Hannah P. v. Avril Haines

80 F.4th 236

Opinion

PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 22-1498

HANNAH P., Plaintiff - Appellant,

v.

AVRIL D. HAINES, in her official capacity as Director of National Intelligence, Defendant - Appellee,

and

MARK EWING, in his personal capacity McLean, VA, Defendant.

Appeal from the United States District Court for the Eastern District of Virginia, at Alexandria. Leonie M. Brinkema, District Judge. (1:16-cv-01030-LMB-IDD)

Argued: May 3, 2023 Decided: August 15, 2023

Before GREGORY, THACKER and QUATTLEBAUM, Circuit Judges.

Affirmed by published opinion. Judge Thacker wrote the opinion in which Judge Quattlebaum joined. Judge Gregory wrote a dissenting opinion.

ARGUED: Timothy Bosson, BOSSON LEGAL GROUP PC, Fairfax, Virginia, for Appellant. Caroline D. Lopez, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellee. ON BRIEF: Isaiah R. Kalinowski, BOSSON LEGAL

2

GROUP PC, Fairfax, Virginia, for Appellant. Brian M. Boynton, Principal Deputy Assistant Attorney General, Alisa B. Klein, Civil Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C.; James R. Whitman, Chief of Litigation, Jared S. Hatch, Associate General Counsel, OFFICE OF THE DIRECTOR OF NATIONAL INTELLIGENCE, Washington, D.C.; Jessica D. Aber, United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Richmond, Virginia, for Appellee.

3

THACKER, Circuit Judge:

Hannah P. 1 (“Appellant”), a former employee of the Office of the Director of National Intelligence (“ODNI”), asserts that ODNI violated the Family and Medical Leave Act of 1993 (“FMLA”), 29 U.S.C. § 2601, et seq., by delaying her leave request and not hiring her for a permanent position. The district court determined that Appellant failed to meet her burden of proof to demonstrate that she was not selected for the permanent position “by reason of” ODNI’s FMLA interference. Id. § 2617(a)(1)(A)(i)(I).

For the reasons explained below, we conclude that the record supports the district court’s finding that Appellant’s non-selection for the permanent position was the result of the hiring official’s poor impression of Appellant as a prospective employee and Appellant’s attendance problems prior to the FMLA interference. Therefore, we affirm.

I.

A.

In March 2011, ODNI hired Appellant for a five-year term as a Senior Systems Analyst. In September 2011, Appellant was diagnosed with recurrent major depressive disorder (“depression”). Appellant immediately informed at least two of her supervisors of her diagnosis, but she did not request any accommodations at that time. Following her diagnosis, Appellant sought treatment from a licensed clinical counselor and a psychiatrist who prescribed Appellant medication.

1

Pursuant to a protective order, Appellant is identified by her first name and last initial.

4

Throughout her employment with ODNI, Appellant generally received outstanding performance evaluations. Because of Appellant’s consistent high-level performance, Stephanie O’Sullivan (“O’Sullivan”), the Principal Deputy Director of National Intelligence, chose Appellant to lead the intelligence community in the coordinated response to the Edward Snowden unauthorized disclosures (the “Snowden Assignment”). This high-stress, high-profile assignment lasted from November 2013 through January 2015, and required Appellant to work long hours, meet tight deadlines, and deal with onerous demands.

To accommodate the changes this new role required, Appellant was put on a “maxi flex” schedule. J.A. 298. 2 The maxi flex schedule requires an analyst to work 80 hours over a two-week period but does not dictate the exact hours that the analyst must work per day. This schedule allowed Appellant to choose her working hours so long as she completed her assignments on time. Appellant continued to operate on a maxi flex schedule after the Snowden Assignment ended, with no initial concerns raised by her supervisors.

Throughout 2015, Appellant attempted to secure permanent employment at ODNI in anticipation of her term-limited position ending in March 2016. Toward that end, in February 2015, Appellant applied for two permanent positions within ODNI but was not selected for either position.

2

Citations to the “J.A.” refer to the Joint Appendix filed by the parties in this appeal.

5

Until early 2015, Appellant had been able to successfully manage her depression symptoms. However, as her time-intensive work on the Snowden Assignment was ending, Appellant began to experience more acute symptoms, including insomnia, which led to difficulty getting out of bed in the morning. Her symptoms made it “very difficult to function,” and she struggled to arrive at the office at the same time as her co-workers. J.A. 299.

Although she was still operating on the maxi flex schedule, Appellant arrived to work well after normal business hours and racked up numerous unplanned absences. There were days when Appellant was “extremely late,” sometimes arriving after 2:00pm, and her co-workers began to perceive her schedule to be “erratic.” J.A. 221. On other occasions Appellant was unreachable for hours, often missing and failing to return repeated phone “calls to her home.” Id. When Appellant’s supervisors were able to reach her, they noted that “she seemed lethargic or unconcerned” about her lateness and absences. Id. They also noted that her demeanor was “sad, very flat, and almost trance like.” Id.

Ultimately, on March 19, 2015, one of Appellant’s supervisors met with Appellant to address her attendance issues and set expectations for working hours. Together, Appellant and her supervisors developed a plan to reconcile Appellant’s depression with ODNI’s staffing needs. According to that plan, Appellant was to arrive to work by 10:00am, and if she was going to be absent or late, Appellant was to contact one of her supervisors in advance by either email or phone. If Appellant had not arrived at work or contacted a supervisor by 11:00am, a supervisor would call her to determine if or when she would arrive. Although Appellant understood this plan was developed to be “an

6

accommodation” for her depression, there is no evidence in the record that any of her supervisors mentioned the FMLA to her while developing this schedule. J.A. 300.

From March 23, 2015 to March 27, 2015, the first week after Appellant and her supervisors agreed to this schedule, Appellant was on previously scheduled leave to work on a home renovation project. However, upon her return to the office, Appellant almost immediately began disregarding the plan.

On Tuesday, March 31, 2015, Appellant emailed her supervisors just before noon to let them know that she would not be coming into the office that day because she was “swamped with contractor stuff.” J.A. 266. And on Wednesday, April 1, 2015, Appellant did not arrive at work by 10:00am nor did she call or email her supervisors as agreed in the plan. Instead of calling Appellant at 11:00am to locate her, as per the plan, one of Appellant’s supervisor’s called Appellant at 12:30pm, informed her that the accommodation plan was not working, and told her they would need to develop a new accommodation plan. Appellant arranged to meet with the supervisor on April 9, 2015, to discuss alternative plans.

In preparation for her April 9, 2015 meeting, Appellant met with both her psychiatrist and counselor to discuss what course of action they would recommend for her going forward. Appellant’s medical care providers recommended that she take four to six weeks of leave to combat her depression. That same day, Appellant requested four weeks of leave to allow her to treat her depression and return to the office “before the next round of studies started” in late May. J.A. 111. Appellant testified that she requested that her leave “start as soon as possible,” meaning within “half a day or a day” to enable her to

7

finish up outstanding assignments. Id. at 114. Appellant’s supervisor, however, claimed that Appellant did not say when exactly she wanted her leave to begin.

Instead of granting her leave request, Appellant’s supervisors presented Appellant with a referral to the Employee Assistance Program (“EAP”), an employee counseling service offered to ODNI employees. Appellant’s supervisors told Appellant that her attendance at an EAP counseling session the next day, April 10, 2015, was mandatory. Appellant’s supervisor did not mention FMLA leave to Appellant during the April 9, 2015 meeting.

Appellant attended the April 10, 2015 session with the EAP counselor as directed.

During the meeting, she asked the EAP counselor what she needed to do to be able to go on leave, but her EAP counselor told Appellant that such decisions were up to management. On April 13, 2015, Appellant’s supervisor advised Appellant that they could move forward on Appellant’s request for leave, but the supervisor was “heavily emphasizing that it would be annual leave,” as opposed to FMLA leave, because that was the only kind of leave the supervisor believed he could authorize. J.A. 303. However, at that point, Appellant informed her supervisor that her leave request was “on hold,” without further explanation. Hannah P. v. Coats, 916 F.3d 327, 335 (4th Cir. 2019).

Despite Appellant’s continued participation in the EAP, her attendance problems persisted. For example, on April 13, 2015, Appellant emailed her supervisors at 10:58am to inform them that she would arrive to work by 11:30am. Similarly, on April 14, 2015, Appellant emailed her supervisors at 11:08am to inform them that she would arrive to work

8

by 12:00pm. But that day, Appellant’s supervisors were not able to confirm her arrival to work until after 1:50pm.

On April 21, 2015, Appellant renewed her request for four weeks of medical leave.

Appellant’s supervisors approved that request on Friday, May 1, 2015, and Appellant’s leave began on Tuesday, May 5, 2015. However, despite having knowledge of Appellant’s FMLA qualified disability, ODNI never notified Appellant of her right to take FMLA leave. Although it was against ODNI’s FMLA policy, Appellant’s supervisors required her to use her annual leave to account for four-fifths of her four week leave period, allowing Appellant to take one day of sick leave per week on days she was seeing her medical care providers.

After returning from leave, Appellant’s performance significantly improved, and her attendance was nearly flawless. On June 9, 2015, soon after returning from leave, Appellant interviewed for the Program Mission Manager Cyber (“Cyber”) position, which was a permanent ODNI position. The members of the interview panel unanimously selected Appellant as the most qualified candidate. On June 17, 2015, the panel’s unanimous recommendation was provided to the Chief Management Officer, Mark Ewing (“Ewing”). At this time, Ewing was “involved in approving all hiring actions” at ODNI and was “the hiring authority” with respect to the Cyber position. J.A. 71. Appellant’s application then stalled for several weeks pending Ewing’s determination.

In late June 2015, Appellant was informed that she had been selected for the Cyber position by the interview panel. However, on July 7, 2015, ODNI Human Resources informed Appellant that she was not selected for the Cyber position, based solely on

9

Ewing’s decision. In fact, while later discussing her interview with a member of the interview panel, the member noted that she had no constructive feedback for Appellant, as she “interview[ed] very, very well.” J.A. 259.

