Your patients are telling their doctors everything. Your queries don't see any of it.

Type a question in plain English. PopLens reads clinical notes alongside structured EHR data and surfaces the patients your current queries walk right past.

POPLENS
30-day readmission risk · cohort
High risk
47%
Patient 10089342
Heart Failure · Discharged 14 days ago · ZIP 21201 (ADI 87)
"...expressed concern about getting to follow-up appointments, relies on neighbor for transportation..."
High Risk
Patient 10156832
HF, DM2 · Discharged 31 days ago · ZIP 21223 (ADI 93)
"...no reliable transportation, missed last cardiology visit, unable to arrange ride..."
High Risk
47 matched patients · sorted by risk score

Care teams at health systems and ACOs pull patient lists every week. They keep missing the same people.

Queries only see what got coded

ICD codes capture the diagnosis. Lab values capture the result. Neither one touches the note where the patient said they can't afford their meds.

Notes are a black box

Housing instability. Transportation issues. Missed appointments. It all gets written down. Your queries never see it.

Risk shows up after the fact

Readmission reports tell you who came back. They don't help you get there before it happens.

Trial recruitment is a manual slog

Half of eligibility criteria live in notes - functional status, patient history, contraindications. Research coordinators review thousands of charts by hand to find 30 enrollable patients.

Who uses it

Two problems. One query interface.

Care Management

Find the patients your queries are walking past

A diabetic patient has been splitting her insulin doses because she can't afford refills. Her doctor wrote it in the note three visits ago. Her ICD codes say controlled. Your current platform says she's fine. She is not fine.

"Diabetic patients whose notes mention cost barriers to medication or missed doses"

Surfaces patients invisible to any structured query - ranked by clinical risk and social vulnerability score.

Relevant for

  • ACOs on value-based contracts
  • Health system care management teams
  • Payers managing chronic disease populations
Clinical Trial Recruitment

Screen for eligibility in minutes, not weeks

Eligibility criteria are half structured (age, ICD codes, lab values) and half narrative (no cognitive impairment, documented adherence challenges, no recent cardiovascular events). Manual chart review to find 30 enrollable patients takes a research coordinator two weeks.

"T2DM patients aged 40–70, HbA1c 7.5–10, not on insulin, whose notes don't mention cardiovascular events or cognitive impairment"

Candidate list with eligibility-relevant note excerpts surfaced. Coordinator reviews 30 names instead of 3,000 charts.

Relevant for

  • Academic medical center research sites
  • CROs managing multi-site enrollment
  • Sponsors with under-enrolling trials

How it works

01

Ask the way you think

Type the question a clinician would say out loud - "Diabetic patients with HbA1c above 9 whose notes mention cost barriers to medication." No SQL. No filters. No analyst request.

02

One search, everything included

Diagnoses, labs, medications, encounter history, clinical notes - searched together in one pass. Results come back in seconds, not a week later.

03

Ranked, with context attached

Every result includes the note excerpts that matched, a social risk score from public SDOH data, and a suggested outreach message. Open the chart only if you need to.

Validated on MIMIC-IV

Real de-identified data from Beth Israel Deaconess Medical Center

0
Admissions
0
Clinical notes
0
SDOH datasets enriched at ingestion
Seconds
Average query time

Social risk data from the CDC Social Vulnerability Index, USDA Food Access Research Atlas, and Area Deprivation Index is joined automatically at ingestion by ZIP code. Your team doesn't configure it. It's just there.

MIMIC-IV is a publicly available dataset. PopLens deploys on your EHR infrastructure.