10 Real-World Case Studies of Implementing AI in Healthcare

1. Moorfields Eye Hospital – the best case study of AI in healthcare

Moorfields is the world’s oldest eye hospital. Eye health professionals there had to analyze over 5,000 optical coherence tomography (OCT) scans per week to spot and diagnose severe eye conditions like diabetic retinopathy or age-related macular degeneration (AMD). However, these eye scans could take a long time for manual analytics, which affects early detection and diagnostics.

Solution & Result

In 2018, Pearse Keane, a consultant ophthalmologist at Moorfields, came to DeepMind for AI solutions.

Moorfields and DeepMind collaboratively developed an AI tool that can identify more than 50 eye diseases as accurately as top eye professionals. The tool was trained with almost 15,000 OCT scans from 7,500 patients and real referral decisions. It uses deep learning algorithms to detect the various anatomical elements of an eye and create a 3D image that shows the thickness of retinal tissue through near-infrared light (the figure below).

An example of an OCT scan, showing the thickness of retinal tissue in a patient’s eye.

The software can even offer clinical advice based on the different signs of eye conditions in the scans. As a result, its recommendations were considered 94% accurate as diagnostics by top eye professionals.

The software even explains how it came to its decisions. This helps doctors and nurses trust and use its recommendations more carefully.

In addition to early detection, AI algorithms can help predict disease progression. Google’s DeepMind conducted a test to evaluate how its AI model could forecast the high risk of an eye converting to exudative AMD. The model automatically segmented different types of tissues identified in the eye scans and observed their changes over time. As a result, the model successfully predicted that the eye would likely worsen at least 2 visits before signs of exAMD became clear.

an example of how AI can predict the high risk of an eye converting to exudative AMD

2. HCA Healthcare

HCA Healthcare is one of the US’s largest healthcare systems. It manages a vast network of hospitals, surgery centers, and other healthcare facilities across the United States and the United Kingdom. HCA Healthcare focuses on patient care and community health.

With numerous points of entry, HCA Healthcare must manually sift through pathology reports, physician schedules, and referrals to detect newly diagnosed cancer patients and cases. This not only proved time-wasting but also made resource allocation ineffective. As a result, there were delays in identifying cancer patients, hindering them from timely diagnosis and treatment.

Solution & Result

HCA Healthcare chose Azra AI as a comprehensive solution to this problem. Azra AI is a SaaS (Software-as-a-Service) clinical intelligence platform that uses AI technology to automate oncology workflows. It’s widely used by more than 250 US-based hospitals and cancer centers like Inspira Health, HCA Healthcare, and the University of Pennsylvania Health System.

Here’s how Azra AI can help:

Early Cancer Detection: Azra AI uses AI to analyze pathology reports and spot possible cancer patients in real-time.
Surface Incidental Findings: Azra AI detects incidental findings (e.g., undiagnosed cancers) in radiology reports. This helps discover cancers that could be overlooked during other screenings.
Cancer Registry Automation: The software extracts key information from medical records and automatically fills in over 50 certain fields. This saves time and minimizes errors due to human entry.
Nurse Navigator System: The software is directly integrated into the Meditech Electronic Health Record (EHR) of HCA Healthcare. This allows HCA’s nurse navigator team to automatically receive diagnosis results and track patients through treatment.
Centralized Data Platform: By centralizing HCA’s data into one platform, Azra AI provides real-time analytics to manage cancer patient volumes efficiently.

As a result, HCA Healthcare decreases their time from diagnosis to the first treatment by 6 days and saves over 11,000 hours for manually reviewing and reading pathology reports. Azra AI also allows HCA’s care team to spend more time navigating and coordinating patient care (65%). For this reason, HCA Healthcare added over 10,000 new oncology patients within 14 months.

3. Duke Health

Duke Health is a world-class academic healthcare organization affiliated with Duke University. It includes a large network of hospitals, research centers, and educational institutions like Duke University Hospital, or Duke Clinical Research Institute. This healthcare facility, therefore, offers a variety of medical services, including clinical care (e.g., primary care or complex surgeries), research, and training.

Solution & Result

Since 2019, the facility has used GE Healthcare’s Command Center Software to streamline its operations and improve hospital-wide visibility. Particularly, this AI platform helps them track patient flow, manage capacity (e.g., bed availability or staffing levels), and predict future patient demands. This gives their care teams more time to focus on supporting and caring for patients.

Now, Duke Health continues to adopt GE Healthcare’s new functionality, known as Hospital Pulse Tile. According to Kristie Barazsu, Associate COO at Duke University Hospital, this new tool provides a sustainable solution to Duke Health’s back-end operations. In other words, Hospital Pulse Tile provides hospital leaders with real-time operational insights to make data-driven decisions.

