Innovating for data-driven decisions in healthcare industry

Healthcare industry is one of the last bastions to fully enter the digital age - Rod Hochman, President and CEO of Providence

Cure and care in the digital age involves a series of choices the industry makes consistently that lead to better or poor healthcare systems. Innovation, especially in healthcare, starts with transformative decision-making. Given the high degree of personalization in healthcare, innovative and informed decision-making in the health care sector would thus involve the three Ds – Diversify, Deconstruct and Digitize. This relies heavily on data.

Let’s begin with collecting data 

Digital innovation in healthcare can create systems that aid faster and reliable decisions. This begins at the source of the data collection process. Many countries employ popular digital mechanisms such as Electronic Health Records (EHRs) or Electronic Medical Records (EMRs) to capture medical information at various sources. Over the years, radiology and laboratory scans have contributed to a large collection of images. In the recent decades, Internet of Things (IoT), Internet of Hospital Things (IoHT) and Internet of Medical Things (IoMT) have penetrated the healthcare space, connecting the patient with the healthcare staff through wearables and similar devices. These connected devices become yet another source of digital information. 

Challenges in the path of digital innovation in healthcare

The data capture mechanisms discussed above form a sure and firm step towards transformative digital innovation for informed decision-making in healthcare. This data, then, needs to be converted into a consumable format so that it becomes medical intelligence. There are many challenges to this. 

• Data in large amounts - Organizations require the ability to crunch large amounts of data at sustainable costs and practical storage capacities. 
• De-identified data - Another barrier towards intelligent use of data is that the data sets are usually available in de-identified forms to follow regulatory compliance. While protecting data privacy is of utmost importance, appropriate techniques will then need to be employed to make this partial or masked data useful.
• Missing data content – Incomplete data brings with it partial knowledge of the real-world situation, resulting in misleading insights.
• Out of context data - Data that could be taken out of context brings issues that need advanced processes so that accurate information is derived out of it. 
• Dealing with variety - Defining meaningful relationships between structured and unstructured data sets is a space for digital innovation. Unknown relationships between available data sets present challenges in terms of missing out on useful insights. There are vast unknowns as of today within the Healthcare data captured within the systems.

So, how do we continue to innovate for better decisions?

Considering these challenges, digital innovation for impactful data-driven decisions in healthcare becomes a highly complex proposition. Healthcare organizations have begun taking two fundamental steps to deal with them. 

They are: 

• Movement towards a secure cloud infrastructure.
• Adoption of advanced analytical processes such as Artificial Intelligence (AI) or Machine Learning (ML) guided by the Quadruple Aim.
  While the former can help in laying down a scalable and affordable digital infrastructure foundation; the latter can help in creation of timely insights. This two-pronged strategy can help deal with the many challenges that were discussed earlier and to steadily make progress in innovating for improved decision-making.

Movement towards a secure cloud infrastructure

Organizations providing healthcare technologies are moving towards building advanced computing platforms that serve as the go-to cloud architecture for many complex workloads. Such platforms become foundational to tackling the issues around large volumes of data that are sourced from a variety of different systems. Reliable de-identification algorithms are applied on data pooled into the platform. Not only do they reuse available methods from marketplace but also utilize logic that are developed inhouse to close on any data security gaps for both structured as well as unstructured data. Building cloud-enabled platform of such a nature offers immense potential to support advanced research and intelligent digital services in healthcare.  

Adoption of advanced analytical processes such as Artificial Intelligence (AI) or Machine Learning (ML) guided by the Quadruple Aim

For better decision-making it is pivotal to handle the challenges posed by missing data, ambiguous or unknown data relationships as well as data received out of context from variety of data sources. Based on techniques of Natural Language Processing, Representational Learning, Machine Learning and Deep Learning, adoption of Artificial Intelligence promises to deal with these issues and offers supplemental information through digital systems to enable faster and better decisions from humans. The insights developed through these processes can be weaved into system-level touchpoints. Both the development as well as seamless integration of insights and recommendations at appropriate decision-making moments for patients and caregivers during their day-to-day activities will be a point of differentiation for healthcare systems. As an example, understanding a system predicted value for the duration of hospital stay for a patient at the point of arrival in a care facility – allows decision makers to detect needs early and assign additional nursing care or critical support in a timely manner. Another example is to predict the overall expense that a facility is likely to incur by the end of the month. This system prediction at the beginning of the month, allows room for better planning and risk management. Inter-weaving AI into decision-making points in the lives of users is a space where vast innovation will be seen in the forthcoming times for healthcare.

Apart from the technical know-how the Quadruple Aim needs to be used as the strategic compass that guides the development of AI processes. It is an all-inclusive way to optimize health system performance.  The involvement and contribution of healthcare experts as well as end users (both patients and caregivers) acts as a key parameter for the successful implementation of AI solutions and to overcome many of the challenges posed by incomplete and ambiguous data. It leads to the understanding that merging human expertise with digital technology makes room for powerful solutions. These solutions are to be responsibly applied for better patient and caregiver experiences as well as improved healthcare outcomes at affordable costs. 

In conclusion, the journey of innovation for data-driven decisions in healthcare is a long one. It comes with many challenges and learnings. It demands responsibility and patience. It is however also a rewarding and purpose-driven journey making it easier on the effort and worth the wait.

The article has been written by Preetha Kumar, Director – Data Analytics, Providence India