The Arizona Department of Health Services' (ADHS) Childhood Lead Poisoning Prevention Program (CLPPP) has conducted surveillance and coordinated care with clinicians and families of children with elevated blood lead levels (EBLLs) for nearly 2 decades. CLPPP identifies approximately 350 children annually with EBLLs. While lead-based paint remains a major source for Arizonan children, cultural and nontraditional sources also contribute significantly. Blood lead test results are reportable per Arizona Administrative Code R9-4-302, allowing ADHS to identify high-risk areas for targeted blood lead screening for children. CLPPP receives approximately 65 000 results annually from commercial laboratories and health care providers using point-of-care systems (POCS), such as LeadCare® II. CLPPP utilized Centers for Disease Control and Prevention's (CDC's) database, Systematic Tracking of Elevated Lead Levels And Remediation (STELLAR), beyond CDC's support of the software. CDC developed a new system, Healthy Homes Lead Poisoning Surveillance System (HHLPSS), intended to replace and expand surveillance capacity; however, ADHS was unsuccessful in implementing HHLPSS within the needed internal time frame due to resource and expertise constraints. CLPPP shifted focus and undertook a unique approach by introducing an environmental health morbidity into Arizona's communicable disease surveillance system, Medical Electronic Disease Surveillance Intelligence System (MEDSIS). MEDSIS was chosen because it had established partial capacity for blood lead surveillance, in-house expertise to incorporate additional components, and could be implemented quickly.
The database transition required adaptations to blood lead reporting and data management to align with MEDSIS-specific formatting and processing. As such, CLPPP had an opportunity to evaluate data management processes and determine whether process efficiency could be improved in a new system. This article aims to evaluate changes in data management processes pre- and posttransition with the objective of improving efficiency, data quality, and surveillance efforts.
ADHS CLPPP, MEDSIS, and Electronic Laboratory Reporting (ELR) teams reviewed data management processes, data flow, and surveillance efforts in the former system. Key strategies for process change were selected on the basis of feasibility to implement and potential for impact on program operations (Figure 1).
The selected strategies included data standardization and validation, data automation, and centralized data capturing of surveillance activities. CLPPP used existing MEDSIS features (ELR, direct entry into MEDSIS by reporters, case management tracking, and customizable disease-specific observations) to implement the process changes and built more robust data reporting tools and validation methods informed by reviewing former data management processes.
After 1 year posttransition, CLPPP evaluated the newly implemented process changes (Figure 2) to determine effectiveness. Indicators to evaluate changes included data receiving methods and automation, data processing time, patient address completeness, ability to house statutorily required data, duplicate reporting and tracking efforts, and standardization of surveillance activities.
Results and Discussion
Reviewing STELLAR-specific data processing and management before the MEDSIS transition allowed CLPPP to establish more efficient methods of operation, leading to better data quality and surveillance. Many of the findings of the former system related to the inconsistencies in reporting formats and data completeness. Changes mainly impacted POCS reporters because data quality issues related to commercial laboratories were largely resolved by implementing health level 7 (HL7) reporting.
Comparing data management processes pre- and posttransition
Acceptability and usability
The pretransition review revealed the extensive time CLPPP spent cleaning data to meet former database formatting needs (0.5 FTE). This led to the automation of data processing and receiving in an effort to decrease resources dedicated to data management. From 2015 to 2016, 97% (n = 62 614) of blood lead test results were reported electronically via e-mailed spreadsheet and 3% (n = 1937) by fax or mail; all results reported required some degree of manual importation, such as multistep reformatting and processing. Posttransition, data processing and importation required 0.02 FTE, a decrease of 96%. Moreover, 99% (n = 51 288) of results (N = 51 507) were received electronically, 89% (n = 45 745) of which were processed automatically into the system: 77% (n = 39 540) were imported automatically via HL7 messaging, and 12% (n = 6205) entered directly into MEDSIS by reporters. Eleven percent (n = 5543) of results, received by spreadsheet, fax, or mail, needed some degree of manual importation. In addition, reporting of results by POCS users improved markedly, with 100% of reporting facilities reporting electronically posttransition compared with 91% pretransition. This is largely due to the adaptation, optimization, and standardization of reporting options for POCS users: direct entry into MEDSIS and submission by validated spreadsheet.
CLPPP focused on the quality of data received, reducing data management efforts. Pretransition, data formatting was not strictly defined and CLPPP could not easily identify duplicate test results. Posttransition, CLPPP devoted additional effort to ensure standardized, accurate, and complete reporting: reports with errors or missing required fields were rejected and returned to the reporter for correction and resubmission.