This feature, accordingly, compares historical and real-time data to explain important operational metrics. For example, suppose the current number of patients admitted to emergency rooms exceeds daily admissions over the past 6 months. In that case, it can point out an increase in demand or a possible crisis.

Besides, Hospital Pulse also visualizes these key metrics to detect bottlenecks and ensure operational efficiency. It also tailors dashboards to align insights with different leadership levels (e.g., executive teams or nursing units).

As a result, Duke Health has made impressive improvements in its operations. For example, GE Healthcare has increased 6% in overall productivity, reduced 50% in temporary labor demands, and decreased 66% in time from bed request to assignment.

4. University Hospitals

University Hospitals (UH) is one of the leading healthcare organizations in Ohio that focuses on quaternary care and medical research. Its vast network consists of 21 hospitals, over 50 health centers and outpatient facilities, and more than 200 physician offices.

The healthcare facility is committed to offering the highest-quality patient care and helping its care team handle medical issues effectively. Therefore, it looked for a hyper-accurate AI solution that can smoothly integrate its infrastructure of current hospitals and outpatient locations.

Solution & Result

To address its existing demand, University Hospitals chose Aidoc’s proprietary aiOS™ – a unified operating system – for its 13 hospitals and dozens of outpatient locations in Cleveland. Aidoc is well-known for its healthcare solutions that use AI technology to detect crucial findings and prioritize urgent cases by analyzing medical images (e.g., CT scans or X-rays).

With Aidoc, University Hospitals’s care team can instantly access key patient data that is centralized in one healthcare system. Besides, they can provide faster diagnosis and treatment of serious diseases and complications (like pneumothorax, aortic dissection, or pulmonary embolism).

Particularly, when a patient visits a UH facility and experiences a CT scan for an injury, Aidoc will use FDA-cleared AI algorithms to analyze the scan. By spotting expected and unexpected findings, these algorithms can help radiologists assess patient images quickly and prioritize emergencies. Further, they flag all conditions for the care team to review, ensuring no cases will be missed. This will speed up diagnostics and treatment, enhancing patient outcomes.

5. Johns Hopkins Medicine

Johns Hopkins Medicine is a healthcare and medical research leader headquartered in Baltimore. The healthcare organization has a large network of top-ranked hospitals, suburban healthcare and surgery centers, and care locations – all renowned for innovations in surgical techniques and specialties like oncology or cardiology. Besides clinical care, Johns Hopkins Medicine also focuses on biomedical research and education.

Solution & Result

To improve in-hospital and outpatient care as well as help healthcare staff avoid excessive alarms, the facility has long invested in AI technologies.

Below are several AI projects being deployed at Johns Hopkins Medicine:

inHealth

This is Johns Hopkins’s strategic approach to improving precision medicine. To support inHealth, the healthcare facility has founded 16 Precision Medicine Medicine Centers of Excellence to research different chronic diseases like prostate cancer or multiple sclerosis. Further, JHM has used PMAP (Precision Medicine Analytics Platform) to collect and analyze data from different sources securely.

Since 2015, JHM has integrated Microsoft’s Azure and analytical tools to advance its discoveries in personalized healthcare through AI. According to Paul B. Rothman, M.D., CEO of JHM, integrating Azure will enhance JHM’s ability to develop innovative treatments for its patients.

Aging Research

Further, Johns Hopkins Medicine has conducted research on aging diseases. Its research teams leverage AI technologies (like computer vision) to detect early signs of diseases (e.g., stroke or Alzheimer’s). Let’s take a look at how they can do so:

Patient Messaging

Clinicians at Hopkins saw the number of patient messages via email and patient portals increase by almost 3X from late 2019 (pre-COVID) to now. This increased communication overwhelmed clinicians as traditional clinical workflows were not developed to process this growing volume.

Therefore, Hopkins uses AI technology to automate 30-40% of response tasks. By analyzing incoming patient messages, AI helps clinicians create draft responses to the increasing volume of daily inquiries. This saves time and minimizes burnout.

Ambient Scribing

This technology uses machine learning to transcribe clinical conversations (e.g., ER history or outpatient visits) in real-time. Combined with large language models (LLMs) like GPT, ambient AI can automatically create clinical documents required by complex healthcare systems and regulatory standards.

6. Safoni

Sanofi is another impressive case study of using AI in healthcare. This Paris-headquartered company is renowned for developing, producing, and delivering pharmaceuticals, vaccines, as well as patient care. It focuses on specialties like infectious diseases (e.g., polio or influenza), immunology, and diabetes.

Solution & Result

Sanofi has sought innovative ways to improve people’s health and operational efficiency.

Drug Discovery and Development

Sanofi has partnered with various biotech organizations like Insilico Medicine or Owkin to integrate AI capabilities into drug discovery and development.

