Summary
finding an abuse of discretion where an insurer concluded that a salaried employee did not work the required 32 hours per week by demanding only written evidence of hours worked, even though insurer knew that written records would not establish the actual number of hours the employee worked
Summary of this case from Schindler v. Unum Life Ins. Co. of Am.Opinion
Case No. 06-CV-3910 (PJS/RLE).
March 4, 2008.
Michael B. Chase, Chase Law Office, for plaintiff.
Sarah J. Gorajski and Hal A. Shillingstad, Flynn Gaskins Bennett, L.L.P., for defendant.
MEMORANDUM OPINION AND ORDER
Defendant The Guardian Life Insurance Company of America ("Guardian") issued a group term life-insurance policy to the Paynesville Area Heath Care System ("Paynesville"). That policy provided coverage for "active full-time employee[s]" of Paynesville. Bowser Aff. Ex. A at 112 [Docket No. 25] (hereinafter "AR"). Charlotte Granite, the wife of plaintiff Dennis Granite, was a Paynesville employee. After Ms. Granite died of breast cancer on July 10, 2005, Mr. Granite submitted a claim for life-insurance benefits. Guardian denied the claim on the sole ground that Ms. Granite was not an "active full-time employee" of Paynesville. Mr. Granite now brings this action against Guardian to recover benefits pursuant to § 502(a)(1)(B) of the Employee Retirement Income Security Act ("ERISA"), 29 U.S.C. §§ 1001 et seq.
This matter is before the Court on the parties' cross-motions for summary judgment. For the reasons set forth below, the Court denies Guardian's motion and grants Mr. Granite's motion in part. Mr. Granite's claim for benefits is remanded to Guardian for reconsideration of whether he is entitled to benefits under the life-insurance policy.
I. BACKGROUND
As noted, the central question in this litigation is whether Ms. Granite ever became an "active full-time employee" of Paynesville and thus became covered under the Guardian life-insurance policy. The policy does not define the term "active full-time employee," but the policy does define "full-time" as follows:
Full-time means the employee regularly works at least the number of hours in the normal work week set by the employer (but not less than 32 hours per week), at his employer's place of business.
AR 117. The policy also provides as follows:
An employee must be actively at work, and working his or her regular number of hours, on the date his or her coverage is scheduled to start. And he or she must have met all of the conditions of eligibility which apply to him or her. If an employee is not actively at work on his or her scheduled effective date, we will postpone the start of his or her coverage until he or she returns to active full-time work.
Sometimes, a scheduled effective date is not a regularly scheduled work day. But an employee's coverage will start on that date if he or she was actively at work, and working his or her regular number of hours, on his or her last regularly scheduled work day.
AR 114.
Paynesville hired Ms. Granite as a full-time physician assistant in late 2004. AR 548. Although Ms. Granite's effective date of employment was December 27, 2004, it appears that her first day of work was January 3, 2005. Id.; AR 567. Ms. Granite's duties included seeing patients, charting, and administrative work. AR 548. As a salaried employee, Ms. Granite did not punch a time clock and was not required to keep track of how many hours she worked. Id.
Unfortunately, shortly before she began working for Paynesville, Ms. Granite was diagnosed with an aggressive form of breast cancer. AR 301. She underwent a mastectomy in early December 2004, and then received chemotherapy and, later, radiation therapy. AR 269, 294-95, 301-02. Despite these treatments, Ms. Granite's health continued to deteriorate, and she stopped working altogether by April 2005. AR 567.
After Ms. Granite's death, Mr. Granite submitted a claim for life insurance benefits to Guardian. AR 620. Paulette Hagen, Paynesville's director of human resources, signed the claim form on Paynesville's behalf. AR 621. By signing the form, Paynesville certified that Ms. Granite "has been a full-time, active employee for whom premiums have been paid." Id.
After receiving the claim, Guardian's claim analyst, Janine Bowser, sought information from Paynesville to verify that Ms. Granite had indeed been an "active full-time employee." Specifically, on August 10, 2005, Bowser sent a letter to Hagen stating, in relevant part:
In order to continue the processing of the Group Life Claim for Charlotte Granite, please submit payroll and attendance records for Charlotte Granite from December 1, 2004 through April 13, 2005. If attendance records are not available, please provide us with work documents signed and dated by Charlotte Granite certifying active full time employment (i.e. letters, memos work orders).
AR 608. Significantly, Bowser sought only written evidence of Ms. Granite's work schedule in this letter.
Hagen responded by sending Bowser the following documents: (1) Paynesville's September 7, 2004, offer letter, signed and dated by Ms. Granite, in which Paynesville offered Ms. Granite the position of physician assistant with a 38-40 hour workweek, AR 566; (2) a "Caregiver's Statistical Report," which lists the number of patients Ms. Granite saw each day for the period January 3, 2005 through April 13, 2005, AR 567; and (3) a "Caregiver's Scheduled Events Report," which provided detail with respect to each of Ms. Granite's patient appointments, including the scheduled time and length of each appointment. AR 566-607. On the Caregiver's Statistical Report, Hagen had calculated the number of hours Ms. Granite spent with patients each day. AR 567. Critically, though, the number of hours that Ms. Granite spent with patients each day was not the same as the number of hours that Ms. Granite worked each day, as Ms. Granite's duties extended beyond seeing patients.
