Clean Team Protocol

This Protocol governs the handling of competitively sensitive information between ACME Health and Cumberland Valley Health Plan during the period between signing of the definitive agreement (February 13, 2023) and closing (September 29, 2023). During that period the two organizations remain, in law, independent competitors in overlapping Tennessee markets. This document establishes the information barrier, the Clean Team that operates behind it, the categories of data that may not cross it, and the form in which analysis may be released to the integration program. Issued by the Office of the General Counsel, ACME Health; countersigned by counsel for Cumberland Valley. Effective March 6, 2023 and automatically dissolved at closing. Annex A was appended at program close on April 30, 2025 and is dated separately; the Protocol itself is reproduced as issued.

1. Purpose, Authority and Effective Period

A signed merger agreement does not make two companies one. Between signing and closing, ACME Health and Cumberland Valley Health Plan continue to operate as separate legal entities, continue to compete for the same employer groups and the same provider contracts, and continue to be treated by antitrust regulators as competitors. Everything the integration program needs to know about Cumberland Valley — its rates, its membership, its provider economics — is precisely the information two competitors are not permitted to exchange.

This Protocol resolves that conflict. It does not resolve it by waiting: a program that learns nothing until closing has no plan on Day 1. It resolves it by routing sensitive information to a small, contractually isolated group that analyzes it and returns conclusions rather than data.

2. The Legal Basis — Why Two Merging Companies Stay Apart

Two distinct legal regimes operate in the pre-close window, and they are frequently confused.

2.1 The waiting period

The Hart-Scott-Rodino Act requires notification of qualifying transactions to the federal antitrust agencies and imposes a waiting period before the parties may close. For this transaction the waiting period expired on May 30, 2023 without a second request. Separately, because both parties are regulated insurers, the Tennessee Department of Commerce and Insurance had to approve a Form A change-of-control filing, which it did on September 12, 2023.

2.2 The conduct rule

Independent of any filing, Section 1 of the Sherman Act prohibits agreements between competitors that restrain trade. Two health plans exchanging current premium rates, negotiated provider reimbursement, or forward bid strategy is a per se problem regardless of whether they have a signed merger agreement. The merger agreement is not a defense; if the deal fails to close, the parties have simply exchanged competitive intelligence.

2.3 Gun-jumping

Gun-jumping is the acquirer exercising operational control over the target before it owns it, or the parties behaving as a combined entity before they are one. It is a conduct offense, not a filing offense, and it is what this Protocol is chiefly designed to prevent. Section 8 lists the specific prohibitions that applied to this program.

3. Competitively Sensitive Information: Definitions

Not all target information is restricted. Over-restricting is its own failure — it starves the plan and pushes discovery past close, where it costs far more. The categories below were agreed by both counsel and are the operative list.

4. Clean Team Composition and Independence

The Clean Team is provided by Harpeth Clean Team Services under a fixed-price statement of work. No member is an employee of either party.

4.1 Why a third party rather than ACME employees

A Clean Team may be staffed with employees of the acquirer who are walled off from commercial decision-making. That model works where the acquirer is large enough to have people with no line responsibility in the overlapping market. It was rejected here for three reasons:

• Overlap. ACME and Cumberland Valley compete directly in Middle Tennessee commercial group business. Nearly every ACME employee with the domain knowledge to analyze Cumberland Valley's book also has a role touching that market.
• Contamination is permanent. An employee who reads a competitor's rate table cannot unread it. If the deal had failed to close, ACME would have had rate-informed people back in its own pricing function — an unmanageable position.
• Evidence. A third-party team under a fixed-price SOW produces a clean, auditable record of who saw what and when. That record is what you produce if the barrier is ever challenged.

5. The Wall: Technical and Physical Controls

6. Permitted Outputs — Conclusions, Not Data

The Clean Team's product is a numbered Release Memorandum. What may appear in one is defined by form, not by topic.

