What variables are included in SSA’s CDR profiling model and how have they changed over time?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive summary
The Social Security Administration’s continuing disability review (CDR) profiling model uses a mix of demographic, diagnostic, programmatic and earnings-related variables—what the agency calls a “profile score”—to decide whether to send a beneficiary to a full medical review (FMR) or to a mailer . Public reporting and agency documents show the model has been augmented and operationalized over decades (GAO), tweaked via rulemaking (Federal Register 2019), and its inputs and priorities have shifted modestly with technology and policy priorities .
1. What the profiling model looks like today: demographic, diagnostic, programmatic and earnings inputs
Contemporary descriptions of SSA profiling note core inputs include basic demographics (date of birth/age, gender), program tenure measures (date of program eligibility, years receiving benefits, diary category), primary disability diagnosis/impairment and medical-diary type (MIE/MIP/MINE), recent earnings or pre‑eligibility earnings quartile, prior CDR history and whether a beneficiary was sent a mailer—variables that feed into the computed “CDR profile score” used to rank cases for FMRs . SSA’s public data pages and research papers also show the Waterfall/profile files include date of birth, gender, diary type, CDR profile score, mailer receipt and clinical diagnosis fields—underscoring that the model mixes administrative, clinical-category and labor-market markers .
2. How the profile score is used operationally—thresholding and two-path processing
The profiling output is not a final decision but a triage score: SSA profiles cases to estimate likelihood of medical improvement, sends higher‑likelihood cases to State Disability Determination Services for full medical review, and uses mailers for lower‑profile cases; this two‑path approach has been central since SSA adopted computerized profiling in the 1990s and was highlighted by GAO as a cost‑effective shift away from universal FMRs . Researchers and SSA studies repeatedly note that the profile score appears in econometric models as a predictor and interacts with time‑since‑FMR and other covariates when studying outcomes [1].
3. How variables and emphasis have changed over time: incremental updates, rule changes and tech shifts
The model’s architecture has evolved incrementally rather than being rebuilt publicly: GAO’s historical review documented the initial move to computer profiles and mailers to prioritize FMRs , the Federal Register rulemaking in 2019 formalized changes to diary categories and reiterated profiling as the mechanism to decide FMR vs mailer , and more recent practitioner commentary and POMS guidance reflect greater reliance on electronic records and attention to mental‑health conditions in CDR focus . SSA research papers and administrative files show researchers and agency analysts add interaction terms and time‑varying hazard ratios (for example, letting the CDR profile score have different effects by year after an FMR) when modeling outcomes—indicating analytic sophistication in how the same underlying variables are used over time [1].
4. What’s debated, hidden or uncertain: incomplete public disclosure and policy incentives
SSA publishes descriptions of inputs but not a full, public variable‑by‑variable specification or the exact weights used to compute the profile score, creating uncertainty about model mechanics; GAO and SSA notes describe the broad categories used but stop short of a full recipe . Observers point out the profiling system serves operational priorities—targeting scarce DDS resources and reducing improper payments—so cost‑management incentives are implicit in how variables are selected and thresholds set; this policy framing is explicit in GAO’s appraisal of the profiling shift . Researchers therefore rely on administrative extracts (Waterfall/profile files) and modeling exercises to infer the model’s behavior and to test sensitivity [1].
5. Bottom line and limits of reporting
The CDR profiling model uses demographic (age, sex), diagnostic (primary impairment, diary category), programmatic (years in program, prior CDRs, mailer status) and earnings variables (recent or pre‑eligibility earnings measures) to produce a profile score that determines mailer vs FMR routing; changes over time have been evolutionary—new diary categories and procedural rules in 2019, greater use of electronic records and shifting attention (e.g., to mental‑health reviews), and more complex analytic treatments of the score in research—while the precise formula and weights remain nonpublic in SSA releases . The reporting reviewed documents inputs and operational shifts but does not provide a full, year‑by‑year public log of variable inclusion or the exact score algorithm, so definitive claims about every change over time cannot be made from the available sources .