Building a Bias Record

1. Plant the standard

The standard you fail to assert is the one the court picks for you.

• Lead with the inference-of-bias standard: a fair inference of financial conflict makes bias cognizable — no proof of a corrupt mind required.
• Cite Hangarter as an application at the extreme end, not as the yardstick, so the defense cannot win by distinguishing its facts.
• Name “dishonest selection” as one route to liability, not the measure — and insist the court rule under the standard you named.

2. Build the money factor — from the expert, not the insurer

The financial relationship is concrete, countable, and documented. The documents are in the expert’s own files.

• Subpoena the expert and the review vendor directly: 1099s, billing records, invoices, client lists.
• Establish the two metrics over a multi-year window — compensation magnitude and frequency of retention.
• Establish the insurer-side share of the practice — the percentage that signals dependence, not ordinary volume.
• Do not rely on a hopeful interrogatory to the insurer; a number you can authenticate from a third party cannot be waved away.

3. Build the pattern — in your own case

Pattern evidence is unanswerable when it is your record, and dismissible when it is borrowed from another lawsuit.

• Propound discovery into other insureds’ claim files involving this expert, in the instant action.
• Develop the expert’s outcome record — how often the expert’s opinion favored the insurer — as this case’s admitted facts.

4. Build the methodology record — how the opinion was actually made

Money shows the motive; the methodology shows the mechanism. Relational data alone rarely carries the day.

• Document the departures in this expert’s own work product: a paper-only review that overrides physicians who examined the claimant; selective engagement that answers the evidence the expert can rebut and ignores the rest; reliance on stale or superseded assessments.
• Surface the outcome-driven trail where it exists — the iterative questioning that narrows the expert’s task until the wanted answer appears (Hinds), and the retention letter or referral instructions that prime the conclusion before the review begins (Hangarter). These communications often sit outside the claim file; request them by name.
• Hold the expert to the evaluative baseline: a reviewer who contradicts a treating physician must explain why, and a demand for “objective” proof the policy never required is itself an irregularity.
• Heed the Bagramyan lesson directly: relational data with no methodology evidence is the weaker case. The financial relationship supplies the motive; the methodology supplies the mechanism that ties it to the result.

5. Build the no-measures record — the system the insurer never built

An inference of bias the insurer never tried to prevent is an inference that stands.

• Make the insurer’s expert-selection and oversight system its own discovery target: did it track reviewer outcomes, rotate assignments so no expert grew dependent, and wall off the adjuster who wanted the denial from the choice of who supplied it?
• Anchor the asymmetry from Demer: the absence of safeguards does not by itself prove bias — it leaves the inference you have already raised unrebutted. The insurer “could have maintained records of its reviewers’ findings … but it did not,” and so “missed an opportunity to negate” the conflict.
• Measure that silence against the benchmark that already exists by rule: in the disability market, ERISA requires claims be “adjudicated in a manner designed to ensure the independence and impartiality” of the decision-makers, and forbids selecting or paying an expert based on “the likelihood that the individual will support the denial.”
• This is the factor that still reaches automated claims handling: when an algorithm selects the expert — or is the expert — there is no methodology to cross-examine and no witness to depose. Only the design remains. Demand it: selection criteria, outcome logs, and the data behind automated picks.

6. Invoke the presumption — and weaponize the evasion

The burden-shift solves the claimant’s structural problem: the proof of bias sits in the insurer’s files.

• Make the threshold showing, then argue the rebuttable presumption: the burden shifts to the insurer to prove the expert was actually neutral.
• Treat every “overbroad / unduly burdensome” objection to a volume request as grounds for a motion to compel.
• Treat every “we don’t keep track” admission as grounds for an adverse inference — a sophisticated insurer that could easily track its expert’s outcomes should not profit from the gap.
• Close the loop: an insurer that cannot identify one pro-insured result cannot carry the burden the presumption places on it.

7. Protect the record

Evidence the court can ignore is evidence you did not really put in.

• Get rulings on objections to your key declarations; do not leave the centerpiece in limbo.
• Preserve each factor argument in the opening brief — arguments raised late are forfeited.
• Build the timeline so the bias record is complete before the summary-judgment cutoff, not after.