Page 47 - JCBP-1-1

Journal of Clinical and
Basic Psychosomatics Emotional exhaustion and work–family conflict in physicians

were utilized to gather comprehensive data regarding 2.3. Statistical analysis
participants’ general information and sociodemographic One-way ANOVA for continuous variables and chi-
attributes, encompassing variables such as sex, age, squared tests for categorical variables were used to compare
marital status, education level, hospital grade, categories differences between groups because all continuous
of hospitals, department affiliation, work experience, and variables were found to have normal distributions,
professional title. To ensure methodological consistency, confirmed by the Kolmogorov–Smirnov one-sample test
all surveyors underwent a standardized training program. (all P > 0.05). The correlations between EE and WFC were
Throughout the measurement process, consistent guidance also examined using univariate linear regression analysis. In
was provided using both individual and collective accordance with the advice of the STROBE declaration ,
[28]
measurement approaches. Finally, the questionnaires the outcomes of unadjusted, minimally adjusted, and
were effectively reclaimed and meticulously stored for fully adjusted regression analyses were simultaneously
future reference. Each hospital and volunteer completed a investigated. The matched odds ratio needed to vary by at
formal consent form after fully understanding the aim of least 10% to be modified when covariance was included
the study. A total of 3000 questionnaires were distributed, in the model . The non-linear correlations were also
[24]
out of which 2705 were returned (response rate: 90.2%). found using generalized additive models (GAM). A two-
Due to incomplete data, 175 questionnaires were excluded, piecewise linear regression model was used to calculate
leaving 2530 analyzable questionnaires. the threshold effect in terms of the smoothing plot if a
2.2. Measures of EE and WFC non-linear correlation between WFC and EE was seen.
The inflection point was determined using the recursive
The Chinese Maslach Burnout Inventory Questionnaire technique with the highest model likelihood when the ratio
(CMBI) , which has three dimensions: EE, DP, and between EE and WFC became visible in a smoothed curve.
[25]
decreased personal accomplishment (RPA), was used to Finally, subgroup analyses were conducted using stratified
measure EE. There were 15 items in the CMBI scale, and linear regression analysis, and the interaction between
each dimension contains five items. The EE dimension different subgroups was examined using likelihood ratio
encompasses a set of five distinct indicators: (i) “I am very tests. All statistical analysis was performed using the R
tired;” (ii) “I am worried that work will affect my emotions;” statistical software program (http://www.r-project.org,
(iii) “I often feel exhausted;” (iv) “At the end of a day’s The R Foundation) and Empower-Stats (http://www.
work, I feel extremely tired;” and (v) “Recently, I have been empowerstats.com, X&Y Solutions, Inc., Boston, MA).
feeling a bit depressed.” Each item on the questionnaire Statistically significant P-values (two-tailed) were defined
was given one of seven grades, with seven being the highest as those <0.05.
and one being the lowest (where one is “never” and seven
is “every day”). The three CMBI subscales have Cronbach’s 3. Results
coefficients of 0.817, 0.765, and 0.670, respectively . 3.1. Baseline characteristics of participants
[26]
The Chinese version of the WFC Scale, which has two A total of 2530 participants, consisting of 1,356 (54.07%)
factors: WFC and FWC, was used to evaluate the impact of males and 1152 (45.93%) females, were included in our
[27]
work on family conflict . The WFC scale had 18 elements, study. The baseline characteristics of participants are
with nine items for each dimension. The items in the WFC listed in Table 1, which reveals that physicians in China
scale include: (i) “My work prevents me from participating have varying levels of education, and certain individuals,
in family activities;” (ii) “I spend too much time on work, particularly older physicians, may have qualifications
which means I don’t have enough time to participate in limited to the junior college level.
family activities;” (iii) “When I return home from work,
I usually feel emotionally drained and unable to fulfill my 3.2. Univariate analysis
family obligations;” and so on. The scores for each item The univariate analysis results are shown in Table 2.
ranged from one (never) to five (always). The Cronbach’s WFC, FWC, undergraduate (β = 1.99, 95% confidence
α coefficients of the WFC and FWC scales are 0.856 and interval [CI]: 1.33 – 2.66 vs. reference), postgraduate
0.744, respectively . (β = 3.02, 95% CI: 2.22 – 3.82 vs. reference), specialized
[16]
Both the CMBI scale and the WFC scale were self- hospital (β = 1.61, 95% CI: 0.90 – 2.33 vs. reference),
rated. Before the test, all fieldworkers received unified and work experience of 5–10 years (β = 0.85, 95%
methodological training. The repeated evaluation of the CI: 0.06 – 1.65 vs. reference) were all positively correlated
scales showed that the intraclass correlation coefficient of with EE (all P < 0.05). Conversely, female (β = −0.93,
the fieldworkers was >0.80. 95% CI: −1.48 – −0.38 vs. reference), secondary hospital

Volume 1 Issue 1 (2023) 3 https://doi.org/10.36922/jcbp.1013