自述工作滿意度及身體症狀與預測長期請病假之相關性研究

Abstract

前言：長期請病假不僅是與經濟相關的重要議題，亦是醫療健康上的重要問題。長期請病假可能造成員工的職場邊緣化並且與未來永久性失能有相當地關係。因此，為了預防長期請病假以及其後續造成的員工失能結果，找出具有長期請病假傾向的高風險族群並預測其發生的相關因子是相當地重要。本研究旨在探究與一年內及兩年內長期請病假之發生的相關危險因子，包含工作滿意度以及與健康相關之身體自覺症狀，並發展一可用於辨識長期請病假族群的預測模型。 研究方法：來自製造工廠的資料用於探討與長期請病假之相關危險因子。分析步驟分別為單變項羅吉斯迴歸分析、逐步汰選羅吉斯迴歸分析以及多變項羅吉斯迴歸分析。另使用ROC (receiver operator characteristic)曲線繪製，提供預測模型的建議切點值以及對應之敏感度、特異度。 研究結果：性別、年齡(大於34歲)、工作年資、疾病史、用藥史、手術史、吸菸、中度飲酒、從不運動、工作滿意度以及身體自覺症狀與一年內的長期請病假之發生相關。而性別、年齡(介於28-34之間、大於34歲)、工作年資、疾病史、用藥史、手術史、吸菸、中度飲酒、從不運動、工作滿意度以及身體自覺症狀與兩年內的長期請病假之發生相關。根據ROC曲線分析結果，工作滿意度以及身體自覺症狀納入一年內長期請病假之預測模型後，是具有顯著地影響力。當預測模型的切點為0.021時，辨識一年內及兩年內長期請病假發生之敏感度皆為80.7%；特異度則分別為44.4%及45.3%。 結論：在追蹤觀察期間，自述工作滿意度以及自覺身體症狀與長期請病假有顯著地相關。發展及早預測長期請病假之工具是可行且有意義的。

Introduction : Long sick leave is an economic as well as a medical problem. It tends to marginalize the employee form the workplace and is associated with risk of future disability pension. To prevent long sick absence and subsequent transition permanent disability, it is imperative to recognize predictive factors in order to screen employees at risk for long sick leave. The aim of this study is to investigate the extent to which variables including job satisfaction and somatic symptoms are useful in identifying at high risk of long sick leave, and to develop a predictive model for identifying employees at high risk of long sick leave within one and two years separately. Material and Methods: Data from manufacturing factories were used to identifying factors to be associated with an increased risk of long sick leave. The analytical procedures univariate logistic regression, backward stepwise logistic regression, multiple logistic regression were successively applied. Sensitivity and Specificity of cut-off points were determined by the receiver operator characteristic (ROC) analysis. Results: Within one year follow-up, results suggested that gender, age(>34), work tenure, disease history, medical history, surgery history, current smoker, moderate drinker, never exercise, self-reported job satisfaction and self-reported somatic symptoms were associated with long sick leave. Within two years follow-up, results suggested that gender, age(28-34; >34), work tenure, disease history, medical history, surgery history, current smoker, moderate drinker, never exercise, self-reported job satisfaction and self-reported somatic symptoms were associated with long sick leave. Based on ROC curve analysis, it revealed that areas under curve were significantly different (p-value=0.0159) within one year follow-up. It proposed that self-reported job satisfaction and somatic symptoms were significant predictors for long sick leave. A potential cut-off point of 0.021 on the predictive model in a sensitivity score of 80.7% within one year follow-up and 80.7% within two years follow-up and a specificity score of 44.4% within one year follow-up and 45.3% within two years follow-up. Conclusions: This study showed that self-reported job satisfaction and self-reported symptoms were significantly associated with long sick within follow-up period. It also proposed it was possible and valuable to develop an instrument for early identification of employees at risk for long sick leave.