Efforts dedicated to data standardization led to stringent formatting and content validation, particularly among address fields. While full address is typically available to reporters, it remains the most underreported required field. This may be because reporters do not fully recognize the value it adds to public health surveillance and evidence-based practices. In the former system, approximately 10% of records were missing critical patient address fields, whereas less than 3% of records were missing this information over 1 year posttransition. In addition, through content validation, CLPPP ensured that ZIP code, city, and county for Arizona residents were reported by POCS users, resulting in improvement of field completeness.
The pretransition review also found that 2 large medical facilities were reporting test results that were already reported by commercial laboratories. However, the former database was not structured to capture the performing laboratory's information, inadvertently allowing duplicate records. After CLPPP transitioned to the new system, the facilities were no longer reporting duplicate results.
These efforts led to an improvement in the quality of data received and a decrease in time spent by CLPPP to process results and maintain a clean database. Eliminating duplicate reporting by health care providers was instrumental in successfully improving data management efficiency for both CLPPP and the reporting facilities.
Case identification and creation were manual processes in the former system and could lead to human errors and delayed public health action: all reports were reviewed for elevated results to be entered manually for case creation. This challenge was addressed posttransition by filtering test results automatically based on predefined criteria, primarily elevated results for new or existing cases, and prioritized and flagged for review by CLPPP. Results that do not meet the criteria are labeled as “MEDSIS Unused Data” and do not display by default but are still stored and searchable within the database for case testing history and future analyses. Results received through validated spreadsheet or HL7 are reviewed and processed through this filter.
To determine whether the filtering process improved efficiency, CLPPP reviewed the amount of time from specimen collection to case creation. In the former system, the median time between these 2 dates was 6 days (IQR = 4-8) compared with 4 days (IQR = 3-6) in the new system. While report date is the preferred indicator to measure program timeliness, this date was often not captured in the former system. Since this indicator is captured automatically in MEDSIS, CLPPP also evaluated time from report date to case creation posttransition and determined the median time was less than 1 day (IQR = 0-1).
Pretransition, CLPPP's former operations utilized 6 separate spreadsheets to track essential elements for surveillance efforts, in addition to housing data in STELLAR. STELLAR also limited the type and format of information that could be stored, leading the team to strive for centralized and standardized data capture. Provider information, a statutorily required field, was only stored in STELLAR if manually entered. Only 2.5% of results had provider information in STELLAR compared with 100% of results in MEDSIS.
Implications for Policy & Practice
- The data management system implemented by Arizona's CLPPP can be adapted by other programs for improved data quality and data flow.
- This article identifies key strategies for efficient blood lead surveillance and data management: standardization and validation of incoming data, automation of data processing and importation, and centralization of data capture for surveillance activities.
- A validated spreadsheet is a successful tool for receiving standardized and complete data from POCS users.
- Collaborating and building strong internal relationships within the health department allowed CLPPP to identify partners and expertise that were essential to implementing effective strategies for improved surveillance.
- Building rapport with reporters was critical to establishing open communication and encouraged reporters to be receptive and responsive to program and reporting needs.
- This article contributes to the dialogue of establishing and adapting surveillance practices for emerging public health needs in a time of limited resources.
During the initial phases of the transition and deployment of new strategies, CLPPP devoted additional effort to provide technical assistance and trainings for POCS reporters to ensure standardization and completeness of data. An immeasurable, but valuable, outcome of the invested time was the enhanced relationship between CLPPP and reporters, resulting in improved data quality and timeliness of reporting.
Reviewing the former data quality and management processes allowed CLPPP to identify improvement opportunities for data management in a new surveillance system. Key strategies chosen to implement in the new system included standardized and validated data, automated data processing and importation, and centralized data capture of surveillance activities. CLPPP successfully adapted surveillance methods, using these key strategies, to a communicable disease database by utilizing and enhancing established database elements and processing systems. Evaluating the new data management process 1 year after implementation demonstrated the remarkable improvements achieved.
High data quality is essential for timely evidence-based practices and effective public health interventions. The improved efficiency achieved through automation and improved data quality will allow CLPPP more time and ability to explore surveillance data in ways not previously capable for the program, resulting in a better understanding of lead sources and challenges affecting Arizona's children. The strategies implemented in the new data management process can be used to develop best practices for other childhood lead poisoning prevention programs across the country.