Webinar 5 Pre-work.docx (42.9 KB)

1. We have a data system for linkage to HIV care and data to care activities but the system wasn’t developed to collect all variables needed to evaluate these programs. The current system cannot accommodate the changes we require. This has resulted in using the system for some variable collection in addition to collecting variables in an access database. We are now working to create a new database where these variables can be collected and analyzed from the same site. 
2. Across many HIV systems there are multiple “test dates” for labs and test results. Each system defines these dates as something different. For example, HIV Surveillance would define the test date as the specimen collection date whereas other programs would use the date the result was received. 

1) At our organization, there are multiple electronic health records (EHRs) systems being utilized across medical/behavioral healthcare, social services, and public health divisions. The EHRs lack interoperability, limiting the ability to access the patient’s care throughout the care continuum and providing public health surveillance for the community. Therefore, our organization implemented the Health Information Exchange to consolidate patient information into a single platform. Viewing a patient’s information in one platform improves patient care by reducing redundancies in medical care, improves communication and coordination of care, and helps the organization understand its population from a public health perspective. 
2) As we have been working on revising our daily situational awareness report, we have encountered several different definitions of 'extreme heat' - with slight, but important, differences among how it is defined by the local shelter network, our health officer, and various county and city emergency organizations.

1. At Minnesota, we utilize a couple of flat files for our hospitals to report cases. Each one of these was built differently, based on what the partner indicated they could and could not do. Because of this, we are unable to build one route for each of these different files, which is costly and time consuming even though they have similar information. It also makes it difficult on end users to understand why something is or is not reported in a certain way based on who is reporting. To compensate, work on building one specification began, but what one group finds as satisfactory may not for another.
2. There are two areas where we've experienced this for our project. One is in defining what an electronic laboratory report (ELR) is. Since end users do not see raw messages, it is often difficult for them to understand there are nuances and difference from the flat files we receive and the ELRs since they display similarly in the surveillance system. The other definition we've realized can differ is on when program areas want to do lab audits, who can fulfill the auditing process, only the lab, any one at the facility or the infectious preventionist group. Because information can differ slightly with different providers and within LIS/EHR systems, information reviewed from the lab audits can differ.

Thanks Ann, and the MN team!

Maryland Department of Health HIV Surveillance Team:
1. Regular database matches are completed to share relevant data between systems. We are always improving these processes and adapting as the data systems change.
2. a. The most common terminology difference, that we see in the data that we receive, is sex at birth and gender.
2.b. N/A

Thanks Veronica and MD HIV team!

Virginia Rabies Data Management Team:
1. The Virginia Department of Health has experience with this, where sharing data is challenging due to lack of coordinated governance, although we are striving to improve this.  We have existing systems that may be redundant, instead of leveraging solutions that would work across divisions and the agency.  This includes VEDSS for most infectious diseases, Maven for HIV/STD, VIISS for the immunization registry, WebVISION for patient scheduling and care (used by localities), the state lab’s laboratory information system (does generate ELR), and EHD database used by environmental health for tracking rabies bites, animal control and PEP.  In addition, localities like Fairfax may have their own systems used to capture this information that do not communicate with those of central office VDH.  Analysis of comorbidity, or patient care across VDH is difficult to assess.   

2.  We have not experienced many terminology challenges in our rabies data management project, thankfully.  The team has been communicating regarding which fields in the Fairfax rabies report that central office VDH needs.   Most of these fields are understood and defined.  There is sometimes explanation needed on understanding each other’s (informatics, business) processes and workflows, including email reports, feeds, and the rabies data reconciliation process.  Data governance would allow us to more successfully evaluate the options for data management, e.g., knowing more about the EHD database and its capabilities.

Did the content get cut off Rebecca?

Hey Bonny, I've updated my post  - thanks!

Got it. Thanks!

1. At Columbus Public Health many of the public health programs track their own data, without any way to link participants across programs. For example, there are two big home visiting programs housed within the Maternal & Child Health section, serving overlapping populations. These programs use different data systems, making it difficult to ensure that a woman is only enrolled with one program at a time. Also, both programs track smoking status as an indicator, and some of the women enrolled in home visiting might also be enrolled in the Baby & Me Tobacco Free program, which uses breath tests to assess smoking status (which is more reliable than self-report) but they use a different data system that doesn't link to either home visiting data system.
2. Within this example, 'delivery' is comparable to the term 'served', which can mean different things to different programs or in different contexts. 'Warehouse management' would be similar to a program referral, 'sales' could be program enrollments, and 'support' could be successful program completion. Data governance could set definitions for each of these terms, so that different programs can understand each other and indicators (e.g. number of women served) can be compared across programs. 