One of the most striking collaborations is with Insilico Medicine. Insilico Medicine was born with the transformative idea of combining bioinformatics and deep learning to discover disease pathways and life-saving therapeutics.

Since their agreement was signed in November 2022, the joint R&D teams of Sanofi and Insilico have used PandaOmics and Chemistry42 (Insilico’s proprietary biology GenAI tools) to analyze complex biological data and forecast which molecules can work. This process helps identify potential new drugs that specifically target transcription factors (“undruggable” proteins controlling gene activity). In other words, these AI platforms open an avenue to discover novel treatments for specific oncology diseases.

Further, Sanofi has initiated an internal program called BioAIM (Biologics x AI Moonshot). This program aims at transforming how the facility identifies and develops new drugs, particularly nanobodies and antibodies.

Also, AI plays a key role in mRNA vaccine research. To ensure an mRNA vaccine will arrive at the right cells in your body and generate proteins against diseases, a protective coat is necessary to carry it safely. This coat is known as a “lipid nanoparticle” (LNP). R&D teams would design AI predictive models to identify which LNP works best for a specific vaccine.

With AI capabilities, Sanofi speeds up the R&D team’s research process from months to days while fostering drug discovery by 20-30%.

Operational Efficiency

Sanofi collaborated with Aily Labs to develop plai, a customized AI solution that makes hospital operations more effective.

For clinical operations, plai helps detect and recruit suitable patients for clinical trials. Plus, it aids R&D teams in setting up more convenient trial sites to encourage subjects from diverse backgrounds to participate.

In manufacturing and supply, plai can forecast 80% of possible shortages in inventory. This enables Sanofi to take prompt action to secure the supply chain. Sanofi also develops an internal AI solution that automates quality evaluation and optimizes the use of raw materials for manufacturing.

7. Humber River Health

Humber River Health (HRH) is the first digital hospital in Canada. It uses modern technologies like AI and robotics to streamline operations as well as ensure accurate diagnosis and treatments even for remote patients.

The hospital has a mission of improving inpatient care through innovations in areas like cardiac care or mental health. Therefore, it has invested in fully digital infrastructure and research to address existing healthcare challenges in Canada, particularly overcrowded emergency rooms and long wait times.

Solution & Result

The most transformative AI application in HRH is the adoption of robotics in procedures. One typical example is using the da Vinci Surgical System which enables surgeons to perform minimally invasive procedures with improved accuracy. Let’s see how it works:

Moreover, HRH’s surgeons also leverage the ROSA® Knee System to support their knee replacement procedures. The system provides real-time data and 3D imaging to help surgeons customize surgeries for each patient’s unique anatomy. This ensures precise alignment of the knee implant.

Similarly, Intellijoint HIP® is also widely used at Humber to enhance hip replacement procedures. This surgical robotics system reduces the need for additional imaging equipment (e.g., fluoroscopy) and X-rays during surgery. Therefore, the surgical process will become much simpler while enabling accurate cup placement and leg length adjustments.

8. Boston Children’s Hospital

Boston Children’s Hospital (BCH) is one of the world’s biggest pediatric training and research hospitals. It offers a wide range of medical services to those aged 0-21 and even adults who need pediatric care.

Boston Children’s Hospital has deployed AI initiatives across various operations. According to John Brownstein, Chief Innovation Officer at BCH, its motivations for using AI lie in improving the quality of patient care, boosting financial ROI, and ensuring responsible AI usage.

Solution & Result

So, how are AI technologies being deployed in Boston Children’s Hospital? Let’s take a look:

Research

BCH established the Image, Informatics and Intelligence (i3) Lab to advance healthcare through the following innovations:

Algorithm Development: The i3 lab develops machine/deep learning algorithms and medical imaging analytics tools for image segmentation, prediction, and other purposes. For instance, they created unbiased DRAMMS group-wise registration algorithms to combine all brain scans from different people into a single, average atlas.

Algorithm development in BCH for the brain atlas

Computational Neuroscience: The research team collects a vast volume of brain data, including MRI scans (e.g., the brain’s structure) and non-MRI data (e.g., genetic information) to track how the brain’s structure and functions change over time. They also identify how male and female brains differ, and how two brain hemispheres work. This helps with discovering more accurate treatments for brain disorders.

Translational/Clinical Research: The lab studies brain abnormalities with different levels of severity (e.g., malnutrition, atrophy, or psychiatry).

Hospital Admissions

BCH has partnered with different AI solution providers like SMART Health IT to build POPP (Prediction of Patient Placement) – a predictive model based on its expertise and data. This real-time model can forecast incoming admissions from the Emergency Department (ED), enabling the proactive coordination of resources (like hospital beds or equipment). Surprisingly, the model achieves more than 90% accuracy in its predictions.

