Data Science

The Value of Computer Vision in Healthcare

Reviewing medical images

Artificial intelligence solutions and computer vision, in particular are accelerating the pace of change in healthcare. Best of all, the widespread adoption of visual AI strategies for healthcare has only just begun. As a subfield of artificial intelligence, visual AI enables computers to "see" and process images and videos faster, more accurately, and more cost efficiently than a highly trained medical practitioner.

Computer vision systems for healthcare are also designed with HIPAA compliance in mind. For security purposes, these systems can deploy “on the edge.” This means that they run on processors with no connectivity to the cloud—making them less vulnerable to security breaches.

Computer vision technology is quickly offering the following benefits to the healthcare industry:

  1. Accurate and efficient imaging analysis
  2. Smart operating rooms
  3. Better patient identification
  4. Increased healthcare safety
  5. Accelerated medical research

We'll take a closer look at each of these benefits and also touch on ROI opportunities.

1. More Accurate and Efficient Imaging Analysis

Computer vision can improve both speed and accuracy when analyzing medical imaging: recognizing hidden patterns and making diagnoses with fewer errors than human professionals. A study in Nature found that visual AI systems were more accurate than human radiologists when analyzing mammograms for signs of breast cancer, reducing both false positives and false negatives. What's more, workloads were reduced by 88% when the AI system and providers worked together.

This technology can also act as another pair of “eyes” to support accurate diagnoses. This will be especially critical in the coming years as demand for medical image analysis continues to outpace the number of skilled human workers capable of providing these services. The U.S. Bureau of Labor Statistics projects a 9% increase in the number of radiologic and MRI technicians in the United States by 2028, but analysts expect that the digital pathology market will grow by 11.8% before 2027. As we face the increasing shortage of radiologic and MRI technicians, this technology can 1) reduce your medical imaging workloads, 2) reduce labor costs, and 3) ease your staffing shortages despite this explosive growth in demand.

2. Smart Operating Rooms

A typical EHR system may require surgical nurses to make up to 100 clicks to document a surgical procedure. Computer vision systems can eliminate the need for this manual effort through direct observation and documentation, reducing or eliminating the need for human input. This advancement lets doctors and nurses spend more time on patient care, helping reduce stress for providers and improving outcomes for patients.

Smart operating rooms with this technology can also prevent errors. An estimated 1,500 surgeries every year result in a foreign object being left in a patient. Computer vision systems can keep track of surgical supplies and tools to protect against injuries caused by so-called "retained surgical bodies" (RSBs). By warning providers of RSBs, computer vision systems for healthcare help ease a significant source of stress during the operation while significantly improving patient care and surgical outcomes.

3. Better Patient Identification

Cases of mistaken patient identity are unfortunately all too common in healthcare. A study by the Ponemon Institute found that 86% of nurses, physicians, and IT practitioners have personally seen or known of a patient being mistaken for someone else. While these errors are usually caught early, they can be dangerous or even deadly: If not corrected, patients can end up taking the wrong medications, receiving the wrong tests or treatments, or undergoing the wrong surgery. The Ponemon study also estimated that the average hospital risks $17.4 million in losses per year because of patient identification errors, including personal injury or malpractice claims.

That's exactly where computer vision comes in, helping prevent these cases of mistaken identity. Fully HIPAA-compliant, AI-powered facial authentication systems can recognize patients with an extremely high degree of accuracy in just a fraction of a second. Identification avoids patient misidentification while allowing providers access to correct medical records. These systems offer multiple layers of security for patient identity and access.

4. Increased Healthcare Safety

In 2019, U.S. hospitals recorded 221,400 work-related injuries and illnesses and a rate of 5.5 work-related injuries and illnesses for every 100 full-time employees. This is almost twice the normal rate for private industry as a whole. While many of these accidents and injuries are unavoidable, this technology can prevent many common worker injuries at medical facilities. They can reduce patient injuries as well. Moreover, when unavoidable injuries do occur, visual AI systems can send immediate alerts to appropriate personnel for faster response times.

