Dispute Resolution — Mediation Before Litigation

Two numbers frame every partnership dispute, and most partners have never seen them side by side. Drafting a thorough agreement while everyone still gets along runs $15,000 to $40,000 in legal fees. A single contested exit without a valuation formula routinely burns $150,000 to $500,000 in combined legal and expert costs. The same disagreement, in other words, costs roughly 10 times as much to settle in crisis as to prevent in calm weather—and the version that reaches judicial dissolution destroys going-concern value outright, because the panel, the referral relationships, and the trained staff all evaporate while the lawyers argue about who owns them.

Practices that skip the drafting do not avoid the cost. They pay it anyway, at the worst possible moment, in the worst possible forum. So the last operating system every agreement needs is the one that processes its own failures, and it is built as a ladder with the cheapest rung first.

Rung one is structured discussion, and it resolves more disputes than the two rungs above it combined. The clause requires the complaining partner to state the dispute in writing, requires a partners’ meeting to be noticed and held with minutes taken, and imposes a defined cooling period before anything escalates. The reason this works is unglamorous: a large share of partnership disputes are information failures. Two partners disagree about whether the practice can afford a hire, and the disagreement dissolves the moment both are reading the same monthly financial package. Writing the dispute down is itself a filter, because a grievance that will not survive a paragraph rarely survives a mediation.

Rung two is mediation, and the agreement makes it mandatory. A non-binding session with a neutral must occur before any binding process may begin. Costs are split. Settlement discussions are privileged, which is what lets partners say the thing they actually want without it reappearing in a deposition. The mediator is not a judge and issues no ruling; the work is facilitated negotiation, and its real function is to separate the dispute from everything that has accreted on top of it. In the rural two-physician practice whose deadlock ran 14 months, the mediator spent two sessions doing exactly that—dividing the actual question, sell to the hospital or build toward an internal succession, from more than a year of accumulated operational grievance. Once separated, the fight was small enough to negotiate.

Pre-wire the selection. That same practice spent six weeks choosing a mediator, because the agreement named none, and choosing a neutral in the middle of a dispute is itself a dispute. Name the provider organization or the selection mechanism in the document, the way a well-drafted valuation clause names the appraiser process rather than the appraiser.

Rung three is binding arbitration, and it is the rung that requires the most deliberate design. Arbitration is faster and cheaper than litigation and it is private, which matters to a practice whose patients read the local paper. It also waives the right to a jury trial and narrows appeal rights close to nothing, which is a real trade rather than a technicality. A group that wants those advantages should size the process to a practice rather than to a merger: a single arbitrator, limited discovery, a one-day hearing target, and a 90-day timeline from demand to award. Name the governing rules—the American Arbitration Association’s, JAMS’s, or a set the parties specify—and name the governing law and the seat. Prevailing-party fee provisions discourage recreational grievance, and they cut both ways, which is the point.

The carve-outs matter as much as the process. Injunctive relief for covenant breaches stays in court, because a practice cannot wait a quarter for an award while a departed partner works through the panel. The deadlock machinery of article 17 stays out of arbitration too, since a deadlock is a transaction rather than a dispute; the shotgun clause and the appraisal buyout are self-executing by design and belong nowhere near a hearing. And a clause drafted too broadly can sweep in matters that were never partner disputes at all. A payer audit, a licensure board proceeding, a government investigation, and a malpractice claim are not disagreements between partners, and an arbitration provision that appears to cover them creates confusion at the exact moment clarity is worth the most. Draft the perimeter so the agreement governs disputes among the partners and leaves everything else where the law puts it.

One definitional point saves an enormous amount of trouble. Distinguish a dispute from a governance decision. A partner who is outvoted at a properly noticed meeting on a matter the agreement assigns to a majority has lost a vote, not suffered a dispute, and the ladder is not an appellate court for votes that went the wrong way. Say so in the document. Otherwise the dispute clause becomes a veto with extra steps, and the tier framework the practice built stops functioning.

One honest coda belongs after the ladder, because the ladder presumes the partnership is worth repairing. Sometimes it is not. One unworkable partner makes the whole group unworkable, and a dispute process that keeps resolving the same partner’s conflicts is measuring the problem rather than solving it. The answer there is the buy-out machinery, not a better mediator. The halfway house deserves particular caution: converting the exiting partner to an employed salary-plus-production arrangement preserves the panel and the production, but it only works if he will not stay bitter. A resentful former partner still in the building, now with no fiduciary duty and nothing left to lose, is a larger problem than the one being solved. Take the temperature before choosing that design, and where the honest answer is resentment, pay for the clean exit.

Red flags. A dispute-resolution clause that is generic boilerplate with no specifics—no mediator, no rules, no timeline, no carve-outs. No mediation requirement, so the first formal step is a complaint. An arbitration clause drafted so broadly it appears to cover regulatory and licensure matters. No pre-selected mediator or arbitration organization, leaving a six-week selection fight inside every dispute. And governing law and jurisdiction left undefined, which converts a $20,000 disagreement into a $60,000 argument about where to have it.

The ladder is another provision best written while nobody knows whether they will someday be the complainant or the accused. Mediation looks obviously reasonable in a partner retreat and looks like surrender in a crisis, which is precisely why it gets chosen at the retreat. The cheapest lawsuit is the one that never gets filed. Mediation first.