Objectives:

The key question of this report, as posed by the Social Security Administration, asks, "Do the current criteria cited in SSA's Listing of Impairments ("Listings") for chronic renal failure correlate with an inability to work for 12 consecutive months?" The current Listings assume that all patients on dialysis have an impairment severe enough to prevent them from doing substantial gainful activity. Our goals were to determine whether existing evidence supported or refuted this assumption and to determine if there were alternative criteria that might be more predictive of inability to work.

Renal failure may be "acute," occurring from a sudden injury or illness, such as a blow to the abdomen, bacterial infection, or drug overdose or "chronic" as a result of conditions, such as glomerulonephritis, diabetes, hypertension, and heart disease. When chronic renal failure (CRF) becomes severe enough that replacement therapy (dialysis or kidney transplant) is required, the functional diagnosis is end-stage renal disease (ESRD).

Approximately 110 out of every 100,000 people are diagnosed with ESRD, and the United States Renal Data System (USRDS) estimated that more than 300,000 individuals in the United States had ESRD as of 1997. The average patient with ESRD has a survival time between 19 and 47 percent of the age-, sex-, and race-matched U.S. population.

Search strategy:

This project was divided into two phases. For Phase 1, we sought evidence in the published literature. We searched 27 databases (including MEDLINE® and EMBASE®) for relevant information. Most of these databases were last searched in late 1998, at the end of Phase 1 of this contract. Searches of the World Wide Web were also conducted.

Individual patient data of the USRDS Dialysis Morbidity and Mortality Study (DMMS) Wave 2 database were analyzed in Phase 2 of this contract

Selection criteria:

The search strategies identified 3,492 documents, books, and World Wide Web resources. A total of 503 documents were ordered and read in full. Fourteen studies were identified that analyzed predictors of employment, all of which attempted to correlate physiological, functional, and/or psychological factors with employment status.

Data collection and analysis:

Limitations in the 14 studies identified in Phase 1 precluded using them to evaluate the current criteria cited in the SSA Listing of Impairments for CRF. Therefore, we examined data in the DMMS Wave 2 special study of the USRDS. Patients in the DMMS Wave 2 study were followed prospectively for 1 year, and assessed physiologically, functionally, and psychologically at the beginning and end of the 1-year period. More than 300 variables were included in this database of 4,026 patients. Outcome variables included self-reported ability to work and work status.

We conducted numerous de novo statistical calculations on these data to determine the external validity, construct validity, reliability, and reproducibility of statistical analyses that could be performed using this database. We also conducted an illustrative analysis, including missing data imputation, data recoding, regression analysis, and assessment of diagnostic efficacy.

Main results:

These analyses demonstrated that neither the published literature nor the DMMS Wave 2 data could be used to answer the key question posed by SSA. The primary limitation of the DMMS database arose because, although 4,026 patients were initially included, approximately 43 percent were over age 65 and not eligible for our analysis, and an additional 37 percent were lost to followup. This reduced the data set to a small number of patients and a large number of potentially relevant variables. This, combined with the different results obtained with randomly selected halves of the database rendered the results of our analyses unreliable.

Conclusions:

Currently available evidence does not allow us to answer the key questions of this evidence report. Answering these questions would require a large-scale prospective study of patients who are followed rigorously for at least 1 year to monitor their functional, physiological, and disability status.