October 12, 2023
Background
In the spring of 2023, we (DATIS) discovered duplicate client records were being created when NLO data submissions were integrated with historical data in the OSP database. Investigation into the issue revealed large-scale data quality issues with client OHIP numbers being submitted. Client OHIP # is the only unique client identifier in the OSP database, but having an OHIP # is not mandatory for service. Per the OSP Data Dictionary, data entry guidance for client records with no OHIP # is to submit a unique Client ID. Further investigation revealed that a large number of records submitted to the OSP database do not contain a truly unique Client ID, and identified this to be the cause of the issue. Please see the OHIP Number Data Quality Issue report shared with the OSP program in June where details of the analysis, immediate data cleaning process, and interim solution were detailed. In the interim, we have asked that any changes to a client's OHIP #, or Unique Client ID (where OHIP # is absent) be communicated to DATIS prior to the bi-monthly data submissions until a final resolution is put in place. This has been a necessary process to ensure that the client records submitted are integrated with the corresponding client records in the OSP database.
Resolution
Based on the limitations of the OHIP # and Unique Client ID data elements in the OSP database (outlined in the OHIP Number Data Quality report linked above), it is necessary for incoming client records to include a persistent unique identifier that can be used at the programmatic level. In consultation with the CoE, the addition of a new data element, OSP Client ID, will resolve this issue. This new data element is in addition to the current Unique Client ID populated by NLO and Low Intensity Service Providers, and Client ID 2 populated by iCBT providers.
Rationale
Although this will be a new data element in the OSP MDS, we would like to leverage the unique identifier assigned to each client in your vendor system. This unique identifier may not show on your application screens, but is available at the database level.
The current OHIP # and Unique Client ID data elements are not meeting the needs of the OSP database for various reasons (as detailed in the report linked above). When implemented, the OSP Client ID will accompany all client records submitted to the OSP database providing a persistent unique client identifier by which all incoming client records will be validated against.
Requirements
The new OSP Client ID will be an enforced, mandatory data element for each submission to the OSP database - records without an OSP Client ID will be rejected. To reduce the possibility of two vendor systems generating the same OSP Client ID for different clients, the ID will need to be:
- Auto-generated by the vendor system
- Auto-generating the ID will ensure it is not manually constructed and cannot be manually edited.
- Unique to each client
- Alphanumeric - containing a combination of letters and/or numbers
- Where a vendor system auto generates an ID that includes a space or special character (such as a hyphen), it will be removed during the data ingestion process (as is currently done with OHIP #s).
The OSP data specs will not prescribe a format for creation of the ID to allow the vendor systems to continue using their current process and allowing more variability in the IDs produced across the vendor systems.
How it will work/How OSP Client ID will solve the problem
As described in the OHIP Number Data Quality Issue report, all incoming records go through an ETL (extract, transform, load) process that looks to match them with existing records for a client in the OSP database. This process will continue with the client OHIP # being the first identifier used to match to existing records (if records exist). In the event that an incoming client OHIP # does not match to any existing records, the ETL will use the persistent (un-changing) OSP Client ID as the contingency. If in these cases, the OSP Client ID is matched to an existing record, the ETL logic will add the incoming data to that client's record.
Note:
- Client OHIP # will continue to be the truly unique client identifier in the OSP database and is expected to be submitted when available.
- At this time, the current data elements Unique Client ID and Client ID 2 will remain part of the OSP data set. Data in these fields will continue to be submitted to the OSP database.
Implementation
The following steps will be taken to implement the OSP Client ID:
- Presentation at the October 16th Data Working Group including discussion.
- Extract of select data elements, including the OSP Client ID for initial load in the OSP database.
- Integration of the OSP Client ID in the OSP database.
- Bi-monthly data submission including the new OSP Client ID data element.
Following the Data Working Group meeting on October 16th, detailed requirements and information outlining the process will be released.
