COCA Conference Call –Update on BioSense

NOTE: This document is provided for historical purposes only and may not provide our most accurate and up-to-date information. The most current Clinician's information can be found on the Clinician Home Page.

(September 26, 2006)

Clinician Outreach and Communication Activity (COCA) Conference Call Update on BioSense
September 26, 2006
Lynn Steele, MS, CIC

Diana Hadzibegovic:Thank you, Laurie. Good afternoon. Thank you for joining us for today’s Clinician Outreach and Communication Activity conference call. Today we will be talking about BioSense.

Objectives for today’s call: understand the vision and approach for building BioSense; be familiar with the 2006 implementation goals and progress toward those goals; have increased awareness of the evaluation plans and communication channels now in place.

We are pleased to have Lynn Steele on the call today to discuss these objectives. Lynn Steele is the director of the Division of Emergency Preparedness and Response in the National Center for Public Health Informatics at the Centers for Disease Control and Prevention. In her current position as division director, she is also the BioSense program lead. We are pleased to have Ms. Steele today.

Lynn, you may begin.

Lynn Steele: Thank you, and thank you everyone for taking the time to be on this call.

I spoke to this group about a year ago. It was I think in August or September, as we were just beginning to talk about BioSense and the ability to link hospital data or healthcare data in real time. So I'm happy to talk to you about our progress to date and where we’re going. And probably the most useful part of this for you will be to answer your questions.

But just to talk a bit about BioSense which is really CDC and HHS’s attempt to develop a national biosurveillance capacity.

We’re increasingly using the term biosurveillance as it defines traditional health surveillance but with a timeliness aspect -- so the ability to exchange information and have information available in real-time for a number of purposes: to have a platform for situational awareness, health situational awareness, in this capacity; also to help understand and work on an event early, so early event detection or early event management.

There has been some use of these types of systems or some attempts to use computerized data for initial event detection or to identify increases in specific syndromes, syndromic surveillance. While those have promise, initial event detection and syndromic surveillance are not the primary purpose for connecting health data for the use by public health.

So we have a lot of work to do to clarify the terminology and to make sure we’re all on the same page about what such a system and the capability should be able to do in the early part of what might be an unfolding public health event. And we’ll be talking more about that when I talk about how this fits into the work of the American Health Information Community.

So some of what we've learned over the year -- as you know, we connected real-time hospital data from at least 10 cities by the end of 2005. And we began displaying that data in the early part of this year and we continue to look at the best ways to display clinical data in real time from these initial hospitals.

What we've learned from some of the initial users of this real-time data is that we need to really continue to look at this as a beta development phase, and just to be very clear, that we very much consider both the BioSense program and the application to be in development. And we’re working with the initial users, both healthcare users and public health users, to help firm up our plans for the development. And that’s really an important process, and how we get input is critical to the success of this program.

We also have learned that we need to be clear about BioSense’s mission and vision and our approach to achieving that vision. We have a lot of intense work that we’re doing this fall to move toward that clarification of the vision and mission. And we’ll talk a little bit more about evaluation. Evaluation is really what we need to serve and refine the BioSense program in this application.

We've also heard that local communities and local public health departments who have their own syndromic surveillance systems or their own capability for linking hospitals and public health in place are struggling with how to fit BioSense into their overall strategy for local surveillance. And this continues to be a challenge.

We have increasingly over this year adopted an approach to connect not just hospitals to BioSense or to CDC for display of data by state and local public health, but also to adopt an approach to connect a system of systems, so recognizing there are some good syndromic surveillance or public health systems and relationships that are already in existence. And we should capitalize on that from a national perspective.

One of the things that’s critically important that we heard is that BioSense and CDC has the potential to acquire national data and additional sources of national data that might not be available locally. And that is the priority for us -- to get and distribute those data, for state and local public health use where it really only makes the most sense for CDC to get and distribute that data, so that they don’t have to be distributed 50 times. And we’ll talk about that a little bit more and where we are in the progress of those national data sources.

