HUD published an HMIS Data Quality Report to help ensure the integrity of the data being reported over. The report is broken up into seven sections on the standalone report and comprise questions 5 and 6 within the HUD Annual Performance Report (APR). The guidance followed to produce these reports comes from pages 17 -19 of the HMIS Reporting Glossary. The drill-down feature in HomelessData.com is especially helpful with this report as a list of records that need attention can be generated quickly and easily.
Data Quality Report Q1 / APR Q5
The Report Validations Table reviews the data set for all people served, and identifies the major population categories that are addressed in HUD reports.
Data Quality Report Q 2-4 / APR Q6A, B, and C
Questions 2 through 4 on the Data Quality Report, and questions 6A, B and C in the APR are checking for data completion rates. The “error rate” determines the percentage of applicable records for which a client did not know or refused to provide information, for which data was not collected, or for which there were known data issues. Data issues include dates of birth in the future, invalid social security numbers
Data Quality Report Q 5 / APR Q6D
These questions ensure that all of the data that is required to calculate chronic homelessness is available. The Count of Total Records refers to the total number of adults whom are heads of households. In the example below, 46 people, or 7% of the total records are missing the “approximate date started”. This element is seeking to capture when a person started their homeless episode.
Data Quality Report Q6 / APR Q6E
Timely data entry is critical to ensuring data accuracy and completeness. This section identifies how quickly project entries and project exits are entered into the HMIS after they occur.
Data Quality Report Q7 / APR Q6F
Data quality includes maintaining an accurate picture of the number of clients that are being served. For street outreach and night-by-night shelters it is a common issue that client records remain left open long after a client was last seen. This hinders the community’s ability to generate accurate performance measurement and inflates the numbers. The “# of Records” is a count of all clients that were active during the reporting period according to the project entry and exit dates. The “# of Inactive Records” column is a count of the number of clients within the report range that did not have a contact (outreach) or did not stay in the shelter for 90 days or more.