The problem list with onset dates
Every condition with the date it first appears, who recorded it and the page, so the report can say how long each illness has run.
onset dateA1c, eGFR, weight and cognition across 6 years, grouped by condition, each value tied to its page.
Record review for life expectancy reports, built for the medical underwriter who writes them. Medrecords AI organizes the full medical history by condition and date: diagnoses, lab and vital sign trends, treatments, functional status and comorbidities, every value cited to its page. You spend your hours on the estimate. The estimate and the report are yours.
Updated 21 Nov 2026
| Built for | Medical underwriters and physicians who write life expectancy reports |
| What it reads | Office notes, specialist consults, hospital records, lab panels, imaging, pathology and pharmacy printouts |
| By condition | Each condition with its onset date, treatments, latest status and the pages behind each |
| Lab trends | A1c, eGFR, UACR, albumin, BNP, PSA and other results by date in 1 table |
| Vitals | Weight, BMI and blood pressure at every visit, so a slow weight loss is visible |
| Function and cognition | ADL notes, falls, mobility aids, memory testing and who helps at home |
| File gaps | Referrals and tests ordered with no result in the file |
| Speed | About 1 hour per case |
| Price | From 10 cents per deduplicated page, as low as 5 cents with volume |
| Security | SOC 2 and HIPAA, BAA at signup |
| Trial | First case is free on us |
| What it does not do | It does not produce a life expectancy, a mortality multiple, debits or a report opinion |
In a sample 342-page, 2-packet production, Medrecords AI found 11 duplicate pages. You pay for unique pages only.
eGFR 31 mL/min. Albumin 3.2 g/dL. A1c 6.9%.
How long a typical case takes after upload.
Medrecords AIThe A1c level at or above which a test result indicates diabetes.
NIDDK, A1C testAn eGFR under 60 may mean kidney disease. A UACR over 30 mg/g may be a sign of it.
NIDDK, CKD testsFound in a sample 342-page, 2-packet production before billing.
Medrecords AIMedical underwriters and reviewing physicians at life expectancy report providers who have to read every page of an insured's file before writing a word. If you spend the first half of each case building a lab table and a problem list by hand, it is built for you. If your files are short, you need it less.
No. It never produces a life expectancy, a mortality multiple or any debit. It organizes the records the estimate rests on: each condition with its dates, the lab and vital sign trends, the treatments and the function notes, each value with its page. You weigh them, write the estimate and sign the report.
A life expectancy report is only as good as the history under it, and the history arrives as a stack of unsorted pages. These are the 6 things Medrecords AI pulls out for the writer, each with its page so the report can cite it.
Every condition with the date it first appears, who recorded it and the page, so the report can say how long each illness has run.
onset dateA1c, eGFR, UACR, albumin, BNP, PSA, blood counts and any other result, by date, whichever lab ran it.
lab trendEvery weight, BMI and blood pressure reading by visit, so 21 lb lost in a year reads as a trend, not a footnote.
weight lossMedications with doses and changes, surgeries, radiation, dialysis talk and treatment the insured declined.
treatmentADL notes, falls, mobility aids, memory scores and who manages medications, from every source in the file.
functional statusMissing Records Detection lists referrals, consults and tests the notes mention with no record, so you can ask before you sign.
missing recordsThe same fictional insured, 1 tab per condition. Each row is a dated finding with its page, ready to quote in the report.
| Date | Source | Finding | Page | Flag |
|---|---|---|---|---|
| 11 Apr 2019 | Lab report | A1c 7.2%. eGFR 58 mL/min. | p. 14 | first values |
| 2 Mar 2021 | Lab report | A1c 8.1%. eGFR 49. UACR 210 mg/g. Insulin added. | p. 52 | lab trend |
| 15 Aug 2023 | Lab report | A1c 7.6%. eGFR 38. UACR 340 mg/g. | p. 118 | lab trend |
| Oct 2024 | No nephrology record | Referral 20 Jun 2024, no note in file. | none | missing record |
| 4 Feb 2025 | Lab report | eGFR 31. Albumin 3.2 g/dL. | p. 176 | latest values |
| Date | Source | Finding | Page | Flag |
|---|---|---|---|---|
| 7 May 2019 | Oncology consult | Stage IA, ER positive. Lumpectomy planned. | p. 19 | onset date |
| 28 May 2019 | Operative report | Left lumpectomy, margins clear, 0 of 2 nodes. | p. 23 | cited |
| 9 Sep 2022 | Oncology note | No evidence of disease. Anastrozole continued. | p. 96 | cited |
| Date | Source | Finding | Page | Flag |
|---|---|---|---|---|
| 11 Apr 2019 | Primary care note | Drives, lives alone, independent. | p. 13 | baseline |
| 12 Feb 2024 | Neurology note | MoCA 22 of 30. Daughter fills pill box. | p. 139 | functional status |
| 20 Jun 2024 | Primary care vitals | Weight 131 lb, down 21 lb in 12 months. | p. 147 | weight loss |
Medrecords AI does not produce a life expectancy, choose a mortality table, apply debits or credits, or sign anything. It puts what the records say in order, with pages. You review the history, check any value against its page, and write and sign the report under your own method.
