PSYCHOMETRIC ANALYSIS OF TWO SCALES MEASURING PATIENT SATISFACTION IN PHYSICAL THERAPY. Susan E. Roush*; A. Michelle Jones; Melissa Nassaney Physical Therapy Program, Univ. of Rhode Island, Kingston, RI PURPOSE: The purpose of this study was to compare the psychometric properties of the Physical Therapy Outpatient Satisfaction Survey (PTOPS) and the Patient Satisfaction Questionnaire (PSQ), specifically addressing missing data, ceiling effect, reliability and validity. BACKGROUNDS/SIGNIFICANCE: Measuring patient satisfaction is an important issue in physical therapy (PT). The emerging tools to measure this construct vary in psychometric sophistication and theoretical grounding. Two of the most developed tools are the PTOPS and the PSQ. These tools, however, were developed from different frames of reference and no data are available to compare/contrast their psychometric properties. Such data are needed to guide those interested in measuring patient satisfaction to match their needs with the most appropriate instrument. SUBJECTS: Subjects (n=156) from five outpatient clinics in RI provided data on both the PTOPS and PSQ. Clinics had no prior experience with either measurement tool. Mean subject age was 52.5 years (s.d.=15 yrs) with a range 18-88 years; mean educational level was 13.9 years. METHODS AND MATERIALS: The PTOPS is a 34-item survey that measures four aspects of patient satisfaction: Enhancers, Detractors, Location and Cost. The PSQ is a 20-item, unidimensional satisfaction survey. Both surveys have a five-choice Likkert format and are designed to be used with PT outpatients. Subjects at the identified clinics were approached during non-treatment time and asked to participate in the study. Informed consent was obtained and subjects anonymously completed both surveys, which were administered in alternating order. Subjects received a token gift (value = $0.50 - $1) for completing the surveys. ANALYSES: Data were entered into the computer program Statistical Package for the Social Sciences (SPSS) and a 100% error check was made. Both instruments were scored according to their respective protocols. Missing data were examined in two ways: across (sub)scales and per item. Four cases demonstrated missing data that met deletion criteria of >2 items on any PTOPS subscale or >4 PSQ items; these were dropped from the data set, leaving 152 cases. Other missing data were replaced with the mean subscale score for the PTOPS, and survey mean for the PSQ. Considered per item, 20% and 21% of subjects did not answer PSQ items #16 and #24, respectively. No individual PTOPS item was missing data from more than five subjects. The data set of 152 cases was analyzed: descriptive statistics and Cronbach alphas were calculated on the four PTOPS subscales and the PSQ data. Additionally, regression analyses assessed each scale's ability to predict an external satisfaction criterion measure. RESULTS: The PTOPS Enhancer subscale and the PSQ demonstrated a ceiling effect with 1.45 and 1.02 S.D.s separating their mean and their scales' maximum value, respectively. A minimum two S.D. separation is recommended to ensure adequate variability of responses. PTOPS subscale internal consistency ranged from .601- .862; PQS internal consistency was .939. Finally, the PTOPS accounted for 13% of the variance in an external satisfaction criterion measure while the PSQ accounted for 7.3%. CONCLUSIONS: The PTOPS is a generally reliable scale with marginal reliability seen on the Cost subscale; the PSQ is highly reliable. Users of the PSQ should be cautious of the ceiling effect which limits obtainable variability, a critical limitation. Additionally, PSQ items #16 and #24 may be problematic due to their high rates of missing data. The ceiling effect of one of the PTOPS subscales is also a concern, although it allows for more variability than the PSQ and only affects one of four subscales. Neither tool accounted for a high percentage of variance in the external satisfaction criterion measure, although the PTOPS demonstrated greater concurrent validity with this measure. Those interested in measuring patient satisfaction should be aware of each tool's strengths and limitations. Further research is indicated to understand this important construct in PT and how to measure it. FUNDING SOURCE: This project was supported by the Graduate Programs Fund of the College of Human Science and Services at the University of Rhode Island. KEYWORDS: psychometrics, measurement, Patient satisfaction Copyright 2009 by the American Physical Therapy Association. Requests for reprints should be directed to the corresponding author of the article. Educators, students, and other academic customers may receive permission to reprint copyrighted material from Physical Therapy (ISSN 1538-6724) by contacting the Copyright Clearance Center Inc, 222 Rosewood Dr, Danvers, MA 01923. Other types of customers who want permission to reprint should contact the APTA Editorial Office, Attn: Physical Therapy.