---
name: conducting-mortality-reviews
language: en
description: Structures mortality case reviews with root cause analysis and system improvement recommendations. Use when conducting M&M reviews, analyzing adverse outcomes, or documenting mortality cases.
tags:
  - process
  - hospital-medicine
metadata:
  author: casemark
  practice_areas:
    - Hospital Medicine
    - Internal Medicine
  document_types:
    - Process Documentation
  skill_modes:
    - Process Management
---

# Conducting Mortality Reviews

Structures mortality case reviews with root cause analysis and system improvement recommendations for quality improvement in hospital medicine.

## Why This Skill Exists

Mortality and Morbidity (M&M) conferences are a cornerstone of medical quality improvement, peer review, and professional accountability. In-hospital mortality rates range from 1.5-3% for general medicine services, and a significant proportion of inpatient deaths — estimated at 10-15% — involve potentially preventable contributing factors related to diagnostic delays, treatment errors, communication failures, or system gaps. The Joint Commission requires hospitals to analyze sentinel events and near-misses, and CMS Conditions of Participation mandate an ongoing quality assessment and performance improvement (QAPI) program.

Mortality reviews serve dual purposes: individual case learning and system improvement. When conducted properly under the protection of peer review statutes (state-specific), mortality reviews are legally privileged and cannot be used in malpractice litigation. This protection enables candid analysis of failures. However, the privilege only applies when reviews follow institutional peer review bylaws, are conducted through an authorized committee, and documentation is appropriately managed. Hospitalists who lead or participate in mortality reviews must understand both the clinical methodology and the legal framework.

---

## Checkpoint A: Pre-Draft Intake (Mandatory)

Before conducting a mortality review, confirm:

1. Is this review conducted under the **institutional peer review committee** with appropriate legal privilege? *(Default: Verify with medical staff office or legal department)*
2. What is the **scope** of the review — single case, thematic review, or all-mortality screening? *(Default: Per institutional mortality review protocol)*
3. What **documentation** is available — complete medical record, autopsy results, family complaints, event reports? *(Default: Obtain full medical record at minimum)*
4. Was an **autopsy** performed? *(Default: Check with pathology; request if death was unexpected and consent was obtained)*
5. Who will **present** the case and who will **participate** in the review? *(Default: Presenting physician + department members + invited specialists)*
6. What **classification system** will be used to categorize the death? *(Default: Use Goldman classification or institutional equivalent)*
7. Is there a **family complaint or litigation** associated with this case? *(Default: If yes, coordinate with risk management before review)*
8. What is the **timeline** — how soon after the death is the review conducted? *(Default: Within 30-60 days)*

### Documents to Request

- Complete medical record (admission to death, all notes, orders, lab results, imaging)
- Nursing flowsheets and medication administration records
- Consultant notes and recommendations
- Code/RRT documentation (if applicable)
- Autopsy report (if performed)
- Event or incident reports related to the case
- Risk management file (if applicable — may be privileged)
- Family communication documentation
- Prior hospitalization records (if readmission or known to the institution)

---

## Step 1: Case Timeline Reconstruction

Build a detailed chronological timeline of the clinical course:

```
MORTALITY REVIEW TIMELINE

Patient: [De-identified identifier per peer review protocol]
Age/Sex: [Demographics]
Admission date: [Date]
Death date: [Date]
LOS: [Days]

Timeline:
[Date/Time] — [Event/Decision/Finding] — [Provider] — [Outcome/Response]
[Date/Time] — [Event/Decision/Finding] — [Provider] — [Outcome/Response]
...

Key inflection points identified:
1. [Time point where clinical trajectory changed]
2. [Time point where alternative action might have altered outcome]
3. [Time point where system factor contributed to delay or error]
```

Focus on decision points, transitions of care (handoffs, transfers, weekends), and timing of interventions relative to clinical changes.

---

## Step 2: Apply the Goldman Mortality Classification

| Class | Definition | Example |
|-------|------------|---------|
| **Class I** | Major diagnostic error discovered at autopsy — treatment would have changed and potentially altered outcome | Missed pulmonary embolism as cause of death; missed aortic dissection |
| **Class II** | Major diagnostic error discovered at autopsy — treatment might not have changed outcome | Missed metastatic cancer in patient with terminal multiorgan failure |
| **Class III** | Non-diagnostic issues — related to treatment, timing, or system factors | Delayed antibiotic administration; missed escalation; handoff failure |
| **Class IV** | Death was anticipated and unavoidable given the clinical scenario | Expected death from terminal illness with appropriate comfort care |
| **Class V** | Insufficient information to classify | Incomplete record, no autopsy, unclear clinical picture |

---

## Step 3: Root Cause Analysis (RCA)

For each identified contributing factor, use the "5 Whys" method and categorize using the RCA taxonomy:

**Contributing factor categories:**

| Category | Examples | Investigation Questions |
|----------|---------|----------------------|
| **Diagnostic** | Missed or delayed diagnosis, wrong diagnosis, failure to use indicated tests | Was the differential diagnosis complete? Were test results followed up? Was clinical reasoning documented? |
| **Treatment** | Wrong medication/dose, delayed treatment, procedure complication, unnecessary procedure | Was the treatment evidence-based? Was timing appropriate? Were protocols followed? |
| **Communication** | Handoff failure, consultant miscommunication, inadequate documentation, failed escalation | Was I-PASS used? Were recommendations acknowledged? Was the family informed? |
| **System** | Equipment failure, staffing inadequacy, ED boarding, ICU bed unavailability, EMR usability | Was the system capable of supporting the clinical need? Were resources available? |
| **Patient factors** | Delayed presentation, non-adherence, comorbidity burden, patient/family decision | Were patient factors modifiable? Was the care team responsive to patient limitations? |
| **Monitoring** | Failure to rescue, missed deterioration, alarm fatigue, inadequate vital sign frequency | Were early warning scores calculated? Were escalation criteria followed? Was monitoring appropriate for acuity? |

**For each contributing factor, document:**
```
Contributing Factor #[N]:
- What happened: [Factual description]
- Why it happened: [Root cause — go 5 levels deep]
- Classification: [Diagnostic / Treatment / Communication / System / Patient / Monitoring]
- Preventability: [Definitely preventable / Possibly preventable / Not preventable]
- Recommended action: [Specific, actionable, assignable improvement]
```

---

## Step 4: Classify Overall Preventability

| Classification | Definition | Action Required |
|---------------|------------|-----------------|
| **Definitely preventable** | Clear error or omission that directly caused or significantly contributed to death | Immediate corrective action, system change, peer review follow-up |
| **Possibly preventable** | Alternative actions might have changed the outcome, but uncertainty exists | System review, process improvement consideration |
| **Not preventable** | Death occurred despite appropriate care, or was expected given terminal condition | Document as learning case; confirm care was appropriate |

---

## Step 5: Generate Improvement Recommendations

For each identified contributing factor, create a specific, actionable, measurable recommendation:

**Recommendation template:**
```
Recommendation: [Specific action]
Type: [Individual learning / Education / Policy change / System improvement / 
      Equipment / Staffing]
Owner: [Department, committee, or individual responsible for implementation]
Timeline: [By when should this be implemented]
Measurement: [How will we know if the recommendation was effective]
```

**Avoid vague recommendations:**
- BAD: "Improve communication during handoffs"
- GOOD: "Implement mandatory I-PASS verbal handoff with receiver read-back for all ICU-to-floor transfers by Q3 2025. Measure: Audit 20 handoffs/month for compliance."

---

## Checkpoint B: Post-Draft Alignment (Mandatory)

After completing the mortality review:

1. Was the review conducted under **peer review privilege** with appropriate documentation protections?
2. Were all **contributing factors** identified with root cause analysis?
3. Is the **preventability classification** justified and documented?
4. Are **improvement recommendations** specific, actionable, and assigned?
5. Has the review been presented to the **appropriate committee** with documentation per institutional bylaws?

---

## Quality Audit

- [ ] Review conducted within 30-60 days of death (per institutional protocol)
- [ ] Full medical record reviewed including nursing notes, consultant notes, and MAR
- [ ] Chronological timeline reconstructed with key decision points identified
- [ ] Goldman classification (or equivalent) applied
- [ ] Root cause analysis completed for each contributing factor
- [ ] "5 Whys" methodology used to identify true root causes (not surface-level factors)
- [ ] Contributing factors categorized (diagnostic, treatment, communication, system, patient, monitoring)
- [ ] Preventability classification assigned with justification
- [ ] Improvement recommendations are specific, measurable, and assigned to an owner
- [ ] Recommendations include a timeline for implementation
- [ ] Case presentation is de-identified per peer review protocol
- [ ] Documentation is stored under peer review protection (not in general medical record)
- [ ] Follow-up mechanism exists to track recommendation implementation
- [ ] Trends across multiple mortality reviews are analyzed (not just individual cases)

---

## Guidelines

- Mortality reviews are educational and improvement-oriented — they must never be punitive or blame-focused
- All documentation produced during peer-review-protected mortality reviews must be clearly marked as privileged and stored separately from the medical record
- If a case involves potential litigation, coordinate with risk management BEFORE the review — some findings may affect privilege
- Focus on system factors, not individual blame — most errors result from system failures that set individuals up to fail
- Present cases with a complete timeline before discussing what "should have been done" — hindsight bias is the greatest threat to fair analysis
- Track recommendations longitudinally — a mortality review without implementation follow-up is an exercise in documentation, not improvement
- Aggregate trends across multiple reviews (e.g., "3 of last 10 mortality cases involved handoff communication failures") are more powerful than individual case findings
- Invite non-physician team members (nursing, pharmacy, case management) when their perspective is relevant to the contributing factors