Infectious Disease Monitoring

Further, BCH also confronted a high demand in pediatric beds due to large population surges in RSV & influenza infections. That’s why BCH combined machine learning, expert forecasts, spatiotemporal models, and big data to develop a predictive model for infection hospitalization.

Fine-Tuned LLMs

BCH refined large language models for different healthcare use cases. One typical example is leveraging these AI tools to help nursing staff access and understand clinical policies or protocols just-in-time.

Moreover, BCH deploys a Bot Builder that allows people from any department to develop their own AI bots, whether it’s for serving a cardiac ICU, taking patient histories in the ED, or tailoring patient documentation.

9. Diagnostikum (Linz, Austria)

Diagnostikum is an Austria-based group that operates leading radiology institutions in four locations: Vienna, Schladming, Linz, and Graz.

The healthcare facility provides a variety of medical imaging services for diagnostics and treatments, especially for breast cancer. These services range from magnetic resonance tomography and sonography to digital mammography and low-dose X-ray radiation.

For years, radiologists at Diagnostikum have faced challenges like a growing workload, a serious staff shortage, and the increasing complexity of diagnostic tools. This raised a need for using AI to handle these problems while ensuring effective patient care.

Solution & Result

Since 2021, Diagnostikum has chosen the AI-Rad Companion of Siemens Healthineers for chest CT imaging. This AI tool uses machine/deep learning algorithms to automatically analyze CT scans and spot abnormal patterns (e.g., lung nodules or pneumonia).

Once the AI-Rad Companion Chest CT has finished its calculations, all the results are sent to radiologists for rigorous review. Then, these findings will be delivered to the PACS (Picture Archiving and Communication System) for storage, retrieval, and management.

With this AI tool, Diagnostikum saves much time and resources for both radiographers and patients. In the below example of a patient who needs a thoracic aorta examination, AI helps Diagnostikum remove unnecessary steps like testing the creatinine level before the assessment or preparing the IV line and contrast injector.

How AI-Rad Companion Chest CT helps Diagnostikum in aorta examination

Here’s how the AI-Rad Companion Chest CT works in this example:

Use deep learning algorithms to generate a 3D model of the thoracic aorta and detect its 11 key measurements in a contrast-enhanced CT scan (Figure 1) and a non-contrast CT scan (Figure 3).
Define the aorta’s centerline and measure the diameter at nine positions along the centerline (Figure 4).
Create a DICOM SR and DICOM image and highlight results using color coding to help radiologists easily detect abnormalities (Figure 2).

10. University of Florida Health

Now, you’ve come to the final case study of AI in healthcare in today’s list: University of Florida Health. This world-class healthcare system consists of 11 hospitals and hundreds of outpatient places across Florida. It promotes health through excellent patient care, innovative research, and dedicated education in different specialties, like pediatrics, cancer, or neurology.

Solution & Result

To deliver optimal diagnostics and therapies, UF Health’s researchers are developing AI-powered predictive systems that help clinicians make informed decisions and monitor patients effectively.

One typical project at UF Health is the adoption of AI to evaluate a patient’s condition, movement, and room environment, especially in the ICU. With pervasive AI, busy doctors and nurses can track even the smallest cues that indicate pain or discomfort.

Tyle Loftus, M.D., UF Health’s acute care surgeon and researcher, also highlights the importance of AI after procedures. Particularly, AI predicts postoperative complications that sometimes can be missed by human error or resource shortage. This helps doctors decide whether a patient after surgery should be moved to the ICU with specialized care and high-frequency surveillance.

Transforming Your Healthcare with Designveloper’s AI Solutions

These case studies have clarified how healthcare organizations worldwide are using AI to improve patient care and operational efficiency. With continued advancements in AI, we expect to see a wider adoption of this modern technology in the medical industry, especially for identifying diseases and developing new drugs.

However, your AI journey might become more challenging if you don’t have appropriate AI strategies and the right solution. So, Designveloper is here to facilitate AI integration into your healthcare services.

Here’s why you should consider us as a long-term partner:

Proven Track Record

Since our foundation in 2013, people at Designveloper have carried out 100+ successful projects in different segments.

One typical healthcare project is ODC. We developed a comprehensive telehealth platform that connects patients and healthcare providers across France.

At the peak of the COVID-19 pandemic in the first half of 2020, our team processed 1 million patient information and COVID-19 vaccination standard forms daily. In 2021, Designveloper was honored by the Ministry of Health as a rising healthcare startup for its contributions during the pandemic.

End-to-End Development

We offer full life-cycle support, from ideation to deployment and maintenance.

Modern Tech Stacks & Tools

We have extensive expertise in applying 50 different technologies to our projects. From backend and databases to frontend and UI/UX design, we can turn your idea into working software.

Also, we leverage communication tools like Trello and Jira to help you and the whole team track the project closely.