COVID-19 has boosted data-driven healthcare. There has been a boom in telehealth and innovation in data technology supporting pandemic response initiatives, including genomic sequencing, testing and tracing, and the remote monitoring of COVID-19 patients.
Healthcare providers have also found new value in their electronic medical record (EMR) systems. EMRs have supported telehealth, providing easy access to patient data in new settings to maintain quality of experience, and helped providers to pivot their operations, such as setting up remote clinics or new COVID wards.
Data-driven healthcare approaches, like value-based care, predictive analytics, and AI, promise further benefits. But providers will have to work hard to sustain innovation. Continued success will require cultural change in data literacy and collaboration with suppliers, as well as technologies like interoperability and data management.
Trust in data needs to be a focus
We all know that there is value in data. With analytics tools, healthcare providers can unlock that value by linking data across different systems and analysing it in new ways to inform decision-making or ‘operationalise’ the data.
But poor data quality is holding many organisations back. A recent study1 by Sage Growth Partners found that only 20 per cent of healthcare executives in the U.S. fully trusted the data they rely on to make decisions, with 64 per cent saying it was ‘somewhat credible.’ This is also a big problem for Asian healthcare organisations.
The study found that 53 per cent of healthcare executives say poor data quality reduces their decision-making ability. It also impacts their ability to identify gaps in care, meet quality metrics, and optimise the revenue cycle. For a healthcare organisation with 2000 beds, Sage Growth Partners calculated the average cost of poor data quality to be US$12.9 million per year.
Factors leading to poor trust in data
Talking to healthcare executives in Asian hospitals, we have identified several factors contributing to poor trust in health data. One is the way information systems are used. If a system is not fully adopted by staff, or data entry is difficult and people take shortcuts, then the data is not trusted.
For example, when users of a clinical application or EMR module are under pressure, they may choose the first item on a drop-down list rather than taking the time to select the correct item. When this data is analysed later, it will become evident that it is skewed or inaccurate, and clinicians will not trust it to make decisions.
Inconsistent data is also a widespread problem. Data may make sense within the system it was created, but not when compared with data from another system. For example, a radiology test and a lab test on the same patient may be correct when viewed in the proper context. But if the patient demographic data does not match, how can you be sure it’s the same patient?
Another common factor is around naming. Suppose the names of medications, consumables, or suppliers are not the same across different systems. In that case, it is difficult to trust the data when it is aggregated for analysis unless further work is done to make the data consistent.
Build trust through data management
The challenge for most healthcare organisations is to move from somewhat trusting their data to fully trusting their data. That can be achieved by “harmonisation” or creating a single source of truth for data across the organisation and beyond.
Smart data fabric technology can greatly assist this process. It can unify, harmonise and analyse data across a healthcare enterprise with far less time and effort than manual processes. It can also support a wide range of analytic capabilities – including self-service business intelligence, natural language processing, and machine learning – to make it easier and faster for healthcare organisations to process and share larger volumes of data, and ensure it can be relied upon for analysis.
These data management capabilities are essential to encourage greater use of data and build trust in it. They can assist clinicians and other carers in leveraging healthcare data and understanding the best treatment for an individual patient. This could be achieved by applying AI algorithms trained using a complete and accurate patient data set.
For example, Mercy Radiology2 in New Zealand, working with Ferrum AI3, has deployed AI algorithms that use machine learning to improve over time with use. Operating in a second read capacity for radiologists, the use-cases are in lung nodules on CT scans and limb fractures on X-ray. The algorithms have improved the quality of reporting, and “there has been positive engagement from clinicians as well,” says Lloyd McCann, CEO of Mercy Radiology and Head of Digital Health for Healthcare Holdings Limited.
Provide value to engage clinicians
The positive engagement of clinicians is an important point. Improving the quality of healthcare data, and trust in it, requires the active engagement of clinicians. They need to see the link between the quality of data captured by clinical systems and the quality of outcomes achieved.
Clinician engagement cannot be taken for granted. Doctors and nurses are often very unselfish and generous. However, they are also very busy people, and if you want to engage them in improving data quality, you must be able to answer the question, “What’s in it for me?”.
To do that, healthcare providers need to find ways to deliver data back to clinicians in ways that help them do their jobs. This could be as simple as using visualisation tools to allow clinicians to better understand diagnostic pathways or outcomes for a particular patient. Or it could be as complex as feeding their data into machine learning algorithms to create diagnostic support tools.
Marie Kondo your data!
Another strategy is to consider which applications provide the most value to clinicians and are most likely to have good quality data. In January, I had the opportunity to speak with the digital health community at the Asia Healthcare Analytics Summit4. One tip I gave was to Marie Kondo your data! If a system doesn’t bring you joy, then consolidate it or replace it with one that does.
