mimic-iv-patient-analysis

name

mimic-iv-patient-analysis

description

Comprehensive strategy for analyzing individual patient records in MIMIC-IV EHR database and generating high-quality, diverse QA pairs. Use this skill whenever the task involves analyzing a specific patient's clinical data from MIMIC-IV (or similar EHR databases), querying across hospital and ICU tables, and submitting QA pairs that cover the patient's full clinical story — diagnoses, procedures, medications, care trajectory, and outcomes. Trigger when you see tasks like "Analyze patient <ID>", "Generate QA pairs for patient", or any patient-centric EHR exploration task.

MIMIC-IV Patient Analysis: Comprehensive QA Generation

Goal

Systematically explore a patient's complete clinical record and submit diverse, high-quality QA pairs covering all meaningful clinical domains. Target 25–40 QA pairs for complex patients (multiple admissions, ICU stays, rich medication history); 15–20 for simpler cases. More QA pairs of focused scope are better than fewer compound ones.

Database Overview

27 tables with two prefixes:

• hosp_ — hospital-level data (diagnoses, procedures, medications, admissions, labs)
• icu_ — ICU-specific data (stays, inputs/outputs, procedures, events)

Three metadata tables: table_comments, column_comments, column_documentation

Start with get_database_info to confirm available tables.

Key Column Names (Common Pitfalls)

Incorrect column names are the #1 cause of failed queries.

Table	Use This	NOT This
hosp_prescriptions	drug, starttime, doses_per_24_hrs	medication, start_date, frequency
hosp_pharmacy	medication, route, frequency	drug
hosp_emar	medication, event_txt, charttime	route, dose_val_rx
hosp_d_icd_diagnoses	long_title	description, title
hosp_d_icd_procedures	long_title	description
icu_icustays	los	length, length_of_stay
hosp_omr	subject_id, chartdate, result_name, result_value	hadm_id, charttime, result_unit
hosp_drgcodes	description (own column, no JOIN needed)	joining a separate dictionary
hosp_hcpcsevents	hcpcs_cd, short_description	joining hosp_d_hcpcs on hcpcs_cd
hosp_poe	order_type, order_subtype, ordertime	order_name
hosp_transfers	careunit, intime, outtime, eventtype	unit, transfer_type
hosp_services	transfertime, curr_service, prev_service	starttime
hosp_microbiologyevents	spec_type_desc, org_name, ab_name, interpretation	specimen_type, organism_name

Critical drug column distinction: hosp_prescriptions uses drug (orders). hosp_pharmacy and hosp_emar use medication (dispensed/administered). Using medication in hosp_prescriptions will always fail.

Critical: hosp_labevents does NOT exist. Use hosp_omr for outpatient measurements. Use icu_inputevents/icu_outputevents for ICU lab-like data.

Core JOIN Patterns

-- hosp_prescriptions: drug frequency across all admissions
SELECT drug, COUNT(*) as cnt FROM hosp_prescriptions
WHERE hadm_id IN (SELECT hadm_id FROM hosp_admissions WHERE subject_id = <sid>)
GROUP BY drug ORDER BY cnt DESC LIMIT 20

-- hosp_pharmacy: dispensed medications (uses medication, not drug; no subject_id column)
SELECT medication, route, frequency, COUNT(*) as cnt
FROM hosp_pharmacy
WHERE hadm_id IN (SELECT hadm_id FROM hosp_admissions WHERE subject_id = <sid>)
GROUP BY medication ORDER BY cnt DESC LIMIT 20

-- hosp_omr: query directly by subject_id (outpatient measurements)
SELECT chartdate, result_name, result_value
FROM hosp_omr WHERE subject_id = <sid> ORDER BY chartdate

-- ICD diagnosis with readable title
SELECT d.hadm_id, d.seq_num, d.icd_code, d.icd_version, dt.long_title
 hosp_diagnoses_icd d
 hosp_d_icd_diagnoses dt  d.icd_code  dt.icd_code  d.icd_version  dt.icd_version
 d.subject_id  sid