In making his decision, Ewing said that he “consider[ed] [Appellant’s] erratic attendance . . . mid-January through April.” J.A. 86. Ewing’s explanation is consistent with a June 30, 2015 email he sent his supervisor, O’Sullivan, in which Ewing stated that he was “concerned about hiring [Appellant]” because “her recent performance is not consistent with a potentially good employee.” Id. at 289. Ewing was apparently wary that even after Appellant had been given a flexible schedule by her supervisors and sent to counseling with the EAP, “her late attendance at work . . . continued.” Id. According to Ewing, he was “informed that EAP concluded [that Appellant] does not have a medical problem, rather she is a disciplinary problem.” Id. Ewing ultimately recommended against hiring Appellant for the Cyber position because he believed Appellant “approached permanent employment as an entitlement” despite her “consistent history of issues” and “recent attendance issues suggest[ing] she is more than a disciplinary problem.” Id. at 290.

After receiving Ewing’s June 30, 2015 email, O’Sullivan “walk[ed] over to talk to Ewing about what he had written about [Appellant], largely to figure out[,] are we pas[t] the absence issue and what’s the next step.” J.A. 149. O’Sullivan testified that she thought “two . . . [to] four weeks” of improved performance after Appellant returned from her leave would have “resolved” Appellant’s “absence issue.” Id. at 149–50. However, in late June, she was not sure if Appellant had completed that stretch, so she did not push Ewing to make the hire. This is consistent with Ewing’s testimony that “[i]f we had seen some

10

pattern of [] positive conduct after the 9th of April, [considering Appellant for the Cyber position] would have been a real possibility.” Id. at 87. But when asked directly if he would have hired Appellant for the Cyber position “if [Appellant] had just eliminated her attendance and reporting issues in April of 2015, after she started meeting with the EAP counselor,” Ewing responded, “I really don’t know, because . . . we would need some time to understand that the conduct has actually been corrected. [And] I’m not sure that in April or May, we would have.” Id. at 89–90. Ewing testified that he had made up his mind not to hire Appellant as of June 30 or July 1, 2015.

In early June 2015, Appellant also applied for another permanent position with ODNI, and she was asked to come in for an interview after she was rejected for the Cyber position. However, Appellant testified that she did not go to the interview because she was “encouraged by Kelly G. [the Deputy Assistant Director of National Intelligence for Systems and Resource Analysis] to withdraw [her] name from contention for that position” given that Ewing “would also likely reject [her] for the second position,” which “could get embarrassing.” J.A. 121–22.

Appellant continued to reach out to former supervisors to discuss possible opportunities in ODNI. She sent her resumé to two of her former supervisors, one who had switched offices within ODNI and another who had moved to the private sector, but neither followed up with her. Appellant also applied for 27 government jobs during the spring and early summer of 2016 but was not offered an interview for any of them. Unable to find employment within ODNI or related agencies, Appellant finished out her term at

11

ODNI on March 27, 2016. Unemployed, Appellant decided to pursue a career in real estate.

B.

On August 12, 2016, after exhausting her administrative remedies, Appellant filed a civil action against ODNI, alleging discrimination, failure to accommodate, performance of a wrongful medical exam, and unlawful use of confidential medical information in violation of The Rehabilitation Act of 1973, 29 U.S.C. § 794; and interference and retaliation in violation of the FMLA, 29 U.S.C. § 2601. After the close of discovery, ODNI moved for summary judgment on all counts. The district court granted ODNI’s motion and entered a corresponding order of final judgment on July 27, 2017. Appellant appealed that ruling on August 15, 2017.

On February 19, 2019, we affirmed the district court’s order granting summary judgment with respect to all of Appellant’s Rehabilitation Act claims and her claim of FMLA retaliation. See Hannah P. v. Coats, 916 F.3d 327, 332 (4th Cir. 2019). However, we reversed the grant of summary judgment on Appellant’s FMLA interference claim, concluding that a “reasonable jury could find that [Appellant’s] disclosure of her depression and her April 9, 2015 request for psychiatrist-recommended leave was sufficient to trigger [ODNI’s] responsibility to inquire further about whether [Appellant] was seeking FMLA leave.” Id. at 346–47; see also id. at 333 (“[B]ecause a genuine issue of material fact remains as to whether [Appellant] provided notice of her disability and interest in FMLA leave sufficient to trigger [ODNI’s] duty to inquire, we hold that summary judgment as to [Appellant’s] FMLA interference claim was not warranted.

12

Accordingly, we vacate that part of the district court’s judgment and remand [Appellant’s] FMLA interference claim for further proceedings.”).

Upon returning to the district court, the case was tried in March 2021 as a bench trial. See 28 U.S.C. § 2402 (mandating bench trial). At the conclusion of trial, the district court denied closing arguments stating, “[I]t’s obvious that there is definitely liability because there was clearly a violation of the FMLA. I mean, there’s just no question about it.” J.A. 214. The district court found that Appellant had “put [ODNI] on sufficient notice that she was requesting leave guaranteed by the FMLA and that [ODNI] did not respond by making [Appellant] aware of her FMLA rights and promptly allow her to take leave.” Id. at 312.

The district court then instructed the parties to submit post-trial briefs to address damages. When Appellant’s trial counsel asked whether the parties needed to discuss liability in their briefing, the court responded, “I am giving that to you right now. I am finding that the FMLA interference claim has been established. . . . So just address damages -- or remedy. I mean the full range of remedies.” J.A. 219.

Appellant sought damages equal to the value of the annual leave she took; the value of the sick leave that she took during the weeks after she first put ODNI on notice of her need for leave; back and front pay for the Cyber position, for which she claims she would have been selected had she taken leave when she first requested it; and equitable relief. On December 30, 2021, the district court issued a Memorandum Opinion awarding Appellant $15,083.20 in compensatory and liquidated damages for the annual leave disallowed in May 2015 but denying all other relief sought.

13

In denying Appellant’s claim for damages related to the Cyber position, the district court held that Appellant had “not carried her burden of showing that [ODNI’s] FMLA interference proximately caused her non-selection for the Cyber position.” J.A. 315–16. To reach its conclusion, the district court held that FMLA “losses must be direct” because “the Fourth Circuit does not allow consequential damages in FMLA cases.” Id. at 315 (internal quotation marks omitted). The district court further determined that Appellant’s “attendance issues after April 9, 2015 -- the date on which she first requested leave -- were not the principal cause of her non-selection for the Cyber position.” Id. at 316 (internal quotation marks omitted). Rather, according to the district court, Appellant’s “non- selection for the Cyber position was a result of many intervening factors, including [Appellant’s] attendance problems before April 9 and Chief Management Officer Ewing’s poor impression of [Appellant] as a prospective employee, which was significantly independent of the FMLA interference.” Id. Therefore, the district court resolved, “[I]t would be too speculative to conclude that Ewing would have selected [Appellant] for the Cyber position if she had been allowed to take leave on April 9.” Id. at 318.

On January 27, 2022, Appellant filed a motion for reconsideration challenging the legal standard of causation the district court applied to the Cyber position damages. On March 18, 2022, the district court denied Appellant’s motion, emphasizing that the burden was solely on Appellant to prove causation of the Cyber position damages and that the wages needed to be a “direct” result of the FMLA interference. The district court entered a final judgment on April 4, 2022, and Appellant filed this timely appeal.

14

II.

We employ a “mixed standard of review” when judgment results from a bench trial.

Va. Elec. & Power Co. v. Bransen Energy, Inc., 850 F.3d 645, 654 (4th Cir. 2017). We review the district court’s legal conclusions and allocation of the burden of proof de novo. Id. (legal conclusions); Everett v. Pitt Cnty. Bd. of Educ., 678 F.3d 281, 288 (4th Cir. 2012) (burden of proof). And we review the district court’s factual findings for clear error. Va. Elec. & Power Co., 850 F.3d at 654.

III.

A.

The FMLA provides employees who have qualifying medical conditions the right to take up to 12 weeks of leave during a 12-month period. 29 U.S.C. § 2612(a). Employees who take qualifying FMLA leave also generally have the right “to be restored by the employer to the position of employment held by the employee when the leave commenced” or to “an equivalent position with equivalent employment benefits, pay, and other terms and conditions of employment.” Id. § 2614(a)(1)(A)–(B). Leave taken under the FMLA “shall not result in the loss of any employment benefit accrued prior to the date on which the leave commenced.” Id. § 2614(a)(2). However, an employee’s restoration rights are limited, such that a restored employee is not entitled to “any right, benefit, or position of employment other than any right, benefit, or position to which the employee would have been entitled had the employee not taken the leave.” Id. § 2614(a)(3)(B).

“Because the FMLA grants valuable leave and restoration rights to eligible employees,” Throneberry v. McGehee Desha Cnty. Hosp., 403 F.3d 972, 977 (8th Cir.

15

2005), it also secures these rights by making it “unlawful for any employer to interfere with, restrain, or deny the exercise of or the attempt to exercise, any right provided under” the FMLA, 29 U.S.C. § 2615(a)(1). A violation of this provision creates what is commonly known as an FMLA interference claim. Yashenko v. Harrah’s N.C. Casino Co., 446 F.3d 541, 546 (4th Cir. 2006). The Supreme Court has clarified that the “cause of action” for FMLA interference is “set out in § 2617.” Ragsdale v. Wolverine World Wide, Inc., 535 U.S. 81, 89 (2002).

Thus, to succeed on an FMLA interference theory of recovery,

[A]n employee must prove, as a threshold matter, that the employer violated § 2615 by interfering with, restraining, or denying his or her exercise of FMLA rights. Even then, § 2617 provides no relief unless the employee has been prejudiced by the violation: The employer is liable only for compensation and benefits lost “by reason of the violation,”

§ 2617(a)(1)(A)(i)(I), for other monetary losses sustained “as a direct result of the violation,” § 2617(a)(1)(A)(i)(II), and for “appropriate” equitable relief, including employment, reinstatement, and promotion, § 2617(a)(1)(B).