Understanding an event on a large scale is really the utmost important public health.

So probably the most important concept of BioSense and its implementation and perhaps one of the things that’s been most misunderstood is that BioSense provides the ability to have cross-jurisdictional view on what’s going on. So the ability to have a single place that can look across geographic boundaries is the most important concept, that the data are available to local, state, and federal public health as well as healthcare officials at the same time.

From a hospital user’s perspective, we’ve realized acutely from our users that it’s the infection control and hospital epidemiology communities who are most engaged in the use of these systems for their intersection with state and local public health.

The most important capability for hospitals seems to be the ability to connect with public health in order to have the same information that public health has in order to be able to interpret what might be going on in the community with public health in real time.

So one of the things that we’ve heard clearly from the hospitals who are some of our initial participants is that they want to not only see their own data, and as you know, BioSense allows jurisdiction-specific right into what data can be viewed by what user. So currently, a hospital could only view the data that has been submitted to the BioSense program from their hospital or from their hospital system.

But what we've been hearing is that hospital officials, epidemiologists want to see how their data fits into a community picture. So we need to develop a plan for providing that kind of aggregate view of the community back to the hospital users.

So those are some of the things that we've heard most loudly from our initial users from BioSense.

So to tell you where we are on the progress - oh, one more thing I wanted to mention is, I said that in order to connect more hospitals more quickly, we are also looking at this system of systems so that we’re connecting more hospitals who are already connected to state and local public health from an existing system. So, some state and local public health have ESSENCE or RODS or a proprietary system that the state public health agency has developed.

And we believe being able to share that data as part of the BioSense application, getting basic data such as foundational data, chief complaint data, (ICD)-coded diagnoses, those are the foundations that most of the systems are built on - are important to framing that national picture of situational awareness.

So we’re currently working with some health departments across the country to see about getting their existing systems data into similar standards as the other BioSense data. We’re working with Indiana, North Carolina, Ohio, and Seattle.

These efforts continue to be accomplished under the recruitment leadership of one of our contractors, the Constella Group led by Wayne Myers. So if those names are familiar to any of you.

We’re interested in working with other systems that can be implemented more quickly.

There’s some urgency to get at least some clinical data visible from as much as the country as possible as move towards our influenza preparedness activities at HHS and CDC.

So just to tell you again where we are today.

We have 40 hospitals who are currently, as of today, providing real-time data to CDC through the application.

We have another over 700 hospitals who are in some form of implementation. And our goal is to have 350 hospitals with data visible through the application by the end of 2006, and we’re well on track to meeting that goal.

We’ve worked hard to get hospitals who are part of large healthcare systems this year to participate in BioSense. Many of them are going through technical assessments or some form of the implementation process right now.

Beyond the end of December 2006, CDC has described a stretched goal, and that is to have 3,500 hospitals connected in some way to this national biosurveillance picture by the end of 2009. So that would mean our target goals for the end of 2007 are about 1,450.

We developed a somewhat mixed model of the types of data and the depths of data that we’ll be receiving in BioSense. We are very committed to having data that includes those data most useful to public health such as microbiology results and other clinical data. And we’ll be working toward implementing that type of data and doing the developmental work for proof of concepts on how those different types of data will and are useful to public health.

We’ve also made it a priority to get real-time data from our long-term partners in BioSense and that is the Veterans’ Administration and Department Of Defense healthcare facilities. They’ve been reporting data to BioSense for about three years in the national picture and want to move toward a more real-time format in order to get information on the treatment of VA and DOD patients in real time.

And we’ll be moving their data into the format to make it comfortable to the real-time hospital data by the end of the year and moving toward getting more real-time data from both of our federal partners over the coming year. Currently we receive coded diagnostic data and procedure code data from those facilities, and we’ll continue to do that.

And we’re doing this - a similar initiative for large commercial laboratory data. Our goal this year is to begin connecting three of the largest commercial laboratories by the end of the year. We’re in good negotiation with two of them, LabCorp has been a partner of CDC’s for a long time, and we’ll have test orders and results as part of the new BioSense application. We’ve also been working with some of the other large commercial laboratories.