 ...  hosp_services s
 hosp_admissions ha  s.hadm_id  ha.hadm_id
 ha.subject_id  sid


 ic.stay_id, ic.hadm_id, ic.intime, ic.outtime, ic.los, ic.first_careunit
 icu_icustays ic
 hosp_admissions ha  ic.hadm_id  ha.hadm_id
 ha.subject_id  sid


 di.label, (ie.amount)  total, ie.amountuom
 icu_inputevents ie  icu_d_items di  ie.itemid  di.itemid
 ie.stay_id  stay_id   di.label, ie.amountuom   total 


 di.label, (oe.value)  total, oe.valueuom
 icu_outputevents oe  icu_d_items di  oe.itemid  di.itemid
 oe.stay_id  stay_id   di.label, oe.valueuom


 di.label, (pe.value)  total_minutes, pe.valueuom
 icu_procedureevents pe  icu_d_items di  pe.itemid  di.itemid
 pe.stay_id  stay_id   di.label, pe.valueuom

When a query fails with "no such column", check column_comments:

SELECT column_name, comment FROM column_comments WHERE table_name = '<table>'

Systematic Exploration Order

Phase 1 — Foundation (always first)

1. Patient demographics: hosp_patients → age, gender, date of death
2. Admissions overview: hosp_admissions → count, dates, admission types, insurance, discharge locations, in-hospital deaths. For many admissions, query total count first, then fetch in batches.
3. Diagnoses: hosp_diagnoses_icd JOIN hosp_d_icd_diagnoses → primary and comorbid conditions. Use OFFSET to paginate if results are capped.
4. Procedures: hosp_procedures_icd JOIN hosp_d_icd_procedures → surgical and clinical interventions
5. ICU stays: icu_icustays (JOIN through hosp_admissions) → LOS, care units, timing

Phase 2 — Care Context

6. Clinical services: hosp_services → service transitions per admission
7. Prescriptions (ordered): hosp_prescriptions → GROUP BY drug ORDER BY COUNT(*) DESC for most-ordered drugs. Use drug column, not medication.
8. Pharmacy (dispensed): hosp_pharmacy → GROUP BY medication ORDER BY COUNT(*) DESC for most-dispensed drugs with route/frequency detail. This complements prescriptions and is often more clinically specific.
9. DRG classifications: hosp_drgcodes → billing severity and mortality risk (description is inline, no JOIN needed)
10. Transfers: hosp_transfers → intra-hospital care unit movement sequences
11. Microbiology: hosp_microbiologyevents → organisms, antibiotic sensitivities (always include ab_name and interpretation columns for resistance patterns)

Phase 3 — Clinical Depth (when ICU stays exist, do steps 12–13; otherwise pursue as relevant)

12. ICU inputs/outputs: For each ICU stay, query icu_inputevents and icu_outputevents by stay_id → aggregate (GROUP BY di.label, SUM(amount)) to identify key medications, vasopressors, fluid totals, and urine output. For extended stays (>5 days), also check icu_ingredientevents for nutritional formula totals.
13. ICU procedures: icu_procedureevents → ventilation duration (sum of minutes), dialysis, vascular access lines and their durations
14. Outpatient measurements: hosp_omr → weight, BMI, blood pressure trends over time
15. eMAR: hosp_emar → actual medication administrations with GROUP BY medication, event_txt ORDER BY COUNT(*) DESC
16. Provider orders: hosp_poe → COUNT(*) GROUP BY order_type for order distribution
17. HCPCS events: hosp_hcpcsevents → billed services/procedures; identify observation vs. inpatient billing patterns

Phase 4 — Synthesis

18. Identify clinically interesting patterns: readmission intervals (days between discharge and next admission), disease progression, care escalation, discharge destination evolution, per-admission diagnosis complexity (diagnoses count per hadm_id)
19. Look for cross-cutting themes: recurrent infections with same/different organisms, resistance evolution, DRG severity trajectory, ICU readmissions, pivotal admissions that marked a turning point
20. Check for allergy/implant documentation: search ICD codes Z88x (drug allergies) and Z95-Z96 (implanted devices) for additional clinically meaningful facts

Aggregation tip: When a table returns truncated results, use COUNT(*) first, then GROUP BY for summary, and OFFSET to paginate. Prefer compact aggregate queries over many sequential offset queries.

