The Indiana Network for Patient Care (INPC), a city-wide clinical informatics network, has the ability to track data from physician offices and medical facilities across Indiana, providing public health officials in the state with early warning of outbreaks of influenza and other communicable diseases. It will also let them know about weather-related health conditions or food-borne illnesses, enabling immediate actions including the alerting of appropriate medical personnel and policymakers.
INPC is among the highest-volume health information exchanges (HIEs) in the United States, according to the Indiana University School of Medicine. Created by the Regenstrief Institute, an international leader in electronic medical records (EMRs) and HIE development and operations, the network handles, on an average day, half a million secure transactions of clinically important data—including medical histories, laboratory test results, medication records, and treatment reports—in a standardized, electronic format. Regenstrief Institute investigators have also developed a system that helps public health officials generate alert messages that can be delivered in a rapid and targeted manner in contrast to the majority of states, which still use ad-hoc paper-based system rather than the electronic exchange of information.
Recently, Healthcare Informatics Assistant Editor Rajiv Leventhal had a chance to speak with Shaun Grannis, M.D., director of the Indiana Center of Excellence in Public Health Informatics at the Regenstrief Institute, about how INPC tracks data and warns officials of outbreaks. The center, established with Centers for Disease Control and Prevention (CDC) funding, builds upon the Indiana Network for Patient Care's ability to securely exchange health information when and where needed.
How does the system work to specifically track diseases?
We use routinely collected registration data (age, gender, reason for visit, etc.) that comes from emergency departments (EDs) and healthcare organizations. So if an ED sees five gastrointestinal (GI) disorder cases in a day, it’s not a big deal. But if it sees 15 cases, that is unusual and it triggers an alert. We had one case when a group of people came into an ED complaining of GI illness—nothing was reported to public health but there was clearly a large spike in the data. It turned out a number of people had been shopping in the same supermarket with poor food handling practices. Our system identified the outbreak. This data is collected at the point of care in real time, received by INPC in real time, and forwarded onto public health for analysis. By aggregating this data and looking at average trends, you can detect unusual events and target them. We are able to assess the whole population using the INPC to feed that information back to clinicians so they can customize the care of individual patients.
Is INPC unique in its ability to track diseases and predict outbreaks?
Our system began operation in 2004, and Indiana is one of the first to develop something like this to scale. There are an increasing number of these systems across country, but ours has one of the most consistently sustained systems, as we have developed a very deliberate strategy to reuse existing data that is already collected instead of asking people to reenter or enter new information, which places a burden on healthcare providers. Many surveillance systems have people separately enter the visit information in a separate system maintained by public health. And with the advent of meaningful use over the last couple years, that has actually steered people into the approach that we have been running for the last eight years— use standard based transactions with routinely collected data to send to public health.
Primary healthcare providers and public health professionals in most other states operate in separate spheres. Why does Indiana integrate the two fields?
Integration makes the process much easier to operationalize and sustain, and frankly, public health and primary care providers have a much better view of what is happening in the community. Public health can inform primary care in many ways, starting with medical decision making. If influenza levels are high and patients are coming in with upper respiratory tract infections, for example, providers might be more likely to attribute those infections to influenza rather than to a bacterial infection. Or if there are diarrhea outbreaks in day care organizations, public health can advise providers through alerts that if you are seeing diahhreal illness in a young person, consider ordering a stool culture. Even further, public health can inform primary care where influenza vaccines can be provided, where their inventory is, and where resources are to receive a vaccination. There are various ways this information can be used.
How do these alerts work?
Public health seeks to retain that communication role. We provide the situational reports to public health and they can provide updates, at the frequency of their desire, on the current status of a disease in the community. Public health is managing that message and process. In addition to the detection side of things, we have built a broadcast messaging system for public health. We have taken our system designed to deliver results to 19,000 physicians in Indiana and adapted it to allow public health to inject and deliver messages seamlessly and directly to all providers in the state.
What are some challenges INPC has faced?
The primary challenge, and I think this is the main challenge in health IT in general, is the constant variation and change in clinical vocabulary. Also, maintaining systems like these requires constant monitoring, and fundamentally, it takes resources to become sustainable. If you don’t build these processes into larger systems where multiple people can benefit from the infrastructure, it’s not clear how you can sustain it. The volume of the data isn’t the issue, but getting knowledge out of it requires semantic interoperability, and that is where the real challenge lies.