The value there to local communities is that CDC can receive the data from these large commercial laboratories and make the population affected or the citizens affected, their information available to the relevant local public health jurisdictions.

As you know, many of the commercial laboratory data are not data from hospital labs, are not data that goes through - for hospitalized patients, but rather it would be our first foray into getting information on relevant microbiology results from an ambulatory care setting made available to local communities.

So that’s one of those national data sources that makes sense for CDC to get and distribute for state and local public health.

As many of you know, we've also been working with Poison Control Center data, and the ability to display that data over the coming months through the BioSense application, through the work and collaboration with the American Association of Poison Control Centers. And we’re on the right timeline to be able to do that.

As far as other national data sources, we’re working on pilots to be able to look at over-the-counter drug sale data that’s been aggregated by the University of Pittsburgh and the National Retail Data Mart to pilot that and make that available through this national application for state and local public health.

Another real-time data source that we’re looking to pilot for usefulness by local public health is to look at pharmaceutical claims or prescriptions filled. In Atlanta is a company who processes about 75% of prescriptions, so to look at very specific claims that we think would be useful for public health such as antimicrobials and antivirals.

We’re also working with the American College of Emergency Physicians in a capability to be able to look at hospital utilization through Web-based systems such as those that are implementing the NEPN standards or HAVE standards, so those that are reporting emergency department storage capacity.

Currently, BioSense gets information from hospital IT systems on census data, but it’s a limited data set that we are able to get. And there are some push-to-capture data that is useful locally and used locally to help describe diversion and other capabilities that this data currently reside in Web-based data entry system and should be available for display alongside other relevant health information, and again, available for local public health use.

We need to have further discussions with our other federal partners at HHS and HRQ and HRSA around the ability to aggregate this data and make it available alongside the BioSense data. But, you know, doing that in a way that is contributing to the efforts of these other federal partners who also have responsibility for tracking these resources as part of emergency response.

So we’re working to try to be a partner in that area but also to provide those data alongside, as I said, the other biosurveillance data for state and local public health use.

So I will answer more questions about specifically our progress when we get to the question and answer. I want to talk to you briefly about our evaluation -- it’s one of the things that we’ve heard most clearly, that we need to make sure that this is a good federal investment, that we’re able to make the case for the usefulness and the approach of what it is we’re doing.

So we’re doing a number of things to evaluate this program in our approach. First of all, we wanted to make sure that we’re applying industry best practices in how we’re developing the technology around BioSense and the operations of the program.

The Gartner Group, is an independent IT consulting firm, has been working with us. They did a 30-day assessment and are helping us to implement some changes that based - as necessary. So, some of these quick hits that they identified are - they’re helping us to improve the program especially as related to our practices - business practices as well as assessing the technical environment.

We have also recently funded, awarded cooperative agreements to four universities to help us look at various aspects of evaluation related to BioSense including some data validation, utility, usability, and cost effectiveness.

The Office of Management and Budget as part of the federal government performance assessment activities or the part has really described a way that forces us to evaluate BioSense not just related to who is using the system or how many people are using this system or how connected nationally we are as this biosurveillance capability, but rather to really look at how such a national asset will improve how we practice public health.

So specifically, what we’ll be evaluating is how timeliness of public health response changes as a result of having a real-time clinical data system to begin investigations to rule out whether an investigation is needed. So we will be looking to make that part of our evaluation framework through this cooperative agreement. And I will be happy to answer more questions about our evaluation.

The last thing I wanted to talk about is just how all of this fits into the American Health Information Community. As many of you know, the American Health Information Community or AHIC is really the group that was charged by the President to advance his or the nation’s health IT agenda.

It’s being convened by the Secretary of Health and Human Services and the Office of the National Coordinator of Health Information Technology at HHS, who have been working to come up with both the longer-term five to ten year plan for advancing health IT as well as some very short term working groups, activities that can make progress in specific areas within the next year or two.

