Key takeaways
- Always ask for intent-to-treat outcomes. Completer-based numbers are not comparable across vendors.
- Ask who employs the prescriber and how they are compensated. The answer predicts prescribing behavior.
- Request a redacted sample of the actual reporting package before signing, not a screenshot.
On outcomes
Are your reported results intent-to-treat or completer-based? This is the highest-yield question in the entire evaluation. The gap between the two is routinely large enough to reverse a ranking. If a vendor cannot produce intent-to-treat numbers, treat their outcomes as unusable for comparison.
What is your 12-month and 24-month persistence rate? Persistence is the variable that determines whether spend converts into outcomes. Vendors that lead with weight loss and avoid persistence are telling you which number is weaker.
Has any of this been externally reviewed? There is a real difference between a peer-reviewed trial, a peer-reviewed observational study, a vendor white paper, and a marketing figure. Ask which one you are being shown, and get the link.
What happened to members who discontinued? Weight regain trajectory after discontinuation is the least-marketed and most decision-relevant data a vendor holds.
On the clinical model
Who employs the prescriber, and how are they compensated? If prescriber compensation is tied to volume or to program enrollment, the clinical model has a structural bias toward initiation. This single question explains a large share of the variance in prescribing behavior across vendors.
What are your written escalation criteria from lower-cost agents? 'We practice step therapy' is a posture. Written, time-bound criteria are a protocol.
What is your titration schedule and how is it monitored? Look for scheduled clinical contact during titration specifically, since that is where discontinuation concentrates.
How do you screen for eating disorders and mood effects before initiation? Very few vendors have a real answer. The ones that do have thought harder about the whole problem.
Describe your deprescribing protocol. Ask for the actual clinical document, not a description of one.
On money
What exactly is at risk under your performance guarantee? Most guarantees put a fraction of the vendor's own fee at risk while excluding drug spend — which is 80%+ of your cost. A guarantee that cannot touch your dominant cost line is a marketing instrument.
How is the baseline for any savings guarantee constructed? Negotiate the baseline methodology before the guarantee percentage. Baseline construction moves the outcome far more than the headline number.
What is your fee if a member enrolls and never engages? Engagement floors and enrolled-member billing can produce substantial spend on members who did nothing.
Show me the all-in cost for a member who starts, titrates, and discontinues at month five. This is your modal expensive member. Vendors rarely model it, and it is the scenario that most often breaks the business case.
On data and exit
Can we see a redacted copy of the actual quarterly reporting package? Not a screenshot, not a sample dashboard — the real artifact a comparable client receives.
Will you provide member-level engagement data on a claims-compatible key? Without this you cannot independently verify a single thing the vendor tells you.
What happens to member clinical records at termination? Continuity of care at exit is both an ethical obligation and a practical transition risk.
What is the notice period, and what are the early-termination economics? A 36-month term with 180-day notice is a three-and-a-half-year decision. Price it that way.
Put this to work
The RFP toolkit turns these guides into a scoring rubric and question bank you can send to vendors this week.