Bowser forwarded these documents to her supervisor for review. Her supervisor responded:
Cannot pay this claim. This employee did not meet the hourly work requirement of 32 hours per week, per the attendance records. Never worked full-time from the effective date of coverage 1/1/05 through the last day worked 4/13/05. Any exceptions?
AR 556.
The supervisor was, of course, wrong in asserting that "the attendance records" had indicated that Ms. Granite "did not meet the hourly work requirement of 32 hours per week." As noted, the records that Hagen supplied would permit Guardian to calculate the number of hours that Ms. Granite spent with patients, but it would not permit Guardian to calculate the number of hours that Ms. Granite spent at work. Bowser's supervisor had no basis for concluding that Ms. Granite had not worked 32 hours per week.
To her credit, Bowser did not drop the matter, but instead called Hagen to ask for additional information about the number of hours Ms. Granite worked. AR 555. Not to her credit, Bowser continued to insist on written evidence. Hagen again told Bowser that, because Ms. Granite was a salaried employee, no written evidence of the number of hours that she worked existed. AR 550, 552. Also, Hagen again informed Bowser that the statistical reports only tracked the time Ms. Granite spent with patients and that Ms. Granite spent the balance of her workday completing paperwork and performing administrative duties for a total of 38 to 40 hours of work per week. AR 555. Finally, Hagen sent Bowser a letter stating:
As a follow-up to our phone conversation, Ms. Granite was hired full-time with the Paynesville Area Health Care System effective 12/27/04.
Ms. Granite's responsibilities were that of a Physician Assistant; seeing patients, patient diagnosis, patient follow-up, charting, and [miscellaneous] administrative type work.
With this being said, Ms. Granite was in a salaried exempt position, did not punch a time clock, nor was she required to track hours worked. She came into this position 12/27/04 according to the terms of her offer and acceptance letter dated 9/7/04.
AR 548.
At this point, Guardian knew that, if it wanted to determine how many hours Ms. Granite worked each week, it would need to expand its search beyond written evidence — by, for example, interviewing Ms. Granite's supervisors or co-workers, or by interviewing Mr. Granite. Guardian inexplicably chose not to pursue such evidence. Instead, Guardian obtained Ms. Granite's medical records from December 1, 2004 through the date of her death. After analyzing nearly 200 pages of medical records, Guardian located a single reference to Ms. Granite being employed "part time" in a note dated February 2, 2005, AR 290, as well as several later discussions of Ms. Granite's desire to go back to work and her doctors' advice to take it easy, AR 402, 294. Ms. Granite's medical records also indicated that she suffered from nausea related to her chemotherapy. AR 275, 299. Contrary to Guardian's assertions before this Court, however, see Def.'s Mem. Supp. Summ. J. 5, these records did not state that the nausea caused Ms. Granite to miss work, AR 275, 299.
Based solely on this evidence — that is, based solely on written records that did not show the number of hours that Ms. Granite worked — Guardian denied Mr. Granite's claim for benefits on the ground that she had not worked enough hours. AR 252-55. Specifically, in a letter dated October 19, 2005, Guardian informed Mr. Granite that it was denying his claim because, according to the "Caregiver's Statistical Report," Ms. Granite worked less than 32 hours per week. AR 253-54. Guardian made this statement even though it had been told — clearly and repeatedly — that the "Caregiver's Statistical Report" only provided information about the number of hours that Ms. Granite spent with patients and not about the number of hours that Ms. Granite spent at work. Guardian also cited Ms. Granite's February 2, 2005 medical record referring to her "part time" employment. AR 253.
The letter went on to inform Mr. Granite that he had the right to appeal and of the type of evidence he would need to prevail on that appeal:
Should you wish to appeal this decision, you need to state your reasons for appeal in writing and provide Guardian with any proof (i.e. payroll and attendance records) that Ms. Granite was actively at work on a full time basis on or after January 1, 2005.
AR 255. An attached document entitled "ERISA — New Procedures for Life Extended Life Benefit Claims" further informed Mr. Granite of the appeal procedure: "Submit your formal request for reconsideration in writing within 180 days of the date of the attached letter with the additional medical and/or other information mentioned within the letter." AR 256.
Mr. Granite timely requested review. AR 248. Mr. Granite did not provide further documentation of Ms. Granite's work schedule. Instead, he pointed out, as had Hagen before him, that Ms. Granite was not an hourly employee who punched a time clock, and thus no payroll or attendance records documented the number of hours she worked. AR 249. Mr. Granite also noted that Paynesville paid Ms. Granite for full-time work and that she performed work not reflected on the "Caregiver's Statistical Report." Id.
On February 15, 2006, Granite affirmed its original decision to deny benefits. AR 238. Guardian again concluded, based on the documentation previously submitted, that Ms. Granite never worked 32 hours per week. AR 240. Guardian dismissed Mr. Granite's assertion that Ms. Granite worked full-time because Mr. Granite "submitted no compelling proof to substantiate [his] claim." Id. Guardian also cited Ms. Granite's medical records as support for its conclusion that "it would have been unlikely that Ms. Granite would have been able to perform work on a full-time basis." Id.