So one of those areas that the AHIC has been focused on since last fall is biosurveillance. And the biosurveillance workgroup’s specific charge has to then to make recommendations so that within a year, essential ambulatory emergency department visit utilization and laboratory results data from electronically-enabled health care delivery can be transmitted and standardized to authorize public health agencies within 24 hours.

So that is the very specific charge. The workgroup made recommendations to the secretary at the end of May suggesting that there needs to be additional discussion in regards to a minimum dataset for biosurveillance.

So BioSense is very closely watching the recommendations and participating in the recommendations development around what should comprise, what are those data elements that are absolutely necessary for informing event detection and management, situational awareness, outbreak and response management, et cetera.

The steering group that’s looking at this minimum dataset will look at the recommendations of the biosurveillance work from the past eight months and come up with recommendations that will be sent to the health information technical standards panel to come up with the national recommendations for how such data need to be standardized.

So all of these efforts, BioSense as well as the AHIC work, is moving us toward some standard infrastructure, standard vocabulary, and messaging. So we think BioSense is in the right alignment with all of that because BioSense actually had to produce a data standard as we began implementation last year.

In some ways, this project and this program have been helping to drive those standards, helping the IT standards panel and others to recognize where there are gaps and standards, and to make recommendations for how industry may fill those gaps in health IT standards.

So again, we are following the work of AHIC closely. We’re looking for those recommendations for moving forward. I’m one of the representatives on the biosurveillance data steering group and there’s been very good discussions about what are the right amounts of data for enhancing this capability of sharing clinical information with public health.

Those are the things that I just thought I would brief you on. And I think I’ll turn it back over to Diana and open up for questions.

Diana Hadzibegovic:  Thank you Ms. Steele. Great set of information.

Laurie, we can start Q&A session.

 

Question: I realize you’re not getting into the syndromic surveillance but I’m just curious whether there’ll be another call on what the plan for eventually incorporating syndromic surveillance might be for this. I know CDC has some projects going on with different beta test. I think the US-Mexico border project has one product in test mode.

Lynn Steele: I didn’t mean to imply that we’re not engaged in syndromic surveillance. Certainly we have an obligation to and we need to come up with the best methods to look at all of this data that we’re aggregating. So we are looking and continuing to use both chief complaint and diagnostic data to analyze data by putting it into the 11 syndrome categories that have been developed by CDC and others.

We’re further looking at how to make those syndromes more refined by taking the 11 syndromes and making them more granular. We now are working with about 78 what we’re calling subsyndromes that would allow you not just to look at a syndrome that was related to gastrointestinal illness but would get down to the granularity of diarrhea or bloody diarrhea, those things that certainly we’ve tested and used over the past two weeks with the E. coli outbreak.

So, yes, we are working and continue to try to advance the science, the best way to group data into syndrome categories. But in many cases, we believe that the primary value is when you know something is going on to be able to build a flexible interface that can query the data, look for additional cases, not just to look for increases in syndrome categories that would reflect the statistical anomaly or statistical increase. We still have a lot of work to do in that area.

Q (cont): Thanks.

 

Question: Lynn, my question relates to how the activities that are funded out of the coordinating office on Terrorism Emergency Preparedness, particularly with respect to the surveillance capacity development that BT and emergency preparedness grants, states that are supporting -- how is that being coordinated with the development of BioSense you alluded to, other syndromic surveillance systems that they’ve already put in place, and much of that initial and ongoing development is funded out of that BT cooperative agreement.

Lynn Steele: Right. Well, you know, we’re funded out of the office of Terrorism Preparedness and Emergency Response as well. They manage all of the terrorism allocation that comes to CDC about 1.8 billion or so, a billion - almost a billion of which goes out to state in local and public health.

So we do work closely with that coordinating office as the primary funder of all of these initiatives. And it’s really their leadership that tries to help us to make sure that we’re coordinating and collaborating for the best use of these resources.

