Health data intelligence specializes in healthcare analytics and evidence-based business decision support for hospitals and health systems. It includes doing data trending, disease diagnostic applications, clinical trial data, and data warehousing.
Health intelligence often incorporates tools and methods from AI to capture data in the process of performing care management, utilization management, assessments, and appeals.
The shift toward telehealth practices — initially instituted as a stopgap measure in response to COVID-19 — yielded measurable healthcare improvements.
Unlike a traditional doctor visit where a small number of discrete measurements are taken in a clinical setting (along with a few modifying factors like age, weight, and gender), telehealth practices allow for the collection of large quantities of high-quality data via wearable devices.
Smartwatches are used to continuously monitor a patient’s blood pressure and pulse rate from home, where they are free from the stress of being in a clinical setting.
Besides the benefit of more data (which means more information), remote monitoring reserves the use of clinical resources, so they can be prioritized and made available to the patients who need them most. This allows for an increase in the scope of what can be addressed, and the number of people who are treated as early as possible.
The continual advancement and application of AI and ML to analyze and extract useful and actionable information from all this collected data has initiated a spiral of progress. The more these “smart” algorithms are developed and refined, the more devices will be used. More devices mean more high-quality data being generated to facilitate the further development of ML and AI algorithms.
Data collected in this way also removes the bias traditionally present in medical data. Often, people receiving tests (and generating data) in a clinical environment present symptoms of some condition. This means that the medical data may be biased toward symptomatic people rather than asymptomatic (i.e., healthy) people. Diagnostic programs will be able to identify these differences and provide a much better understanding of a population’s health as a whole.
Specialized clinicians and nurse practitioners focused on managing a patient’s disease condition use dedicated disease diagnostic applications.
In the case of heart failure, for example, a patient might have an implantable cardiac device that provides daily feedback to the patient, caregiver, and clinician through a variety of sensors or device inputs, along with a patient diary that captures the patient’s daily information.
The data gathered might trigger any of a set of alerts to notify the nurse practitioner or clinician when the patient is approaching a preset level, thereby facilitating a patient physical visit, patient telehealth visit, or other appropriate interaction.
Producing actionable information from a variety of data sources requires integration of numerous applications or subsystems. These can span embedded software, medical devices, secure communications, mobile applications, and cloud systems.
As a result, disease diagnostic application developers will require expertise in data integration, interoperability, security, and the ability to find and utilize the islands of data that currently exists in healthcare systems.
Rethinking Screening Programs
The NHS offers a screening program for abdominal aortic aneurysms. Currently, all men aged 65 are invited to undergo a Doppler ultrasound.
(No screening is offered to women due to lower prevalence, and consequently cost effectiveness.)
During the screening, if the aorta measures less than 3cm, no further treatment is carried out. If the aorta is larger than 3cm, further monitoring or treatment is arranged.
Rather than undergo this traditional screening (which many men won’t do), individual assessments of arterial health could be done remotely using peripheral measurements of pressure and flowrate obtained by a wearable device.
In this way, any person (male or female) who is found to be at high risk of having an aneurysm could be invited to undergo the Doppler ultrasound.
This approach would significantly increase the reach of the current screening program.
Among the types of data gathered during clinical trials, data around patient engagement has been impacted with particular strength by digitization in forms such as wearables and social media. These technologies have made it both easier and more enticing for patients to participate in the tracking of their own vitals.
Moreover, the expanding use of wearables and the ease of their use combine to make it more common for patients to engage with clinical trials from their homes. The competition for attention in the digital world is intense, but the higher and more consistent patient engagement is, the more valuable the data will be.
Leveraging digitization in the pursuit of clinical trial data will require crosspollinated expertise in areas such as UIX design, digital health and clinical trial management platforms, and the use of AI and ML to unlock the value in the data gathered.
The digital transformation of the Healthcare industry is generating ever-increasing swathes of data that will ultimately improve diagnosis, treatments, and outcomes.
Currently, the three main sources of this data come from genomic studies, patient data collected by clinical facilities, and mobile/wearable devices. And while data overload is a huge topic in Healthcare, the problem is industry agnostic.
Data management challenges are currently facing all industries and it all centers around data accessibility — how to make data centrally accessible so it can be quickly accessed from a single source and analyzed to produce valuable actionable information.
The ability for enterprises to establish 360-degree views of key business data entities — such as customer, product, account, and transaction — has already become critical to running the business and is among the biggest technology trends shaping (or disrupting) multiple industries.
Enterprises often struggle because they have only fragmented views of their key business data entities due to siloed business units and product lines, mergers and acquisitions, etc.
The ability to achieve a mature, functional, scalable enterprise data platform (wrangling, lakes, serving, etc.) will be a basic requirement for success.
Healthcare companies have an opportunity to realize the new requirements for an enterprise data platform and embrace it along with their other product initiatives.
The data platform needs to be seen and managed as if it were an internal product, which means taking a disciplined approach to product management, architecture, SDLC, DevOps, quality, etc.
It’s not something that “data science folks” do in the back room apart from engineering and product-team processes.
Enterprises that embrace this elevation of enterprise data platforms as products will excel — and enterprises that don’t will fall behind.
The need for enterprise data platforms (centralized or mesh) has spurred a broad spectrum of open-source and commercial technologies that help enterprises accelerate implementation.
These technologies include real-time/batch ingestion/ storage, data pipelines, polyglot storage, catalogs, data mastering, and data warehousing functionalities. Mainstream cloud providers offer these technologies in PaaS form, so the enterprise doesn’t need to self-manage or host them, but the enterprise must still do the work of integrating and customizing on top of these technologies.
Healthcare companies that proactively establish enterprise data platform capabilities will be able to leverage 360-degree views of the data they need to run their business and build new insights and monetization on top of that data.
They can do it all while addressing the data governance concerns — data rights, licensing, privacy, anonymization, and sovereignty — critical for the Healthcare and Life Sciences industries.