II. ANALYSIS A. Summary Judgment Standard
Summary judgment is appropriate "if the pleadings, the discovery and disclosure materials on file, and any affidavits show that there is no genuine issue as to any material fact and that the movant is entitled to judgment as a matter of law." Fed.R.Civ.P. 56(c). A dispute over a fact is "material" only if its resolution might affect the outcome of the suit under the governing substantive law. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). A dispute over a fact is "genuine" only if the evidence is such that a reasonable jury could return a verdict for either party. Ohio Cas. Ins. Co. v. Union Pac. R.R., 469 F.3d 1158, 1162 (8th Cir. 2006). In considering a motion for summary judgment, a court "must view the evidence and the inferences that may be reasonably drawn from the evidence in the light most favorable to the non-moving party." Winthrop Res. Corp. v. Eaton Hydraulics, Inc., 361 F.3d 465, 468 (8th Cir. 2004).B. Plaintiff's Claim
The parties dispute two aspects of Guardian's denial of Mr. Granite's claim: (1) whether, under the policy, Ms. Granite was required to work a 32-hour week before coverage attached, and, if so, (2) whether Ms. Granite ever worked a 32-hour week. The Court will address these issues in reverse order.1. Guardian's Factual Finding
Guardian interprets the policy to require that, before life-insurance coverage attaches, an employee must work 32 hours in one week. Because it found that Ms. Granite never worked a 32-hour week, Guardian found that coverage never attached. As explained below, the Court finds that Guardian abused its discretion in interpreting the policy to require an employee to work a 32-hour week before coverage attaches. But even if the Court agreed with Guardian's interpretation, the Court would nevertheless remand this case because Guardian also abused its discretion in concluding, on the basis of the evidence before it, that Ms. Granite never worked 32 hours in a week.
The policy gives Guardian the discretion to make benefit determinations, and thus the Court would ordinarily be limited to determining whether Guardian's conclusion that Ms. Granite never worked a 32-hour week was supported by "substantial evidence, that is, `such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.'" King v. Hartford Life Accident Ins. Co., 414 F.3d 994, 999 (8th Cir. 2005) (en banc) (quoting Donoho v. FMC Corp., 74 F.3d 894, 900 n. 10 (8th Cir. 1996)). But Mr. Granite argues that the Court should apply a less-deferential standard of review under Woo v. Deluxe Corp., 144 F.3d 1157 (8th Cir. 1998). In Woo, the Eighth Circuit held that a claimant may be entitled to review under a less-deferential standard if he can present "material, probative evidence demonstrating that (1) a palpable conflict of interest or a serious procedural irregularity existed, which (2) caused a serious breach of the plan administrator's fiduciary duty. . . ." Id. at 1160. Mr. Granite argues both that Guardian had a conflict of interest and that Guardian's failure to adequately investigate his claim for benefits was a serious procedural irregularity. The combination of these circumstances, Mr. Granite contends, warrants a less deferential standard of review.
The policy provides as follows: "The Guardian is the Claims Fiduciary with discretionary authority to determine eligibility for benefits and to construe the terms of the plan with respect to claims." AR 224.
Whether an insurer in Guardian's position operates under "a palpable conflict of interest" — and, if so, whether that insurer's decisions are therefore entitled to less deference — are difficult questions that are now pending before the United States Supreme Court. See Glenn v. MetLife, 461 F.3d 660 (6th Cir. 2006), cert. granted, 128 S. Ct. 1117 (U.S. Jan. 18, 2008) (No. 06-923) (granting certiorari on the questions "Whether the Sixth Circuit erred in holding . . . that the fact that a claim administrator of an ERISA plan also funds the plan benefits, without more, constitutes a "conflict of interest" which must be weighed in a judicial review of the administrator's benefit determination . . .?" and "If an administrator that both determines and pays claims under an ERISA plan is deemed to be operating under a conflict of interest, how should that conflict be taken into account on judicial review of a discretionary benefit determination?"). Fortunately, though, this Court need not reach those questions, because Guardian's investigation of Mr. Granite's claim was so clearly inadequate. Under Eighth Circuit precedent, "[t]he lack of a thorough investigation by a fiduciary can result in a serious procedural irregularity requiring a less deferential standard of review." Sahulka v. Lucent Techs., Inc., 206 F.3d 763, 769 (8th Cir. 2000) (citing Woo).
Guardian began its investigation by asking Paynesville for written records of Ms. Granite's work schedule. Specifically, Guardian requested "payroll and attendance records" and, if those were not available, "work documents signed and dated by Charlotte Granite certifying active full time employment (i.e. letters, memos work orders)." AR 565. Paynesville responded with patient-care records and other documents from Ms. Granite's employee file.
Asking for only written records regarding the number of hours that Ms. Granite worked may have been a reasonable place for Guardian to start its investigation. But, after getting those records, Guardian should have realized that the records reported only the hours that Ms. Granite spent with patients, and not the hours that Ms. Granite spent at work. Even if this was not immediately apparent to Guardian, Paynesville's Hagen informed Guardian's Bowser at least three times that Ms. Granite worked more hours than were reflected in the patient-care records. Hagen also made it clear that no further documentation was available to establish Ms. Granite's precise work schedule: "Ms. Granite was in a salaried exempt position, did not punch a time clock, nor was she required to track hours worked." AR 548.
