Back to Annual Meeting

Back to Annual Meeting

APHA Scientific Session and Event Listing

Recorded presentation

Manuel Cifuentes¹, Rebecca Gore, PhD¹, Jon Boyer¹, Jamie Tessler¹, Angelo D'Errico, MD¹, Patrick Scollin², Laura Punnett¹, and The PHASE in Healthcare Research Team³. (1) Department of Work Environment, University of Massachusetts Lowell, One University Ave., Lowell, MA 01854, 978-934-3132, cif-vill@comcast.net, (2) Department of Community Health and Sustainability, University of Massachusetts Lowell, 3 Solomont Way, Suite 3, Lowell, MA 01854, (3) Center for Public Health Research and Health Promotion, University of Massachusetts Lowell, 3 Solomont Way, Lowell, MA 01854

Because of methodological concerns about self-reported psychosocial exposures we examined their agreement with data from the U.S. D.O.L. O*NET database. We obtained questionnaires from 340 workers in 24 hospital and nursing home jobs (minimum 5 respondents per job). Job characteristics were extracted from O*NET and exposure scales (demand/control (DC) and effort/reward (ER) ratios) were constructed. All data were continuous scales. Four measures of association were defined and stratified by working or not in a healthcare-specific job: job-level O*NET-survey correlation, linearity of the association (multilevel regression), under or over-estimation of survey scores (difference from predicted values), and dispersion of survey values around O*NET values (absolute value of difference). Finally, differences were modeled looking for explanatory variables. Healthcare-specific-stratum showed better agreement than non-healthcare-specific stratum; decision latitude, psychological demands, job strain, rewards and ER ratio showed strong associations at the job-level (Spearman correlation coefficients = 0.58, 0.76, 0.55, 0.47 and 0.88 respectively) and using multilevel linear regression (regression coefficients = 0.37, 0.57, 0.43, 0.08, and 0.82). Differences showed that there was not significant over or under-estimation; job- strain and ER-ratio were the most dispersed around the O*NET value (medians of absolute value of differences = 18.8 and 35.8). Association with explanatory variables varied according to whether the job was or not healthcare specific. Despite likely misclassification, O*NET-derived indicators of psychosocial exposure show positive associations with worker survey data when taking economic sector in account. Utilization of additional explanatory variables to adjust for under or overestimation could eventually lead to improved levels of concordance.

Learning Objectives:

• Participants will learn methods to study agreement between psychosocial measures from different levels and sources; and will learn about the potential value of the national O*NET database as a source of information about job psychosocial characteristics.

Keywords: Survey, Health Care Workers

Presenting author's disclosure statement:

Any relevant financial relationships? No

Recorded presentation