Kidney function rarely falls in 1 report. It slides across 6 years and 4 labs. Medrecords AI pulls every result into 1 table by date. The NIDDK says an eGFR under 60 may mean kidney disease and a UACR over 30 mg/g may be a sign of it, and the A1c threshold for diabetes is 6.5%. How the slope bears on the estimate is your judgment.
| Date | Test | Result | Page | Flag |
|---|---|---|---|---|
| 11 Apr 2019 | eGFR | 58 mL/min | p. 14 | first value |
| 2 Mar 2021 | eGFR and UACR | 49 mL/min and 210 mg/g | p. 52 | lab trend |
| 15 Aug 2023 | eGFR and UACR | 38 mL/min and 340 mg/g | p. 118 | lab trend |
| 4 Feb 2025 | eGFR | 31 mL/min | p. 176 | lab trend |
| 2 Mar 2021 | A1c | 8.1% | p. 52 | peak |
| 4 Feb 2025 | Albumin | 3.2 g/dL | p. 176 | latest value |
An older insured often brings several conditions at once, and they interact. Medrecords AI lists each one with its onset, latest status and the pages behind both, then the function notes that sit across all of them. Blood pressure readings sit beside the NHLBI definitions for reference. Which conditions drive the estimate is yours to decide.
| Condition | Onset | Latest status | Pages | Flag |
|---|---|---|---|---|
| Type 2 diabetes | Before 2019 | A1c 6.9% on insulin, Feb 2025 | p. 14, p. 176 | comorbidity |
| Chronic kidney disease | Apr 2019 | eGFR 31, Feb 2025. Nephrology note missing | p. 14, p. 176 | comorbidity |
| Breast cancer, stage IA | May 2019 | No evidence of disease, Sep 2022 | p. 19, p. 96 | cited |
| Mild cognitive impairment | Feb 2024 | MoCA 22 of 30. Help with medications | p. 139 | functional status |
| Weight loss | 2023 to 2024 | 21 lb in 12 months | p. 147 | weight loss |
The NAIC model regulation defines life expectancy as the mean number of months the insured can be expected to live, determined considering medical records and experiential data. The NAIC model act asks for a description of who determines it and the information the projection is based on. Medrecords AI handles the medical records part, with a page for every value. The experiential data, the method and the number stay with you.
No. It never produces a life expectancy, a mortality multiple or any debit. It organizes the records the estimate rests on: each condition with its dates, the lab and vital sign trends, the treatments and the function notes, each value with its page. You weigh them, write the estimate and sign the report.
That page is for the provider's case managers, who sort a policy file before an offer. This one is for you, the medical underwriter writing the report, and centers on lab trends, conditions and function you quote in it.
Yes. Each condition gets its own list of dated findings with pages: onset, tests, treatment and latest status, so a 3-condition insured reads as 3 clean threads.
Every result in the file: A1c, eGFR, UACR, albumin, BNP, PSA, lipids, blood counts and the rest, by date in 1 table, whichever lab ran them.
Every value links to the page it came from, so you can check it and cite the source page. What goes into the report and how you word it is yours.
Yes. Memory test scores, ADL notes, falls, mobility aids and who manages medications are listed by date with pages.
Yes. Record Q&A answers questions like when the last eGFR was drawn, with the page cited, so you do not scroll the whole file for 1 value.
About 1 hour per case on average. Exact duplicate pages are set aside before billing, so you pay for unique pages only.
Yes. Files are handled under a BAA signed at signup with SOC 2 controls. They never train an AI model, and they are deleted on request. The insured's authorization under 45 CFR 164.508 is what released the records.
From 10 cents per deduplicated page, as low as 5 cents with volume. Duplicate pages are free, and the first case is free on us.
Book a short demo. After it we open trial access on one of your own cases, the first case free on us, with the history back by condition and page-cited.
Every rule, regulation, statute and published standard named on this page is linked below to the text the body that issues it puts out, read on the date shown. Where a standard is sold rather than published, the link goes to its publisher's page for it.