At this point, Guardian knew that the records supplied by Paynesville would not give Guardian a basis for determining whether Ms. Granite ever worked a 32-hour week, and Guardian knew that it would have to expand its investigation beyond written records. And yet Bowser's supervisor told Bowser that Guardian "[c]annot pay this claim" because Ms. Granite "did not meet the hourly work requirement of 32 hours per week, per the attendance records." AR 556. The supervisor's treatment of Mr. Granite's claim was almost recklessly cavalier; as already noted, Guardian at that point did not have "attendance records" — it had patient-care records — and Guardian did not have any way of knowing whether Ms. Granite "[met] the hourly work requirement of 32 hours per week. . . . "
Bowser, at least, recognized that Guardian needed to do more investigation. But Bowser continued to be almost obsessive in pursuing only written evidence, even though she had been told that written evidence of Ms. Granite's hours did not exist. Bowser inexplicably failed to interview Mr. Granite or a single person who worked with Ms. Granite. In other words, Bowser did nothing to collect the best evidence available as to the number of hours Ms. Granite worked, even though that evidence was readily available.
Rather than making a brief phone call to Mr. Granite or to one of Ms. Granite's co-workers, Bowser instead devoted hours to securing and reviewing Ms. Granite's medical records. Nothing in those medical records gave Guardian any basis for concluding that Ms. Granite had never worked a 32-hour week. Indeed, virtually nothing in those medical records — aside from a single, ambiguous reference to Ms. Granite working "part time" — even discussed the number of hours that Ms. Granite worked. And yet Guardian concluded that the medical records, combined with the patient-care records, gave it a sufficient basis to deny Mr. Granite's claim.
Guardian's attorneys — who, of course, were in no way responsible for Guardian's treatment of Mr. Granite — gamely suggest that, when Bowser spoke with Hagen on the telephone, Bowser might have asked for all relevant information, and not just relevant information that was in writing. The record refutes this suggestion, though. Bowser first asked for payroll and attendance records. Bowser did not ask, in the alternative, for any other type of evidence, as might have been expected. Instead, Bowser made the odd alternative request for "work documents signed and dated by Charlotte Granite certifying active full time employment," reflecting Bowser's fixation on written evidence.
The remainder of Hagen's interactions with Bowser, and Bowser's notes of those conversations, also supports the view that Bowser sought only documents and cared not at all about non-documentary evidence. Hagen promptly responded to Bowser's initial request for documents. Hagen also called Bowser to point out that no further documents were available, and followed up that call with a written letter emphasizing that Ms. Granite was not required to track her hours. AR 550, 548. In Bowser's own notes of these discussions, she seems to equate "further info" with the type of information generated by "punch[ing] a clock." AR 550. Given Hagen's diligence in following up with Bowser, there is no reason to suppose that Hagen would have failed to respond to a request for the names of people who would have personal knowledge of Ms. Granite's work schedule. And Bowser herself never sought information from one obvious source of knowledge of whose identity she was well aware — Mr. Granite. From all this, it is obvious that Bowser never sought any non-documentary evidence of Ms. Granite's hours before denying Mr. Granite's claim. Indeed, Guardian's October 19, 2005 denial letter essentially admits as much: "In order to document Ms. Granite's full-time employment status, we requested copies of attendance records, and/or payroll records, etc." AR 253.
Guardian's failure to seek other relevant evidence of Ms. Granite's hours makes this case very similar to Woo. In Woo, the plaintiff suffered from health problems and resigned from her job. Woo, 144 F.3d at 1159-60. Later, the plaintiff was diagnosed with scleroderma, a rare and difficult-to-diagnose disease. Id. at 1160. Both a scleroderma expert and the plaintiff's treating neurologist concluded that, in retrospect, the plaintiff suffered from scleroderma before she quit her job. Id. Nevertheless, the insurer failed to have a scleroderma expert review her claim. Id. at 1161. The insurer argued that it had no duty to do so because her treating physicians did not mention her disability or diagnose her with scleroderma before she quit her job. Id. The Eighth Circuit held that, in light of the later medical evidence of the plaintiff's condition, the insurer had a duty to have a scleroderma expert review her claim. Id. at 1161.
Similarly, here Guardian was repeatedly told that written evidence of the hours worked by Ms. Granite did not exist and that the patient-care records reflected only hours spent with patients. Rather than seek other types of evidence, though, Guardian doggedly insisted on documentary evidence that it had been told did not exist. When no such evidence was forthcoming, Guardian treated Ms. Granite's patient-care records as affirmative proof that she never worked the required number of hours: In its initial denial letter, Guardian stated that "[i]n fact, according to the `Caregiver's Statistical Report,' Ms. Granite worked [the hours listed on the Report]." AR 253 (emphasis added). This assertion was false, and Guardian knew it was false.
After denying Mr. Granite's claim based on a false assertion about an inadequate record, Guardian then compounded its mistake by informing Mr. Granite that, to appeal, he needed to provide "any proof (i.e. payroll and attendance records) that Ms. Granite was actively at work on a full time basis on or after January 1, 2005." AR 255. "I.e." is the abbreviation for the Latin phrase "id est," which means "that is." Black's Law Dictionary 762 (8th ed. 2004). By using the restrictive phrase "i.e. payroll and attendance records," Guardian was telling Mr. Granite that, in order to succeed on appeal, he had to provide "payroll and attendance records." Not surprisingly, Mr. Granite could provide no such proof, and Guardian denied his appeal.
