Fall 2019​

In today’s technology-connected world, things move pretty quickly. But in the world of public health, diseases often move faster than the data that are used to diagnose, treat and prevent illness. On top of reporting delays, public health data are often incomplete or difficult to electronically intermingle with other data to gain insights into pressing public health problems. The solution? Investing in public health infrastructure and resources to rapidly deliver data to public health and clinical decision makers.

By Nancy Maddox, MPH, writer

On September 3, The Washington Post reported that health officials are “sounding the alarm” about rising rates of congenital syphilis—a preventable disease that leads to stillbirth or neonatal death for 40% of affected babies and potential blindness or other serious health consequences for survivors. According to a US Centers for Disease Control and Prevention (CDC) epidemiologist quoted in the article, “There’s a system failure somewhere. It’s either at the healthcare level or public health level, but somehow we are failing these women.”

One problem? Almost all the data cited in the article dates to 2017.

“The diseases are moving faster than the data,” said Michael Iademarco, MD, MPH, repeating a statement common in public health today. He said, “All you have to do is read [CDC’s] Morbidity and Mortality Weekly Report and you’ll see that there’re lots of reports—robust, scientifically sound, public health reports—but they rely on data from two, three or even five years ago. And that’s not quick enough. For something like the opioid crisis, looking at data from even two months ago might not be timely enough.”

On top of reporting delays, public health data are often incomplete or difficult to electronically intermingle with other data to gain insights into pressing public health problems. The Missouri State Public Health Laboratory, for example, has a high-volume HIV/STD specimen submitter that, without changes to its standard electronic test order, is unable to note on its electronic test request form whether or not the patient is pregnant—a piece of missing information with direct relevance for preventing congenital syphilis, as well as perinatal HIV transmission.

In fact, the antiquated state of public health data systems is well known; since 2014 Iademarco, director of CDC’s Center for Surveillance, Epidemiology and Laboratory Services, has been helping to implement a strategy to achieve “newer, faster, smarter, better” public health surveillance.

“We’ve accomplished a lot in the past five years,” he said. “But it’s just a start. ... Congress is looking at this and agrees we need data modernization.”

Just this past spring, Congresswoman Rosa DeLauro (D-CT), chair of the House Appropriations Subcommittee on Labor,  Health and Human Services (HHS), and Education, included $100 million in this year’s House-passed HHS funding bill to support what some are calling a public health data transformation. And in June, Senators Tim Kaine (D-VA), Johnny  Isakson (R-GA) and Angus King (I-NH) introduced the Saving Lives Through Better Data Act (S.1793), authorizing CDC to continue improvements to its own data systems and to fund efforts to modernize public health data systems in state, local, tribal and territorial public health agencies.

In a statement to Lab Matters, DeLauro said, “We must move away from outdated, paper-based [public health] systems to create a true digital network capable of transferring data in real time, which will connect resources to make common investments for the future. CDC currently stores data in 120 silos, limiting data sharing. The enterprise effort in our bill will have benefits across the public health spectrum including foodborne illness, influenza, antibiotic resistance, lead poisoning, opioids and Zika, and will bring CDC and public health into the 21st century with shared data platforms that are interoperable, accessible, and [able to] provide data in a way that supports timely action.”

APHL and partners are pushing for a billion-dollar federal investment—$500 million for CDC and $500 million for CDC grants to US jurisdictions and Tribal nations—to be paid out in $100 million installments over ten years. And a joint statement from APHL, the Council of State and Territorial Epidemiologists (CSTE), Healthcare Information and Management Systems Society (HIMSS), and National Association for Public Health Statistics and Information Systems (NAPHSIS) highlights six areas where this investment is urgently needed:

  • Enhancing the security and analytical capabilities of the National Notifiable Disease Surveillance System (NNDSS), which collects data to inform outbreak response and to create a snapshot of health in the US.
  • Expanding and simplifying electronic case reporting (eCR) from clinicians to public health agencies to enable timely interventions to prevent the spread of “notifiable” diseases and conditions—so-called because clinicians are legally required to notify health authorities about these infectious diseases and conditions of public health concern. (About 75 nationally notifiable diseases and conditions are tracked by NNDSS. Additionally, each state maintains its own list of state notifiable diseases and conditions, which may also be reportable to the patient’s local health agency.)
  • Enhancing syndromic surveillance to provide near real-time data on hospital emergency department visits for ongoing monitoring of community health threats, such as opioid overdoses.
  • Enabling secure, interoperable, real-time reporting and exchange of electronic vital records system data—i.e., birth outcomes and causes of death—to inform public health activities, such congenital syphilis prevention programs.
  • Strengthening laboratory information management systems (LIMS), without which laboratories cannot electronically report data to health authorities and healthcare providers.
  • Developing a public health data science workforce able to maximize the security and utility of public health data, such as the reams of information generated by technologies like whole genome sequencing.

