John Johnston appeals a district court 1 order finding that Prudential Insurance Company of America ("Prudential") did not abuse its discretion when it terminated his long term disability benefits. We affirm the district court's order.

I. Background

Johnston was an Enterprise Storage Engineer in the computer department at Commerce Bancshares, Inc. ("Commerce"). As part of an employee welfare benefit plan (the "Plan"), Commerce provided its employees long-term disability ("LTD") insurance from Prudential.

In July 2013, Johnston became unable to continue working due to complications from hydrocephalus, which ultimately led to surgery to remove a colloid cyst from his brain. Johnston filed a claim for LTD benefits with Prudential. Dr. Kala Danushkodi, Johnston's treating physician, submitted a statement that Johnston had "cognitive impairment / moderate to severe" and was unable to return to work due to the impairment.

Prudential sent Johnston a letter approving his claim for LTD benefits in November 2013. Prudential also requested the results of two neuropsychological examinations "for the ongoing review of your claim and benefits beyond December 31, 2013." It further advised that it would "periodically review your claim, and request or obtain information, to ensure that you meet all eligibility requirements."

After receiving and reviewing the results of Johnston's examinations, Prudential staff noted that one of the tests was not valid due to Johnston's inconsistent performance. After Johnston underwent an additional surgery in March 2014 to place a shunt in his head, Prudential decided that another neuropsychological evaluation was needed to determine whether he continued to be disabled.

Neuropsychologist Dr. Robert Denney examined Johnston in June 2014. Dr. Denney used multiple tests for the validity of Johnston's responses, both "embedded" and "free-standing." 2 He was unable to *714 determine whether Johnston was cognitively impaired because Johnston failed almost all of the validity tests. Dr. Denney opined that two of the free-standing validity tests indicated that Johnston "was actively attempting to perform poorly." In a supplemental addendum to his report, Dr. Denney reviewed the data from two of Johnston's previous examinations. He did not change his conclusions about Johnston because one examination had failed validity indicators, while the other examination had inconsistent results that suggested invalidity.

Based on Dr. Denney's report and addendum, Prudential terminated Johnston's LTD benefits as of September 1, 2014. It determined that Johnston had failed to support his claim that he was still unable to work due to cognitive impairment.

Johnston appealed the termination of his LTD benefits. In support of his appeal, he submitted a statement from his therapist, Dr. Marcia Meyer, explaining that he was exhausted by Dr. Denney's tests and that he was unable to maintain the focus and concentration needed for his job.

After reviewing Johnston's appeal, Prudential sought a second neuropsychological examination. Dr. Michelle Zeller examined Johnston in June 2015, and she reported that he failed all nine validity measures on the tests she administered. She concluded that he was attempting to appear more impaired than he actually is, and she stated that she was unable to determine his level of impairment. Dr. Zeller explained: "Failure on any one of these measures would raise the possibility of negative response bias, suboptimal effort and/or symptom exaggeration. Failure on all nine, however, is compelling evidence of suboptimal effort."

Prudential upheld its denial of LTD benefits, and Johnston sued Prudential under 29 U.S.C. § 1132 (a)(1)(B), a part of the Employee Retirement Income Security Act of 1974 ("ERISA"). On cross-motions for summary judgment, the district court granted summary judgment to Prudential and denied Johnston's motion. Johnston filed a motion to reconsider, which the court denied on the basis that it did not contain any new evidence or arguments not previously available. Johnston timely appealed.

II. Standard of Review

"We review the district court's adjudication of [an ERISA] claim de novo, applying the same standard of review to the plan administrator's decision as the district court." McClelland v. Life Ins. Co. of N. Am. , 679 F.3d 755 , 759 (8th Cir. 2012). "When an ERISA-qualified employee benefit plan grants the plan administrator the discretion to determine whether a claimant is eligible for benefits, review of the administrator's decision is for an abuse of discretion." Id. "The administrator's decision should be affirmed if it is reasonable, meaning it is supported by substantial evidence." Green v. Union Sec. Ins. Co. , 646 F.3d 1042 , 1050 (8th Cir. 2011). "Substantial evidence is more than a scintilla but less than a preponderance." Id. "[W]hen a conflict of interest exists because the plan administrator is both the decision-maker and the insurer, 'we take that conflict into account and give it some weight in the abuse-of-discretion calculation.' " Nichols v. Unicare Life & Health Ins. Co. , 739 F.3d 1176 , 1181 (8th Cir. 2014) (quoting Carrow v. Standard Ins. Co. , 664 F.3d 1254 , 1259 (8th Cir. 2012) ).

