While the IC’s research organization looks into adding security to cloud environments, in the here and now, intelligence agencies are sharing more data.
The federal government has operated the data.gov website since 2009 as a repository for data generated by federal agencies. The website lets users find data, tools, and resources to conduct research, develop web and mobile applications, design data visualizations and more.
Yet data analytics is not often at the forefront of discussions around federal IT, at a time when cybersecurity, data breaches and cloud adoption are grabbing headlines. However, analytics can enhance government services and, depending on the agency gathering insights from data, save lives.
Recently, FedTech spoke with Steve Bennett, director of the global government practice at SAS. Bennett is the former director of the National Biosurveillance Integration Center within the Department of Homeland Security. He is a strong proponent of the use of analytics in government to effect positive change, and he discussed how analytics is being used in public health and how it influences information sharing among government agencies.
BENNETT: Advanced analytics are providing government health agencies with new access to integrated and reliable population health data that can greatly improve their ability to manage public health issues. Analytics provides information agencies can act on immediately to mitigate public health threats, such as infectious disease or the opioid epidemic.
Let’s look at the latter. Combating the opioid epidemic requires a rapid and multipronged approach, reaching beyond the capabilities of any individual government agency or mission program. However, traditionally, information sharing between agencies and on-the-ground public health workers has been a challenge, for many reasons.
First, the Health Insurance Portability and Accountability Act prevents medical professionals from sharing patient records with epidemiologists and on-the-ground public health organizations. Another reason is that reliable research is expensive and time consuming. Additionally, publicly available population health information is complex and fragmented. Many opioid users are unable or unwilling to come forward, and can be unreliable and difficult to locate. Lastly, the policy changes that could mitigate some of these challenges happen gradually.
Ideally, we could assemble fragmented information, by integrating and cleansing publicly available health data from a cross-section of organizations that collect and store it. Advanced analytics then could enable analysts to better predict at-risk populations and discover insights into the health and behavioral patterns of opioid abusers. Resources could be redirected from time-consuming manual data processing toward the development of a holistic strategy to combat the opioid crisis. This is just a single scenario where advanced analytics might be applied to improve the public health environment.
BENNETT: Traditional approaches for epidemiological surveillance of population health were not designed to fully exploit today’s explosion of data clues created by an infectious disease event. Such an event creates a Big Data cauldron of both meaningful and meaningless data that varies in complexity and substance. This includes any and all information collected from local, federal and international health agencies, the public health and medical communities, the academic research community, pharmaceutical vaccine makers and even “nontraditional” health data sources, such as social media.
Advanced analytics are designed to rapidly integrate and cleanse disparate data sets, spotlight previously unseen correlations and elevate the most meaningful insights. Agencies can better characterize an infectious disease event and streamline the generation of more comprehensive and accurate data analyses. Laying the foundation for a centralized public health information infrastructure, advanced analytics will help detect early warning signs of infectious disease events more quickly, support deployment of a timely and integrated on-the-ground response strategy and generally enable faster, better decision-making.
BENNETT: Absolutely. Advanced analytics can help provide early warning and situational awareness by leading analysts down a previously unseen trail of breadcrumbs, spotlighting patterns and anomalies along the way. It enables them to detect and characterize anomalies much earlier than traditional methods, which could significantly improve the nature and speed of response.
It’s especially exciting to consider how advanced analytics could completely overhaul the way we track, understand and communicate emerging infectious diseases, like Zika virus, in regions that lack well-developed public health infrastructure.
BENNETT: Individual government entities collect all kinds of information that may seem useless on its own, but can become meaningful alongside other data. Unfortunately, the sheer size of government, combined with stove-piped communications and behaviors can hinder cross-governmental information sharing.
Advanced analytics, however, can help bridge the knowledge gap between government entities, by integrating information automatically, selecting the most important correlations and connections, in real time, as new information is acquired. Now, an individual organization or program can access, augment and draw new insights from streaming interagency data, which has been traditionally difficult to analyze, and equip themselves with the most up-to-date, comprehensive and actionable information possible.
Consider how first responders could leverage real-time, streaming data around everything from drug trafficking routes to regional ambulance calls related to overdoses. They could engage in better resource deployment, such as equipping themselves with naloxone, which can reverse the effects of an overdose. What if developing countries could access a birds-eye view of an infectious disease outbreak? Not only could they get ahead of the disease, but they could also use that information to help populations better understand the severity of the outbreak and prepare effectively.
BENNETT: Many people around the world today have greater access to the internet than to a hospital. While a difficult data set to analyze, social media nevertheless provides an additional source of data clues to consider alongside information that’s more traditionally collected.
Combined with other information, social media can be a resource for understanding disease dynamics in global populations, such as rate of spread, or emergence of new symptoms. For instance, perhaps a handful of medical reports from a rural hospital detail new cases of pneumonia, which may appear isolated at first glance. By identifying geo-tagged Tweets that detail similar cases in the area, the rural hospital will have additional insights to determine how widespread an issue this could be. Advanced analytics connects these seemingly separate signals to paint a much more comprehensive picture of the issue at hand.
BENNETT: Yes, I think in some instances that is true, especially in the intelligence community or data-rich agencies like the Department of Commerce and NASA. These types of organizations base their missions on data, and that data needs to be boiled down — or analyzed — into actionable information.
That said, government still has a long way to go. The use of advanced analytics for public health initiatives is a great example of a domain that could reap huge benefits by adopting these capabilities. Envision an agency being able to predict where the next outbreak of a crippling virus will occur and being able to prepare the facilities in those areas with targeted information – or better yet, an increased supply of treatments or vaccines — to get out in front of the epidemic.
BENNETT: The use of analytics within government will only become more refined and effective, as the barriers to adoption crumble. Thankfully, we’re starting to see cultural changes occurring across agencies, to the point where interest in, and desire for, analytics and the associated toolsets are becoming more “mainstream.” This will continue.
I think agencies will begin to also understand that it’s not just about collecting more data, but utilizing the data they already have in more meaningful ways. The first step will then be to put a more focused effort on data digitization, management and cleansing. You can’t perform meaningful analyses on bad data that’s still on paper in a filing cabinet, or spread over 1,000 spreadsheets on employees’ hard drives.
Public health agencies will continue to integrate visual analytics that graphically depict the data in a way that can be easily understood by anyone, not just by data scientists and experts. By nature, humans look for visual patterns of data to glean intelligence and make sense out of the information. This democratization of analytics through sound effective data visualization will drive wider adoption.
However, the future is not without its challenges. While cloud technologies have made it a bit easier in the past few years, it’s still very hard to get anything with an “on” switch procured and installed in a government environment. Once installed, it can be even harder to keep up with the technology as it matures and improves. These challenges, while not insurmountable, are a remaining obstacle to seeing some of the benefits that advanced analytics can bring to government and public health.
Going forward, the adoption of more advanced capabilities, such as machine learning and powerful predictive analytics, will help automate a lot of the analysis process in coming years. I am optimistic that thanks to continuing advances in analytics, our best people will be able to spend more of their valuable time fighting illness and disease outbreaks, limiting their spread earlier in the disease cycle, and reducing the number of people affected.