We have been working with them in order to understand who - how state and local health departments are using those funds to contribute to their PHIN, their Public Health Information Network preparedness activities. And we’ll continue to ensure that we’re providing similar guidance in how to direct those funds for individual early event detection capabilities or syndromic surveillance capabilities and how that fits into this national biosurveillance picture.

So I’m not sure if I specifically answered your question. You know, there is a recommendation to ensure that there’s interoperability of IT systems across the preparedness work of all state and local public health. I mean, that is really the framework of the Public Health Information Network, to ensure the capabilities for how you communicate and alert during an emergency is part of that, how we are able to manage and track people who receive counter-measures as part of an emergency response, and this capability for early event detection or health situational awareness is also part of that.

So it’s all part of ensuring that we are able, you know, interoperable capacity to have these same functions, how to make sure that we’re best using the resources that come directly to CDC for BioSense versus those that go to the states I think is something that we continue to work on and are looking to the coordinating office for terrorism preparedness to help us with that.

Did that answer your question?

Q (cont): Yeah, that helps. Thank you.

 

Question: I’m calling from the Office of the Surgeon-General in Rockville, Maryland. I was curious how you are dealing with the privacy issues and those sorts of related issues with information that’s, you know, you’re looking to come from ambulatory care centers and that sort of thing.

Lynn Steele: That’s an excellent question.

You know, we’re very sensitive to the privacy issues. Certainly every - we have data - CDC has data sharing agreements with all of the healthcare institutions that - for whom - with whom we’re sharing data that describes how those data will be used -- broadly, for public health purposes.

We are very clear in the use of those data specifically for public health. And as you know, public health has an exclusion under HIPAA in that data can be shared with public health where there may be need for public health intervention. So those data sharing agreements make it clear that the patient level data will not be shared with entities other than public health.

We also are very keyed-in to the sensitivities about identifiers with those data, so you may know that the BioSense data does not contain obvious patient identifiers.

There’s a randomized linker number that helps us to track a single patient over time within a single healthcare institution or healthcare entity that linking back to the actual patient or the patient can only be done by local public health with the collaboration of the hospital.

So it’s not completely the identified data, of course, because we would - we need to have the capability to analyze data geographically and receive ZIP code data. But we are very sensitive about ensuring that not all of the data that we receive are linked to a patient name.

So we have a lot of work to do still to address the privacy issues. We were ready to convene and begin talking with that community but have been really directed by the AHIC who is now going to be convening a workgroup specifically on privacy and confidentiality.

Because the issues around patient privacy have not just been raised from the biosurveillance work but also from the electronic health record activities which is really looking to advance the exchange of laboratory result data among physicians and the consumer empowerment group which is really looking at how you would reproduce especially a patient pharmaceutical record in event of a loss like we saw with Hurricane Katrina.

So all of those activities and how we share -- more broadly share -- electronic health information require that we really take a national look at these concerns and how to best address the privacy and confidentiality issues.

Q (cont): Well, thank you.

Lynn Steele: But just - I want to be really clear because I get this question a lot.

CDC takes very seriously our responsibility to look at this data and to make date available for state and local public health use for the need for a public health intervention.

We are also very clear that we want to be able to share some information in order to form a national picture through the National Bio - the NBIS, the national integration system that the Department of Homeland Security is developing. But we will not be sharing patient level data nor would Homeland Security or others expect that level of data sharing.

So we will - we believe it’s important to be able to share human health data or unusual findings of biosurveillance with our other federal partners, but it won’t be at the level of individual patient data and everyone is in agreement with that.

Q (cont): So if a finding was referred back to a local public health agency, they would simply follow their own rules and regulations and not deal with it?

Lynn Steele: Correct. I mean, it’s not so dissimilar to what happens now.

In some jurisdictions, the hospital would fully expect local public health to contact them to get additional information on the case that might need investigation or about additional patients.

In some cases, local public health and hospitals or healthcare facilities have a good relationship in that they know and trust that local public health are only going to re-link the patients where there’s a compelling reason to do so, and so have made that link available to local public health. But those are, as you said, those are local agreements and local regulations.

Q (cont): Thank you.

