preparation and execution of the will, in connection with other facts and circumstances, supports a presumption of undue influence. Mary's analysis of this factor in isolation ignores Dean's instruction that each case be decided based upon the combination of facts established in that case. Thus, her argument fails.

¶ 45 Mary also claims that the 2006 will is consistent with Haviland's desire, stated 10 years earlier, to provide substantial benefit to her. While it is true that Haviland provided for Mary in earlier estate plans, it is also true that those plans devised the majority of his estate to his children and various charities. For this reason, Mary's claim that Haviland wished to provide substantial benefit to Mary is insufficient to rebut the presumption of undue influence.

¶ 46 Lastly, Mary argues that the trial court erred by relying on a pattern of asset depletion because that factor is not listed among the Dean criteria. But nothing in Dean suggests that its list of suspicion-raising factors is comprehensive or exclusive. To the contrary, Dean specifically states that the circumstances identified are "the most important" and that "other considerations" may be taken into account.34 Moreover, a decade-long campaign of financial exploitation corroborates a finding of undue influence, especially where the beneficiary engaged in a large number of self-dealing transactions that exhausted a substantial estate without any credible explanation.

¶ 47 Conclusion of law 11 states that clear, cogent, and convincing evidence establishes that at the time of the 2006 will Mary exerted undue influence over Haviland. We agree. There is clear, cogent, and convincing evidence in the record that Mary "`controlled the volition of the testator, interfered with his free will, and prevented an exercise of his judgment and choice.'"35

Attorney Fees

¶ 48 Mary asks this court to reverse the trial court's award of attorney fees and costs. She makes this request contingent on her success in this appeal. Because she does not prevail, we affirm the trial court's award of fees and costs in favor of the respondents.

¶ 49 The respondents request attorney fees on appeal and ask that the fees be paid by Mary rather than the estate. RCW 11.96A.150 and RCW 11.24.050 grant this court discretion to award fees in this case. Because it is equitable to do so, we grant their request. The respondents are awarded their reasonable fees and costs upon compliance with the applicable Rules of Appellate Procedure.

CONCLUSION

¶ 50 Because the trial court properly applied the Dean factors when analyzing the undue influence claim asserted against Mary and because substantial evidence in the record supports the trial court's written findings of fact and the conclusions that flow from them, we find no error and affirm.

WE CONCUR: DWYER, C.J. and BECKER, J.