What the work involves

You will read AI-generated clinical review outputs and judge them the way you would judge a first-level reviewer's work. That means checking whether an admission recommendation actually maps to InterQual or MCG criteria, whether a continued-stay rationale is supported by the documentation provided, whether observation versus inpatient status holds up under the Two-Midnight Rule, and whether a discharge plan reflects realistic post-acute options. Tasks typically arrive as a case packet plus a model output; you rate it, flag the specific clinical or regulatory error, and write a short structured rationale a non-clinician trainer can follow. Some projects ask you to write reference determinations from scratch, or to compare two model responses and defend your preference.

What the platform screens for

  • Verifiable credentials. Active unrestricted RN license, or MD/DO for physician-advisor tracks. CPUR, ACM, or CCM helps.
  • Criteria fluency under follow-up. Expect to be pressed on how you apply a specific InterQual subset or MCG guideline, not just whether you have used them.
  • Regulatory precision. Two-Midnight Rule, CMS Conditions of Participation, condition code 44, payer-specific UM timelines.
  • Written clarity. Your feedback is training data. Vague criticism ("this doesn't seem right") has no value; specific criticism ("cites the wrong severity of illness criterion; documentation supports observation only") does.
  • Calibration. Reviewers who mark everything wrong, or everything acceptable, get filtered out quickly.

Logistics

Fully remote and asynchronous. Most contributors take 10–20 hours per week, with some projects offering more during ramp periods. Work is hourly contract through Mercor, invoiced against logged and reviewed task time. Onboarding usually includes a calibration set scored against expert consensus before paid work begins; rate placement within the $100–150 band tends to track licensure type, certification, and calibration performance. Nothing here is guaranteed — project volume shifts with the lab's research priorities.