About Administrative Data
Administrative data is information that service organizations or government organizations collect from programs, including services provided and client demographic information. Typically, organizations use this data for record-keeping rather than research, but it can be a rich source of information for understanding the scope of homelessness, trends, and demographics of people experiencing homelessness. Administrative data is the most powerful when organizations create a shared database that pulls information from multiple sources.
“I have not always been homeless, I was not born this way. But I will always be human.”
– Charlotte (via Invisible People)
Steps To Collect Administrative Data
- Length: 3 months to 3 years
- Involve citizen and non-government stakeholder participation, including people with the lived experience of homelessness at each step to ensure that principles of inclusive, community-owned data are incorporated
- Identify potential partner organizations; pinpoint what data points organizations currently collect; determine additional data points that need to be collected; ensure that all required homelessness client data is included (Thomas & Tweed, 2016)
- Establish a data governance structure, including an oversight body.
- Determine privacy protection methods and data security approaches
- Privacy protection methods are determined to ensure the privacy and security of data vary depending on the intended use of data and the local and national legislation around sharing and processing personal data (Thomas & Mackie, 2020)
- Data security approaches are (1) processing of data, i.e., aggregation, (2) anonymizing or de-identify data, or (3) using personal data with added measures to ensure minimization of disclosure risks, e.g., a ‘split file’ method of sharing data.
- Determine the most appropriate form of a centralized database
- A central data repository has the data process in pushing into a shared platform for different organizations to enter data.
- A federated data repository has a more single-use approach to pulling data for only specific uses. Unlike a central permanent repository, the data extracted from federated systems upon request for a data broker to extract data.
- Establish data sharing agreements or Memorandums of Understanding between partner organizations
- Determine and implement informed client consent processes and forms
- Solidify data quality control measures
- Ascertain whether to include a capture-recapture component
- Implement and regularly review quality control measures to ensure that data meets a high level of quality. Clean data as needed before inputting it into the database.
- Ensure partner organizations are regularly entering their updated data (quarterly at a minimum)
- The data governance oversight body should meet at least annually to review and update data sharing agreements, Memorandums of Understanding, client consent processes and forms, and data quality control measures. Incorporate citizen and non-government stakeholder participation, including people with the lived experience of homelessness at each point of review.
- At least annually, data analysts should review the data collected to analyze and publish information on the total number of people experiencing different types of homelessness, including demographic characteristics, housing and services provided, and other relevant trends and information.
- These reports should be used by stakeholders to determine additional funding, housing, and service provision needs.
Determining a data governance structure, data sharing agreements, Memorandums of Understanding, client consent processes and forms, and establishing a database can be time consuming.
Service providers and government organizations must collect data regularly.
Data must be cleaned to ensure accuracy, consistency, and adherence to data standards, and can delay the data collection process.
Characterizing people experiencing homelessness and trends in homelessness using population-level emergency department visit data in Ontario, Canada
Data on people experiencing homelessness often come from time- and labour-intensive cross-sectional counts and surveys from selected samples. This study
uses comprehensive administrative health data from emergency department (ED) visits to enumerate people experiencing homelessness and characterize
demographic and geographic trends in the province of Ontario, Canada, from 2010 to 2017.
A classification model of homelessness using integrated administrative data: Implications for targeting interventions to improve the housing status, health, and well-being of a highly vulnerable population
This study evaluated a linked dataset of administrative records in Massachusetts collected from 2011-2015. The purpose of this study was to show a specific approach to identifying homelessness and people at risk of homelessness (health-related issues, like mental illnesses, substance use disorders, etc.) using integrated administrative data.
Questions About Admin Data
For questions, please email email@example.com.
This collection of resources contains IGH’s sample policies and procedures, to help you choose an enumeration method. Browse by clicking on the links provided to you.