Defining the Question: Case Management Systems in Social Services
The use of computer-based case management systems in social services creates a tremendous opportunity for practitioners who want to answer questions that identify gaps, monitor performance, and document improvement. This involves analyzing discrete bits of case management data to transform them into meaningful, actionable information. Doing so successfully requires a specific, well-defined question.
For example, the analysis for a question like, “Did all workers contact each of their clients last month?” is very different from the analysis for “Across all workers, was there an average 95% compliance level for required contacts last month?” To answer the first question, we have to look for a case that was open last month but was not contacted:
- List all of the cases open last month.
- Identify if any of these cases did not have a contact.
To answer the second question we have to:
- List all of the cases open last month.
- Identify which of these cases required a contact. This may involve looking at case plans, case type, status (e.g. runaway, unable-to-locate), or other data items.
- Determine who was responsible for making contacts for each case.
- For each worker, count the number of cases that required a contact.
- For each worker, count the number of cases that required a contact and had a contact.
- For each worker, divide the number of cases requiring a contact by the number with a contact.
- Calculate the average percentage for all workers.
In practice, federal and local statutes and regulations may make the analysis even more complex. One monthly contact analysis created by NCCD staff used over 65 different data elements to determine contact compliance.
While this may seem daunting, the key is to be as specific as possible when forming the question. One helpful method for developing a specific question is to identify three important component parts:
Time: There will always be a time component, because all practice takes place in time. (If not, revise the question to include one, even if it is “For as long as we have data for …”).
Cohort: There will always be a cohort, or some set of elements being counted. In social services, these will likely be cases, clients, workers, units, or some other grouping. The cohort is usually represented in the denominator of any percentage that is calculated.
There will often be limits on your cohort, such as “cases open the entire month.” It is important to identify each of these limits and understand what information will be needed to add or exclude elements from the cohort.
Measure: Lastly, there will always be something about the cohort to count. In social services, these are likely to be demographics, events, or activities. The measure is usually represented in the numerator of any percentage that is calculated. As with the cohort, there will be limits on the measure that determine whether or not a specific demographic, event, or activity is counted.
For the question “Did workers in Unit X contact a higher percentage of clients last month than the same month a year ago?” these components break out as:
Time: Last month plus the same month a year ago.
Cohort(s): This question has two cohorts that are compared: Clients in cases open and assigned to Unit X during each timeframe. There are two limits here: Each case needs to be both open during the timeframe and assigned to Unit X during the timeframe. Note that we want to create a percentage of clients contacted, so we will use clients as the cohort, not workers or contacts. As a result, we also make an assumption that the cohort is limited to clients assigned to Unit X. The question should be revised to make this explicit.
Measure: Number of clients with a contact by a worker assigned to Unit X. The question is not specific about the type of contact, so any contact by a worker assigned to Unit X with the client will count. Again, if something different is intended, the question should be revised to state it explicitly.
Once you have identified these components and understand what information is needed, the next step is to identify which data elements provide that information. Case management system data is usually organized in small, linked fragments. To use this data, we need to bring these fragments together in a coherent way that reflects the state of practice for our cohort and timeframe. This is often referred to as “mapping” the data elements to the question. When doing this, it helps to have access to documentation and/or someone who is knowledgeable about how the data system is designed.
In many cases the exact data needed may not be in the case management system. When this happens, you will need to find proxies for the necessary information. Often, this is data that indicates an event happened, but doesn’t document the event itself. For example, an open foster care placement with no active setting may indicate that the child ran away from placement. Of course, the same data may also represent a foster child on a trial home visit. It is up to you and your knowledge of the case management system and agency practice to understand what is a good proxy and what assumptions you are making when using it.
Once data elements and their links are identified, the data is extracted and loaded into a tool that can help with the analysis. For simple questions, you may be able to use the same query tool for the extract and the analysis. More complicated questions may require additional tools like a spreadsheet or statistics program.
Whether you are going to extract data and analyze it yourself, or have others do it for you, it is important to understand the question and state it as specifically as possible. The better you understand and communicate what you are asking for, the more accurate and useful the answer will be.
case management UK services must work hand in hand with the HR services so that they can work together for the benefit of those facing hard times due to serious medical issues or temporary disabilities.