The upgraded PMN platform includes a newly designed concept for Model Adapters that can be “plugged into” the software. The plug-in design allows for development of new features that can be added without impacting other parts of the system. This new type of architecture separates the concerns of the network platform from the details of the requests (i.e., queries) that travel through it. The result is a network that forms a tunnel through which requests and responses travel.
PopMedNet currently supports a number of plug-in request models:
- Menu-driven queries summary table queries that execute against Summary Table databases
- Menu-driven ESP queries that execute against ESP databases
- File distribution queries used to exchange documents between the portal and DataMarts
- SQL distribution queries used to distribute raw SQL between the portal and DataMarts
- SAS queries used to distribute SAS program files between the portal and DataMarts
- Request Metadata queries to search the metadata of previously submitted queries.
- Organization Metadata queries to search the metadata of Organizations and DataMarts across a PMN network.
- Queries composed from a number of external applications, such as I2B2 and hQuery, that use a native interface to compose a query and submit it to a PopMedNet network.
Menu-Driven Summary Table Queries
Menu-driven request types are preset parameterized queries where the investigator uses a visual wizard to set query parameters. Menu-driven summary table queries execute against summary tables. These tables provide summary counts of individuals by period, age groups, and sex. The summary counts include information on medication (e.g. number of dispensing, users, and days supplied), diagnoses (e.g. number of individuals with the diagnosis), procedures, and the overall data partner population.
The summary queries are grouped into three request models:
- Prevalence – prevalence counts of diagnoses, procedures, and drug utilization
- Incidence – incidence counts of diagnoses and drug utilization
- Most Frequently Used – top XX events and members within a specified query type (drugs, diagnoses, or procedures) within the prevalence tables
The screenshot below is an example of a prevalence ICD-9 3-digit diagnosis query. In this example, an investigator can select 3-digit ICD-9 codes, observation period, care setting, age stratification, and sex stratification.
The screen shot below shows an ICD-9 Diagnosis code wizard. The investigator is led through the process of searching and selecting ICD-9 diagnosis codes:
Menu-Driven ESP Queries
The Electronic Support for Public Health (ESP) Query Builder is a set of requests that operate against the ESP database. ESP (link to http://www.esphealth.org/) uses a set of complex algorithms to identify selected diseases of public health concern. The ESP algorithms are based on the ESP data model that is a standardized representation of EHR-based encounter and patient demographic information. The ESP platform enables automated extraction of data from EHRs into a format suitable for disease surveillance activities.
The requests supported by ESP Query Builder are:
- ICD-9 Diagnosis – used to build ad-hoc custom health measure queries
- Reportable Diseases – used to select one of a list of reportable disease reports that are produced periodically on the ESP servers
The screenshot below is an example of an ICD-9 diagnosis query. In this example, an investigator can select ICD-9 diagnosis codes, observation period, age range, gender, race, and report selectors.
The screenshot below is an example of a reportable disease query. In this example, an investigator can select a disease, observation period, age range, gender, race, and report selectors.
File Distribution Queries
File distribution queries allow investigators to distribute any type of file to the network DataMarts. Files may contain queries, processing instructions, or any other machine or human readable file content. Files transmitted may often be data files, programs, or work plans. Files are received and approved by the DataMart Administrator and processed by hand according to local policies or instructions of the receiving site. The screenshot below is an example of a file distribution query:
SQL Distribution Queries
SQL distribution queries allow investigators to distribute raw SQL code from the portal to the network DataMarts where it can be executed within the DataMart client. This request type is data model agnostic.
The screenshot below is an example of a SQL distribution query. In this example, an investigator can paste the SQL query string into the box provided.
SAS queries allow investigators to distribute SAS program files to the network DataMarts. Program files are received and executed by DataMart Administrators from within their local DataMart Client.
The screenshot below is an example of a SAS query:
Note that to support a PopMedNet SAS request a Data Mart must have a SAS license.
Request Metadata Queries
This request type gives PMN users the ability to query the metadata of previously submitted queries from within the Query Tool.
Organization Metadata Queries
Information about Organizations, their DataMarts and Users, the patient data they have available and data models they support, and the PMN Groups and Projects they participate in can be called organizational metadata. Organizations, Groups, Projects, DataMarts, and Users are the key PMN application entities. Metadata will be available on their existence and usage within the network as well as information captured in a questionnaire that contains information that is beyond what is needed by the network to function.
External Query Composers
In addition to its own “native” request models, PopMedNet also supports the following external query composers.
Note that support of these query composers are part of several Department of Health and Human Services Office Of National Coordinator Query Health Pilots and are not yet available in production. PopMedNet will be enhanced over time to include new request models and query composers.
PopHealth is an open-source quality measure reference implementation Query Composer that empowers healthcare providers to perform Stage 2 Meaningful Use quality measure reporting and promotes easier submission of quality measures to public health organizations. The PopHealth application can be run as a Query Composer by an investigator within a PopMedNet network.
The screenshots below demonstrate the PopHealth Query Composer.
The investigator selects the PopHealth Query Composer and is redirected to the PopHealth Query Composition application:
The investigator builds up the query request:
Once the request is processed by the Data Provider Data Mart Administrators, results are returned to PopMedNet.
i2b2 (Informatics for Integrating Biology and the Bedside) is an NIH-funded National Center for Biomedical Computing based at Partners HealthCare System. The i2b2 Center is developing a scalable informatics framework that will enable clinical researchers to use existing clinical data for discovery research and, when combined with IRB-approved genomic data, facilitate the design of targeted therapies for individual patients with diseases having genetic origins. The i2b2 application can be run as a Query Composer by an investigator within a PopMedNet network.
The screenshots below demonstrate the i2b2 Query Composer. For more information, view the PopMedNet-i2b2 Integration for ONC Query Health Pilot demonstration video.
The investigator selects the i2b2 Query Composer and completes the query metadata information in PopMedNet:
PopMedNet launches the i2b2 Query Composer and the investigator constructs the query and sends off the query request to the data providers: