Philadelphia HMIS


Best Practice | August 11, 2006


In 1991 Philadelphia initiated a data system to keep track of shelter use, utilizing the City’s existing information technology. It initiated the program because it needed much more specific data on shelter use for billing purposes, and because it needed better information to improve program design. Information on homeless people and shelter use has been kept since that time, with periodic improvements in hardware, software and methods of data collection.


The Philadelphia system covers City-funded programs. The data that is currently collected about homeless clients is used to allocate resources, to conduct performance based contracting (and thus to look at client outcomes), and to look at trends in numbers and demographic characteristics.

The City is constantly improving its HMIS system. Currently it is shifting to a system that better integrates data on homeless clients with data on other human service programs via what will be a social services data warehouse. This is in part to implement Mayor Street’s vision for a collaborative case management system.

Data Elements and Process

The system is collecting two types of information:

• Client demographic information
• Shelter stay history

With respect to confidentiality for specific groups (HIV/AIDS, domestic violence, etc.), the decision about how to handle the data in such cases is made through a due diligence process to determine an appropriate protocol.

Philadelphia’s system is based upon centralized intake. Potential users of the City shelter system come to one of two sites (or several off-hour sites). One site is for families and single women, and the other is for men. At these sites an intake is performed, and clients are assigned to a shelter. (The intake centers do not share their intake information with the shelters, so most shelters conduct another intake when the client arrives.) The demographic data, therefore, comes from the intake centers.

The shelter stay history is received from the shelters. The shelters submit a summary report at the end of each month. This provides information on how many nights each individual has stayed at the shelter. This is the basis of billing for the shelter, and of course also provides the City with data.

People in transitional housing are not included in the data system being discussed here. There is another, similar, system that captures data from the City-funded transitional programs. However, many of the clients in the transitional system arrived there via the shelter system, and thus their demographic data is captured by the shelter data system.

Agency Participation

All of the City-funded shelters participate. It is estimated that the majority of people who become homeless are included in the system.


Seven full time staff people work in the department that handles the data. Their job is to make sure the data comes in and is analyzed, maintain the mainframe, keep track of the hardware (some 300 computers), and ensure that everyone is trained. Staff includes a director, a programmer, a business analyst, and hardware technicians. In addition, there are several hundred people who collect data.


The City pays for this system with its own funds supplemented by state and federal resources. Anyone who is involved in collecting or working with the data receives training arranged by the Information Technology office.

Analysis and Interaction with Other Systems: The Information Technology office analyzes the data generally. However others, including researchers, also analyze the data. When data is given to researchers, either the clients are asked to sign research protocols, or the City blinds the data (make it anonymous). When the data is placed in the data warehouse, the City will be able to cross analyze it with other systems within the warehouse. At present this can only be done by cross-referencing identifiers, which is done under the provisions noted above.

Lessons Learned

• Having data helps with fundraising. Not only is the City able to provide ready data to funding processes such as the Continuum of Care, but grantors like the idea that the City is paying attention to how the money is being used.
• Centralized intake is helpful for data collection and purity, but you have to have good systems in place for security, because you have a larger universe of users.
• When setting up a system, do as much planning as you can. Find out in detail what businesses processes you need, what could be improved by automation, and what could not. Don’t look at the software first and then try to fit your needs into it. Decide what your needs are first, and then go shopping for software.
• When dealing with software and hardware vendors, make sure that the roles and responsibilities are clearly delineated (installation, maintenance, etc.) The same is true after the system is up.
• Once the system is in place, make sure the technology people understand how the users are using it. This will ensure that the system is enhanced to meet real needs.
• Recognize that social services data is very fuzzy. Technology people who have not worked in the human services field will find this difficult at first.

For more information, Contact

Matthew Berg
Information Technology Manager
Office of Emergency Shelter and Services
City of Philadelphia