The vast majority of worker and patient injuries at medical facilities happen when staff members fail to use required safety equipment, or when they fail to comply with safety protocols related to sterile processing. Here’s how this technology can dramatically reduce injuries related to these failures:

  • Accurately identifying when employees fail to use appropriate safety equipment: Medical and laboratory staff who fail to use gloves, goggles, masks, face shields and protective clothing when safety rules require it, significantly increase their chances of injury. Computer vision systems can monitor staff at all times to ensure they are wearing the required safety equipment at all times.
  • Accurately identifying sterile processing failures: When medical staff members do not adhere to stringent device cleaning, disinfection and sterilization standards, the risk of surgical site infection increases dramatically. Surgical site infections have a 3% mortality rate and an associated annual cost of $3.3 billion. Computer vision systems can accurately monitor and track whether medical personnel are adhering to sterile processing protocols. This reduces instances of sterile processing failures while holding staff members accountable when such failures occur.

Injuries related to smoke and fire are also important concerns at medical facilities. With the ability to immediately identify the subtle signs of smoke and fire, computer vision systems can trigger evacuation alarms and notify emergency response teams earlier than traditional detection systems. Beyond smoke and fire, computer vision for fall detection can identify when a worker or patient falls, instantly notifying medical teams, and record the instance for legal and insurance purposes.

5. Accelerated Medical Research

Last but not least, computer vision for medical research helps healthcare organizations speed up the process of investigating and testing new treatments and technologies. Here are a couple of examples to illustrate how visual AI can assist with medical research:

  • Faster and more accurate cell counting and typing: According to a study on cell counting methods, 71% of 400 researchers used a hemocytometer to count cells—even though the process is labor-intensive and prone to misuse and user bias, causing inaccurate results. In contrast, computer vision systems for cell counting can count and type cells 100-times faster than human workers. This allows researcher facilities to achieve faster results—potentially researching and releasing new medications faster and more economically.
  • Faster and more accurate cohort identification: Medical researchers can use interactive visual AI systems to speed up the process of visually analyzing and exploring clinical data. For example, one study showed how computer vision systems can assist with cohort identification in cases of pediatric asthma patients. The study showed that these systems offer a faster, more efficient, and more accurate means of identifying cohorts than traditional approaches.

These are just two examples applying this technology in medical research. Ultimately, this advanced AI technology can perform virtually any task that would normally require human vision, human analysis, and understanding in a research or laboratory setting.

The Benefits of Computer Vision in Healthcare

Leveraging computer vision solutions for these healthcare use cases offers ROI benefits for doctor’s offices, hospitals, outpatient surgical centers, medical labs, medical research centers and other healthcare-related facilities. These ROI benefits include:

  • Fewer patient readmissions: By allowing doctors, medical staff, and lab workers to provide more accurate diagnoses and better care, these solutions help medical facilities achieve better outcomes for their patients and fewer readmissions.
  • Fewer accidents and injuries and safer medical facilities: By helping medical facilities monitor whether employees are using appropriate safety equipment and following sterile processing protocols—and by identifying smoke, fire, fall accidents—these solutions help medical facilities reduce accidents, injuries, and unintended surgical site infections.
  • Lower insurance costs: By dramatically improving the safety of medical facilities—and the accuracy and quality of medical care—these systems reduce the risk of workers’ compensation and medical malpractice claims. This results in lower insurance costs for medical facilities.
  • Decreased labor expenses: By assisting healthcare facilities and medical labs to achieve greater speed and accuracy when completing a variety of tasks, this technology streamlines workflows and decreases labor costs—empowering organizations to accomplish more with fewer staff members on the payroll. 

Sponsored content. The views and opinions expressed in this content or by commenters are those of the author and do not necessarily reflect the official policy or position of HIMSS or its affiliates.