---
title: "Retraining the Algorithm: How Expert Eyes Correct EMR Errors in Maternal Care"
description: This study highlights the importance of human intervention in correcting EMR errors to improve machine learning predictions in maternal care, enhancing patient safety.
image: https://blog.jmir.org/hubfs/StoryTap%20Blog%20-%20Jennifer%20Glance.png
---

[![JMIR Publications logo](https://blog.jmir.org/hs-fs/hubfs/JMIR-Publications_logo_2020_original.png?width=550&height=109&name=JMIR-Publications_logo_2020_original.png "JMIR Publications logo")](http://www.jmirpublications.com)

Open main menu Close main menu

[Research News](https://blog.jmir.org/tag/research-news) [Artificial Intelligence](https://blog.jmir.org/tag/artificial-intelligence)

# Retraining the Algorithm: How Expert Eyes Correct EMR Errors in Maternal Care

![Liana Ramos, Marketing Associate](https://blog.jmir.org/hs-fs/hubfs/Liana%20Ramos.jpeg?width=45&name=Liana%20Ramos.jpeg) 

[Liana Ramos, Marketing Associate](https://blog.jmir.org/author/liana-ramos)

Published Date: 22 July 2026 11:41 a.m.

[Share this blog post on Twitter](https://twitter.com/intent/tweet?text=I+found+this+interesting+blog+post&url=https://blog.jmir.org/retraining-the-algorithm-how-expert-eyes-correct-emr-errors-in-maternal-care) [Share this blog post on Facebook](http://www.facebook.com/share.php?u=https://blog.jmir.org/retraining-the-algorithm-how-expert-eyes-correct-emr-errors-in-maternal-care) [Share this blog post on LinkedIn](http://www.linkedin.com/shareArticle?mini=true&url=https://blog.jmir.org/retraining-the-algorithm-how-expert-eyes-correct-emr-errors-in-maternal-care)

![Retraining the Algorithm: How Expert Eyes Correct EMR Errors in Maternal Care](https://blog.jmir.org/hubfs/StoryTap%20Blog%20-%20Jennifer%20Glance.png)

Postpartum hemorrhage (PPH) remains a leading cause of maternal morbidity worldwide, with blood transfusion rates hovering around 2% to 2.7%. While discussing these risks is a vital component of antenatal informed consent and shared decision-making, translating historical medical records into an accurate personal risk profile is notoriously difficult.

| --- **Key Takeaways** |
| --- |
| The Danger of Flawless Software with Flawed Data: Machine learning models trained on raw EMR data can appear statistically superior while producing completely inaccurate clinical predictions, as seen when the algorithm misidentified vaginal deliveries as unspecified cesarean sections due to data labeling errors. |
| **Essential Human Intervention:** Manual physician chart review is mandatory for cleaning clinical data, as it allows the model to shift its feature importance rankings to reflect actual clinical reality rather than structural archiving flaws. |
| Improved Clinical Utility: Once retrained on physician-labeled data, the model achieved high accuracy metrics (PR-ROC of 0.78) and was effectively integrated into a clinical decision support application that assists obstetricians in prenatal risk assessment. --- |

Listen to this blog post

4:07

---

In a recent [retrospective cohort study](https://formative.jmir.org/2026/1/e82424) published in [*JMIR Formative Research*](https://formative.jmir.org/), researcher **Jennifer Glance and colleagues** from [**Kettering Health**](https://ketteringhealth.org/) highlight a critical bottleneck in maternal health informatics. They demonstrate that while machine learning models can accurately predict which patients will require a peripartum blood transfusion, the software is only as good as the human eyes reviewing the underlying data.

### When Flawless Software Meets Flawed Medical Records

To explore how delivery modes and maternal indications affect blood transfusion rates, the research team analyzed a 10-year retrospective cohort of 62,521 births. They trained an extreme gradient boosting (**XGBoost**) machine learning model on a sample of 1,734 patient records using two separate approaches: one using raw, structured data pulled directly from the Electronic Medical Record (EMR), and another using that same dataset after it was manually cleaned and reviewed by a physician.

On paper, the model trained on raw EMR data actually looked superior, logging a slightly higher precision-recall curve. However, testing the model in a live clinical interface revealed a glaring problem: its clinical predictions and feature importance rankings were completely inaccurate.

The issue stems from a common electronic archiving flaw. In raw EMR databases, vaginal deliveries are routinely labeled as an unspecified indication for a cesarean section. Because the machine learning algorithm lacks clinical context, it misidentified these unlabeled files, mixing up delivery modes and completely miscalculating which patient variables actually drove the hemorrhage risk.

### The Power of Human Context in Retraining Clinical AI

To resolve this data pollution, a physician performed a manual chart review, tracking down 120 unlabeled entries and explicitly correcting misclassified elective delivery records based on real clinical notes.

Once the model was retrained on this clean, physician-labeled dataset, the structural math aligned perfectly with live clinical logic:

- **Ecosystem Integrity:** The precision-recall receiver operating characteristic curve (PR-ROC) held strong at **0.78**, paired with a balanced F1 safety score of **0.809**.

- **Clinical Accuracy:** Feature importance rankings shifted to reflect clinical reality, correctly placing the active delivery mode (cesarean vs. vaginal) as the primary risk vector, followed directly by specific medical or surgical complications.

Using this validated model, the team built a prototype clinical decision support application. The interface translates complex algorithmic probabilities into clear risk tiers (low, moderate, or high) based on the Youden J statistic. This gives obstetricians an intuitive, workflow-integrated tool to help guide shared decision-making conversations during prenatal checkups.

|   | In this video, Dr. Jennifer Glance from Kettering Medical Center presents a retrospective cohort study evaluating machine learning models to predict postpartum hemorrhage (PPH) requiring a blood transfusion. |   |
| --- | --- | --- |

 

Why JMIR?