“Why can’t that data flow intelligently?”

One area that has received much attention of late is eCR. Historically, case reporting has relied on a physician or nurse recognizing a notifiable disease, such as tuberculosis or measles, extracting patient data from paper files, and then filling out a form and faxing, mailing or e-mailing the form to health authorities at the federal, state and/or local levels. At the receiving end, public health personnel had to manually enter the data into their notifiable disease databases and follow up with providers to secure missing information—a tedious process with room for error.

Needless to say, many notifiable diseases have been significantly underreported.

Today, with 96% of hospitals and 78% of physicians using electronic health records (EHRs) for their patients, Iademarco asked, “Why can’t that data flow intelligently to the health departments that need it to take action?”

The federal Digital Bridge initiative aims to jumpstart that electronic data flow, beginning with a handful of core  notifiable conditions—chlamydia, gonorrhea, pertussis, Salmonella infection, Zika virus and hepatitis C (an optional add- on)—in seven jurisdictions taking part in a CDC-funded pilot project. Iademarco said, “It’s not just about counting cases, but getting useful information about the cases. ... With interoperable systems, you’re able to focus on data quality.”

Shan He, PhD, a senior medical informaticist at Utah’s largest healthcare system, is among those implementing the project at ground level. Her employer, Intermountain Healthcare, operates 24 hospitals and over 160 clinics in Utah, plus a hospital in Idaho.

She said, “The goal of eCR compared with electronic laboratory reporting [which conveys laboratory test results] is to provide all that additional clinical information, including symptoms and even some treatment initiated by the providers,” to reduce the burden on public health case investigators and Intermountain staff.

Additionally, with eCR, reports for some diseases may be triggered solely by symptoms before laboratory results are available. Pertussis, for example, is reportable in Utah based on the classic disease presentation: coughing for at least two weeks, sleep apnea and patient age under two years old.

In such cases, said Jason Barnes, a senior health informaticist at the Utah Department of Health and He’s Digital Bridge partner, “we’re just relying on the providers to report and hope they remember.” However, since the start of automated eCR, he said, “already we can tell that case reporting for pertussis has increased.”

Timely, accurate data is important, Barnes said, “because if we have a misunderstanding of the [pertussis] disease burden for the state, we won’t be allocating the appropriate resources for public education and vaccinations.” Other pilot conditions might trigger mosquito abatement (Zika virus), food safety investigations (salmonellosis) or contact tracing (chlamydia and gonorrhea, both of which are reportable in Utah based on a test request alone).

The challenge now, said Iademarco, is to scale up from five or six notifiable conditions to 75, from seven pilot projects to the nation and—the biggest challenge— “for all states to [electronically] connect to their healthcare facilities.”

From an informatics standpoint, eCR depends on 1) identifying reportable events, 2) identifying appropriate report recipients and 3) validating, translating and transmitting data messages, as necessary, so they flow seamlessly from EHRs to public health databases. Rather than create an informatics infrastructure from scratch, Intermountain is using the APHL Informatics Messaging Service (AIMS)—a cloud-based system that is infinitely scalable and fully compliant with the Health Insurance Portability and Accountability Act and Federal Information Security Modernization Act.

Dari Shirazi, APHL’s health IT manager, said “The knowledge and skill of the AIMS team is unparalleled in the [health IT] industry. I like to say, ‘AIMS was built by public health professionals for public health professionals.’” A distinct advantage of the APHL platform is that it can reconfigure messages from one transport protocol to another. “You can use whatever you use,” said Shirazi. “So it’s easy for all our partners.”

Among other things, the platform is being used to transmit data on biothreat agents and influenza directly to CDC and to house data on antibiotic resistant bacteria, rabies and emerging infections “so CDC can create customized reports and export the data in any format they want.” AIMS software also analyzes select data in the cloud; for example, assessing flu data for vaccine likeness, genetic novelty and other characteristics.

