Add healthcare KPI analysis tutorial using synthetic inpatient admissions data

[Rifa-111]

Overview

This PR adds a new tutorial demonstrating how to compute healthcare KPIs (length of stay, readmission rate, and bed occupancy) using synthetic inpatient admissions data.

The tutorial supports users working with EHR-like tabular data alongside medical imaging workflows. Synthetic data avoids any patient-identifiable information and enables reproducible experimentation for benchmarking and educational use cases.

Notebook contents

• synthetic data generation
• age, diagnosis, and hospital site cohort filtering
• KPI computation
• KPI visualisation
• discussion and reproducibility notes

Implementation details

• New notebook: tutorials/health_kpi_analysis.ipynb
• Dependencies: pandas, matplotlib, seaborn, health-analytics-toolkit >= 0.1.0

Notes

• No PHI or real patient data is used.
• Notebook runs on CPU-only environments.
• Happy to adjust folder placement, naming conventions, or formatting to align with MONAI contributor guidelines.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a new Jupyter Notebook tutorial demonstrating healthcare KPI computation from synthetic patient data. The notebook covers data generation via a health analytics toolkit, KPI calculations (average length of stay, 30-day readmission rate, bed occupancy), visualization of results, and includes reproducibility and extension notes.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Description check	❓ Inconclusive	The description covers implementation details, notebook contents, and dependencies. However, it doesn't follow the template structure with 'Fixes #' issue reference or 'Types of changes' checkboxes as required.	Add 'Fixes #' section referencing the related issue and include the template's 'Types of changes' checkboxes to match repository standards.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change: adding a healthcare KPI analysis tutorial with synthetic inpatient data. It's concise, specific, and directly reflects the changeset.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@monai/tutorials/health_kpi_analysis/health_kpi_analysis.ipynb`:
- Around line 4-8: The first cell is currently defined as a code cell
("cell_type": "code") but contains markdown content, so change its "cell_type"
value to "markdown" for that cell (the one with "execution_count": null and
empty "outputs") and ensure the notebook cell's "source" remains the same so
headings render correctly; you can also remove "execution_count" and "outputs"
entries if present for cleanliness.

🧹 Nitpick comments (2)

monai/tutorials/health_kpi_analysis/health_kpi_analysis.ipynb (2)

47-55: Add a deterministic seed for reproducibility.
Line 81-82 generates synthetic data but the tutorial claims reproducibility. Add a seed (verify the toolkit’s preferred API).

💡 Suggested fix (adjust per toolkit docs)

 import pandas as pd
 import seaborn as sns
 import matplotlib.pyplot as plt
 import health_analytics_toolkit as hat
+import numpy as np

 sns.set(style="whitegrid")

-df = hat.generate_synthetic_patients(n=2000)
+np.random.seed(42)
+df = hat.generate_synthetic_patients(n=2000)

Also applies to: 81-82

---

147-149: Verify the 30‑day readmission window.
Line 148 computes readmission_rate without an explicit window, but the text says “30‑day.” Confirm the default or pass the window explicitly.

[coderabbitai]

⚠️ Potential issue | 🟡 Minor

Make the intro cell markdown.
Line 4-8 defines a code cell but the content is markdown, so headings render as comments.

💡 Suggested fix

-        {
-            "cell_type": "code",
-            "execution_count": null,
-            "metadata": {},
-            "outputs": [],
+        {
+            "cell_type": "markdown",
+            "metadata": {},

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	"cell_type": "code",
	"execution_count": null,
	"metadata": {},
	"outputs": [],
	"source": [
	{
	"cell_type": "markdown",
	"metadata": {},
	"source": [

🤖 Prompt for AI Agents

In `@monai/tutorials/health_kpi_analysis/health_kpi_analysis.ipynb` around lines 4
- 8, The first cell is currently defined as a code cell ("cell_type": "code")
but contains markdown content, so change its "cell_type" value to "markdown" for
that cell (the one with "execution_count": null and empty "outputs") and ensure
the notebook cell's "source" remains the same so headings render correctly; you
can also remove "execution_count" and "outputs" entries if present for
cleanliness.

[ericspod]

Hi @Rifa-111 thanks for this contribution. We keep tutorials in a separate repo here, please propose your notebook there as a PR and then we can review. Any details about placement and conventions we can discuss there and sort out any details.