6.1 The general rule

A release may state a conclusion, a range, a rank order, a variance, or a directional finding. It may not state, or permit arithmetic recovery of, an individual rate, an individual account's terms, or a named provider's reimbursement.

6.2 Worked examples

6.3 Release register

Fourteen Release Memoranda were issued between March and September 2023. Selected entries:

7. Request, Review and Release Workflow

1. Raise. A workstream lead submits a Clean Team Information Request to the IMO stating the decision the information will support. Requests that cannot name a decision are returned — curiosity is not a basis for release.
2. Triage. The IMO checks the request against the Section 3 categories. Open-category items are routed directly to Cumberland Valley and never reach the Clean Team.
3. Counsel review. Both General Counsel review restricted-category requests. Either may refuse.
4. Analysis. The Clean Team performs the work inside the walled environment.
5. Form review. H. Tillinghast reviews the draft release for reconstruction risk — whether the stated conclusion, combined with what the recipient already knows, would permit recovery of a restricted figure.
6. Release. Issued as a numbered memorandum to a named distribution list. Onward circulation is prohibited.
7. Register. Logged in the release register, which is retained as the evidentiary record.

8. Gun-Jumping: Prohibited Pre-Close Conduct

The following were prohibited to both parties until closing. Each was a live question raised by a workstream during the pre-close period.

9. Dissolution at Close and Handover

At closing on September 29, 2023 the antitrust basis for the barrier disappeared: the parties became one entity and could no longer conspire with themselves. The Protocol terminated automatically. The wind-down sequence:

• Access revoked. Clean Team credentials to the segregated data room were disabled at close of business on the closing date.
• Holdings transferred. The catalog of restricted holdings (CT-014) and the full working file set transferred to ACME's General Counsel, not to the IMO.
• Personal undertakings survive. Harpeth personnel remain bound for five years. Dissolution of the barrier does not release them.
• Register retained. The complete release register and access logs are retained as the evidentiary record of compliance.
• No Clean Team member joined the integration program. This is asserted in the program roster and was a condition of the Harpeth engagement.

10. Annex A — Program Consequence: What the Barrier Cost Us

The barrier was correctly designed and correctly operated. It also imposed a cost, and that cost shaped the entire program. This annex records both.

10.1 The deal-model estimate was built blind, by law

The integration estimate carried in the deal model — a base of $42.0M with 15% contingency — was prepared without any ability to inspect member-level data or to test Cumberland Valley's records against ACME's. In estimating terms it was an AACE Class 5 estimate: concept screening, expected accuracy roughly −20% to +50% at best. The high contingency was not conservatism; it was the arithmetically correct response to a legally imposed information deficit.

10.2 What could not be known, and when it surfaced

Two change requests trace directly to constraints recorded in this Protocol. CR-003, which proposed staffing data stewards offshore to relieve the clerical review queue, was rejected on the Form A processing conditions described in Section 9 — the standing note issued on October 2, 2023 was the basis. CR-004, the three-month TSA extension, followed from the identity-resolution scale that Section 6.2 records as unknowable before closing.

10.3 What we would do differently

• Commission a Clean Team data-quality assessment earlier and scope it wider. CT-006 flagged internal duplicates in May 2023 and the program logged it as a watch item rather than a planning assumption. The Clean Team could not measure cross-entity overlap — but it could have profiled the target's own file far more deeply than it was asked to, and a stronger signal in May would have shaped the re-baseline in January instead of the crisis in April.
• Size contingency against the barrier, not against the program. The 15% figure was benchmarked against integration programs generally. A better method sizes contingency against the specific list of things the barrier prevents you from knowing — which is a list you can write down on day one, and which this program did not write down until after the fact.
• Write the "dissolution is not release" note before closing, not three days after. Several workstreams assumed closing opened everything. It did not, and the assumption survived long enough to generate a change request that had to be rejected.

Related artifacts: 1 — Integration Charter · 6 — Regulatory Approval Tracker · 7 — Due Diligence Findings · 31 — Data Profiling Report · 44 — Change Control Log