Lynn Steele: You’re welcome.

Question: About a year ago, the program goals were to connect up to four data sources in each of the 31 bio watch cities I think by the end of 2007 for about a total of 120 data sources. And today’s goals are 350 by the end of this year and 3,500 by the end of 2009. So I have two questions.

One is, what principally drove the significant change in scope? And two, where do the bio watch city coverage perspectives figure into your plans.

Lynn Steele: Those are good questions.

And in some ways, the - I'm not sure the program goals have changed that much; they’ve just been clarified. So, first, in regards to the bio watch cities.

There was some interest in bio watch being - having the capability to look for environmental - potential environmental exposures to Category 1, Category 2 terrorism agents. There was a lot of concern that if environmental signals wouldn’t go off, that we had a real-time human health data source to look for clinical cases of disease.

So there was some desire to link those cities first that had bio watch capabilities to have the ability to look for human health data in real time.

That’s still a goal obviously there is the bio watch cities are those most populated metropolitan areas in the country, so especially as we move the VA and the DOD data to real time, we will have at least a sentinel hospital or more in each of the bio watch cities. We’re very close to having at least one health - real-time health data source in all of those cities. So that really hasn’t changed.

I think you said something else about four or five of the facilities in bio watch cities. The way we targeted our recruitment efforts last year -- if we really believe that we want to have a sentinel hospital in each of those cities, we would look for one of the hospitals that has the top - one of the top five emergency department volumes for that city. So that if there were something going on, you’d be more likely to pick up cases of human illness by having one of the top receivers of healthcare in those population centers. So I’m not sure that that has changed.

As far as the original contract that talks about four data types from - I can’t remember what it was. Can you repeat that -- 120…

Q (cont): Yes, I think it was up to four data sources in each of the 31 bio watch cities…

Lynn Steele: Right.

Q (cont): …that comes about 120 I think.

Lynn Steele: Right. And, you know, the definition of a data source was not really clear. So we’ve been trying to look at both healthcare facilities and hospitals. When we talk about having 350 hospitals by the end of the year, we’re really only talking about probably about 50 connections because many of those hospitals are linked because they’re part of a large healthcare systems or now they’re already part of a large conglomerate or a large biosurveillance - date-based bio surveillance or syndromic surveillance system.

So we really haven’t changed the scope. We’ve looked at the number of individual hospitals that we get information from, but the number of connections still remains to be about 50.

Q (cont): Okay. I guess the essence of my question is, at the end of the day, we really need to have broad and deep coverage throughout the country, but in the near term which to your mind is more important, the broadness of coverage to get, you know, as much as geographic reach as we can or going after the aggregate databases and getting as much data as we can?

Lynn Steele: Well, it’s really that we’ve been told and I think we would all believe I think on this phone that we want broad coverage. We’re giving up a little bit of the deep coverage now to get a broader picture of what might be going on, especially as it would relate to influenza-like illness related to a flu pandemic.

So to be able to have broad coverage to get a little bit of data from a lot of hospitals in order to be able to look at hospitalizations, look at clinical laboratory data that might be related to - for influenza, et cetera is more important at this point than getting deeper clinical data.

Some of those data types for which we’re still having to make the case for their - how they will be most useful, how they will be analyzable for public health.

Q (cont): Thank you.

Lynn Steele: So, yes, in this development of this broad capability or national capability for biosurveillance, we’re trying to do a little bit of both, to have enough hospitals where we have deep clinical data that we can really learn how to use that data, but to in the interim have a little bit of clinical data but from a lot of coverage. And when we talk about 3,500 hospitals by the end of 2009, that’s what we’re referring to.

Q (cont): Thanks.

 

Diana Hadzibegovic:  Thank you, Lynn; this was a great presentation and we were glad to have you again on our COCA conference call.

I would like to remind any listeners who did not have a chance to ask a question, to email the question to coca@cdc.gov. Please visit our COCA Web site for more information for clinicians.

Thank you and stay tuned for our next COCA conference call. Good bye.