Forecasting the risk of WRMSDs in home care nurses

• Author(s): Carneiro, P. ; Braga, A. C. ; Barroso, M. P.
• Date: 2012
• Persistent ID: http://hdl.handle.net/1822/20449
• Origin: RepositóriUM - Universidade do Minho
• Subject(s): WRMSDs; Nurses; Home care; Logistic regression
• Document Type: Conference Object
• Language: English

Description

Studies regarding work related musculoskeletal disorders (WRMSDs) in nurses have been carried out mostly in hospitals or in other institutional contexts. Information about this theme in providing home-based care is scarce. The main goals of this work are the characterization of musculoskeletal complaints in nurses who work at the Health Centers of the northern Portugal and that provide home-based care, the identification of the main risk factors present in the homecare context and the development of statistical models to forecast the risk in the same context. The principal methodology used in this work was a questionnaire developed in electronic format which was based on the “Standardized Nordic Questionnaire” for the analysis of musculoskeletal symptoms. It were used univariate models of binary logistic regression to estimate the risk of WRMSDs present in the practice of home-based care and also to assess which risk factors that could contribute to the appearance of complaints in the lumbar region in the professionals who provide homecare. The body areas with more musculoskeletal complaints are the back and the shoulders. The nurses who provide home care have nearly triple chance of having musculoskeletal complaints in the lumbar region than their counterparts of Health Centers (OR=3.19 (p<0.05), 95% Confidence Interval [1.26; 8.08]). We obtained various statistical models for forecast the risk of having low back complaints in home care nurses. From all of them was selected the one that presented more stability and reliability. The model performance was evaluated by ROC (Receiver Operating Characteristic) analysis yielding a value for the area under the ROC curve of 0.889 (p<0.05). This value reveals a high discriminating power, that is, the model is able to correctly forecast the complaints in the lumbar region in 88.9% of cases.