A smaller set of systems suppliers with broader capabilities will simplify your operations and make it easier to harmonise your data. While a smart healthcare data fabric enables data harmonisation, automated processes will be easier to set up, and creating a complete dataset will be easier if all your systems are well adopted and trusted.
Finally, you should aim to minimise manual processes. Use automation to do the heavy lifting and employ your staff to identify new data insights and operationalise them to be used day in and day out with minimal effort.
Adopt a “growth mindset”
The global pandemic increased the willingness and the appetite to change and opened the way for innovation. The challenge for leaders of healthcare organisations is to sustain this enthusiasm and encourage a “growth mindset” where people are motivated to find and adopt new and better ways of doing things.
Reflecting on what his organisation learned from the remotelysupported deployment of a new EMR system, Lloyd McCann described the need to embed a change culture.
“Through COVID-19, we’ve come to learn that seeing change as a change management process is not enough. Through the virtual go-live experience, for example, we learned that the factors for success are mostly driven by human teams,” said McCann.
“We’ve got to build and embed a change culture across the health system and across the organisation. Because only when people are comfortable that change is part of the new normal will we start to derive the true benefits and value of digital health approaches.”
Improve people’s data literacy
For organisations to change and grow to adopt data-driven healthcare, people need to understand the potential of technology. Organisations can start by assessing their levels of data literacy. That will give you an understanding of people’s capacity to adopt new technology and participate in developing data-driven care initiatives.
Then, you can determine what levels of literacy are required to drive change and benefit from new technology investments. After setting those goals, you can put learning and mentoring programs in place to get there. Leaders should take a tailored approach. For example, some staff may need basic training in how to use digital devices.
But even very digitally literate people like systems developers may benefit from new skills. For example, InterSystems recently conducted a summer internship program for five information technology students from Charles Darwin University. We taught the students about FHIR®5, HL7’s newest clinical data interoperability standard, and how to use it within the new TrakCare Interoperability Toolkit to submit to and retrieve and display information from the InterSystems TrakCare unified healthcare information system.
Using this toolkit, the students were able to quickly develop a prototype for a mobile phone app to help hospital ward nurses to capture and monitor patients’ vital signs. The app uses FHIR to make it easy for nurses to capture patient data at the bedside and to integrate with data already entered in the TrakCare EMR system to quickly alert them about any deterioration in a patient’s condition.
“Disruptive” solutions for better care
Healthcare leaders should also look for innovative data-driven solutions from suppliers, including health tech companies and start-ups, medical technology and device companies, pharmaceutical companies, and others. By working collaboratively with suppliers, healthcare providers can take advantage of global innovations and learn from their experience.
InterSystems is working with one company, RxMx®, that develops apps and portals around lab testing to ensure the safety of specialty medications. Built on the InterSystems IRIS for Health™6 data latform, RxMx’s Chameleon platform7 provides automated risk management to keep patients safe throughout complex specialty treatments while integrating with labs and other vendors in real time.
We will see many more ‘disruptive’ solutions that provide better care for patients and make lives easier for clinicians. Health providers will increasingly integrate their systems with those of medical device manufacturers and use big data and AI for improved diagnosis and predictive analysis. And we will see more innovative care solutions from start-ups using healthcare data standards like FHIR that make them easier to integrate.
Build interoperability at the beginning
At the moment, healthcare providers can struggle to integrate innovative solutions into their workflows and systems. For example, more and more medical devices are collecting data, whether in the hospital or the home. These are new sources of valuable data, but healthcare providers can find it challenging to use them for analytics and data-driven care.
Manufacturers and start-ups need to consider how devices can integrate with existing data infrastructure using standards-based interoperability support. For example, many remote monitoring devices currently require their own data infrastructure. Because of the cost involved, this is a considerable barrier to providers approving business cases.
If products include standards-based interoperability, they can leverage existing infrastructure to make them easier and more cost-effective to deploy. A modern interoperability standard like HL7 FHIR, which works securely via the Internet, could make a real difference to a product’s viability.
InterSystems believes it is essential to prioritise interoperability and data cleansing so that data is usable in machine learning and other innovative, data-driven solutions. Our company’s cloud-first data platforms also solve the speed and scalability problems that healthcare must overcome to manage the exponential growth in data it produces.
During the pandemic, breakthroughs in digital health have given providers a once-in-a-lifetime opportunity to innovate and improve how care is delivered. To take advantage of it, they need to sustain their efforts through cultural change and embrace enabling technologies like interoperability and data management.