Header background

AIOps observability adoption ascends in healthcare

As healthcare digitizes and modernizes, it has started to see the merits of AIOps observability, enabling IT teams to identify and address issues automatically.

AIOps observability has become a critical need in healthcare.

Every day, healthcare organizations across the globe have embraced innovative technology to streamline the delivery of patient care.

During the early months of the COVID-19 pandemic, this trend was undeniably apparent. Many hospitals adopted telehealth and other virtual technology to deliver care and reduce the spread of disease.

But the push toward digital experiences was thriving long before the advent of the pandemic. Over the past decade, the industry moved from paper-based to electronic health records (EHRs)—digitizing the backbone of patient data. As patient care continues to evolve, IT teams have accelerated this shift from legacy, on-premises systems to cloud technology to build, test, and deploy software and fuel healthcare innovation.

The healthcare industry is embracing cloud technology to improve the efficiency, quality, and security of patient care. As a result, these organizations are finding they need AIOps observability to keep track of their increasingly complex systems.

Four challenges to achieving AIOps observability

While providers embrace the benefits of AI and other emerging technologies, there are several challenges providers should consider as they navigate change, including the following:

  1. Overwhelming complexity. Today’s enterprise environments are dynamic and complex. With containers, microservices, applications, and other components that are constantly broken down and rebuilt as part of the software development lifecycle (SDLC), IT teams struggle to identify true issues and deliver software that enables doctors and nurses to deliver care. This is a critical challenge: When software breaks, finding the root cause of the problem may take time, fuel finger-pointing among teams. The accompanying outage can disrupt patient experiences or, worse, threaten patient care.
  2. Scale and growth amid change. With new staff, patients, and merger-and-acquisition activity pervasive in healthcare, IT teams have to manage and update environments that constantly change. Amid so much change, teams can no longer rely on time-consuming, manual approaches for building, testing, operating, and upgrading software. They need automated AIOps observability approaches based on real-time, contextualized data.
  3. Emerging security threats. Cybercriminals have targeted healthcare. As a result, security has to be integrated within the SDLC natively to protect health information, to continue to innovate, and to ensure excellent patient care. To do so effectively, security cannot be treated as a barrier to innovation; it should be a pillar of it. Integrated security approaches is especially important as organizations increasingly shift to a DevSecOps mindset.
  4. Compliance. Health Insurance Portability Accountability Act (HIPAA) violations are costly, and they can undermine a hospital’s reputation with its patients. It’s not just HIPAA that providers must worry about; there’s the payment card industry, Sarbanes-Oxley, Occupational Safety and Health Administration, and other regulations to consider. This means that an organization could face regulatory fines and penalties if a software vulnerability, downtime, or misconfiguration leads to loss of PHI, security issues, or patient injury.

Artificial Intelligence for IT and DevSecOps

COVID-19 placed new pressure on medical practitioners and put digital healthcare at the forefront. It’s been a high-stakes, high-intensity few years for an industry that already faced attrition as well as burnout among clinical staff.

With so much at stake, the directive for IT and security teams became even more concrete: clinicians need systems that are available at any time and from anywhere, they cannot experience outages, and they cannot be vulnerable to cyberattacks. Strained IT and security resources teams to think creatively to meet the challenge.

Simply adding more resources or working longer hours would not suffice. Further, as organizations adopted cloud computing, it opened the doors to newer technology, but it inadvertently increased the pressure for teams to navigate more change, manage more data, and implement new systems flawlessly. This perfect storm of challenges has led to the accelerated adoption of artificial intelligence, including AIOps. It has also increased the need for AIOps observability.

AIOps (or “AI for IT operations”) uses artificial intelligence so that big data can help IT teams work faster and more effectively. Gartner introduced the concept of AIOps in 2016. Since then, the term has gained popularity. Observability is the ability to determine the status of systems based on their outputs. AIOps observability gives IT teams the ability to determine the status of all the constituent parts of their increasingly complex AIOps environments.

