Journal of Healthcare Management - January/February 2014 - (Page 39)

t H e v alue of p atIents ' H andwr Itten c o MMents on HcaHps s urveys ta B l E 2 incidence of Patient Comments and Number of Words by Coded Category Mean Number of Words Number % 473 80.3 .00 .000 Positive 17 2.9 9.71 7.776 Neutral 53 9.0 3.72 8.977 Negative 28 4.8 12.25 15.253 Mixed 18 3.1 16.50 11.618 No comment the surveys contained at least one missing response. One way to handle this problem is to delete all surveys that contain missing data, but as the preceding example indicates, this option would eliminate patients who have answered most or all of the relevant questions in good faith and we would be deprived of their responses. The problem was handled instead by employing a multiple imputation procedure (Graham, 2009) whereby missing values are imputed for each case, and thus a complete data set can be used. The comment score variable is not normally distributed, so dummy codes were created for the analysis, with no comment coded as zero. Thus, the estimated coefficients resulting from the regression analysis effectively compared the positive, neutral, negative, and mixed codes to the no comment category (Cohen, Cohen, West, & Aiken, 2002). The HCAHPS data were aggregated into subscales, resulting in six composite measures: communication with nurses, communication with doctors, responsiveness of hospital staff, SD pain management, communication about medicines, and discharge information (Giordano, Elliott, Goldstein, Lehrman, & Spencer, 2010), along with two individual items relating to the hospital environment: clean room/ bathroom and quiet room. These measures were used as quantitative independent variables in a regression analysis. Ordinary least-squares regression was performed on the HCAHPS measures and the comment score dummy-coded variables. The two dependent variables were overall hospital rating and intent to recommend, which correspond to HCAHPS Questions 21 and 22. Such attitude and intention variables have been used as outcome measures in similar analyses performed in a variety of studies (e.g., Elliott et al., 2012; Otani, Kurz, Burroughs, & Waterman, 2003; Otani et al., 2012; Press Ganey, 2009; Qu et al., 2008; Rathert, May, & Williams, 2011). Results from our study show that negative comments significantly affect the prediction of the two global outcome measures (Table 3). Coefficients 39

Table of Contents for the Digital Edition of Journal of Healthcare Management - January/February 2014

Journal of Healthcare Management - January/February 2014
Contents
Interview With Kenneth R. White, PhD, FACHE, Associate Dean for Strategic Partnerships and Innovation and the University of Virginia Medical Center Professor of Nursing, University of Virginia School of Nursing
Team-Based Care at Mayo Clinic: A Model for ACOs
The Management Springboard: Eight Ways to Launch Your Career as a Healthcare Leader
The Role of a Public–Private Partnership: Translating Science to Improve Cancer Care in the Community Donna M. O’Brien and Arnold D. Kaluzny
The Value of Patients’ Handwritten Comments on HCAHPS Surveys John W. Huppertz and Robert Smith
Can Inbound and Domestic Medical Tourism Improve Your Bottom Line? Identifying the Potential of a U.S. Tourism Market
Success Factors for Strategic Change Initiatives: A Qualitative Study of Healthcare Administrators’ Perspectives

Journal of Healthcare Management - January/February 2014