Analyzing the SSA Disability Evaluation Process

Clinical Center

Investigators

Linked publications, trials & patents

Abstract

Analytics (Obj 1) 1) Expanding SSAs Use of Electronic Medical Evidence Initial SSA decisions are made by examiners with little time to review evidence for a case, usually in the form of lengthy medical records from various healthcare providers. This focus area aims to develop, adapt, and apply methods that can process and leverage electronic medical records with a focus on identifying and characterizing language related to whole-person function. The ability to automatically review the electronic medical records aids SSA disability adjudicators to reach accurate, consistent, timely decisions in accordance with SSA regulations and available medical evidence. In FY23, the NIHs research focused on developing natural language processing (NLP) models for identifying function information related to criteria used in SSAs determinations, both at the initial determination and as part of SSAs continuing disability review (CDR) process. In addition, due to the growing focus on the potential impact of long COVID on disability claims and SSAs processes, we are conducting a rapid review on how artificial intelligence (AI) methods, like NLP, are being used to identify long COVID cases from medical records. NLP Models for Functioning Infor: The aim of this subproject is to extract functioning information from clinical documentation, an underdeveloped area of machine learning and NLP. In FY23, we made progress on developing and refining our NLP models across the four domains of mobility, self-care and domestic life, interpersonal interactions and relationships (social functioning), and communication and cognition. We also focused on extracting functioning information from entire cases rather than individual documents. This facilitates additional research around aggregating and ranking identified information and allows for better comparison with criteria used in SSAs business processes. Work this year also included development of a mapping table so that extracted functioning formation can be automatically aligned and compared with SSA function measures. In FY23, we continued developing models that can extract temporal information for descriptions of function in medical records, which can be used to build timelines that SSA adjudicators could use to evaluate claimants evidence related to function over time. Long COVID Rapid Review: In the wake of the COVID-19 pandemic, a new condition where individuals infected with SARS-CoV-2 continue to experience symptoms months after initial infection emerged - long COVID. Given the prevalence of COVID-19, SSA is expected to receive many new disability claims for individuals with long COVID. In FY23, the NIH is working with SSA to conduct a rapid literature review to identify ways that AI, including NLP, are being used to identify long COVID from electronic medical records. This reflects the way SSA employees are tasked with reviewing evidence to confirm the presence and severity of conditions, and could inform SSAs future use of support tools to help identify such conditions that are difficult to define and include a broad range of symptoms. 2) Strengthening SSAs Employment Support Programs In addition to the disability benefits programs that SSA administers, SSA also oversees a program called Ticket to Work, which provides career development to beneficiaries and supports beneficiaries pursuing opportunities to return to work. In FY23, we continued work on a research project with SSA to develop new ways to characterize individuals and occupations to inform and support return to work efforts. We leveraged data from multiple SSA sources to identify characteristics of beneficiaries who are most likely to participate in the Ticket to Work program and most likely return to work. WD-FAB development (Obj 2) In collaboration with the SSA, the NIH and Boston University developed a comprehensive and efficient assessment instrument called the Work Disability Functional Assessment Battery (WD-FAB). Contemporary models of disability indicate that in order to assess work disability, what individuals can do and what they are expected to do for work must both be assessed. The WD-FAB assesses what individuals can do. It is a 15-20-minute individualized assessment of functional activity that uses Item Response Theory (IRT) and computer adaptive technology (CAT), to select the most relevant test items from a large pool to measure self-reported functional ability. Item-based scoring means respondents do not need to answer all items or the same items to obtain comparative scores and scores are obtained efficiently. 3) Functional Assessment This focus area aims to develop new ways to collect, structure, and interpret functional data for use by SSA. This work will include development of the WD-FAB and methods to assist in interpreting WD-FAB results. WD-FAB instrument development: The aim of this subproject is to finalize the development of the WD-FAB so that it is ready for real-world, applied testing. The instrument now includes over 300 items across eight domains, four of which represent physical function (basic mobility, upper body function, fine motor function, and community mobility) and four of which represent mental health function (communication & cognition, resilience & sociability, self-regulation, and mood & emotions). Functional stages (e.g., low, moderate, high functioning) were developed by content experts to aid score interpretation. To date, the reliability and validity of the WD-FAB have been supported by a variety of evidence from a continuum of studies. WD-FAB Research Study: SSA is currently conducting a pilot study to evaluate the feasibility of incorporating the WD-FAB into their continuing disability review (CDR) process, where SSA periodically reviews current beneficiaries to ensure they continue to meet SSAs definition of disability. This longitudinal study collects WD-FAB data for a sample of beneficiaries currently undergoing a CDR twice in six months. Due to delays in SSAs data collection efforts, only the first wave of data was available for analysis in FY23. The second wave of data was recently finalized; all remaining analyses will be conducted in FY24. Medical vs. Functional Improvement: As part of SSAs CDR process, SSA employees must consider whether a beneficiary has demonstrated significant medical improvement such that they could return to work. However, it an individual may experience medical improvement without functional improvement and vice versa; the actual impact of these respective changes on work outcomes remains unclear. In FY23, we leveraged data available from SSAs Supported Employment Demonstration to examine whether and how medical improvement coincides with functional improvement and whether one shows greater association with work or benefit outcomes. Linking WD-FAB & the Occupational Requirements Survey: One of the key components of assessing the fit between a person and a job is to understand how functional abilities align with job demands. The NIH has previously conducted work outside of the SSA collaboration to align the WD-FAB with the Occupational Information Network (ONET), a database of nearly 1,000 occupations in the U.S. economy. In FY23, the NIH replicated the work linking the WD-FAB and ONET using the Occupational Requirements Survey (ORS), an ongoing data collection effort conducted by the Bureau of Labor Statistics sponsored by SSA. WD-FAB scales are mapped to corresponding requirements in ORS based on the type and difficulty of the activity represented by the requirement. Estimates provided by the ORS for each occupation are used to determine the likelihood that a given activity would be required for the occupation. Based on this, we can generate a WD-FAB profile to indicate the minimum functional ability a person needs to perform that occupation.