III. Analysis

Johnston acknowledges that the Plan gives Prudential discretion to determine *715 eligibility for benefits and that, as a result, an abuse of discretion standard applies. Thus, the question on review is whether Prudential abused that discretion.

The Plan placed the burden on the beneficiary to provide proof of disability, including continuing disability. The Plan states, "We will stop sending you payments and your claim will end on the earliest of the following:" including, among other things, "[t]he date you fail to submit proof of continuing disability satisfactory to Prudential." Because ERISA allows a beneficiary to sue "to recover benefits due to him under the terms of his plan ," a plan may place the burden of proving eligibility on the beneficiary. See Farley v. Benefit Tr. Life Ins. Co. , 979 F.2d 653 , 658 (8th Cir. 1992) (emphasis added) (quoting 29 U.S.C. § 1132 (a)(1)(B) ).

We agree with the district court that the standard of review is dispositive in this case. Johnston presented some evidence from his medical providers that he was disabled. He genuinely had a colloid cyst in his brain, and the Social Security Administration's ("SSA's") reviewers found that he was "disabled" under the SSA's definition of disability. But Prudential also had evidence that Johnston was deliberately exaggerating his symptoms, making it impossible to determine whether he had cognitive deficiencies that rendered him disabled. Prudential's examinations also occurred after the SSA's review of Johnston's condition, meaning that the SSA did not know about Johnston's potential malingering. Thus, as the district court stated, although a court "might have reached a different conclusion" under de novo review, a court could find no abuse of discretion here because "Prudential's decision is still supported by substantial evidence."

Johnston argues that because his evidence indicates he was disabled and Prudential's examining doctors could not determine whether this was true, Prudential needed to introduce new evidence to show that he was not disabled. He cites no authority for the proposition that the burden of proof shifts after an initial determination of disability. Johnston cites Gunderson v. W.R. Grace & Co. Long Term Disability Income Plan , 874 F.2d 496 (8th Cir. 1989) to argue that Prudential should have obtained a vocational rehabilitation opinion. In Gunderson , we stated that even if the beneficiary bears the burden of proof, a plan administrator cannot change its understanding of the same opinions from the same doctors without substantial evidence supporting the new understanding. See id. at 499-500 & n.4. It was undisputed in Gunderson that the beneficiary was disabled for over two years. Id. at 498 . The plan administrator received reports on Gunderson's condition from his treating physician after one year, after two years, and after four years. See id. at 499 . We acknowledged the plan administrator may have had discretion to terminate benefits under the second report. Id. at 500 . Because Gunderson's condition was the same in the third report as in the second report, though, we found no substantial evidence supporting the plan administrator's decision to terminate benefits when it received the third report after continuing them under the second report. Id. We suggested, based on those facts, that the plan administrator should have obtained a vocational expert's opinion before making a different decision on evidence that was otherwise the same. Id. at 499 .

Here, Dr. Denney's report and Dr. Zeller's report support a new understanding of Johnston's prior medical evidence: that he was malingering. This case is different from Gunderson because Prudential had evidence allowing it to reassess the prior evidence the beneficiary submitted. Thus, because Prudential's changed decision was *716 supported by new evidence, as required in Gunderson , and because no authority requires shifting the burden of proof to Prudential, we decline to adopt Johnston's burden-shifting approach.

Johnston also argues that if this termination is upheld, insurers will have an incentive to claim beneficiaries are malingering in order to terminate benefits and save money. This could be a persuasive argument if the determination of malingering were a purely subjective or opinion-based determination, but there are multiple established ways to test validity of a neuropsychological examination. Dr. Denney extensively discussed which tests he administered and how these tests objectively measure validity. On review of the validity tests administered to Johnston, we see no basis to conclude that Prudential's evidence of malingering was subjective or otherwise manipulable by bias.

In sum, under our de novo review, we agree with the district court that Prudential did not abuse its discretion in terminating Johnston's benefits. There was substantial evidence to support Prudential's conclusion that Johnston may have been malingering in the tests he used as evidence to prove disability. As a result, Johnston failed to provide sufficient evidence of continuing disability.

IV. Conclusion

We affirm the district court's order because Prudential's decision was not an abuse of discretion.