4083.0: Tuesday, November 14, 2000 - 9:06 AM

Abstract #13254

MEPS and CAHPS Quality of Care Calibration Results

George Karabatsos, PhD and William P. Fisher, PhD. Biometry & Genetics, Louisiana State University Health Sciences Center, 1901 Perdido Street, New Orleans, LA 70112, 504 568-4405, gkarab@lsumc.edu

This presentation examines the results of several analyses of the MEPS and CAHPS quality of care scales. Data are evaluated for instrument-sample targeting, measurement reliability, and data-model fit. These results will be presented for both the best-fitting and total samples for each survey. Targeting. A scale is on-target if 1) the average or median measure is near the center of the survey's item distribution, and if 2) the range of the item calibration distribution encompasses the range of the sample's measurement distribution. For example, the MEPS quality of care scale is significantly off target, with 4,936 cases, 31% of the best fitting sample of 16,135, returning responses in the highest category on all items. The average measure of 4.1 is about four errors of measurement (1.1 logits) above the center of the item scale (arbitrarily set to 0.0 logits). Reliability. For the MEPS, the all-cases (20,011; 23,559 with perfect extreme scores) modeled measurement reliability was about .65. For the best-fitting subset of 11,034 cases (16,135 with extremes), the modeled measurement reliability was .82 (.77 with extremes). Data-Model Fit. For the MEPS, the average and standard deviation (SD) of the outlier-sensitive mean square fit statistic (outfit) for the total sample were .88 and 1.1, respectively. For the best-fitting subset, the outfit mean (SD) was at .6 (.68). The low average outfit results from the scale's ceiling effect, such that a large proportion of cases to exhibit a spurious internal consistency caused by all responses being recorded in the same extreme category.

Learning Objectives: At the conclusion of this presentation, the participant learner will be able to: 1. recognize at least three new ways of assessing the value of quality of care survey data; 2. discuss the measurement properties of the MEPS and CAHPS Quality of Care scales; and 3. interpret logit measures and errors, and mean square model fit statistics, from the measurement analysis

Keywords: Accountability, Quality of Care

Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

The 128th Annual Meeting of APHA