Ragsdale, 535 U.S. at 89 (emphasis supplied). Whether an employee was prejudiced by their employer’s FMLA interference is simply another way of asking whether the employee suffered losses “by reason of” or “as a direct result” of the interference. 29 U.S.C. § 2617(a)(1)(A)(i)(I)–(II).

B.

In FMLA interference cases where the alleged injury involves an adverse employment action, “an employer may avoid liability if it shows it would have taken the contested adverse employment action regardless of the employee’s FMLA leave.” Roberts

16

v. Gestamp West Virginia, LLC, 45 F.4th 726, 732–33 (4th Cir. 2022). In other words, once an employee has met her burden to prove that an employer’s FMLA interference caused the alleged adverse employment action, the burden shifts to the employer to provide a legitimate, nondiscriminatory reason for the action taken. The employer has the burden to show that it would have made the same employment decision independent of the FMLA interference. But it is the employee who retains the ultimate burden of proof to demonstrate that the employer’s actions did, in fact, interfere with the employee’s FMLA rights and directly cause the alleged harm.

Where, as here, the FMLA interference occurs before the adverse employment action, the employee must first prove that the interference caused the adverse employment action before the burden shifts to the employer to prove it would have taken the same action absent the interference. Thus, the threshold issue we confront today is whether Appellant met her burden to prove that ODNI’s April FMLA interference caused her not to be hired for the Cyber position.

C.

While the district court did not expressly apply the burden-shifting standard described above, it nevertheless correctly noted that Appellant bore the initial burden to prove that her damages were caused “by reason of” ODNI’s FMLA interference. 29 U.S.C. § 2617. This means that Appellant had the burden to establish a causal connection between the ODNI’s FMLA interference and any alleged damages. Only if Appellant had met this burden of causation would the burden then shift to ODNI to show that it would have made the same decision without regard to the interference. Thus, even if the district court were

17

to have properly recognized the burden-shifting standard of causation, for Appellant to succeed on her interference claim, she carried the initial burden of showing that she would have been entitled to the Cyber position “had [she] not taken the leave.” Id. § 2614(a)(3)(B). For the reasons set forth below, we cannot conclude that the district court’s determination that Appellant’s “non-selection for the Cyber Position was neither a direct result of, nor caused by, [ODNI]’s FMLA interference,” J.A. 332, constitutes reversible error.

D.

Here, Appellant argues that the FMLA interference worsened her attendance problems, and that as a result of these attendance problems, she was not selected for the Cyber position. However, the district court found that Ewing’s words and actions provided several reasons for Appellant’s non-selection. For example, Ewing’s June 30, 2015 email described Appellant as a “disciplinary problem” with a “history of issues” who considered permanent hire an “entitlement.” J.A. 289–90. Ewing was of the view that Appellant’s “attendance at work and attitude” issues began in “mid-Jan[uary] 2015” and grew serious enough to warrant a management referral to EAP in April. Id. at 289. Ewing further testified that he considered Appellant’s attendance issues to have begun before the initial April 9, 2015 3 FMLA interference and that he had fully made up his mind not to hire

3

The parties do not dispute the district court’s determination that on April 9, 2015, Appellant put ODNI “on sufficient notice that she was requesting leave guaranteed by the FMLA,” and that “[b]y failing to give [Appellant] the statutorily required notice and by requiring her to use annual leave, [ODNI] is liable for FMLA interference.” J.A. 312.

18

plaintiff by June 30 or July 1, despite having seen several years of good work from Appellant before her attendance problems began. Ewing did not mention Appellant’s perfect attendance record since returning from leave, nor was he even aware that her attendance had improved at all since returning from leave.

Moreover, Appellant’s own testimony was that her supervisor had suggested that she withdraw her application for a position which came open after the Cyber position -- and after Appellant had been back at work for weeks -- because Ewing would be the decision maker and, as a result, it would be unlikely that she would get the job. Thus, as the district court explained, the record contradicts Appellant’s theory that, “without [ODNI’s] interference,” Ewing might have made a different decision because Appellant “would have had sufficient time to demonstrate positive attendance before Ewing made his hiring decision.” J.A. 316–17.

The district court further explained that the reference in Ewing’s June 30, 2015 email to Appellant’s “attendance problem continu[ing] after she was referred to EAP on April 9” did not alter its analysis because, read in its entirety, the email made clear that Ewing’s view was that Appellant’s “attendance . . . and attitude” issues were “‘consistent’ for ‘several months’” and had “gr[own] serious enough to warrant a management referral to EAP in April” -- a “time period [that] predates when the FMLA interference occurred.” J.A. 317 (quoting J.A. 289–90). The district court also rejected Appellant’s attempt to characterize her lost job opportunity as a loss of her “ongoing employment” with ODNI, finding that “her five-year term contract was neither shortened nor changed, she continued to receive the same pay, and she worked to the contract’s end in March 2016.” Id. at 331.

19

“In other words, [Appellant’s] ‘ongoing employment’ was not affected by her non- selection for the Cyber [p]osition; rather, the Cyber [p]osition was a job opportunity that never came to fruition.” Id.

Appellant also argues that because “employers cannot use the taking of FMLA leave as a negative factor in employment actions,” 29 C.F.R. § 825.220(c), she should prevail “merely by proving that ODNI considered her April leave as a ‘negative factor’ in its Cyber hiring decision,” Appellant’s Opening Br. at 25. But Appellant has not pointed to any evidence suggesting that Ewing took Appellant’s actual leave or even her request for leave as a negative factor in his decision not to hire her for the Cyber position.

At bottom, Appellant simply disagrees with the district court’s finding that Ewing’s mind was made up not to hire her for the Cyber position independent of the FMLA interference. But under clear error review, we must affirm factual findings if they are “plausible in light of the [entire] record,” “even though convinced that had [we] been sitting as the trier of fact, [we] would have weighed the evidence differently.” Walton v. Johnson, 440 F.3d 160, 173 (4th Cir. 2006). Appellant cannot avoid the reality that causation of harm is part of her burden of proof, and the district court found that Ewing’s decision was based on Appellant’s conduct before the FMLA interference occurred and therefore was not caused by that interference. Because this finding is not clearly erroneous, the district court properly denied damages based on Appellant’s non-selection for the Cyber position.

E.

Even assuming Appellant met her initial burden to demonstrate that ODNI’s FMLA interference caused her alleged damages, the issue of whether ODNI successfully

20

demonstrated that it would have made the same decision absent the FMLA interference has already been resolved by our previous decision in this case.

According to the law of the case doctrine, the factual findings and legal conclusions made by an appellate court generally bind all subsequent proceedings in the same case, whether in the trial court or on subsequent appeal. See United States v. Aramony, 166 F.3d 655, 661 (4th Cir. 1999). The doctrine applies “both to questions actually decided as well as to those decided by necessary implication[] . . . .” Sejman v. Warner-Lambert Co., Inc., 845 F.2d 66, 69 (4th Cir. 1998) (internal quotation marks omitted).

In this case, we affirmed the entry of summary judgment on Appellant’s Rehabilitation Act and FMLA retaliation claims, despite her contention that ODNI should not have been permitted to rely on her April attendance issues. In this regard, Appellant argued that because ODNI wrongly denied her April 9, 2015 FMLA leave request, she “suffered immense emotional stress during this one month lapse of [ODNI’s] compliance with the law,” Hannah P., 916 F.3d at 337 (internal quotation marks omitted), and her attendance issues thereafter were primarily the result of ODNI’s own bad act.

But as we discussed in our prior decision, Appellant first requested a leave of absence on Thursday, April 9, 2015. Then, on Monday, April 13, 2015 -- just two business days later -- “Appellant withdrew her request without explanation, telling her supervisor that her leave request was ‘on hold.’” Hannah P., 916 F.3d at 337. Appellant did not renew her leave request until April 21, 2015, and her request was approved on May 5, 2015. Accordingly, we previously held that there was not a “one month lapse,” in processing Appellant’s request, as it was “on hold” for nine days, and “the record

21

demonstrates that [Appellant’s] supervisors were actively considering her request for leave during that time.” Id. at 337–38. Therefore, we concluded that Appellant’s argument that her leave request was “improperly delayed” lacked merit and was not supported by the record. Id. at 337. Any delay in granting her leave request was, at least in part, a consequence of Appellant’s own actions.

Appellant also argued in her briefing in the prior appeal that she was prejudiced by ODNI’s FMLA interference because ODNI’s actions caused her to delay her leave from April 9, 2015 until May 5, 2015, and that her absences and tardiness in the interim period were unlawfully used against her in the hiring decision for the Cyber position. But we did not accept this theory.

The purpose of our limited remand as to Appellant’s FMLA interference claim was to address the narrow question of whether Appellant was prejudiced by FMLA interference because she could have chosen to use sick leave alone, rather than a combination of sick leave and annual leave, for her four-week period of leave -- a benefit she valued at approximately $20,000. Hannah P., 916 F.3d at 345–47. We did not send this case back for relitigation of ODNI’s decision not to select Appellant for the Cyber position. The only count that went to trial in this case was limited to alleging that ODNI interfered with Appellant’s FMLA right to take leave to address her mental health issues, not that the interference caused Appellant to lose out on the Cyber position.

We emphasized that “it is not the job of this court to decide whether [ODNI] made the right choice by not hiring [Appellant] for the Cyber position” but “simply to decide whether [ODNI] made an illegal choice.” Hannah P., 916 F.3d at 345. Significantly, our

22

prior decision conclusively determined that ODNI lawfully chose not to select Appellant for the permanent Cyber position. Thus, the only question remaining is whether Appellant was prejudiced by her use of annual rather than sick leave. And even if the district court had appropriately recognized that the burden shifts to ODNI to prove that Ewing would have made the same hiring decision had ODNI not delayed acting on Appellant’s leave request, it is of no moment because we have already held that Appellant failed to rebut ODNI’s “legitimate, nondiscriminatory reason for rejecting her application for a permanent position.” Id. at 336–37. Consequently, Appellant cannot recover damages for her failure to secure the permanent Cyber position.