Under the circumstances, the Court has no trouble concluding that Guardian's insistence on making a decision on the basis of only written evidence, its failure to seek the most reliable evidence of the hours Ms. Granite worked (indeed, its failure to seek any evidence that did not appear in a document), its treatment of Ms. Granite's patient-care records as records of hours worked, its false assertions about the record, and its suggestion to Mr. Granite that on appeal he could present only "payroll and attendance records" all constituted serious procedural irregularities. Mr. Granite has thus met the first part of the Woo test.
The Court also has no trouble concluding that Mr. Granite has shown that these irregularities caused a serious breach of Guardian's fiduciary duty. To meet this second part of the Woo test, Mr. Granite must show that the irregularity had a connection to the substantive decision reached. Barnhart v. UNUM Life Ins. Co. of Am., 179 F.3d 583, 588-89 (8th Cir. 1999). Guardian based its denial of benefits on its finding that Ms. Granite never worked 32 hours in one week. Obviously, this finding is directly connected to the procedural irregularities identified above. The Court therefore concludes that Mr. Granite has met the two-part Woo test.
The next question, then, is how that affects the Court's review of Guardian's decision. The Woo test does not do away with the abuse-of-discretion standard. Instead, the court applies a "sliding-scale" standard in which the Court takes the procedural irregularity into consideration. Woo, 144 F.3d at 1161. The deference given to the administrator's decision decreases in proportion to the seriousness of the irregularity. Id. When a fiduciary's conduct has been particularly egregious, a court may require that the record contain evidence bordering on a preponderance to uphold the fiduciary's decision. Id. at 1162. Although the Court regards Guardian's failures as particularly egregious, the Court need not determine precisely where on the "sliding scale" Guardian has placed itself, for it is clear that, even under the ordinary abuse-of-discretion standard, Guardian's decision cannot be upheld.
Guardian based its determination that Ms. Granite never worked a 32-hour week on the following: (1) the "Caregiver's Statistical Report"; (2) a statement in a February 2, 2005 medical record referring to Ms. Granite's "part time" employment; (3) a statement in a February 28, 2005 medical record referring to Ms. Granite's desire to return to work; and (4) a similar statement in a March 3, 2005 medical record. See AR 240, 253 (listing reasons for denial of coverage). Before this Court, Guardian also cited evidence that Ms. Granite suffered from nausea as a result of chemotherapy. See AR 275, 299.
According to Guardian, the "Caregiver's Statistical Report" establishes that Ms. Granite never worked more than 25.5 hours in a week — the number of hours that, according to Guardian, Ms. Granite worked during the week of January 10, 2005. But as the Court has made abundantly clear — and as Paynesville made abundantly clear to Guardian — the "Caregiver's Statistical Report" reports only the number of hours that Ms. Granite spent with patients, and not the number of hours that she spent at work. Ms. Granite spent 25.5 hours with patients during the week of January 10, 2005. But her duties included more than seeing patients, and thus we can be certain that Ms. Granite spent more than 25.5 hours at work during the week of January 10. How many more we cannot know, because Guardian did not bother to gather any evidence of how the number of hours Ms. Granite spent with patients correlated with the number of hours she spent at work.
Guardian also relies on Ms. Granite's medical records. But those records are largely irrelevant. As noted, Ms. Granite started work on January 3, 2005, and her patient-visiting hours peaked at 25.5 during the week of January 10, 2005. Yet almost all of the evidence in the medical records on which Guardian relies relates to February 2005 or later. Moreover, the evidence in the medical records is extremely weak even as to the number of hours that Ms. Granite worked in February and later. For example, Guardian relies heavily on the February 2, 2005 reference to her "part time" work. But the reference merely notes that "since [Ms. Granite] works part time in the clinic, Thursday is the best day for her to get her chemotherapy." AR 290. The oncologist who noted her "part time" work was not trying to quantify the total amount of time Ms. Granite worked. Instead, the notation merely conveys the fact that Ms. Granite was not working Monday through Friday, 8:00 a.m. to 5:00 p.m. As no one contends that Ms. Granite worked five full days per week, an offhand reference to her "part time" status in this context is hardly evidence of anything. Guardian's heavy reliance on it, when it could have so easily asked for direct information from Ms. Granite's co-workers (or from Mr. Granite himself), speaks volumes about the lack of evidence in the record to support Guardian's decision.
The remainder of the medical evidence is similarly weak. Guardian contends that records in January demonstrate that Ms. Granite missed work due to chemotherapy-related nausea. As noted, however, these records merely state that she suffered from nausea. They nowhere state that Ms. Granite missed work as a result, and Guardian's assertion to the contrary is highly misleading. More importantly, there is no doubt that Ms. Granite saw patients for at least 25.5 hours during the week of January 10, 2005. The vague references to her medical condition give Guardian no basis to conclude that she could not have worked an additional 6.5 hours that week on charting and administrative matters.
