Michael Armstrong, CTO of Authenticx, discusses how tapping into dark data â€“ unused, unknown, untapped data generated through interactions with devices and systems across companies enables organizations to mine and identify previously unknown data for critical decision-making.
The healthcare industry currently generates 30%Opens a new window of the world’s data, which is expected to grow by 6% by 2025. While data acceleration is a huge driver for generating insights, decision-makers face a big problem. It’s not that they lack data; there’s too much of it. And about 55%Opens a new window of that data is dark.
Dark data â€” unused, untapped data generated through interactions with devices and systems across companies â€” remains a healthcare organization’s best-untapped resource. This unstructured data from conversations creates valuable insights into trends, patient emotions and sentiments, drivers, brand detractors, and more.
Tapping into dark data sources allows organizations to identify and mine previously unknown data for critical business decision-making. When powered by AI, dark data analytics generates more nuanced, accurate insights and enables IT professionals to provide the healthcare industry â€”Â which relies heavily on data â€”Â to guide a patient experience, product safety, and public health.Â
Dark Data Challenges
The vast majority of dark data is unstructured. And because it isn’t easy to arrange qualitative information according to a predetermined data model, organizations struggle with leveraging audio files, business documents, call records, emails, log files, photos, presentations, raw data, videos, and webpages.Â
Traditional search methods are often ineffective. Yet many organizations lack the tools to collect, analyze and identify useful data from disparate, siloed systems. And if people aren’t sure where the information they need is stored, or it’s not immediately accessible or connected to data warehouses, it sits dormant and unused.
A limited supply of data science expertise makes it difficult to process unstructured data at scale. The best solution for resolving these challenges is to tap into AI and ML capabilities to mine, define and categorize the data.Â
Value of Dark Data to the Healthcare Industry
With all the challenges associated with capturing and analyzing dark data, why use it? Because dark data provides valuable insight into customer pain points along their healthcare journey and uncovers hidden areas for growth and improvement.Â
By offering a treasure trove of insight into patient perceptions of their healthcare providers and pain points along the customer journey, leaders can use dark data to improve their patients’ experiences. This data offers insight into business operations, including how people respond to brand messaging or changing processes. It also helps identify areas of confusion and knowledge gaps. And when healthcare leaders unlock dark data, medical teams gain access to critical insights within complex medical charts and texts to deliver better care and achieve better health outcomes.Â Â Â
Here’s an example of conversational data analysis in action:Â
The challenge: A life sciences company wanted to understand the total volume of calls in which call center agents did not comply with approved talking points.Â
The solution: The company used speech analytics to analyze over 800,000 call interactions. It flagged about 28,000 (3.3%) calls using incorrect language and carried risk implications for noncompliance.Â Â
The result: After receiving confirmation that risk was present across all brands, the life sciences organization updated its call guides to exclude questionable language from the use by call center agents. The organization also held a company-wide training to update and educate employees on the new requirement. Calls were reviewed to verify compliance with the new guidance, and within weeks of the initial and ongoing data analysis, undesirable language and phrase occurrences had decreased significantly.
Clinical trial design
Clinical trials can take years to plan and conduct. Researchers creating and identifying cohorts rely on data to identify medication, medical histories, and symptoms from trial participants. And while many studies require additional qualifiers â€” social indicators or family histories, for example â€”Â they lack a dedicated field for this information, or the data is recorded via less easily-accessed methods, like conversations. Natural language processing (NLP), however, can actualize this data. Processing data more quickly and accurately accelerates the process of finding eligible trial participants.
Other uses for dark data in the medical industry include medical imaging detection and different bioinformatics workflows like MRIs, ultrasounds, and x-rays to identify anomalies, patterns, and trends.
Solutions for Using Dark Data in Healthcare
Sorting and analyzing dark data to extract insights requires a multistep process and toolkit, including data quality management, discovery, and classification.
Implementing a data quality management plan enables organizations to determine how best to clean datasets to gain maximum value and minimize storage costs. Data discovery enables organizations to identify and analyze useful information and gives visibility into their data landscapes. Data classification identifies a dataset’s value, use, and possible security issues. Automated tools and AI algorithms, including NLP and ML, organize vast amounts of dark â€” or unstructured â€” data into relevant categories.Â
NLP is more effective than traditional search methods for quickly and accurately identifying text and insights. Together with AI, healthcare organizations gain new, exciting opportunities to listen directly to unsolicited customer feedback â€”Â and learn from it.Â
AI detects, sorts, identifies, and classifies data patterns and logic into predictions and inferences, empowering healthcare leaders and teams to deliver more personalized healthcare. It identifies topics of interest and concern and has the power to detect meaningful connections and relationships to aggregate and predict outcomes to help organizations:
- Listen to their customers’ voices.
- Achieve context at scale accounting for all customer voices.
- Proactively identify disruptions, recurring patterns, and positive trends.
- Eliminate silos and build cross-functional teams sharing a single truth source.
- Use statistically significant sample sets to inform data-backed decisions.
- Champion and implement personalization and customization options.
- Gain insight into patient behavior and preferences.
Making the most of dark data requires using a platform rooted in a service approach: a platform capable of harnessing and seamlessly integrating data while simultaneously handling continuous data growth and eliminating performance bottlenecks and decline.Â
The technology to process and manage dark data is here and continues to evolve rapidly. AI unlocks actionable insights gleaned from the voice of the customer â€”Â critical unstructured information living in dark data through phone calls, call logs, emails, and text messages. By leveraging the data they already collect, organizations gain a rich source of insights to address relevant business use cases, answer questions, and provide more personalized, customized patient experiences.Â
How can aggregating and activating dark data offer insights to organizations and their customers? Share your thoughts with us on FacebookOpens a new window , TwitterOpens a new window , and LinkedInOpens a new window . We’d love to hear from you!
Image Source: Shutterstock