IV.

For the foregoing reasons, the district court’s order denying Appellant damages related to her non-selection for the Cyber position is AFFIRMED.

23

GREGORY, Circuit Judge, dissenting:

Hannah P. served with distinction as an analyst for the Office of the Director of National Intelligence (“ODNI”). Her supervisors entrusted her with leading the response to Edward Snowden’s 2013 leak of classified documents, one of the intelligence community’s highest-profile cases at the time. Hannah performed exceptionally well in that high-pressure role.

For years, Hannah successfully managed the symptoms of recurrent major depressive disorder while excelling in her job. But when her depression intensified around the time the Snowden assignment ended, her supervisors began to find fault with her sometimes-unpredictable work hours, even though they had authorized her to keep a nonconventional schedule. Hannah urgently requested medical leave to treat her depression, but, for several weeks, her supervisors refused to grant her request. Instead, they required her to attend sessions with ODNI’s Employee Assistance Program (“EAP”), an internal counseling service, under threat of termination. The EAP counselor quickly concluded that Hannah’s attendance issues were a disciplinary problem, not a consequence of her diagnosed depression.

When Hannah later interviewed for a permanent position at ODNI, the interview panel unanimously selected her as the most qualified candidate and recommended hiring her. But the ODNI officer with final hiring authority refused to select her for the position. In making that decision, the hiring officer relied heavily on Hannah’s absences during the period when ODNI delayed acting on her leave request, as well as the EAP counselor’s opinions about Hannah’s condition.

24

The district court held that ODNI interfered with Hannah’s rights under the Family and Medical Leave Act of 1993 (“FMLA”), 29 U.S.C. § 2601, et seq., by not promptly granting her request for medical leave. But the court rejected Hannah’s claim for damages related to her non-selection for the permanent position, holding that she failed to prove that ODNI’s FMLA interference caused her non-selection.

In affirming the denial of those damages, the majority, like the district court, applies the wrong standard of causation. Fourth Circuit precedent instructs us to apply the “same- decision test” when a plaintiff alleges that her employer’s FMLA interference caused her to suffer an adverse employment action. Correctly understood, that test follows a two-step, burden-shifting analysis: if the plaintiff shows that the interference was a substantial or motivating factor in the employment action, the burden shifts to the employer to prove that it would have taken the same action absent the interference. The majority misapplies the test by holding that the burden shifts to the employer only if the plaintiff first proves but-for causation.

If we apply the correct causation standard, the record compels us to find that ODNI’s interference with Hannah’s FMLA rights caused her non-selection for the permanent position. For that reason, I respectfully dissent.

I.

At this stage, the facts of this case are no doubt familiar. Still, to resolve this appeal, it is critical that we fully understand the events that culminated in Hannah’s non-selection for a permanent position at ODNI.

25

A.

Hannah P. was, by all accounts, a high-achieving employee at ODNI. While serving as a Senior Systems Analyst, she consistently received “[o]utstanding” and “[e]xcellent” performance ratings. J.A. 39. Her supervisors considered her “an invaluable intelligence officer and a future [intelligence community] leader.” J.A. 273.

ODNI’s senior leadership so trusted Hannah’s abilities that they chose her to lead the intelligence community’s coordinated response to Edward Snowden’s disclosure of classified documents in late 2013. Hannah performed exceptionally well in that high-stress role, which continued through early 2015. According to her superiors, her “leadership, poise, and performance were impeccable.” J.A. 273.

Because the Snowden assignment demanded high-pressure and time-sensitive work, often at irregular hours, Hannah was authorized to follow a “maxi flex” work schedule. J.A. 54. That schedule required her to work eighty hours every two-week pay period but gave her a great deal of flexibility in determining when to log those hours. While leading the Snowden assignment, Hannah had to work much later hours than her co-workers; she typically arrived at the office between 10:00 a.m. and 12:00 p.m. and stayed until 10:00 p.m. or later.

As she distinguished herself at work, Hannah was also managing the symptoms of recurrent major depressive disorder, which was first diagnosed in September 2011. In addition to taking medication for her depression, she received ongoing treatment from both a psychiatrist and a licensed clinical counselor. Hannah immediately disclosed her diagnosis to her supervisors, but for several years, she did not request any work accommodations because she “was adequately handling [her] depression at the time with

26

medication and counseling.” Hannah P. v. Coats, 916 F.3d 327, 349 (4th Cir. 2019) (Gregory, C.J., concurring in part and dissenting in part) (quoting record).

Around the time the Snowden assignment ended in early 2015, Hannah began experiencing more severe symptoms of her depression, including insomnia, which made it “very difficult to function” and led her to struggle to arrive at the office at 9:00 a.m. 1 J.A. 299. Hannah continued to arrive at work in the late morning or early afternoon, which was consistent with her maxi flex schedule. Her supervisors, however, claimed that her unpredictable attendance was starting to have a negative effect on her team’s morale.

On March 19, 2015, Hannah’s first-line supervisor spoke to her about her work schedule. The supervisor did not indicate that Hannah needed to return to a standard “9 to 5” schedule. Instead, he and Hannah agreed that Hannah would “either arrive by 10 [a.m.] or call in or email to tell [her supervisors] her plans,” and that her supervisors would call her if she neither arrived nor contacted them by 11:00 a.m. J.A. 293. Hannah had

1

Hannah testified that she believed her worsening symptoms were the result of “the high stress that [she] had from working the Snowden disclosures task for so long,” as well as the uncertainty of knowing that her five-year contract with ODNI “was going to time out in about a year.” J.A. 109–10. In late February 2015, Hannah learned that ODNI had not selected her for two permanent positions to which she had applied. She explained that that news also “fed into the depression” symptoms she experienced during that time period. J.A. 110 (“I just didn’t know that I was ever going to get a permanent position. I was just feeling really helpless and hopeless about the entire situation.”). In addition, Hannah’s team at ODNI experienced a great deal of turnover when she was on special assignment as the Snowden response coordinator, so when that assignment ended, “she essentially came back to different managers and new teammates.” J.A. 273. And on top of those stressors, Hannah’s living situation was unstable in early 2015. At the time, she was making significant renovations to her house, which required her to live in a friend’s basement “for several weeks” in February and March 2015. Id.

27

previously scheduled leave from March 23 through March 27 to deal with home renovations, so the “first test” of the new plan took place the week she returned. Id.

Hannah was not the only one who struggled to follow the new plan. On Tuesday, March 31, Hannah emailed her supervisors at 11:56 a.m. to let them know she was still “swamped with contractor stuff” and would not be coming into the office that day. J.A. 267. There is no indication that her supervisors called after not hearing from her by 11:00 a.m., as the attendance plan instructed. The next day, April 1, Hannah did not arrive at the office or contact her supervisors by 10:00 a.m. Hannah’s second-line supervisor called her and told her the plan “was not working,” even though it had been in place for less than one work week. J.A. 301. The supervisor told her that they needed to develop a different arrangement, and the two of them planned to meet on April 9 to discuss the matter further.

On April 2, Hannah’s supervisors held a meeting with ODNI human resources and employee relations officers to discuss how to address Hannah’s attendance issues. At the time, Hannah’s supervisors were aware of her depression. Despite having that knowledge, the group decided to refer Hannah to EAP for counseling. Although the participants at the April 2 meeting were reportedly “the experts” on ODNI’s human resources policies, J.A. 186, none of them mentioned the FMLA during the meeting. Notably, Hannah’s second- line supervisor and the director of her component at ODNI both testified that they did not even know that ODNI had an FMLA policy at the time.

Ahead of the April 9 meeting with her supervisors, Hannah met with her psychiatrist and counselor, who recommended that she take four to six weeks of leave to treat her

28

depression. That amount of leave would allow Hannah to return to work before her team’s busy season began.

At the April 9 meeting, Hannah requested four to six weeks of leave “to start as soon as possible” and explained that her medical provider had recommended the leave. J.A. 114. Instead of granting Hannah’s request, her supervisors informed her that she needed to attend a mandatory EAP session the next day, April 10. According to Hannah, her component director explained, “We’re not qualified doctors, so we need EAP to tell us whether or not to grant you leave.” J.A. 115. The supervisors not only deferred a decision on Hannah’s leave request until after the EAP session, but threatened to terminate her employment if she did not attend it. At no point during the meeting did they inform Hannah of her rights under the FMLA. Hannah’s first-line supervisor disagreed with the decision to refer Hannah to EAP, believing that ODNI officials should not “substitute their judgment for Hannah’s doctor’s.” Hannah P., 916 F.3d at 351 (Gregory, C.J., concurring in part and dissenting in part) (cleaned up).

Hannah attended the EAP session on April 10 as instructed. During the meeting, the EAP counselor denied that she had the ability to approve leave requests and said that management made those decisions. On April 13, Hannah discussed her leave request again with her second-line supervisor. According to Hannah, he told her that he could approve only two weeks of annual leave without approval from the Office of Medical Services. 2 But when Hannah offered to take a medical exam, the supervisor said the Office of Medical

Hannah’s second-line supervisor testified that he could not recall limiting the

2

number of weeks of leave Hannah could take during this meeting.

29

Services could not examine her because she was a term-limited employee. Hannah reiterated that she needed four weeks of leave, but the supervisor conveyed that it was impossible for her to take more than two. Discouraged, Hannah eventually “just said that the request was on hold so that [she] could get out of there.” J.A. 304.

As April continued and Hannah’s leave request languished, Hannah’s depression and attendance issues “just spiraled downwards.” J.A. 116.