In short, Guardian conducted a sloppy investigation and was willfully blind to the most relevant information; it assembled a woefully inadequate record; and then it drew unwarranted conclusions on the basis of that record. Guardian's decision that Ms. Granite never worked a 32-hour week is plainly not supported by "such relevant evidence as a reasonable mind might accept as adequate," King, 414 F.3d at 999 (citation and quotations omitted), and remand is appropriate, Abram v. Cargill, Inc., 395 F.3d 882, 887 (8th Cir. 2005). Even assuming, then, that Ms. Granite was not covered under the Guardian policy unless she worked a 32-hour week, Guardian abused its discretion in concluding that Ms. Granite had not, in fact, worked such a week.
2. Guardian's Policy Interpretation
Guardian's stated reason for denying Mr. Granite's claim was that, because Ms. Granite never worked 32 hours in one week, she never became an "active full-time employee" of Paynesville, and thus she never became covered under the policy. The parties have focused most of their attention on the dispute over Guardian's factual finding that Ms. Granite never worked a 32-hour week — a factual finding that, in the previous section, the Court found to be an abuse of discretion. Unfortunately, the parties have devoted comparatively little attention to the logically antecedent question of whether Ms. Granite had to work a 32-hour week before coverage attached.
When an ERISA plan gives the plan administrator or fiduciary discretion to interpret the terms of the plan (as does Guardian's policy), courts review that interpretation only for an abuse of discretion. Firestone Tire Rubber Co. v. Bruch, 489 U.S. 101, 115 (1989). Under this standard, the court must determine whether the interpretation is "reasonable" under the five-factor test set forth in Finley v. Special Agents Mutual Benefit Association, Inc., 957 F.2d 617 (8th Cir. 1992). The Finley factors are: (1) whether the fiduciary's interpretation is contrary to the clear language of the plan; (2) whether the interpretation conflicts with the substantive or procedural requirements of ERISA; (3) whether the interpretation renders any language in the plan meaningless or internally inconsistent; (4) whether the interpretation is consistent with the goals of the plan; and (5) whether the fiduciary has consistently followed the interpretation. Id. at 621. The Eighth Circuit has directed district courts to examine all five Finley factors. Lickteig v. Bus. Men's Assurance Co. of Am., 61 F.3d 579, 584 (8th Cir. 1995).
Mr. Granite argues that, because of the procedural irregularities described above, Guardian's interpretation of its policy is entitled to a less deferential standard of review than "abuse of discretion." But Mr. Granite does not explain why procedural irregularities that infect an insurer's factual investigation of a claim should affect the standard of review that applies to the insurer's interpretation of a policy term. Because, in this case, the Court finds that Guardian's policy interpretation cannot be upheld even under the abuse-of-discretion standard, the Court need not address this issue further.
The Guardian policy provides that "[a]n employee must be actively at work, and working his or her regular number of hours, on the date his or her coverage is scheduled to start." AR 114. The policy further provides that "[i]f an employee is not actively at work on his or her scheduled effective date, we will postpone the start of his or her coverage until he or she returns to active full-time work." Id. If that is all that the policy provided, then Ms. Granite would unquestionably be covered. Ms. Granite's coverage under the Guardian policy was scheduled to start on January 1, 2005. AR 254, 556. Although the effective date of her employment was December 27, 2004, it appears that her first day of work was January 3, 2005. AR 548, 567. Thus, coverage would have attached on January 3, 2005.
As Guardian points out, though, the policy continues: "And he or she must have met all of the conditions of eligibility which apply to him or her." AR 114. In other words, for coverage to attach when Ms. Granite showed up to work on January 3, "she must have met all of the conditions of eligibility which appl[ied] to . . . her." And there is no question that one of the "conditions of eligibility" is that Ms. Granite had to be an "active full-time employee. . . ." AR 112. Thus, Ms. Granite was covered on January 3, 2005, only if she was an "active full-time employee" on that date.
Was she? Therein lies the problem. The policy defines a "full-time" employee as one who "regularly works at least the number of hours in the normal work week set by the employer (but not less than 32 hours per week), at his employer's place of business." AR 117. For new employees (such as Ms. Granite), however, "the date . . . coverage is scheduled to start" is generally going to be the same as the date that work is scheduled to start. So how do we determine on an employee's first day of work whether she "regularly works" at least 32 hours per week?
The most obvious approach would be to look at the employment contract. After all, workers are normally classified as either full-time or part-time from the outset of their employment. And a full-time employee would normally expect that if "the date . . . her coverage [was] scheduled to start" was the same as the date she started work, she would be covered from the outset. Under this approach, then, coverage attached to Ms. Granite on January 3, because when she showed up for her first day of work, she was under contract to regularly work at least 32 hours per week.
Guardian, however, takes another approach: It insists that a new employee such as Ms. Granite does not become an employee who "regularly works" at least 32 hours per week until she actually works one 32-hour week. In other words, Guardian relies not on whether an employee is contractually obliged to work 32-hour weeks, but instead on whether the employee has in the past worked at least one 32-hour week.
There are several problems with Guardian's interpretation:
First, if "works" refers to the historical performance of the employee and not her contractual status, then a person could not be regarded as an employee who " regularly works" at least 32 hours per week after working just one 32-hour week. "Regular" means "[c]ustomary, usual, or normal[.]" American Heritage Dictionary of the English Language 1471 (4th ed. 2006). Working a single 32-hour week does not qualify someone as an employee who "regularly" works 32-hour weeks.
