|
| 1 | +# Finding Vitamin C (ascorbic acid) in MIMIC |
| 2 | + |
| 3 | +There is **no dedicated Vitamin C table** in MIMIC-III or MIMIC-IV. |
| 4 | + |
| 5 | +See [issue #1962](https://github.com/MIT-LCP/mimic-code/issues/1962). |
| 6 | + |
| 7 | +## Where exposure usually lives |
| 8 | + |
| 9 | +| Source | Module | What to search | |
| 10 | +|--------|--------|----------------| |
| 11 | +| `prescriptions` | hosp | drug / drug_name_generic containing `ascorbic`, `vitamin c`, `vit c` | |
| 12 | +| `emar` / `emar_detail` | hosp | administered meds with the same name patterns | |
| 13 | +| `inputevents` (III: `_cv` / `_mv`) | icu | infusions if charted as inputs (rare for vit C) | |
| 14 | +| `labevents` + `d_labitems` | hosp | serum ascorbic acid **levels** (not doses), if ordered | |
| 15 | + |
| 16 | +Retrospective papers that report “Vitamin C exposure” almost always derive a |
| 17 | +cohort from **medication orders / eMAR**, not from a first-class vit-C module. |
| 18 | + |
| 19 | +## Practical guidance |
| 20 | + |
| 21 | +1. Start with `prescriptions` / `emar` string filters; review false positives |
| 22 | + (multivitamins, “C” abbreviations). |
| 23 | +2. Decide whether you need **orders**, **administrations**, or **serum labs** — |
| 24 | + they answer different questions. |
| 25 | +3. Do not expect complete capture: OTCs and outside-hospital vitamins are often |
| 26 | + absent. |
| 27 | + |
| 28 | +## What this repository will not add |
| 29 | + |
| 30 | +A curated “official” Vitamin C concept would need ongoing pharmacy vocabulary |
| 31 | +maintenance. Contribute a versioned derived concept with explicit inclusion |
| 32 | +strings if you build one for a paper. |
0 commit comments