In mid-April, the second-line supervisor had a conversation with Hannah’s EAP counselor, during which the EAP counselor disclosed details about her sessions with Hannah (EAP had scheduled additional sessions after April 10). The supervisor relayed those details to Hannah’s component director in an April 22 email. He explained that the EAP counselor had identified “non-medical factors as the primary cause for [Hannah’s] late attendance.” J.A. 278. In the EAP counselor’s opinion, the “primary reason” for Hannah’s attendance issues “originate[d] from her frustration and a feeling of helplessness caused by her unresolved . . . employment status within ODNI.” J.A. 277. Hannah’s depression “ma[de] this situation worse,” but was “under her control based upon treatment plans with her psychologist and psychiatric care providers.” J.A. 278. The EAP counselor also noted that Hannah often “focused on debating the merits of the initial referral and understanding options to end the EAP sessions” rather than “discussing methods to correct her late attendance.” Id. Hannah, the counselor conveyed, was “very concerned that the EAP process will create a paper trail that will adversely impact her future employment and career despite reassurances by EAP to the contrary.” Id.

30

Meanwhile, Hannah continued to follow up with her supervisors about her leave request. After unsuccessful attempts to schedule another meeting, Hannah finally was able to meet with her component director on April 28, nearly three weeks after she first requested leave. Her component director approved the leave request at that time, on the conditions that Hannah attend one final EAP session and sign a Letter of Expectations that addressed attendance requirements moving forward. Hannah complied with both conditions.

Hannah started her four-week leave on May 5. While away on leave, she continued to receive treatment for her depression. Following her return to work on June 1, her attendance was “nearly flawless,” J.A. 119, as was her performance.

B.

Hannah was hired by ODNI on a five-year contract, which ended in March 2016.

In spring 2015, she applied for a permanent position at ODNI: Program Mission Manager Cyber (the “Cyber position”). Hannah was invited to interview for the position on June 9, shortly after she returned from leave. The interview panel unanimously recommended selecting her for the role and forwarded its recommendation to Mark Ewing, ODNI’s Chief Management Officer. Ewing had the final hiring authority for the position.

At the time Ewing reviewed Hannah’s application, he was aware of her attendance issues and referral to EAP. From discussions with Hannah’s supervisors, Ewing also knew that Hannah “was meeting with a psychiatrist[ and] taking medication for depression.” J.A. 79. Remarkably, though, Ewing claimed he did not know that she had a “disability related to depression.” J.A. 79–80 (“I don’t know that just because you’re meeting with a psychiatrist or you’re taking medication for depression means that you have a disability.”).

31

On June 30, after the interview panel inquired about the delay in a hiring decision, Ewing emailed Principal Deputy Director of National Intelligence Stephanie O’Sullivan to share his views on Hannah’s application. In the email, Ewing recommended against hiring Hannah for the Cyber position. To justify that result, Ewing relied on information he had received from Hannah’s supervisors about her “increasingly erratic” attendance in early 2015, her supervisors’ efforts to address her attendance issues, and her referral to EAP in April. J.A. 221. He particularly emphasized that “since referral to EAP, her late attendance at work has continued.” Id. (emphasis in original). Ewing further noted that he was “informed that EAP concluded [Hannah] does not have a medical problem, rather she is a disciplinary problem.” 3 Id. He then concluded:

We have had a consistent history of issues with [Hannah] over many months, despite some apparently solid performance while on the Snowden project.

She has approached permanent employment as an entitlement. Her recent attendance issues suggest she is more than a disciplinary problem. . . .

Given this knowledge, I am concerned about hiring her. People are aware of her recent actions and watching what we do.

My recommendation is that we do not hire her at this time. He[r] recent performance is not consistent with a potentially good employee. However, I defer to your preference and this discussion is solely between you and me. I seek your guidance as to whether or not she should be offered permanent employment.

J.A. 222.

Ewing and O’Sullivan then had an in-person conversation about the matter.

O’Sullivan testified that she left that conversation believing that Ewing would determine

3

In describing the EAP counselor’s opinions on Hannah’s condition, Ewing relied on an April 23 email that Hannah’s second-line supervisor sent to ODNI leadership on the topic.

32

whether Hannah’s attendance issues had been resolved, and, if they were, that Hannah could be hired for the Cyber position. O’Sullivan also stated that she would have been confident the attendance issues were resolved if Hannah had spent two to four weeks back at work post-leave without further issues (which she had by June 30).

Despite his offer to defer to O’Sullivan’s preferred approach, Ewing admitted that he “d[idn’t] recall” checking on Hannah’s post-leave attendance record. J.A. 90. He said he did not believe it was necessary because Hannah already “had several weeks or months of reportedly unpredictable and unreliable attendance at work.” J.A. 96. Ewing testified that he made up his mind not to hire Hannah by July 1, 2015, just one day after his email to O’Sullivan.

ODNI informed Hannah on July 7 that she was not selected for the Cyber position.

In a July 13 email to Hannah’s component director, Ewing claimed that both O’Sullivan and Director of National Intelligence James Clapper disapproved of hiring Hannah “given her recent performance.” J.A. 228. That claim was false; O’Sullivan explained that she and Clapper “weren’t involved in hiring decisions like that” and “wouldn’t have made that determination.” J.A. 151.

Before Hannah’s five-year term with ODNI ended, she applied to numerous other government positions. She was invited to interview for another ODNI role in the summer of 2015, but she withdrew after her component director told her that Ewing “would also likely reject [her]” from that position and that “it could get embarrassing.” J.A. 121. The component director likely gave that advice based on an email she received from Ewing, which acknowledged that Hannah “may continue to apply for openings” but instructed: “please do not suggest an outcome.” J.A. 228.

33

Hannah continued to apply to positions in the intelligence and national security communities, but she did not receive any other offers to interview. In all, she applied for nearly thirty government positions in 2015 and 2016. Hannah’s employment term with ODNI came to an end on March 27, 2016. Her final performance evaluation rated her “outstanding.” J.A. 126.

C.

After exhausting her administrative remedies, Hannah sued ODNI in the Eastern District of Virginia, alleging interference and retaliation in violation of the FMLA and several violations of the Rehabilitation Act of 1973, 29 U.S.C. § 794. The district court granted ODNI’s motion for summary judgment on all counts, and Hannah appealed.

On appeal, this Court affirmed the summary judgment order as to Hannah’s Rehabilitation Act and FMLA retaliation claims but vacated as to her FMLA interference claim. Hannah P., 916 F.3d at 332–33. With respect to the interference claim, the Court held that summary judgment was improper because a “reasonable jury could find that Hannah’s disclosure of her depression and her April 9, 2015 request for psychiatrist- recommended leave was sufficient to trigger [ODNI’s] responsibility to inquire further about whether Hannah was seeking FMLA leave.” Id. at 346. The Court thus remanded the FMLA interference claim for further proceedings. Id. at 348.

On remand, the district court held a bench trial to adjudicate that claim. Hannah’s supervisors testified at trial and admitted they were unaware of the FMLA—and therefore failed to provide Hannah notice of her FMLA rights—when she requested medical leave on April 9, 2015. After hearing that testimony, the district court did not allow closing

34

arguments, explaining, “it’s obvious that there is definitely liability because there was clearly a violation of the FMLA. I mean, there’s just no question about it. . . . It should have been conceded.” J.A. 214–15. The court issued an oral ruling “that the FMLA interference claim has been established” and directed the parties to file supplemental briefings on the issue of damages. J.A. 219.

In a December 30, 2021 memorandum opinion, the district court reaffirmed that ODNI was liable for FMLA interference. It explained that Hannah “put [ODNI] on sufficient notice that she was requesting leave guaranteed by the FMLA and that [ODNI] did not respond by making [her] aware of her FMLA rights and promptly allowing her to take leave.” Hannah P. v. Haines, 577 F. Supp. 3d 429, 443 (E.D. Va. 2021). The court then awarded Hannah compensatory damages for the 110 hours of annual leave ODNI improperly required her to use in May 2015, as well as liquidated damages authorized by the FMLA. Id.

At the same time, the district court rejected Hannah’s claims for other damages and equitable relief. As relevant here, it denied damages related to Hannah’s non-selection for the Cyber position. The court concluded that Hannah “has not carried her burden of showing that [ODNI’s] FMLA interference proximately caused her non-selection for the Cyber position” because her attendance issues after April 9, 2015—the date she requested leave—were not the “principal cause” of her non-selection. Id. at 445. Instead, the court reasoned, Ewing’s hiring decision “was a result of many intervening factors,” including Hannah’s attendance problems before her April 9 leave request and “Ewing’s poor impression of [her] as a prospective employee,” which was “significantly independent of the FMLA interference.” Id. at 445–46.

35

Hannah filed a motion for reconsideration, which the district court denied. The district court entered final judgment on April 4, 2022, and Hannah timely appealed.

II.

On appeal, ODNI does not dispute that it interfered with Hannah’s FMLA rights by failing to promptly grant her April 9 leave request. The sole issue before us is whether the district court, when assessing damages, erred in holding that ODNI’s interference did not cause Hannah’s non-selection for the Cyber position.

We review de novo the legal conclusions supporting a district court’s judgment in a bench trial—including its allocation of the burden of proof. See Everett v. Pitt Cnty. Bd. of Educ., 678 F.3d 281, 288 (4th Cir. 2012).

A.

The FMLA makes it “unlawful for any employer to interfere with, restrain, or deny the exercise of or the attempt to exercise[]” any right provided under the Act. 29 U.S.C. § 2615(a)(1). As the majority recognizes, § 2617 sets out the cause of action for FMLA interference. Ragsdale v. Wolverine World Wide, Inc., 535 U.S. 81, 89 (2002). To prevail in an interference action, an employee must prove (1) that he or she is entitled to leave under the FMLA; (2) “that the employer violated § 2615 by interfering with, restraining, or denying his or her exercise of FMLA rights”; and (3) that “the employee has been prejudiced by the violation.” Id.; see Roberts v. Gestamp W. Va., LLC, 45 F.4th 726, 732 (4th Cir. 2022).

The prejudice element is the only one at issue here. The Supreme Court has explained that an “employer is liable only for compensation and benefits lost ‘by reason of

36

the violation,’ § 2617(a)(1)(A)(i)(I), for other monetary losses sustained ‘as a direct result of the violation,’ § 2617(a)(1)(A)(i)(II), and for ‘appropriate’ equitable relief, including employment, reinstatement, and promotion, § 2617(a)(1)(B).” 4 Id. Because Hannah seeks to recover the wages and benefits she would have received in the Cyber position, we ask whether she lost that compensation “by reason of” ODNI’s interference with her FMLA rights. § 2617(a)(1)(A)(i)(I).

