CORONAVIRUS (COVID-19) RESOURCE CENTER Read More

Using Optical Character Recognition (OCR) to Collect, Manage and Analyze Data and Guide Decision Making during a Public

State: CA Type: Model Practice Year: 2011

Events such as Pandemic Influenza have demonstrated that information systems are critical to managing response and preventing further outbreaks. In addition, information systems allow for tracking of countermeasures, monitoring their effectiveness, identifying resulting adverse events, analyzing client data to ensure that target population is being reached and enhance preparedness for future emergencies. Each fall, Los Angeles County Department of Public Health (LACDPH) implements a seasonal influenza vaccination campaign that includes scheduled outreach vaccination clinics at various community sites throughout the County. This has allowed the LACDPH to test and update its operations for the “Point of Distribution” (POD) model for mass vaccination emergencies. Prior to 2007, data was collected via “vaccine accountability” forms that were manually completed by vaccination staff. In 2007, LACDPH started developing a scan-based data collection system for mass prophylaxis. The system would digitize responses in handwritten text, extract it and place it into a database, while keeping digital copies of the original form completed by the POD clients in an accompanying data repository (that is indexed and retrievable by all fields). The system also has the ability to collect data using scan technology from a centralized location or decentralized from multiple locations to provide maximum flexibility. During the H1N1 pandemic (2009), LACDPH created a client data collection form to capture H1N1-specific eligibility data and CDC required data, as well as data desired by the department for use during at PODs, schools, health centers and community outreaches. Overall, the system processed more than 230,000 forms from October 2009 through February of 2010. The system allowed LACDPH to collect and report doses administered reports to CDC’s Countermeasure Resources Administration (CRA) each week, process data requests from the department, local officials and the media, provide client forms for VAERS follow up, provide copies of forms for the public and authorized medical personnel, and data for presentations and publications. Above all, the system made it possible to collect, analyze and accurately report time sensitive, client-specific data to decision makers and guide planning efforts. In addition, the system enabled LACDPH to examine things such as distance traveled, using GIS, by individuals to get to PODs, effectiveness of POD sites and vaccination rates among various minority ethnic groups to improve outreach during future events.
Events such as pandemic influenza have demonstrated that information systems are critical to managing response and preventing further outbreaks. In addition, information systems allow for tracking of countermeasures, monitoring their effectiveness, identifying resulting adverse events, analyzing client data to ensure that target population is being reached and enhance preparedness for future emergencies. Each fall, Los Angeles County Department of Public Health (LACDPH) implements a seasonal influenza vaccination campaign that includes scheduled outreach vaccination clinics at various community sites throughout the County. This has allowed the LACDPH to test and update its operations for the “Point of Distribution” (POD) model for mass vaccination emergencies. Prior to 2009, mass vaccination data (such as client characteristics) were collected via “vaccine accountability” forms that were manually completed by clinic vaccinating staff and forwarded to an Immunization Program Area Field office for review and correction as required. These forms were then forwarded to the Immunization Program’s Central Office for data reconciliation and entry. Due to the time involved (a typical flu campaign lasts between four and six months in LAC) in processing and analyzing the “vaccine accountability” forms, the reports cover a period of time that lagged the reporting date by at least two weeks. Since client documentation and tracking relied on hand data entry, such a system would be inefficient in providing time critical information to decision makers regarding the effectiveness of their outreach strategy or ensuring that priority groups were being efficiently reached. A delay in data reporting would also shrink the amount of time available to decision makers to make changes to their outreach strategy in order to contain an outbreak during a true public health emergency such as Pandemic Influenza.
Agency Community RolesThe LACDPH is the primary facilitator of the optical character recognition system as a whole. Several programs within the department together make the system's functionality viable. The three major programs involved in the system's planning, development and operations are the Emergency Preparedness and Response Program (EPRP) that is primarily responsible for the Department's readiness during an emergency, Community Health Services (CHS) that is responsible for community outreaches and Points of Dispensing (PODs) and the Immunization Program (IP) that serves as a lead for all vaccination related outreaches. Costs and ExpendituresEvents such as Pandemic Influenza have demonstrated that information systems are critical to managing response and preventing further outbreaks. In addition, information systems allow for tracking of countermeasures, monitoring their effectiveness, identifying resulting adverse events, analyzing client data to ensure that target population is being reached and enhance preparedness for future emergencies. Each fall, Los Angeles County Department of Public Health (LACDPH) implements a seasonal influenza vaccination campaign that includes scheduled outreach vaccination clinics at various community sites throughout the County. This has allowed the LACDPH to test and update its operations for the “Point of Distribution” (POD) model for mass vaccination emergencies. Prior to 2007, data was collected via “vaccine accountability” forms that were manually completed by vaccination staff. In 2007, LACDPH started developing a scan-based data collection system for mass prophylaxis. The system would digitize responses in handwritten text, extract it and place it into a database, while keeping digital copies of the original form completed by the POD clients in an accompanying data repository (that is indexed and retrievable by all fields). The system also has the ability to collect data using scan technology from a centralized location or decentralized from multiple locations to provide maximum flexibility. During the H1N1 pandemic (2009), LACDPH created a client data collection form to capture H1N1-specific eligibility data and CDC required data, as well as data desired by the department for use during at PODs, schools, health centers and community outreaches. Overall, the system processed more than 230,000 forms from October 2009 through February of 2010. The system allowed LACDPH to collect and report doses administered reports to CDC’s Countermeasure Resources Administration (CRA) each week, process data requests from the department, local officials and the media, provide client forms for VAERS follow up, provide copies of forms for the public and authorized medical personnel, and data for presentations and publications. Above all, the system made it possible to collect, analyze and accurately report time sensitive, client-specific data to decision makers and guide planning efforts. In addition, the system enabled LACDPH to examine things such as distance traveled, using GIS, by individuals to get to PODs, effectiveness of POD sites and vaccination rates among various minority ethnic groups to improve outreach during future events. ImplementationLos Angeles County Department of Public Health (LACDPH) began implementation of the optical character recognition system in 2007. The system was implemented in phases and tested each year during the Influenza Points of Dispensing (PODs) Exercises each year. In 2007, in order to test the existing system against the proposed system in order to get buy off from stakeholders within the department – LACDPH tested the optical character recognition system with the help of the system vendor, FormTran. This was done using data from three PODs and tested the ability of the scan-based system vs. hand entry of data in terms of speed and accuracy. The optical character recognition system provided clean data post verification within 2 days vs. the hand entry that took over two weeks. Due to the initial success LACDPH decided to expand the system and test it again with 10 PODs in 2008. During the same time period LACDPH brought in stakeholders from the Department’s Immunization Program (IP), Community Health Services (CHS) and Public Health Information Systems (PHIS) to develop a business plan for full deployment of the optical character recognition system by 2009. The business plan encompassed all aspects of emergency data management – form development, remote form scanning, data verification and data retrieval at the department operations centers (DOC). The business plan called for all stakeholders to jointly provide input on the development of the form; the forms were generated by the Immunization Program for influenza (in addition LACDPH pre-developed forms for all Category A agents). The forms would be remotely scanned from Health Districts into a shared folder from where the Optical Character Recognition System would retrieve/process them. Data verification would occur at the Immunization Program and data could be retrieved by any authorized personnel (in various formats). Over the course of testing LACDPH revised its software license strategy by purchasing server level data verification licenses rather than individual desktop licenses. This shared license strategy proved to be much more cost effective enabled data verification staff to conduct their tasks from their own workstations rather than designated desktops. LACDPH has been working on implementation of the Optical Character Recognition System since 2007. The system was completely implemented in 2009, with minor refinements being made to the business plan in 2010.
Use of Client Data to Enhance Outreach Strategy: A primary purpose of implementation of the Optical Character Recognition system in LAC was to provide decision makers at the Department Operations Center with real time data in the midst of the emergency in order to ensure that the outreach strategy was effective in reaching the target population. As client forms were scanned and data became available - analysts at the DOC began conducting data analysis based on priority groups and vaccination rates among various ethnic groups. Based on the data analysis it was determined that the African American community was grossly under represented in vaccination (only 2.96% of POD clients). Using the data LACDPH developed a new outreach strategy that was tailored specifically towards the African American community. Electronically Maintain Client Data in a Searchable Data Repository: Given that Los Angeles County anticipated a large number of clients during a pandemic an accompanying database was developed to house searchable client data from the Optical Character Recognition System and link the data to a pdf copy of individual client forms.During H1N1 - LACDPH received 32 client form requests for VAERS reporting, client follow up or patient record. LACDPH analysts were able to search the database and successfully retrieve a copy of the client form for follow up on 30 of these occasions. In addition, LACDPH analysts were able to conduct data queries on client data based for information requested by the Board of Supervisors, Department decision makers and City Council members. Report Data as Required by the CDC to the Countermeasures Response Administration (CRA) System on time: During H1N1, the CDC required all states and major cities to report aggregate data to the CRA system on a weekly basis. Given the large number of shots given by Los Angeles County it would be impossible for Los Angeles County to report aggregate data for specific age and vaccine subgroups. LACDPH choose to use the Optical Character Recognition system to report data to the CDC on a weekly basis. LACDPH successfully reported aggregate data to the CDC consistently during the mandatory reporting period.
The use of the optical character recognition system for data collection, management and retrieval has been universally accepted by all emergency management stakeholders in LACDPH. The Emergency Preparedness and Response Program within the department spent almost a year to ensure stakeholder support during the business plan development phase of the system. The system is now used yearly to process all Influenza vaccination forms and is being expanded out for other disease outreaches as well. In addition, plans are also in place to use the system will used during the Biohazard Detection System activation in Los Angeles County. Due to the new licensing structure of the business plan that eliminated individual desktop licenses and implemented server based licenses, the system has proven to be much more cost effective. Since the software and hardware were one time purchases the system will continue to be used during fiscal crisis and without further expenditures, thus assuring the system’s sustainability.