Finding patients in your clinical system who look like they have a condition that was never coded.
Mohammed Hussein — primary care data, EMIS searching and case finding.
What gets found
- Results indicating diabetes, pre-diabetes, chronic kidney disease or hypertension
- Heart failure that was never identified or coded
- Cardiovascular disease diagnosed in correspondence and never coded
- Patients on inhaled therapy with no asthma or COPD code
- Exception reporting candidates, and serious mental illness sitting in remission
What you get back
| Patient | Age | Finding | Conf. | Evidence |
|---|---|---|---|---|
| #4471 | 62 | Two HbA1c at or above 48, no diabetes code | HIGH | 48 and 52 mmol/mol, Aug and Nov 2025 |
| #0932 | 71 | Loop diuretic and raised BNP, no heart failure code | HIGH | BNP 480 ng/L, 4 repeat issues in 12 months |
| #2204 | 66 | Two eGFR below 60, over 90 days apart, no CKD code | HIGH | 54 and 51 mL/min, Mar and Jul 2025 |
| #3892 | 58 | Persistent BP ≥140/90 on three readings, no hypertension code | HIGH | Aug, Oct, Dec; average 152/94 |
| #1180 | 58 | Inhaled therapy, no asthma or COPD code | MED | 6 repeat issues in 12 months |
| #5077 | 67 | QRISK3 ≥20% at last review, no primary-prevention statin | MED | QRISK 24%, reviewed 12 Mar 2025 |
| #2915 | 74 | SMI diagnosis coded, off register (possible remission) | LOW | Depot injection last 2019, no ETH1 |
Scroll the table sideways to see the evidence. Sample rows. Synthetic data, not real patients.
How it works
Searches run on your clinical system, under your own data agreements. Processing happens on local tools, on NHS or practice equipment. You get back a work list: one row per patient, the evidence that put them there, and a rating of how likely the finding is. A clinician at the practice reviews every row before anyone is contacted.
How a rule reads
RULE heart_failure_missing_code v1 ──────────────────────────────────────────────────────────────── population Registered, age 18+ require Loop diuretic ≥3 issues in 12 months require BNP >400 pg/mL or NT-proBNP >1000 pg/mL absent Heart failure (any code in codeset:hf) mitigate Palliative care → demote to LOW, keep on list output Patient, evidence dates, issue count, confidence
Every finding traces back to a rule that reads this plainly. If a clinician disagrees with a candidate, the rule and its evidence sit on the same page.
One cohort, at no charge
Pick a cohort and I will run it as a sample, so you can judge the output against your own register before deciding anything. Uncoded diabetes. Heart failure without HFrEF coding. Or an exception reporting audit across the diabetes, hypertension and cardiovascular indicators.