THanks!

Maryland Team - OP&R
1. There is a legislatively mandated annual report that is required to examine all aspects of opioid use disorder (stats, treatment programs, population characteristics etc.) and make recommendations for the state on improving prevention, response, and data collection efforts across the state. The data needed for this report are all in different internal/external administrations/agencies (~10) each with their own processes of how to share (some more time consuming and cumbersome than others), and what they are willing to share. There are some efforts being made to use Maryland’s HIE as a centralized locale to collect the different data sources.
2. There is perhaps some misunderstanding of what a DUA is and its purpose…there is also not strong advocacy for data sharing from department leadership…which in turn stalls data governance. Clear communication of the purpose of data sharing/DUA and the role of data stewards may help to clarify any misconceptions and streamline the data sharing process.

Florida Disease Control
1.       For our project we are addressing the challenge of linking our reportable disease surveillance system Merlin with our immunizations registry Florida SHOTS. Florida SHOTS is also linked to Vital Statistics, so we are essentially linking three systems together. Within our division, we have at least four different systems that are not connected. Thinking of projects that use data from multiple systems is difficult when the first step is always figuring out how to connect the data between systems. 

2.       In our project, vaccine history is defined and stored differently between Merlin and Florida SHOTS. In Merlin, vaccine history is stored under one profile and comprises all vaccines a person has received. In SHOTS, vaccine history can be stored under multiple profiles based on how the information is received from providers, which creates a challenge in linking the two systems. More generally, in our division we define report date differently between our individual systems.

Tennessee
1. Our integrated surveillance system (NBS) has a master patient record (consolidated profile); however, this only applies to patients that have been entered into the system. Our local/regional/metro epis often have to use multiple systems (EPI -our EMR-, eHARS, PRISM, Vital Statistics, etc) for a complete patient profile, requiring quite a bit of manual linkage (aka "data wranglin' y'all").
2. As the Informatics team, we liaise between IT and communicable disease programs and often encounter "language barriers" that require clarification. Our project focused on trying to bridge that gap, but we still encounter these issues almost daily, one example being the difference between ELR and lab data submitted electronically and imported automatically. A data governance plan may help address these issues, but would need to be at a higher level than our program due to the data sources in separate program areas. 

Question 1.  In Santa Clara County, we are required to use multiple systems to track communicable diseases.  So for example to perform proper reporting for some conditions we are required to start a record in one system, process it another and then back fill data in the first system by hand so that we may report to the State.

Question 2.  Data governance allows you to come up with a standard definition that can be shared across organizations so that business process may run smoother.  For example we are concerned with when a fax is received in the Department compared to the electronic lab report and had performed a study to determine how many faxes we could match to their equivalent electronic report.

 

Some groups used the definition of when a fax was picked up by the data entry clerk while another program used the date the fax was received.  Basically one was operational and looked at when the person actually picked up the fax and the other looked at a technical definition of when did the fax actually come in.  This lead to dome discrepancy in the results so for the current study we defined the definition to be consistent across all programs so that we could avoid mixed results.

Thanks everyone!

Question 1:  In El Paso County, animal bite information is collected in three different systems; a online database housed in the Environmental Health Division, the statewide disease reporting system (CEDRS) , and  REDCap reports and database housed in the Communicable Disease Program.  None of these interface, nor collect all of the same variables, therefore making it very difficult to conduct any data analysis.

Question 2:  Data governance can be used to address terminology standardization challenges by creating rules for standardization of each field of data entry.  

Our project is a prime example of lack of commonality in terminology.  Our public health department is working with our sheriff's office and we each have our own set of terms.  In fact, even the sheriff's office and their sister program housed in the Colorado Springs Fire Department use varied terminology so they are not able to merge medical records for the same patients, leading to a lack of continuity of care for their patients.

Thanks, Marigny!  I hope you saw some parallels with the pre-work submissions we were able to go over.