The Attendant Who Cannot Lift the Mattress

Ranking (highest exposure first):

1. The accommodation denial. This is the cleanest claim and the hardest to defend. A doctor's note plus a denial of accommodation with no written policy and no interactive process is a textbook failure. The employer must engage in an interactive dialogue, not simply say no. 48-hour action: meet with the attendant, document the essential functions of bed-making, and offer a concrete accommodation - swap her to a room assignment that excludes bed-making (bathrooms, dusting, vacuuming) at the same rate of pay, or provide a bed-making tool. Put the offer in writing.
2. The sexual-comment allegation. Harassment exposure grows every day it goes unaddressed, and the employer is on notice. 48-hour action: separate the parties' schedules immediately (no contact pending investigation), open a documented investigation, and interview the complainant and witnesses. Do not promise confidentiality you cannot keep; promise no retaliation.
3. The arm-grab. Serious, but a single incident with a disputed account. 48-hour action: take a written statement from the attendant, remove the supervisor from direct supervision of that attendant pending review, and interview witnesses. Physical contact to move an employee is not a coaching technique.

Why this order: the accommodation claim is already documented and unwinnable if ignored; harassment compounds daily; the arm-grab is serious but narrower.

Must contain:

• A clear statement that the hotel will engage in an interactive process on any request for accommodation, and that the employee must provide documentation from a qualified provider.
• A written list of essential functions for each housekeeping job (room attendant, houseperson, laundry) - this is the document that lets you say no to a request that removes an essential function, and you cannot do it retroactively.
• A defined light-duty assignment menu: which tasks are available (linen room, public areas, inventory, uniform room, laundry folding) and at what rate of pay.
• A time limit on temporary light duty (e.g., 90 days) with a review, and a statement of what happens when it expires.
• Confidentiality of medical information and a named contact (HR, not the supervisor).

Must not do:

• Create a 'light duty' category that pays less for the same hours without a documented reason - that invites a wage claim.
• Allow supervisors to grant or deny accommodations verbally.
• Require the employee to accept the first offer or be terminated; the process is a dialogue.
• Use light duty as a quiet exit for an injured worker; that pattern is discoverable.

Data point 1: 90-day and 6-month survival curves by hire source and by supervisor. If most departures happen in the first 30 days, the problem is selection and onboarding, not the labor market. If departures cluster at 4-6 months under one supervisor, the problem is that supervisor. Hire source matters too: referral hires usually survive longer than 'warm bodies' from a walk-in blitz, which would directly indict the previous manager's method.

Data point 2: exit-interview reasons coded into categories, cross-checked against actual schedules. 'Better pay' is the reason people give; 'I was scheduled 6 days a week and my section was 16 credits' is the reason they leave. Pull the last 20 exits, code the reasons, and compare stated reasons to the attendance and credit records for those employees. If the stated reason and the schedule data disagree, the exit interviews are worthless and the schedule is the real story.

What neither data point tells you: whether the pay rate is competitive in the local market. That is a third question, and it is the one the GM will want answered first - answer it, but do not let it crowd out the two that actually explain 78%.