Science And Tech At Scale

We are seeing significant gains in the use of technology and advanced analytics around us. The ways we live, connect with one another, and interact with our own bodies have been evolving (and sometimes devolving) through technological innovation. Exciting opportunities are emerging to do things faster and more differently than ever before, including in healthcare and life sciences.

At the same time, we’re not seeing the real-world impact of those opportunities fully achieve their promise. Even with data proliferation and innovation happening at scale, improvements in outcomes — health or economic — lag the rate of innovation. This is due to several factors; high among them are the complex, widely distributed processes that encompass new technology, and the inability of humans at the center of those processes to not “scale” as quickly. To achieve the promise of innovation, the following obligations must be met if organizations are going to compete effectively in the future:

Opportunity 1: Building A Stronger Foundation For Improving Outcomes

With humans creating 2.5 million terabytes of data every day — the equivalent of 1.3 billion hours of your favorite streaming HD video — our lives are more connected than ever. Healthcare has an opportunity — some might say an obligation — to better leverage the deluge of data from inside and outside the industry. Infrastructure advancements, including 5G, are beginning to provide the basis for that opportunity, helping bolster connectivity:

In hospitals, 5G networks are boosting connectivity for point-of-care devices and making it easier to transmit large files. Rural providers can more seamlessly connect with tertiary centers to offer remote monitoring, consultation, and surgical capabilities.
In the home, the use of, and satisfaction with virtual care soared during the early days of the COVID-19 pandemic. While utilization has returned to pre-pandemic levels, virtual care is essential for organizations embracing omnichannel care. With better infrastructure providing more reliable and secure connectivity for more of the country, providers and consumers may be able to collaborate more effectively using virtual mediums, helping telemedicine achieve scale.
On the body, the amount of data captured through remote patient monitoring and Internet-of-Things devices continues to rise. With infrastructure improvements allowing for more timely data transmission and analysis, there is real opportunity to make more reliable predictions on the best near-term interventions. Not necessarily because of more data or better analytics, but because the timeliness of the data makes the prediction timespan shorter and leads to the kind of accuracy that makes early warning systems for acute events more achievable and scalable.

Exhibit 1: Level of satisfaction with virtual care compared with in-person visits

Reality check: Do we have the processes and skills to act on the data we already have?

Organizations that rush into technology deployments like virtual care without commensurate changes to operations run the risk of increased provider and patient friction. Technology innovation must be tightly coupled with process redesign. For instance, implementing virtual visits in a hospital must be done in tandem with changes to scheduling and rooming processes that account for different workflows:

Increased data collection and faster transmission on their own aren’t enough. Clinical staff are asked to act on too many things already and giving them more data won’t help. Interoperability improvements may start to help reconcile data from different devices and sources, but organizations will still need a translation layer of people, process and tech that organizes the data, prioritizes the signals, and presents them to providers, all at pace.
In addition, we should reimagine the last 12 inches of care delivery into something that empowers our health workers to intervene at the speed of technology. They need the social (and legal) permission to give advice based on broader, faster-moving sets of data.
Today, data and insights still only move at the speed of humans. To really see the impact of improved connectivity, we’ll need to change that.

Opportunity 2: Creating New Use Cases For Advanced Analytics And Artificial Intelligence

Advanced analytics, artificial intelligence and machine learning are seeing healthy adoption rates and driving improvement in many sectors — whether it’s applying rule-based automation to underwriting, claims, and product offering recommendations in the insurance industry or retail efforts to boost sales and reducing costs.

Healthcare and life sciences are playing catch-up but there are some exciting use cases, especially within scientific research. Just this year, Google’s DeepMind AI team unveiled (and made freely accessible) their predictions for 214 million protein structures — more than 1,000 times what was determined empirically by the scientific community over the last 60 years. Their analytic approaches may have cracked an enduring problem in computational biology, demonstrating that new tech plus existing clinical knowledge can accelerate future research.

Reality check: Accelerating one part of the value chain often binds constraints elsewhere

The application of advanced analytics, AI and machine learning have proven to be very powerful in research and predictive settings across sectors. For that power to accelerate processes in healthcare and life sciences more holistically, we must also adapt downstream processes that leverage the outputs of those approaches. Similar to the challenges with increased connectivity, putting new analytic tools or insights in the hands of clinicians isn’t the entire answer. Those tools must be factored into their workflow without creating new burdens.

Techniques may also accelerate research and discovery in our decade-long drug development process, but if the clinical trial and regulatory environments don’t evolve accordingly, the overall drug development timelines and costs continue to be rate limiters.

Our industry and society must further adapt its norms and standards to realize the full promise of acceleration.

Opportunity 3: Raising The Floor On Data Starts To Rectify Bias

COVID-19 continues to have a disproportionate impact on Hispanic, Black, and Native American populations in the United States. But one positive that emerged during the pandemic was health systems and insurers using data to target vaccine efforts toward at-risk populations. Health systems nationwide stood up vaccine clinics in underserved communities.
There are pockets of success stories elsewhere across the industry, such as Merck tapping into its databases to understand the prevalence of diabetes among the Hispanic population in a certain market, and then using those insights to create a targeted outreach program. Or the West Side United collaborative whose main goal is to close the life expectancy gap experienced by neighborhoods in Chicago’s west side (as high as 16 years), centering itself on health inequity and using hyper-local data collection and analysis to serve that goal.

Reality check: When data is not representative and inclusive, the analysis and use of this data will be inherently inequitable.

Instead of seeking new heights through innovation, we should first raise the floor when it comes to the data we use:

Acknowledge that any data set will be an incomplete, work-in-progress. For example, those with a historical lack of access or well-founded level of skepticism toward healthcare practitioners will always be underrepresented. Acknowledging isn’t accepting — we should continue to responsibly source and share data, including incorporating broader sets of data (such as SDOH) and prioritizing interoperability to paint a fuller picture. We should also be real about the limitations in fairness of existing data sets and scrutinize any insights drawn from them. For instance, accounting for genomic models that are built on mostly European DNA data.
Build data empathy across the ecosystem of data. Across all who handle the data, recognize inherent bias, create transparency around the collection and use of data and, most importantly, humanize the outcomes we’re seeking.

Raising the floor on data represents a mindset shift and a potential first step to ensure the benefits of tech advancements are distributed more equitably.

The opportunities around more connectivity, faster data, and better analytic techniques are real and becoming table stakes. The harsh reality is that without the harder work of redesigning care processes and empowering the health workers and consumers at the center of them, they won’t make a significant difference in healthcare. And without a laser focus on rooting out the bias in data that underpins it all, we risk letting innovation leave behind the most vulnerable among us. Our obligations are just as important to bear in mind.