We already know that Hannah suffered one form of prejudice by reason of ODNI’s FMLA interference. As the district court held, ODNI’s interference caused Hannah to use paid annual leave for her four-week leave in May 2015, when ODNI should have allowed her to use only sick leave. See Hannah P., 577 F. Supp. 3d at 442–43. ODNI does not challenge that part of the judgment. The question is whether the FMLA interference also prejudiced Hannah by causing her non-selection for the Cyber position.

B.

1.

Fourth Circuit precedent provides guidance on evaluating causation in cases where the injury allegedly caused by an employer’s FMLA interference involves an adverse employment action. In such cases, we have held that “an employer may avoid liability if it shows it would have taken the contested adverse employment action regardless of the

4

During the proceedings below, the district court agreed with ODNI’s position that the FMLA prescribes different causation standards for liability (i.e., the prejudice element) and damages. J.A. 333. That position is at odds with Ragsdale, which clarifies that “prejudice” and “damages” are two different ways of framing the same question: whether the plaintiff suffered an injury “by reason of” the FMLA interference. See 535 U.S. at 89.

37

employee’s FMLA leave.” Roberts, 45 F.4th at 732–33; see also Yashenko v. Harrah’s N.C. Casino Co., 446 F.3d 541, 547 (4th Cir. 2006) (same).

My good colleagues in the majority recognize that this “same-decision test” governs Hannah’s non-selection theory of damages, as her alleged injury involves an adverse employment action. They also acknowledge that this test involves shifting the burden of proof to the employer. But as they see it, the burden shifts to ODNI here only if Hannah fully “prove[s] that her damages were caused ‘by reason of’ ODNI’s FMLA interference.” Ante at 16. That is, the majority asserts that Hannah’s damages claim fails at the first step if she cannot establish, by a preponderance of the evidence, that the FMLA interference was the but-for cause of her non-selection. 5 See Ante at 16–17.

This approach dismantles the burden-shifting feature of the test entirely. If the employee must establish but-for causation before the burden of proof shifts to the employer, there would never be a need to shift the burden. Instead, one of two scenarios would occur. In the first, the court concludes that the plaintiff failed to meet her “initial” burden of proving but-for causation. That conclusion alone would preclude recovery, so

5

The majority specifically states that Hannah “carried the initial burden of showing that she would have been entitled to the Cyber position had she not taken the leave.” Ante at 17 (emphasis added and cleaned up). This misstates Hannah’s theory of causation for her FMLA interference claim, which asserts that ODNI’s unlawful delay in granting her leave caused her non-selection for the Cyber position.

In her briefing, Hannah made the alternative argument that her absences during the period of interference in April 2015 should be treated as FMLA-protected leave. But she appears to have waived that theory below. See Opp. to Motion in Limine on Cyber Position, Hannah P. v. Haines, 1:16-cv-1030 (E.D. Va. Aug. 28, 2020), ECF No. 106 (clarifying that she was “seeking the loss of the Cyber position because she could not take FMLA leave” (emphasis in original)).

38

the court would have no reason to consider the employer’s reasons for taking the adverse employment action. In the second scenario, the court concludes that the plaintiff successfully proved but-for causation. There, the burden-shifting step would be equally pointless, as the plaintiff has already established that the injury was caused by FMLA interference, not some independent factor. Put simply, the majority’s approach recognizes that burden-shifting is a key feature of the same-decision test, but then completely erases that feature in practice. What remains is a one-step causation analysis that assigns the entire burden of proving but-for causation to the plaintiff.

2.

To be sure, a plaintiff like Hannah must make some initial showing of causation before the burden shifts to the employer to prove that it would have taken the same adverse employment action absent the FMLA interference. Yet for the second step of the test to have any force, the first step cannot require full but-for causation. I therefore would hold that the burden shifts to the employer if the plaintiff establishes a prima facie case on causation. To do so, a plaintiff must prove that the employer’s FMLA interference was a substantial or motivating factor in the adverse employment action.

This is precisely the initial showing that plaintiffs must make in other contexts where courts apply the same-decision test. First Amendment retaliation claims are one such example. See Mt. Healthy City Sch. Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 287 (1977). At the first step, the plaintiff must show that her protected First Amendment conduct was a “substantial” or “motivating factor” in the adverse employment action. Id. If the plaintiff clears that hurdle, the employer then bears the burden of proving “by a

39

preponderance of the evidence that it would have reached the same decision . . . even in the absence of the protected conduct.” Id.

So, too, with federal statutes that apply the same-decision test. Take, for instance, statutes that prohibit retaliation against whistleblowers who report violations of certain federal laws. See 49 U.S.C. § 31105(b)(1); 49 U.S.C. § 42121(b)(2)(B). A plaintiff first must establish a prima facie case by showing, in relevant part, that her protected activity was a “contributing factor” in an adverse employment action. Greatwide Dedicated Transport II, LLC v. U.S. Dep’t of Labor, 2023 WL 4279314, at *5, -- F.4th -- (4th Cir. 2023) (quoting § 42121(b)(2)(B)); see also § 31105(b) (adopting same standard). If the plaintiff makes that showing, the burden shifts to the employer to prove it “would have taken the same unfavorable personnel action in the absence of that [protected] behavior.” § 42121(b)(2)(B)(ii). Courts also apply this two-step burden-shifting test when determining whether an employer violated the National Labor Relations Act (“NLRA”), 29 U.S.C. § 158(a), by discharging an employee because of union activity. See NLRB v. Transp. Mgmt. Corp., 462 U.S. 393, 401–03 (1983) (upholding National Labor Relations Board’s interpretation of statute to require this test), abrogated on other grounds by Dir., Off. of Workers’ Comp. Programs, Dep’t of Labor v. Greenwich Collieries, 512 U.S. 267, 278 (1994).

Relying on the “substantial factor” standard for the first step of the causation analysis is also compatible with the text of the FMLA. Section 2617—which imposes liability for injuries suffered “by reason of” the employer’s FMLA interference—does appear to contemplate but-for causation. But that does not necessarily mean the plaintiff must shoulder

40

the entire burden of proof. 6 The burden-shifting approach simply creates a workable method of assessing but-for causation in cases where unlawful and lawful factors combined to result in the adverse employment action. In such a case, asking what course of action the employer would have taken in the absence of the unlawful factor requires the court to “engage in a hypothetical inquiry about what would have happened if the employer’s thoughts and other circumstances had been different.” Gross v. FBL Fin. Servs., Inc., 557 U.S. 167, 191 (2009) (Breyer, J., dissenting). Because this question turns on the employer’s subjective considerations, “the employer will often be in a stronger position than the employee to provide the answer.” Id. “All that a plaintiff can know . . . in such a context is that the forbidden motive did play a role in the employer’s decision.” Id.

The two-step burden-shifting test accounts for these problems of proof. Without it, very few plaintiffs could ever prove causation—after all, an employer rarely admits it took an action solely for unlawful reasons. If the plaintiff establishes a prima facie case and “the defendant fails to carry [its] burden, the inference is that ‘but for’ causation . . . has been shown: the plaintiff would not have been harmed had his rights not been violated by

6

I recognize that the Supreme Court and Fourth Circuit have interpreted similar language in different statutes as placing the full burden on the plaintiff to prove but-for causation. Gross v. FBL Fin. Servs., Inc., 557 U.S. 167, 176–77 (2009) (Age Discrimination in Employment Act, 29 U.S.C. § 623(a)); Sheppard v. Visitors of Va. State Univ., 993 F.3d 230, 236–37 (4th Cir. 2021) (Title IX); Gentry v. E. W. Partners Club Mgmt. Co., 816 F.3d 228, 235–36 (4th Cir. 2016) (Americans with Disabilities Act). But those decisions do not control our interpretation of the FMLA. This Court has already applied the second step of the same-decision test when determining causation in FMLA interference actions, see Roberts, 45 F.4th at 732–33, and, for the reasons already discussed, it would never be possible reach that step if the plaintiff bore the burden of proving but-for causation at step one.

41

the defendant.” Martin v. Duffy, 977 F.3d 294, 299 (4th Cir. 2020) (quoting Greene v. Doruff, 660 F.3d 975, 979 (7th Cir. 2011) (Posner, J.)). Or, in the language of the FMLA, the employer’s failure to meet its burden yields the inference that the employee suffered the adverse employment action “by reason of” the employer’s FMLA interference.

That this burden-shifting test can coexist with the but-for causation standard is not a novel concept. In cases where the tortious acts of multiple defendants, or the tortious and non-tortious acts of a single defendant, combined to cause an injury, “the common law of torts has long shifted the burden of proof to [] defendants to prove that their negligent actions were not the ‘but-for’ cause of the plaintiff’s injury.” Price Waterhouse v. Hopkins, 490 U.S. 228, 263 (1989) (O’Connor, J., concurring in judgment). In those scenarios, as here, “leaving the burden of persuasion on the plaintiff to prove ‘but-for’ causation would be [] unfair and destructive of [the law’s] deterrent purposes,” as the plaintiff usually would have no hope of identifying whether a particular causal factor was sufficient. Id. The burden-shifting approach remedies that problem.

Finally, regulations implementing the FMLA’s interference provisions also recognize that the two-step burden-shifting test complies with the “by reason of” standard. In clarifying the scope of prohibited interference, the regulations provide that an employer may not “use the taking of FMLA leave as a negative factor in employment actions, such as hiring, promotions, or disciplinary actions.” 29 C.F.R. § 825.220(c). This expressly adopts the “motivating factor” standard for certain FMLA interference claims. If the plaintiff makes that showing, the employer can then avoid liability by proving that it had an independently sufficient reason for taking the action. Yashenko, 446 F.3d at 547.

42

3.

