Week one: appoint one broker in writing and ask the incumbent, through them, for the actual reason. Carriers frequently give a specific answer when asked directly, and the answer changes the submission. A class exit and a loss-driven decision call for different presentations.
Week one: request current-valued loss runs for five years and the full expiring policy including all endorsements, not the certificate and not the proposal. Endorsements are where the program actually lives.
Week two: build the narrative. For every claim above a threshold you and your broker agree on, write what happened, what changed afterward, and what evidence exists that the change held. This document is the single highest-leverage thing an operator produces in a non-renewal.
Week two: assemble the operational file. Current census and payroll, staffing by shift, agency use and the trend in it, survey history with plans of correction, clinical leadership tenure, and the risk management program in writing.
Weeks three and four: go to market with a complete file. A complete submission early beats a perfect submission late, and in this class a late submission may find the appetite already committed elsewhere.
Before expiration: resolve the tail. If the replacement carries a later retroactive date, the tail on the expiring program is not optional and the window to buy it closes with the policy.
After binding: read the bound policy against the quote. Terms move between quote and issuance more often than operators expect, and the time to find a difference is while the carrier still wants the account.