Intermountain Healthcare employs AIMS for decision support and routing services: “Software on AIMS determines if a case is reportable in Utah and, if so, sends data to the state,” said He. During a two-month evaluation period early this year, AIMS reviewed 26,000 Intermountain electronic case reports and routed about 18,000 to the Utah Department of Health. The healthcare system plans to add additional AIMS functionality to enable eCR outside of Utah, based on the patient’s state of residence and that state’s reporting requirements.

He said, “AIMS reduces technical complexity from the reporter’s side and keeps track of all those cross-jurisdictional reporting rules. We just need to send to AIMS and we’re done.”

Another private sector source of public health data is Quest Diagnostics, a clinical laboratory with service sites across the US, including at select Safeway and Walmart stores. Oriel Hewlett, a senior software engineer for the company, said Quest has been using AIMS for about three years and, during that time, has sent millions of electronic laboratory messages to health agencies nationwide. Quest relies on AIMS to “maintain that working relationship with each jurisdiction” and eliminate the need to reconfigure notifiable disease messages for each recipient, some of whom receive only one or a few messages every year or two.

Said Hewlett, “AIMS is a vital part of our solution.”

Seeking “the perfect world”

Of course, the full public health benefit of a modern health IT infrastructure cannot be realized or sustained without a comparable investment in a public health data science workforce. Keith Higginbotham, IT systems manager at the Alabama Department of Public Health, said, “One of the things most states struggle with is data exchange messaging expertise.” At the moment, he is trying to recruit two business analysts for a syndromic surveillance project and a lower-level IT professional to support the public health laboratory, but has had difficulty finding people with the right skill sets.

Shondra Johnson, PMP, the Missouri State Public Health Laboratory’s facilities and support manager, is part of a three-person informatics team overseeing day-to-day informatics operations. In addition, the Missouri Department of Health and Senior Service—the laboratory’s parent agency—has three integration engine developers who serve the entire department. “When you’re talking millions of messages, that’s not a lot of people to maintain those data-exchange routes or to add anything new onto the system,” Johnson said. “That’s been a very large challenge for us, and it’s not getting better.”

One example of the laboratory’s data challenges is newborn screening (NBS), involving 91,000 specimens/year. Johnson said, “Our larger-volume NBS hospitals keep asking to do electronic data exchange with us, because of the volume. We’ve offered web portal access so they can see the NBS results online, but the results still have to be manually reentered into patient records.” In this case, she noted, “Some of those conditions we’re screening for could be fatal for newborns. Days matter.”

Higginbotham would love to implement electronic NBS data exchange in his state as well. He said, “Right now, if I want to get an electronic NBS order, I have to go to every single one of Alabama’s 50 or so birthing facilities, talk to their IT staff, find out if they’re capable of generating an electronic order for a NBS test, make sure the order that their system can generate makes sense to my lab system and then make sure the results message that my LIMS would generate can be understood and consumed by their EHR system.”

Both Higginbotham and Johnson share the same dream: access to AIMS technology and AIMS data integration specialists to establish and maintain bidirectional data transmission routes with major public health partners.

“If I could just give AIMS our [test] order and results formats and test catalogue and then have them work with all the various EHR vendors and get our messages to work with their systems, that’s the perfect world,” said Higginbotham. “If APHL and AIMS could build interfaces for even the top five EHR systems, I wonder how many hospitals I could quickly onboard.”

Shirazi said APHL has the technical, policy and program management experience to make that dream come true. In fact he said, “This would be an awesome thing to do, if we had the funding to do it.”

To give an idea of the potential staff time AIMS might save, Higginbotham said it took his IT team six months to build an interface with the EHR system used by most Alabama county health department clinics. Yet having established that interface, the public health laboratory now transmits about 1,000 test results to those clinics each day.

Peter Kyriacopoulos, APHL’s public policy director, said the public health data bills now pending on Capitol Hill represent “a very strong expression of congressional support” for public health data modernization. He said, “APHL continues to press the need for Congressional action on data transformation and, on June 28, hosted a Hill briefing on the subject,” with partners, CSTE, NAPHSIS and HIMSS. “The briefing room held 70 people and it was packed—which is a very good sign of the level of interest in this topic.”

Said CDC’s Iademarco, “In the last five years, we’ve developed the strategy [to modernize public health data systems] and the internal leadership to make progress . . . . If we don’t [act now], we’re going to lose our connection to the data. And remember, it is foundational to our public health mission to collect data, analyze data and respond based on that data. If we don’t get in the raging river around this progress, we’re going to be left behind, and that’s going to undermine what public health does. We’ve demonstrated we can do it. But, there’s formidable work ahead of us that needs to be paired with commensurate resources.”