QA Generation Strategy

Coverage Targets

Generate QA pairs across these domains — focus on what's clinically rich for this patient.

Domain	Example question angles
Primary diagnoses & admission drivers	What condition drove each admission? Sequence of complications?
Comorbid conditions	Which chronic diseases appear across all/most admissions?
Surgical/procedural interventions	What procedures were performed, when, and for what indication?
Medication regimen	Most prescribed drugs across all admissions? Dosing details for critical medications?
Pharmacy dispensing	Most frequently dispensed medications with route/frequency details?
Care trajectory	How did admission frequency, sources, and discharge destinations change over time?
Clinical service assignments	Which services managed the patient and when did they transition?
ICU care — overview	What ICU stays occurred, in which units, and for how long?
ICU care — vasopressors	What vasopressors were used and in what total amounts?
ICU care — sedation/analgesia	What sedatives and analgesics were administered during critical care?
ICU care — anticoagulation	What anticoagulation was used during ICU stays and in what volumes?
ICU care — fluid balance	What were total inputs and outputs during the ICU stay?
ICU care — vascular access	What lines/catheters were placed and for what durations?
Infectious complications	What organisms were cultured? Full resistance/sensitivity per organism?
DRG severity	How did DRG classifications and severity scores change over admissions?
Discharge & outcomes	Where was the patient discharged across admissions? In-hospital deaths?
Longitudinal trends	How did weight, BMI, blood pressure change over the observation period?
Transfer & care unit patterns	What was the intra-hospital care unit sequence during complex admissions?
Readmission patterns	What were the intervals between discharge and readmission?
Admission complexity	How many diagnoses per admission? Which admissions were most diagnostically complex?
Nutritional support	What enteral/parenteral nutrition was provided during prolonged ICU stays?

QA Decomposition Rules

The single most impactful way to increase QA quantity and quality is decomposing compound topics into separate focused questions. Apply these rules:

1. Multiple organisms → one question per organism's resistance pattern (when ≥2 organisms with susceptibility data exist). Additionally generate a cross-cutting "which admission had the most intensive infectious workup?" question.

2. ICU stays with rich intervention data → separate questions per medication category:

   • Vasopressors (norepinephrine, vasopressin, phenylephrine, dopamine)
   • Sedatives/analgesics (propofol, fentanyl, midazolam, dexmedetomidine)
   • Anticoagulation (heparin, argatroban)
   • Fluid balance (total inputs vs outputs)
   • Vascular access devices and line durations
   • Nutritional support (TPN, enteral feeds)

3. Long medication list → group by drug class for separate QAs when clinically distinct classes are present:

   • Anticoagulants/antiplatelets
   • Bowel/GI medications (if extensive immobility or opioid use)
   • Cardiac drugs (antiarrhythmics, rate control, vasodilators)
   • Psychiatric/neurological medications
   • Immunosuppressants (post-transplant patients)
   • Pain management regimen

4. Pivotal admissions → generate admission-specific QA for the most complex or clinically significant hospitalization (e.g., "What were the diagnoses and care unit progression during the patient's most severe admission [hadm_id]?")

5. Long observation period with functional changes → split trajectory into sub-questions: e.g., "How did discharge destinations change over time?" AND "What was the timeline from last discharge to death?" as separate QAs.