Second and more importantly, nothing in the Guardian policy provides that a new employee must work a 32-hour week before he or she will be eligible for life-insurance coverage. To the contrary, the policy provides that an employee is covered "on the date his or her coverage is scheduled to start," as long as, on that day, the employee is "actively at work, and working his or her regular number of hours. . . ." If the employee "is not actively at work on his or her scheduled effective date," then determining when coverage attaches requires consideration of whether the "scheduled effective date" is "a regularly scheduled work day." If it is, then Guardian "will postpone the start of [the employee's] coverage until he or she returns to active full-time work." If it is not, then the "employee's coverage will start on [the scheduled effective date] if he or she was actively at work, and working his or her regular number of hours, on his or her last regularly scheduled work day." Presumably, if the employee was not "actively at work . . . on his or her last regularly scheduled work day," coverage would not attach until the employee returns to active full-time work.
As Guardian interprets the policy, all of this language would essentially be meaningless with respect to a typical new employee such as Ms. Granite — that is, a new employee whose life insurance is scheduled to start on or shortly before the date on which she starts work. For such an employee, it would be a waste of time to work through this policy language, as coverage would never start on the "date . . . coverage is scheduled to start" or on the "scheduled effective date." Those phrases would take on an almost Orwellian hue.
Suppose, for example, that the "scheduled effective date" of a new employee's coverage is Friday, February 1, 2008. Suppose further that the employee begins work on that day — that is, on Friday, February 1, 2008. According to the policy, the employee should be covered from the outset, as he or she is "actively at work, and working his or her regular number of hours," on "the date his or her coverage is scheduled to start." According to Guardian, though, this employee cannot be covered until he or she first works 32 hours in a single week.
Suppose, instead, that the "scheduled effective date" of a new employee's coverage is Friday, February 1, 2008, but the employee did not begin work until Monday, February 4, 2008. Assuming that February 1 is "a regularly scheduled work day," the policy provides that Guardian "will postpone the start of [the employee's] coverage until he or she returns to active full-time work." In other words, the coverage will "start" on February 4. According to Guardian, though, when the policy says that the coverage will "start" on February 4, it does not mean that the coverage will "start" on February 4. It means instead that the coverage will start on some future date, after the employee has worked a 32-hour week.
Similarly, suppose that the "scheduled effective date" of a new employee's coverage is Saturday, March 1, 2008. Suppose further that the employee actually begins work on Friday, February 29. Assuming that February 29 is and March 1 is not "a regularly scheduled work day," then the policy informs the employee that his or her "coverage will start" on March 1 if he or she was actively at work on February 29. This employee plainly qualifies. And yet, according to Guardian, this employee's coverage will not start on March 1.
Guardian contends that its interpretation that an employee must complete a 32-hour work week before coverage attaches is supported by the policy's requirement that an employee "must have met all of the conditions of eligibility. . . ." The Court disagrees. Without question, being an "active full-time employee" is a condition of eligibility. But the policy does not define an active full-time employee as one who has worked at least one 32-hour week. It defines an active full-time employee as one who works — present tense — at least 32 hours per week. In other words, in identifying who is a full-time employee, the policy relies on the employee's current status, and not on her past conduct.
If Guardian had wanted to require that new employees work one 32-hour week before they are eligible for coverage, it could easily have done so. Cf. Whiteside v. Metro. Life Ins. Co., 798 F. Supp. 1380, 1383 (D. Minn. 1992) (employee who worked less than one full month before his death was not covered because the benefit plan documents required "one month of continuous full-time employment" before coverage could begin). It did not. To the contrary, it informed a new employee that his or her coverage would start on "the date his or her coverage is scheduled to start," as long as the new employee is "actively at work, and working his or her regular number of hours," on that date.
Other provisions of the policy also make Guardian's interpretation implausible. Guardian essentially argues that the policy imposes a waiting period before coverage attaches — a period during which an employee must put in a 32-hour week. But the policy states that employees are eligible for coverage "after they complete the service waiting period established by the employer, if any." AR 112-13 (emphasis added). The phrase "if any" implies that, if the employer does not establish a waiting period, then no waiting period applies — i.e., that employees are eligible for coverage from the beginning of their employment unless the employer says otherwise. This conflicts with Guardian's reading, under which a waiting period is imposed on all employees by the policy itself.
Guardian's interpretation also conflicts with the requirement that ERISA plans "be written in a manner calculated to be understood by the average plan participant, and . . . be sufficiently accurate and comprehensive to reasonably apprise such participants and beneficiaries of their rights and obligations under the plan." 29 U.S.C. § 1022(a); King, 414 F.3d at 1002. Guardian's interpretation is not even understood by Guardian's lawyers, much less by the average plan participant.
Although § 1022(a) refers only to summary plan descriptions, the Eighth Circuit has cited that section in noting that plans must be written in an understandable manner. King, 414 F.3d at 1002. Moreover, whether or not § 1022(a) applies to plans, the summary plan description in the administrative record has language nearly identical to the policy language defining "When Your Coverage Starts," AR 11, and does not alert employees that they are subject to a one-week waiting period.