There are two main approaches to AIOps observability:

  • Traditional AIOps: Machine learning models identify correlations between IT events. That includes failures in parts of a system that occur at similar times and have a common root cause. These correlations help with troubleshooting issues or optimizing performance, but in many cases, they don’t pinpoint the precise cause of the issue.
  • Causal AIOps: A modern approach to AIOps extends capabilities beyond simple correlation. It uses deterministic AI to enable teams to make precise determinations about the root cause of problems. Thus, instead of merely correlating two or more events based on the time of their occurrence, deterministic AIOps goes deeper by identifying the underlying root cause that has triggered an event. In turn, it makes clear exactly what teams should do rather than providing hints.

As AI becomes increasingly important to care delivery, the healthcare sector is pursuing a national strategy for AI. Let’s explore how some healthcare organizations have used AI to take patient care further.

Healthcare use cases for AIOps observability

Dynatrace healthcare customers use AI and automation to continuously identify software issues and remediate problems before they affect patient care. Below are some of their use cases in more detail.

Continuous care, anytime, anywhere

One Dynatrace customer is one of the top five healthcare providers, owning and operating nearly 200 hospitals and thousands of clinical facilities across the United States. It has transformed into a modern, digital provider and now uses applications to enable remote healthcare. Pediatricians use these applications to monitor the heart rate of child patients. When doctors are away, the app alerts them if the children’s heart rate crosses a threshold, enabling the physician to manage the child’s care remotely.

AIOps observability plays a critical role in this app’s availability. AIOps delivers continuous automation and integrated AI, enabling the IT teams running this app to know in real-time whether problems could disrupt the system’s performance and the precise root cause of the problem. This reduces MTTD/R (mean time to detect/resolve) and limits potential disruption to the delivery of care. IT teams can use the extra time to release new software and work on enhancements to existing systems.

“We achieved levels of availability we’ve never been able to hit before,” said the chief architect of the healthcare provider.

Automated intelligence for better client experiences

Another Dynatrace customer is a health and life insurance provider from the United Kingdom whose core purpose is to make people healthier. To support its objectives, the company operates a digital platform that enables customers to earn points when they exercise or eat healthily. Users can then redeem points for perks and benefits such as cinema tickets and gym memberships.

Over the past few years, the company has undergone a digital transformation, migrating to a hybrid, cloud-native environment built on Amazon Web Services and a microservices architecture. This has granted the speed and agility to accelerate innovation and bring new rewards to its members more frequently. But it has also introduced complexity that makes it difficult to stay ahead of performance problems. Failure to optimize performance for digital services could create poor customer experiences, which in turn could prompt low customer retention rates. What’s more, the time that developers may spend manually tracking down the root causes of performance issues would take time away from their ability to build new features that would increase customer satisfaction.

The company has reinvented its approach to AIOps by consolidating AI and observability in one platform that automatically identifies if members haven’t received the points they’ve earned for physical activities. Their AIOps observability platform can then trigger an alert for its support team to contact the customer and address the problem. In some cases, the software fix can be deployed automatically without human intervention. As a result, members remain engaged with their health programs and can turn potentially negative customer experiences into positive ones.

What’s next for healthcare and AIOps observability

IT, security, and clinical teams will continue to face pressure to do more with less. The trends that make healthcare a complicated industry—constant M&A activity, security threats, and more complex diseases—will persist.

AIOps has become a core technology answer to cut through the complexity of modern cloud activity and consequent alert noise. As healthcare providers consider AIOps solutions, they should evaluate whether traditional AIOps approaches designed for correlation can enable long-term success. Alternatively, they may want to embrace a modern AIOps observability approach that is purpose-built for driving more efficiency, faster innovation, and better patient outcomes.

Learn more about the Dynatrace approach to AIOps and explore the Davis AI engine at the core of the Dynatrace observability platform.

To learn more about AIOps and observability, read the ebook, “Developing an AIOps strategy for cloud observability.”