The authors selected ***JMIR Formative Research*** to present this project because of the journal's focus on the formative evaluation of digital health interventions and medical predictive analytics. As hospital networks rush to implement automated decision support tools, this study serves as an essential warning: raw electronic medical records are riddled with structural inconsistencies, and collaborative physician data-labeling is mandatory to ensure patient-facing AI tools are safe, accurate, and useful.

**Curious to see how physician-reviewed data models are refining maternal risk analytics?** [Watch the video featuring Jennifer Glance](https://www.youtube.com/watch?v=qH6ddJ3bMmU) and [read the full retrospective cohort study](https://formative.jmir.org/2026/1/e82424) to explore the underlying XGBoost frameworks, data-cleaning methodologies, and the digital app architecture designed for point-of-care informed consent.

---

Glance J, Nielson J, Bonnema A  
Evaluating Postpartum Hemorrhage Transfusion Risk With a Machine Learning Model for Informed Consent: Retrospective Cohort Study  
JMIR Form Res 2026;10:e82424  
URL: [https://formative.jmir.org/2026/1/e82424](https://formative.jmir.org/2026/1/e82424)  
DOI: [10.2196/82424](https://formative.jmir.org/2026/1/e82424)

![Liana Ramos, Marketing Associate](https://blog.jmir.org/hs-fs/hubfs/Liana%20Ramos.jpeg?width=150&name=Liana%20Ramos.jpeg)

#### Liana Ramos, Marketing Associate

[Follow me on LinkedIn](https://www.linkedin.com/in/lianaramos/)

Liana Ramos is a Marketing Associate at JMIR Publications, dedicated to building data-driven systems that connect scholarly content with the communities that need it most. With five years of experience at the intersection of media, publishing, and audience behaviour, she handles automated lifecycle journeys, author recognition programs, and record-scale digital events for JMIR. Passionate about storytelling and digital strategy, she focuses on making open access digital health research discoverable, scalable, and highly engaging.

### Leave a Comment

## Related Articles

[![Elastic Tape for COPD: A Rehab Game-Changer?](https://blog.jmir.org/hubfs/Thiago%20Fernandes%20Pinto%20Blog%20Post.png)](https://blog.jmir.org/can-elastic-tape-supercharge-pulmonary-rehabilitation)

[Research News](https://blog.jmir.org/tag/research-news)

### [Can Elastic Tape Supercharge Pulmonary Rehabilitation?](https://blog.jmir.org/can-elastic-tape-supercharge-pulmonary-rehabilitation)

The struggle for a full breath is a daily reality for those living with Chronic Obstructive Pulmonary Disease (COPD). While Pulmonary Rehabilitation (PR) is the gold...

![Reviewed by Kayleigh-Ann Clegg, PhD](https://blog.jmir.org/hs-fs/hubfs/Headshot.jpg?width=45&name=Headshot.jpg) 

[Reviewed by Kayleigh-Ann Clegg, PhD](https://blog.jmir.org/author/reviewed-by-kayleigh-ann-clegg-phd) 

[Read More](https://blog.jmir.org/can-elastic-tape-supercharge-pulmonary-rehabilitation)

[![The Blueprint for Clarity: The Four Pillars of Health Care Dashboard Design](https://blog.jmir.org/hubfs/Heike%20Vornhagen%20Blog%20Post%20v2.png)](https://blog.jmir.org/mastering-the-dashboard-revolution-in-health-care)

[Research News](https://blog.jmir.org/tag/research-news)

### [Mastering the Dashboard Revolution in Health Care](https://blog.jmir.org/mastering-the-dashboard-revolution-in-health-care)

The landscape of health care management is undergoing a digital transformation. From the frantic pace of emergency departments to the strategic boardrooms of national...

![Reviewed by Kayleigh-Ann Clegg, PhD](https://blog.jmir.org/hs-fs/hubfs/Headshot.jpg?width=45&name=Headshot.jpg) 

[Reviewed by Kayleigh-Ann Clegg, PhD](https://blog.jmir.org/author/reviewed-by-kayleigh-ann-clegg-phd) 

[Read More](https://blog.jmir.org/mastering-the-dashboard-revolution-in-health-care)

#### About

- Menu Item One
- Menu Item Two
- Menu Item Three

#### Services

- Menu Item One
- Menu Item Two
- Menu Item Three

#### News

- Menu Item One
- Menu Item Two
- Menu Item Three

Follow us on Facebook Follow us on LinkedIn Follow us on Twitter Follow us on Instagram

```json
{
  "@context" : "https://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Liana Ramos, Marketing Associate",
    "url" : "https://blog.jmir.org/author/liana-ramos"
  },
  "dateModified" : "2026-09-16T19:34:07.440Z",
  "datePublished" : "2026-07-22T15:41:44.000Z",
  "headline" : "Retraining the Algorithm: How Expert Eyes Correct EMR Errors in Maternal Care",
  "image" : [ "https://blog.jmir.org/hubfs/StoryTap%20Blog%20-%20Jennifer%20Glance.png" ],
  "mainEntityOfPage" : {
    "@id" : "https://blog.jmir.org/retraining-the-algorithm-how-expert-eyes-correct-emr-errors-in-maternal-care",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "url" : "https://blog.jmir.org/hubfs/JMIR-Publications_logo_2020_original.png"
    },
    "name" : "JMIR Publications"
  }
}
```