hospital business.” Laurin v. Providence Hosp., 150 F.3d 52, 59-60 (1st Cir.1998).
Samper offers nothing to rebut Providence’s undisputed evidence except for highlighting that Providence’s policy allows for some unplanned absences, and that her absences had exceeded those permitted under the policy in past years without repercussions. She claims, without evidence, that “[t]he impact on staffing levels resulting from an employee’s first absence is the same as the impact from an employee’s twentieth absence,” and clings to our decision in Humphrey v. Memorial Hosps. Ass’n for support, where we noted that “regular and predictable attendance is not per se an essential function of all jobs.” 239 F.3d 1128, 1135 n. 11 (9th Cir.2001).
Our observation that regular attendance is not necessary for all jobs is hardly remarkable when on-site presence is not required for all jobs, a point not lost on our sister circuits. See Waggoner, 169 F.3d at 485 (“In some jobs ... working at home for a time might be an option.”); Jackson, 22 F.3d at 279 (“[O]ther jobs ... can be performed off site or deferred until a later day.”); Carr, 23 F.3d at 530 (“Indeed, in appropriate cases, that section requires an agency to consider work at home, as well as reassignment in another position, as potential forms of accommodation.”). Similarly, in Humphrey, a medical transcriptionist provided evidence that other transcriptionists were allowed to work at home, and therefore, his attendance was not required for performance. 239 F.3d at 1137. However, even when an employee “work[s] at home, ... regular hours on a consistent basis” often remain a requirement. Carr, 23 F.3d at 530.
Samper’s focus on Humphrey, the unusual case, blinds her to the rule. “Except in the unusual case where an employee can effectively perform all work-related duties at home, an employee “who does not come to work cannot perform any of his job functions, essential or otherwise.’ ” Yellow Freight, 253 F.3d at 948 (quoting Waggoner, at 484-85). As the evidence easily establishes, Samper’s engagement with patients is far more direct than that of a medical transcriptionist — although attendance may not be necessary to transcribe details regarding medical treatment, in the context of a neo-natal nurse, it is necessary to provide that treatment in the first place. Not only is physical attendance required in the NICU to provide critical care, the hospital needs to populate this difficult-to-staff unit with nurses who can guarantee some regularity in their attendance.
Turning to the reasonable accommodation analysis, Samper attempts to gild the lily by claiming not that attendance in general is an essential function, but, rather that her proposed variation to the attendance policy constitutes a reasonable accommodation.
Even under a “fact-specific, individualized analysis” of the accommodation, Sam-per’s argument fails. Wong v. Regents of the Univ. of Cal, 192 F.3d 807, 818 (9th Cir.1999). As Providence points out, “Samper never quantified the number of additional unplanned absences she was seeking,” even though she could have done so at any time during her years-long negotiating with the hospital over attendance. As the Seventh Circuit observed in similar circumstances in Yellow Freight and Jovanovic, such behavior suggests that “the only imaginable accommodation” that would satisfy the employee “would be an open-ended schedule that would allow [her] to come and go as [s]he pleased.” Yellow Freight, 253 F.3d at 951 (quoting Jovanovic, 201 F.3d at 899 n. 9). In these cases, the court was “hard-pressed to imagine a manufacturing facility that could operate effectively when its employees are essentially permitted to set their own work hours.” Id. To imagine a NICU facility,