Q&A Spotlight: What are APCDs Learning about Data Quality?
The first step in maintaining a successful APCD is ensuring it is stocked with quality data.
Seven recommendations on how to get there from Freedman HealthCare Health Data Analyst Rik Ganguly:
1. Build multiple tiers of data quality checks to examine the data from different perspectives and at different stages of processing and file creation.
2. Ensure prompt feedback to submitters after each layer of checking.
3. Design a reporting strategy that parallels data maturity. Cleaner data supports more refined and detailed analysis.
4. Be transparent: acknowledge that a large dataset drawn from many disparate sources will have data quality issues. Explain your strategy for tackling them.
5. Create clear communication channels about data quality for all those that touch the data: intake standards, data aggregation vendor’s processes, the analytic vendor’s transformations and insight from the user community.
6. Be sure that, as administrator, you have access to the raw data. Only by working closely and deeply in the raw data can most quality problems be identified and diagnosed. It is not enough to trust this to your data vendor.
7. Work closely with your data submitters. Forging a collaborative relationship will make this a better experience for all.