Admittedly, our decisions in Roberts and Yashenko did not address the first step of the same-decision test. But that is unsurprising when one considers that, in those cases, the adverse employment action was both the alleged FMLA interference and the injury. Roberts, 45 F.4th at 731–33 (plaintiff’s employer fired him while he was on FMLA leave); Yashenko, 446 F.3d at 546–47 (plaintiff’s employer refused to restore him to his previous position when he returned from FMLA leave). When the interference and adverse employment action are one and the same, there is no need to carefully analyze whether the plaintiff has made a prima facie case; the close temporal relationship between the events readily supports an inference that FMLA-protected conduct was a substantial factor in the employment action. In such cases, the only question is whether the employer can prove it would have taken the same action in the absence of that protected conduct. 7

7

As a factual matter, FMLA interference claims like those in Roberts and Yashenko are indistinguishable from FMLA retaliation claims. In both contexts, the plaintiff alleges that she suffered an adverse employment action because of her FMLA-protected conduct. In such cases, a plaintiff can choose to seek recovery under either an interference theory or a retaliation theory.

In the past, this Court has applied the McDonnell Douglas framework to determine causation in FMLA retaliation actions. See Hannah P., 916 F.3d at 347. That test examines the intent of the employer and places the ultimate burden on the plaintiff to prove but-for causation. See Sharif v. United Airlines, Inc., 841 F.3d 199, 203 (4th Cir. 2016). However, regulations implementing the FMLA have clarified that retaliation claims fall under § 2615(a)(1)’s prohibition on interfering with an employee’s FMLA rights—not under § 2615(a)(2)’s separate prohibition on discriminating against employees who oppose unlawful practices, as we previously held. See Fry v. Rand Constr. Corp., 964 F.3d 239, 245 (4th Cir. 2020); see also 29 C.F.R. § 825.220(c) (“The Act’s prohibition against (Continued)

43

Just as importantly, there is no basis in our precedent for limiting the same-decision test to cases where the FMLA interference and the adverse employment action intersect. The causation inquiry in those cases is ultimately the same as the one in cases like Hannah’s, where the FMLA interference precedes the adverse employment action. In both scenarios, courts must determine whether the plaintiff’s employment-related injury occurred “by reason of” an unlawful factor (either the employee’s FMLA-protected activity or the employer’s interference with that activity).

C.

The district court below did not apply the same-decision test, and instead required Hannah to shoulder the full burden of proving that ODNI would have selected her for the Cyber position absent the FMLA interference. 8 That legal error requires us to vacate the

interference prohibits an employer from discriminating or retaliating against an employee or prospective employee for having exercised or attempted to exercise FMLA rights.”).

This shows that the causation standard for retaliation claims should track the causation standard for other interference claims—that is, the two-step same-decision test. As I have already discussed, the regulations themselves indicate that courts should apply that test to retaliation claims, providing that employers may not “use the taking of FMLA leave as a negative factor in employment actions.” 29 C.F.R. § 825.220(c); see also Fry, 962 F.3d at 245–46 (acknowledging that these regulations raised doubts about applying the McDonnell Douglas framework to FMLA retaliation claim, but declining to resolve the question because the plaintiff chose to rely on McDonnell Douglas).

8

The district court also held that damages related to Hannah’s non-selection were “consequential damages” that “the Fourth Circuit does not allow [] in FMLA cases.” J.A. 315 (citing Montgomery v. Maryland, 72 F. App’x 17, 19 (4th Cir. 2003)). Even if Montgomery were binding authority, it conflicts with Supreme Court precedent, which makes clear that employers liable for FMLA interference “are subject to consequential damages,” Ragsdale, 535 U.S. at 87 (citing § 2617(a)(1)), as long as the damages consist of “actual monetary losses,” Nev. Dep’t of Human Res. v. Hibbs, 538 U.S. 721, 740 (2003). (Continued)

44

district court’s judgment. See, e.g., Tatum v. RJR Pension Inv. Comm., 761 F.3d 346, 368 (4th Cir. 2014).

If we apply the correct causation analysis, the record leaves no doubt that ODNI’s FMLA interference caused Hannah’s non-selection.

1.

At the first step of the same-decision test, Hannah can easily satisfy her initial burden of proving that ODNI’s interference was a substantial factor in her non-selection for the Cyber position. ODNI interfered with Hannah’s FMLA rights by delaying action on her leave request from April 9 through at least April 28. Her attendance issues during that period were part and parcel of the interference. It is equally clear that Hannah’s April absences played a key role in Ewing’s decision not to select her for the Cyber position. Ewing himself conceded as much. He testified that he “consider[ed] her erratic attendance beginning sometime in early—in mid-January through April,” and reiterated that the time period went “all the way through April.” J.A. 86. Ewing further stated that hiring Hannah “would have been a real possibility” if he “had seen some pattern . . . of positive conduct after the 9th of April.” J.A. 87. And, in his June 30 email to O’Sullivan, Ewing placed special emphasis on Hannah’s post-April 9 absences, writing that “since referral to EAP, her late attendance at work has continued.” J.A. 221 (emphasis in original).

If the lost compensation is sufficiently attenuated from the FMLA violation, proximatecause rules might bar recovery. But here, there is a direct, unbroken causal chain between ODNI’s delay in granting Hannah leave and Hannah’s non-selection for the Cyber position.

45

When making the hiring decision, Ewing also considered the EAP counselor’s views on Hannah’s condition. In his email to O’Sullivan, Ewing noted that “EAP concluded [Hannah] does not have a medical problem, rather she is a disciplinary problem.” Id. Hannah’s EAP sessions, and the prejudicial report they generated, are inextricably bound up with ODNI’s FMLA interference. Hannah’s supervisors delayed action on her leave request because they believed she first needed to meet with EAP, and the EAP sessions took place while Hannah’s request remained pending. Ewing’s own testimony confirms that the EAP counselor’s views, in addition to Hannah’s April absences, played an important role in his hiring decision.

2.

Because Hannah can establish that ODNI’s FMLA interference was a substantial factor in her non-selection, the burden shifts to ODNI to prove that Ewing would have made the same hiring decision independent of the interference. On this record, ODNI cannot meet that burden.

Here, too, Ewing’s testimony is dispositive. At trial, he was asked if he would have hired Hannah had she “eliminated her attendance and reporting issues in April of 2015.” J.A. 89. Ewing replied, “I really don’t know . . . . we would need some time to understand that the conduct ha[d] actually been corrected. So I’m not sure that in April or May, we would have.” J.A. 90 (emphasis added). With that statement, Ewing admitted that he was unsure whether he would have made the same decision absent Hannah’s attendance

46

problems during the interference period. 9 By definition, that admission prevents ODNI from proving, by a preponderance of the evidence, that it had an independently sufficient reason for not selecting Hannah.

Because Hannah established a prima facie case on the issue of causation, and ODNI failed to prove that Ewing would have refused to hire Hannah had her supervisors promptly granted her April leave request, we can conclude that her non-selection occurred “by reason of” ODNI’s FMLA interference.

III.

In the alternative, my colleagues in the majority hold that the law-of-the-case doctrine prohibits us from considering whether ODNI’s FMLA interference caused Hannah’s non-selection for the Cyber position. That is simply not the case.

The law-of-the-case doctrine requires us to adhere to this Court’s prior factual findings and legal conclusions in the same case. Sejman v. Warner-Lambert Co., Inc., 845 F.2d 66, 69 (4th Cir. 1988). “Although the doctrine applies both to questions actually

9

This statement by Ewing ignores the fact that he did not make the hiring decision until the end of June, which provided an additional month of “nearly flawless” attendance data he could have considered. J.A. 119. It also contradicts Ewing’s other testimony, which indicates that he had no interest in considering Hannah’s attendance after she returned from leave. Specifically, at another point during trial, Ewing claimed that he did not need to review Hannah’s post-leave attendance record because she already “had several weeks or months of reportedly unpredictable and unreliable attendance at work.” J.A. 96. Regardless, the key point is that Ewing relied on Hannah’s April attendance issues when making the hiring decision and admitted that he did not know whether he would have made a different decision in the absence of those issues.

47

decided as well as those decided by necessary implication, it does not reach questions which might have been decided but were not.” Id. (internal quotation marks omitted).

A.

The majority first contends that this Court’s prior decision “conclusively determined that ODNI lawfully chose not to select [Hannah] for the permanent Cyber position” when it affirmed the summary judgment denial of Hannah’s FMLA retaliation and Rehabilitation Act discrimination claims. Ante at 22. In the majority’s view, those rulings preclude us from now holding that Hannah’s non-selection was caused by any unlawful act. But in reality, the retaliation and discrimination claims involved different legal standards and factual questions, and this Court’s conclusion that Hannah’s non-selection was “lawful” in those contexts does not control the outcome of Hannah’s FMLA interference claim.

Let’s begin with Hannah’s FMLA retaliation claim. When reviewing that claim, this Court applied the McDonnell Douglas burden-shifting framework to analyze causation. See Hannah P., 916 F.3d at 347. Unlike the same-decision test, the McDonnell Douglas framework involves three steps. A plaintiff first must establish a prima facie case of retaliation by showing that “(1) she engaged in a protected activity; (2) her employer took an adverse employment action against her; and (3) there was a causal link between the two events.” Id. (quoting Adams v. Anne Arundel Cnty. Pub. Schs., 789 F.3d 422, 429 (4th Cir. 2015)). If the plaintiff makes out a prima facie case, the burden shifts to the defendant to prove it had a legitimate, nondiscriminatory reason for taking the employment action. Id. If the defendant can make that showing, the burden shifts back to the plaintiff to prove that the defendant’s proffered reason was merely a pretext for retaliation. Id.

48

In affirming summary judgment for ODNI, this Court assumed that Hannah could establish a prima facie case of retaliation, but it agreed with the district court’s conclusion that “Hannah did not sufficiently rebut [ODNI’s] legitimate, nonretaliatory reason for not hiring [her] for the Cyber position”—namely, her “attendance issues.” Id. In other words, the Court concluded that Hannah could not prove that ODNI chose not to hire her in retaliation for taking FMLA leave.