As noted, Guardian argues that coverage under its policy would attach at some point after an employee works a 32-hour week. But Guardian's counsel had great difficulty identifying that point. Suppose, for example, that the "scheduled effective date" of a new employee's coverage is Saturday, March 1, 2008. Suppose further that the employee begins work on Monday, March 3, and puts in a 16-hour day on March 3 and a 16-hour day on March 4, for a total of 32 hours. When does coverage attach? At the end of the day on March 4? At the start of the day on March 5? Not until March 10? What if the employee works 10-hour days on March 3, March 4, March 5, and March 6. When does coverage attach? On March 6 at the moment the employee starts his or her 33rd hour? On March 7? Or not until March 10?
Guardian's counsel struggled to answer these questions at oral argument, even though these questions would arise frequently under Guardian's interpretation, and even though these questions would be critically important to new employees. Guardian's counsel ultimately suggested that, after an employee has met the 32-hour requirement, coverage would take effect retroactively to the "scheduled effective date." Thus, according to Guardian, if the employee started on Monday, March 3, and died on Tuesday, March 4, she would not be covered. But if she worked 32 hours during the week of March 3, then she would be retroactively covered back to the "scheduled effective date" of March 1.
Needless to say, an interpretation that requires an employee to live until March 10 in order to have life-insurance coverage on March 1 is preposterous. The point of life insurance is to insure against the risk of death. An employee who works 32 hours during the week of March 3 will by definition not need life-insurance coverage for the previous March 1. Because Guardian's interpretation would, at least in some instances, provide coverage as of the "scheduled effective date" only if the risk insured against fails to materialize, that interpretation is inconsistent with the basic premise of insurance, which is to cover risk.
Guardian cites several cases that, it contends, establish that "actively at work" means actual work, rather than the status of being a worker. But these cases are irrelevant to the point Guardian wishes to make. For example, Guardian cites Todd v. Dow Chemical Co., 760 F.2d 192 (8th Cir. 1985) to argue that actual work performance is a condition precedent to coverage under Guardian's policy. But there is no dispute that actual work performance is a condition precedent to coverage under Guardian's policy — that an employee who never shows up to work never receives coverage. The only question is how much actual work performance is required. In Todd, the effective date of coverage was July 1. Id. at 193. But, because of illness, the employee did no work after May 29. Id. In upholding the insurer's denial of life-insurance benefits, the Eighth Circuit merely held that the policy's requirement that the employee be "actively at work" on July 1 meant that the employee had to work on July 1. Id. at 194. There was no suggestion that the employee had to work more than one day for coverage to attach.
Similarly, Guardian cites Fink v. Union Central Life Insurance Co., 94 F.3d 489 (8th Cir. 1996) for the proposition that an employee must actually work the requisite number of hours to be eligible for coverage. But Fink is irrelevant because it concerns the termination, not the commencement, of coverage. Id. at 491 (noting the overwhelming evidence that the insured was not an "active, full-time employee at the time of his death") (emphasis added). Guardian cannot rely on Fink because Guardian has never claimed that Ms. Granite's coverage ended. Instead, Guardian has consistently argued that Ms. Granite's coverage never began in the first place.
The remaining Finley factor is whether the fiduciary has consistently followed the interpretation. There is no evidence one way or the other on this point, but it does appear that Guardian has consistently followed the same interpretation throughout its processing of Mr. Granite's claim. This factor is not dispositive, however. An unreasonable interpretation does not become reasonable merely because the fiduciary has applied it consistently. Lickteig, 61 F.3d at 585.
Having considered all of the Finley factors, the Court concludes that Guardian's interpretation is unreasonable. It cannot be reconciled with the language of the policy; it fails to satisfy ERISA's requirement that plans be written so as to be understandable to average participants; and it creates the absurdity of life-insurance coverage that is retroactively effective only so long as the employee continues to live. In finding that Ms. Granite was not covered under its policy unless she actually worked a 32-hour week, Guardian abused its discretion.
The Court's conclusion that the policy does not require one 32-hour week of work as a condition precedent to coverage might suggest that Mr. Granite is entitled to an award of benefits. But because Guardian found that coverage never commenced, it did not address whether coverage ended before Ms. Granite died. Guardian reserved the discretion to determine eligibility for benefits and construe the terms of the plan, and thus it is appropriate to remand this case so that Guardian may consider this issue in the first instance. King, 414 F.3d at 1005-06 (when a plan reserves discretion, plan administrators must have the opportunity to apply the proper standard in the first instance).
The Court cautions, however, that Guardian's interpretation of the termination provisions must be a lot more reasonable than its interpretation of the commencement provisions. The policy states that coverage ends on the date "an employee's active full-time service ends for any reason." AR 114. This language is facially absolute, but it is difficult to believe that coverage under the Guardian policy ends anytime an employee fails to work 32 hours in one week. Otherwise, an employee who goes on vacation or takes a couple of sick days would lose coverage. Guardian's interpretation — both of when coverage terminates and when coverage resumes after termination — must be reasonable in light of the Finley factors.
ORDER
Based on the foregoing, and on all of the files, records, and proceedings herein, IT IS HEREBY ORDERED THAT:
1. Defendant's motion for summary judgment [Docket No. 22] is DENIED.
2. Plaintiff's motion for summary judgment [Docket No. 30] is GRANTED IN PART and DENIED IN PART.
3. Plaintiff's motion is GRANTED to the extent that plaintiff's claim for benefits is REMANDED to defendant for reconsideration in light of the foregoing memorandum opinion.
4. Plaintiff's motion is DENIED in all other respects.