OSCAR to CASPER
The Online Survey Certification and Reporting (OSCAR) system was a legacy administrative database of the Centers for Medicare and Medicaid Services (CMS). The OSCAR system was replaced by the Certification and Survey Provider Enhanced Reporting (CASPER) system and the Quality Improvement Evaluation System (QIES) in mid-2012. CASPER/QIES are part of a large relational database operating within CMS' Automated Survey Processing Environment (ASPEN). Outside organizations who bought OSCAR or CASPER from CMS received the data in fixed-length record format, a.k.a., “flat file” format. The 2012 transition from OSCAR to CASPER was welcomed by the research community for two important reasons. First, the OSCAR system was “destructive” inasmuch as new survey data would routinely overwrite older data. Also, as an added bonus to researchers, the CASPER data was better documented than the OSCAR data it replaced.
The administrative purpose of CASPER/QIES survey data is to support the survey and certification function. There are CASPER data for every "institutional" health care provider in the United States that is certified to provide services under either Medicare or Medicaid (or both). Different types of providers report different information during the survey and certification process. As a result, the CASPER/QIES survey data are a very rich source of information for some types of providers and not for others. The number of variables in the file reflects the intensity of the survey process for that provider type. The CASPER/QIES data for nursing homes are extensive, reflecting the depth of the nursing home survey process. A cursory review of the contents of OSCAR or CASPER data may be misleading because it could list variables that are “unpopulated” for that category of provider.
In 2020 CMS staff were required to make changes to the CASPER/QIES data system under very short notice. In the spring of 2020 CMS suspended standard annual surveys of nursing homes in order to put all their resources toward specially focused infection control Covid-19 surveys. There was a scramble among State Survey Agencies to complete the Focused Infection Control surveys of nursing homes so as to get as much of the performance-based funding that was available through the Coronavirus Aid, Relief and Economic Security (CARES) Act supplemental grants for State Survey Agencies. Later, in the fall of 2020, CMS resumed the normal survey process.
The 2020 changes required CMS staff to incorporate the results of the specially-focused infection control surveys into the CASPER/QIES data system, and to do so under very short notice. CASPER data purchased from CMS after November 2020 included two new “parts” containing the infection control survey data. Also, the post-Nov.2020 data included a format change, official variable name changes, and the elimination of the “descriptive” variables.
Format change – all data purchased post Nov. 2020 are provided in comma separated values (.csv) format instead of fixed-length record format. Going forward, input scripts need to be created/modified to avoid inadvertently converting character variables to numeric.
Variable name changes – some of the official variable names changed post-Nov.2020. In those instances where the variable names changed, the new variable names are more intuitive than the old.
Descriptive variables dropped – the new data no longer contain variables with the character string “desc” in the official variable name. If needed, the descriptive variables can be recreated using pre-Nov.2020 data by cross tabulating the associated variable with its description.
The most important difference between the pre- and post-Nov.2020 data is that the newer data, as of this writing, do not come with the comprehensive documentation that was characteristic of CASPER data from mid-2012 to August 2020. In addition to the often-useful descriptive variables, the 2012-2020 CASPER data came with file layout, data dictionary, metadata, and SAS input and labeling code. CMS staff are currently working to fully document the newer data. Until that process is complete, however, we retrofit the older data documentation to the new data. Cowles Research Group have created crosswalk files to track the variable name changes that occurred in late 2020, linking the old documentation to the new files, and we will continue to work with our clients to create whatever additional documentation is needed going forward.