TABLE.

Lessons learned from digitally capturing ART clinical records at health facilities and updating them in a central database

Observed challenges	Means for improvement
1 Data digitization	Modified data collection procedures
• Hastily performed tasks and photography, including inconsistent use of the camera stands or not removing master cards from their protective transparent plastic folder, producing poor image quality and illegibility and causing data entry gaps and errors	• Use fast, duplex sheetfed scanners (with autonomous power) instead of digital cameras, ensuring good image quality and complete (multipage) record scans
• Lack of systematic logging of digitized master cards, and systematic checking for missing master cards during digitizing	• Use barcoded labels (a sticker with ART ID in barcode and standard characters) on all plastic folders (each holds the master cards of one patient). Generate a list of digitized master cards, using a barcode scanner, and use the list to update a file tracking tool. Identify non-digitized records and take action before leaving the facility
• Missing records not always communicated with facility’s care providers who may be able to provide further information	• Before leaving the facility, communicate and investigate data gaps or errors with the facility’s care providers to make corrections where possible and identify patients that need tracing
• Digital image files not named and filed systematically, resulting in near impossibility to find an image to crosscheck information in the whole patient record	• Rename each electronic image file with predetermined structure; ART ID (health facility code and serial number) and digitizing year and month, and save in predetermined electronic folders
2 Quality control	Implement quality control measures
• Lack of systematic checking of the ART ID sequence to identify missing IDs (assigned to no patient) and duplicate IDs (assigned to multiple patients)	• Systematic (monthly for each facility) use of the ‘ART ID sequence check tool’ to identify misformated, missing or duplicate ART IDs and address errors promptly
• Exclusive focus on digitization of most recent master cards (latest patient visits, recorded on one side of the master card), prohibiting a full record history review	• In addition to monthly digitization of most recent master cards, perform periodic (annual) full file scans (all master cards for one individual recorded over time) to ensure data completeness
• Lack of control of data entry accuracy	• Systematic double entry of 10% of randomly selected records. Review trends in errors made and address these with individual data entry clerks
• Lack of systematic and regular supervision of data collection, management, and entry	• Regular on-site supervision of collection processes for tracking users and regular review of data entry patterns and issues between senior data staff and data teams
3 Health system record keeping	Facilitate health systems record management
• Providers at health facilities not maintaining paper records in systematic fashion, introducing complications to the digitizing processes	• Have data team log issues found in reviewing paper records for communication back to health facility staff and Ministry of Health supervision teams
• Patients move between facilities, temporarily receiving ART from a different facility from where their master card is kept. Those visits are not recorded, creating apparent missing or defaulted patients	• Introduce log sheets for transient patients at all facilities. This should facilitate the recording of patient ART ID and information collected for one patient visit, equivalent to one row on the ART card. These log sheets need to be digitized each month, the data updated in the database and the original ART facility informed
4 Under-resourced data management team	Ensure adequate human resources
• Pressure to complete tasks in time by too few data collectors and data entry clerks, resulting in compromised quality	• Adequate human resources for data collection, data entry and validation
• Inadequate supervision as a result of too few senior data staff, resulting in compromised quality assurance monitoring	• Adequate supervisory staff is key to ensuring regular review of processes to ensure quality and complete data
• Main focus on having all sites digitized and captured in time versus completeness and quality of data	• Introduce performance-based incentives to encourage optimal data collection, capture and quality control
5 Inadequate use of data and feedback	Improve capacity and communication
• Limited attention and capacity of ART program managers to use collected data and give feedback about data issues needing attention	• Capacity building among ART program managers to allow them to use the data more regularly in order to identify clinically relevant issues and provide meaningful feedback
• Limited understanding among the data team about how data are used to inform care delivery	• Capacity building among data teams in the role of high quality data in improving quality of care
• Limited feedback to health facility staff about the results of the data collected and how they are being utilized	• Regular provision of reports and findings generated from data provided to both data teams and health facilities, with opportunity for feedback

ART = antiretroviral therapy; ID = identification.