• Kurt Kroenke, Timothy E Stump, Jacob Kean, Tasneem L Talib, David A Haggstrom, and Patrick O Monahan.
• Division of General Internal Medicine and Geriatrics, Indiana University School of Medicine, Indianapolis, IN, USA; Regenstrief Institute, Inc., Indianapolis, IN, USA. Electronic address: kkroenke@regenstrief.org.
• J Clin Epidemiol. 2019 Nov 1; 115: 116-124.

ObjectiveThe 5 SPADE (sleep, pain, anxiety, depression, and low energy/fatigue) symptoms are among the most prevalent and disabling symptoms in clinical practice. This study evaluates the minimally important difference (MID) of Patient-Reported Outcomes Measurement Information System (PROMIS) measures and their correspondence with other brief measures to assess SPADE symptoms.Study Design And SettingThree hundred primary care patients completed a 4-item PROMIS scale, a numeric rating scale (NRS), and a non-PROMIS legacy scale for each of the 5 SPADE symptoms. Optimal NRS cutpoints were examined, and cross-walk units for converting legacy measure scores to PROMIS scores were determined. PROMIS scores corresponding to standard deviation (SD) and standard error of measurement (SEM) changes in legacy scores were used to estimate MID.ResultsAt an NRS ≥5, the mean PROMIS T-score exceeded 55 (the operational threshold for a clinically meaningful symptom) for each SPADE symptom. Correlations were high (0.70-0.86) between each PROMIS scale and its corresponding non-PROMIS legacy scale. Changes in non-PROMIS legacy scale scores of 0.35 SD and 1 SEM corresponded to mean PROMIS T-scores of 2.92 and 3.05 across the 5 SPADE symptoms, with changes in 0.2 and 0.5 SD corresponding to mean PROMIS T-scores of 1.67 and 4.16.ConclusionA 2-step screening process for SPADE symptoms might use single-item NRS scores, proceeding to PROMIS scales for NRS scores ≥5. A PROMIS T-score change of three points represents a reasonable MID estimate, with two to four points approximating lower and upper bounds.Copyright © 2019 Elsevier Inc. All rights reserved.

21 keywords...

hide…