That conclusion reveals absolutely nothing about whether ODNI’s FMLA interference—the delay in granting Hannah’s leave request—caused her non-selection. The only commonality between the claims is the adverse employment action; the alleged cause of that employment action is totally different. Even setting aside the fact that the retaliation analysis involved a different causation standard, the Court’s finding that Hannah’s non-selection was not caused by the FMLA leave she took in May 2015 does not resolve whether it was caused by ODNI’s unlawful delay in approving that leave.

Hannah’s Rehabilitation Act discrimination claim is similarly distinguishable. To establish causation in a Rehabilitation Act discrimination claim, a plaintiff must prove, among other elements, that “she suffered an adverse employment action solely on the basis of her disability.” Id. at 342. Once again applying the McDonnell Douglas framework, this Court held that Hannah failed to present sufficient evidence that ODNI’s proffered reason for her non-selection—her attendance record—“was merely a pretext for discriminating against her on the basis of her depression.” Id. at 342–43 (emphasis added). Whether ODNI refused to select Hannah for the Cyber position “on the basis of her

49

depression” has no bearing on whether ODNI’s delay in approving her leave request caused her non-selection.

At bottom, this Court’s prior decision did not categorically establish that Hannah’s non-selection for the Cyber position was a lawful act. It merely held that her non-selection was not (1) retaliation for her FMLA leave or (2) disability-based discrimination. Those questions are easily distinguishable, both legally and factually, from the one now before us.

B.

The majority next suggests that this Court’s prior opinion prohibited the district court from even finding that ODNI’s delay in granting Hannah’s leave request interfered with her FMLA rights. This is especially unconvincing.

The majority emphasizes this Court’s previous finding that ODNI did not “improperly delay[]” Hannah’s leave request for one month. Id. at 337. For context, the Court made that finding when examining Hannah’s claim that ODNI violated the Rehabilitation Act by failing to provide a reasonable accommodation for her depression. In affirming summary judgment for ODNI on that claim, the Court rejected Hannah’s argument that the “reasonable accommodation that she requested—a leave of absence— was improperly delayed” for one month. Id. It reasoned that Hannah “withdrew” her leave request on April 13 by telling her second-line supervisor that it was “on hold,” and did not “renew[]” the request until April 21. Id. The Court also observed that “the record demonstrates that Hannah’s supervisors were actively considering her request for leave during that time.” Id. at 338.

50

According to today’s majority, those findings prohibit us from treating ODNI’s slow response to Hannah’s leave request as an act that interfered with her FMLA rights. There are several flaws in this reasoning. First, this Court made the findings in a factual and legal context that is very different from the question now before us. In its prior opinion, the Court never found that there was no delay in granting the leave request. Rather, it merely determined that the delay did not violate the Rehabilitation Act, in large part because ODNI offered Hannah other accommodations this Court deemed reasonable and because her supervisors “actively consider[ed] her request for leave” during the delay (and ultimately approved it). Id. at 337–38.

That is quite different than determining whether ODNI interfered with Hannah’s FMLA rights by delaying action on her leave request. Whether or not the Court counts the period from April 13 to April 21, the fact remains that ODNI did not approve Hannah’s leave for weeks after she formally requested it. 10 That delay is more than enough to establish interference. Indeed, the very fact that Hannah’s supervisors “actively consider[ed]” her request for FMLA leave for several weeks—rather than immediately granting it—proves that interference occurred.

Moreover, this Court’s remand left open the possibility that ODNI’s delay could constitute FMLA interference. The opinion stated that “[a] reasonable jury could find that

10

Nor would excluding this period change any of the causation analysis. Many of Hannah’s April attendance issues—in fact, the majority of them—fell outside the window between April 13 and April 21. See J.A. 245 (April 22); J.A. 247 (April 23); J.A. 249 (April 24); J.A. 251 (April 28); J.A. 252 (April 29); see also J.A. 232 (compiled list of April absences). The majority does not grapple with this important detail.

51

Hannah’s disclosure of her depression and her April 9, 2015 request for psychiatrist- recommended leave was sufficient to trigger [ODNI’s] responsibility to inquire further about whether Hannah was seeking FMLA leave.” Id. at 346. That statement invited the district court to treat ODNI’s delay in making that inquiry as one form of FMLA interference—which is exactly what the district court did on remand. See Hannah P., 577 F. Supp. 3d at 443.

C.

As a last resort, the majority asserts that this Court remanded Hannah’s FMLA interference claim solely to address the “narrow question” of whether ODNI’s interference prejudiced Hannah by requiring her to use her annual leave. Ante at 21. The Court’s opinion never imposes an express limitation on the scope of the remand. Nonetheless, the majority reasons, the Court did not address Hannah’s separate theory that the interference caused her non-selection for the Cyber position, so it must have rejected that theory.

The opinion contains no such holding. In vacating summary judgment as to the FMLA interference claim, the Court held that Hannah had demonstrated a genuine dispute of material fact for each element. With respect to the prejudice element, the Court noted that “the record contains evidence that if Hannah had known that the FMLA protected her position, she would have used only sick leave for her leave of absence.” Hannah P., 916 F.3d at 347. Based on that evidence, it concluded that “a jury could find that Hannah was prejudiced by” ODNI’s FMLA interference. Id. Critically, though, the Court’s holding that Hannah created a genuine dispute of material fact on one theory of prejudice does not

52

implicitly reject other theories of prejudice. After all, a factual dispute on one theory is enough to defeat summary judgment.

On remand, the district court correctly recognized that Hannah’s non-selection for the Cyber position remained relevant to her interference claim. In fact, at trial, the district court described her non-selection as “the heart and soul of the damage claim or certainly part of the damage claim in this case.” J.A. 50. The district court adhered to the law of the case set out in our prior opinion, which stated only that this Court “remand[ed] for further proceedings as to [the FMLA interference] claim.” Hannah P., 916 F.3d at 348.

* * *

All told, the majority’s law-of-the-case analysis does not hold water. This Court’s prior holding that Hannah’s non-selection was not the result of FMLA retaliation or disability-based discrimination hardly precludes us from holding that it was casually connected to ODNI’s FMLA interference. Nor did the Court’s decision foreclose the district court from holding that ODNI interfered with Hannah’s FMLA rights by delaying action on her leave request. And, as the district court recognized, the Court left the door open to Hannah’s theory that that interference prejudiced her by causing her non-selection for the Cyber position.

IV.

At its core, this is not a complicated case. As the majority recognizes, Fourth Circuit precedent instructs us to apply the same-decision test to determine whether an adverse employment action was caused by FMLA interference. If we correctly apply the first step

53

of that test, Hannah can prove that ODNI’s interference was a substantial factor in Ewing’s decision not to hire her for the Cyber position. The record shows that Ewing placed particular emphasis on Hannah’s attendance issues during the April 2015 FMLA interference period when he made the hiring decision.

If the same-decision test means anything, that evidence must be enough to shift the burden to ODNI to prove that Ewing would have made the same decision absent the interference. ODNI cannot satisfy that burden here, as Ewing conceded that he “really d[idn’t] know” whether he would have hired Hannah had she eliminated her April 2015 attendance issues. J.A. 90. No part of this Court’s prior decision in this case requires us to reach the opposite conclusion.

Before closing, I feel compelled to make one final point. In the extended discussion of causation standards and burdens of proof that spans the preceding pages, it might be easy to lose sight of the human cost of today’s decision. By affirming the denial of damages for Hannah’s non-selection, this Court sanctions an employment decision that was deeply unfair—and entirely avoidable.

During her time at ODNI, Hannah was an exceptional analyst who excelled in very challenging roles. Yet when the symptoms of Hannah’s depression intensified in early 2015, ODNI completely failed her. Those failures began when ODNI delayed action on Hannah’s April 9 request for medical leave. Instead of heeding the advice of Hannah’s medical provider, ODNI officials forced Hannah—under threat of termination—to submit to EAP sessions that served absolutely no beneficial purpose.

54

EAP’s role in this case is particularly disturbing. The EAP referral was not just an improper substitute for the rights guaranteed by the FMLA; it was a central part of ODNI’s interference with those rights, and it placed Hannah in a very vulnerable position. The EAP counselor quickly decided that Hannah’s attendance issues were not a medical problem but a “disciplinary problem,” in part because of Hannah’s very understandable frustration with the sessions themselves. J.A. 221. Although the EAP is supposedly a “confidential counseling service,” J.A. 187, the EAP counselor disclosed many details about the sessions to Hannah’s supervisors. It is deeply ironic that the counselor, when sharing those details with Hannah’s second-line supervisor, chose to highlight Hannah’s concern “that the EAP process will create a paper trail that will adversely impact her future employment and career despite reassurances by EAP to the contrary.” J.A. 278 (emphasis added). Wasn’t the counselor violating those very assurances by discussing the sessions with the supervisor?

Of course, “create a paper trail” was exactly what the EAP counselor did. The EAP counselor’s report eventually made its way to Ewing, who relied on it when rejecting Hannah’s application for the Cyber position. And in making that decision, Ewing placed special emphasis on Hannah’s absences in April, during the very period when ODNI dragged its feet on her request for medical leave.

Time after time, Hannah’s colleagues—from certain supervisors to the EAP counselor to Ewing—either played doctor or refused to acknowledge the obvious signs that Hannah was suffering from depression, even though they knew about her clinical diagnosis. Their indifference to Hannah’s condition and ignorance of their FMLA obligations directly

55

caused her condition to deteriorate—which Ewing then used as justification to deny her a permanent role at ODNI. Calling their actions shameful is putting it lightly.

Hannah should have received laurels for her exemplary service to our nation.

Instead, she was exiled—unlawfully—by the very community she so ably served. With today’s decision, this Court fails to rectify that “tragic” injustice. Hannah P., 916 F.3d at 348 (Gregory, C.J., concurring in part and dissenting in part). I respectfully dissent.