QA Quality Standards

Strong QA pairs include:

• Concrete values: Exact dates, drug names with doses/totals (e.g., "Heparin 69,193 units"), organism names with full resistance patterns, LOS in days, procedure names with laterality
• Clinical context: Not just the fact but why it matters (e.g., "discharged to rehab, indicating functional impairment")
• Completeness: Full enumeration when there are only a few items; summaries with top items when there are many
• Focused scope: Each QA addresses one specific clinical question — if an answer requires listing medications AND organisms AND procedures, split it into three separate QAs

Anti-patterns to avoid:

• Bundling multiple clinical domains into one question: "What was the patient's ICU care including stays, interventions, and medication/fluid administration?" → split into 4–6 focused questions
• Vague counts without specifics: "19 prescription orders were placed" → name the top drugs with counts
• Redundant pairs covering the same information in slightly different wording
• Schema questions ("What columns does this table have?")

Example: compound → decomposed

• Compound: "What ICU care did this patient receive, including stays, interventions, and medication/fluid administration?"
• Decomposed:
  1. "What ICU stays did this patient have and what were their durations and care units?"
  2. "What vasopressors were administered during ICU stays and in what total amounts?"
  3. "What sedation and analgesia were used during critical care?"
  4. "What was the fluid balance (inputs vs. outputs) during the longest ICU stay?"
  5. "What vascular access lines were placed and maintained?"

High-Value QA Types

These patterns tend to produce rich, specific QA pairs:

1. ICU medication by category: Separate questions for vasopressors, sedatives, anticoagulation, and nutrition — each requires icu_inputevents aggregation with total amounts
2. Antibiotic resistance per organism: One question per clinically significant organism with full sensitivity/resistance pattern from hosp_microbiologyevents
3. Longitudinal trajectory: "How did discharge destinations change over N years?" or "What was the pattern of care escalation?"
4. Readmission intervals: "What were the shortest intervals between discharge and subsequent readmission, and what were the associated conditions?"
5. Specific procedural detail: "What specific approach was used for [procedure] and what was the clinical indication?"
6. Drug frequency across all admissions: "What were the top 10 most prescribed/dispensed medications?" (requires both hosp_prescriptions + hosp_pharmacy)
7. DRG severity evolution: "How did DRG severity and mortality scores change over successive admissions?"
8. Care unit progression during complex admission: "What care units did the patient transit through during their longest hospitalization and in what order?"
9. Fluid balance during ICU: "What were total inputs and outputs during [ICU stay]?" (requires icu_inputevents + icu_outputevents)
10. Advance care planning: "When was DNR status first documented and how consistently was it recorded?"
11. Admission pattern analysis: "What was the distribution of admission types, sources, and frequency over the observation period?"
12. Per-admission complexity: "Which admissions had the most diagnoses and what conditions drove their complexity?"
13. Blood product transfusions: "What blood products were administered and during which admissions?" (search icu_inputevents for PRBCs, FFP, platelets)
14. Vascular access devices: "What invasive lines were placed during ICU stays and for what durations?" (from icu_procedureevents)
15. Observation vs. inpatient billing: "Which admissions were billed as observation stays based on HCPCS records?"
16. Drug class-specific analysis: "What anticoagulant medications were prescribed and what was the clinical context?" (when multiple anticoagulants are used)

Submission Pattern

Submit QA pairs in thematic batches after completing each exploration phase — don't submit one at a time. Interleave: explore a domain → verify data quality → submit 3–6 related QA pairs → continue. This ensures progress is saved and helps maintain thematic coherence.

After completing Phase 3, review your collected QA pairs and ask: "Are there compound QAs I can split into two focused ones? Are there domains I explored but didn't generate QAs for?" Then generate any missing decomposed pairs before final synthesis.

Handling Query Failures

When a query fails:

1. Read the error — it often lists the available columns for that table
2. Correct the column name using the table above and retry once
3. If still failing, check column_comments for the correct schema
4. If a table doesn't exist, use the alternatives listed above

Do not spend more than 2 retries on any single query — move on if data isn't available.