Body image dissatisfaction and lower self-esteem as major predictors of poor sleep quality in gynecological cancer patients after surgery: cross-sectional study

The present descriptive and cross-sectional analysis was carried out over a period of 10 months, in the Mohammed VI Center for the Treatment of Cancers, Ibn Rochd University Hospital in Casablanca, Morocco. A total of 104 women, who had undergone radical surgery for gynecological cancer, were invited to take part of the study, when they were at the oncology department for follow-up counseling. Of the 104 individuals approached, 100 agreed to participate. The inclusion criteria were: 1. 18 years old or older; 2. confirmed diagnosis of primary gynecological cancers (cervical, body uterus or ovarian), 3. T1–T3 tumor stages, followed by hysterectomy/oophorectomy for more than 3 weeks and not more than 6 months; 4. signed written informed consent (for unschooled patients a verbal consent followed by a fingerprint were used as a consent proof); 5. Willing and able comply with the study conditions. Exclusion criteria were: 1. not born in Morocco; 2. T4 tumor stage; 3. unable to understand and speak the Moroccan Arabic dialect; 4. Cognitive impairment or hearing loss; 5. history of any psychiatric disorder; 6. previous diagnosis of primary cancer in a different location.

Oncologists were randomly invited women who consulted for a check-up to take part of the study. The researcher verbally presented the study to the patients and woman who did not wish to participate were excluded. The researcher read each question and its answers, and then ticked the selected answer. The average time dedicated to fulfill the questionnaire was between 20 and 30 min. Patients were recruited consecutively until the sample size was sufficient. Socio-demographic and medical data were directly collected from patients and supplemented by medical records (Additional file 1).

Descriptive statistics were performed to describe socio-demographic and medical characteristics participants. Verification of normality and homogeneity of variances was carried out by a Shapiro–Wilks test. These conditions were not met in our study. As a result, Fisher’s exact bilateral test to compare percentages, and the Spearman rank correlation coefficients were calculated to study the correlation between two quantitative variables. The difference between the groups for a continuous variable was tested by the Mann–Whitney test (in the presence of two modalities) or the Kruskal–Wallis test (for comparison of several averages). To identify predictors of sleep quality, the multiple linear regression was used including the variables most correlated with the PSQI score. To determine the unique impact of mental health variables on sleep quality, three sets of predictors have been introduced in stages in a hierarchical multiple regression model; these sets of predictors were: demographic (age), social (medical coverage system) and mental (anxiety, depression, body image dissatisfaction, self-esteem). Demographic variables were included in the first step. The medical coverage system was related to the quality of sleep in our sample and was therefore included as a potential social confounding factor in the second step. In the third step, we entered the scores of anxiety, depression, dissatisfaction with body image and self-esteem.

The Socio-demographic and clinical characteristics of the participants were represented in the Table 1.

Sixty-six % of the patients reported symptoms of anxiety; 59% reported severe symptoms of depression, and 54% reported body image problems.

Thirty-nine % reported very poor subjective sleep quality, and 14% reported somewhat poor quality. For sleep onset latency, 35% needed more than one hour to fall asleep, and 20% needed between 31 and 60 min to fall asleep. The sleep duration did not exceed 5 h per night for 42% of the participants, with a median of 5 h per night (Table 2).

Forty-seven % of participants had usual sleep efficiency less than 65%, while the latter exceeded 85% in 18% of participants. The sleep disorders in the participants were explained by several reasons, mainly citing: waking up in the middle of the night or early in the morning (68%), getting up to use the bathroom (62%), having nightmares (42%), feeling hot (38%), having breathing problems (35%), having pain (26%), feeling cold (18%) and coughing and snoring loudly (16%). Only 21% of the participants claimed the use of sleep medication and 28% had daytime dysfunction. Finally, 78% of the participants were considered to be poor sleepers (Fig. 1).

Higher PSQI scores were significantly associated with younger age (p = 0.001, r = −0.32). PSQI scores significantly tended to be higher among unmarried women (single/divorced/widowed), those who had no children, those who had Ramed, those who worked, those who had lower economic status and women with time since diagnosis did not exceed 1 year (p < 0.05) (Table 3).

Higher PSQI scores were positively correlated with higher symptoms of anxiety (p < 0.001), depression (p < 0.001), and body image dissatisfaction (p < 0.001), and negatively correlated with lower self-esteem (p < 0.001) (Table 4).

The main predictors of sleep quality according to multivariate analysis were: medical coverage system (p = 0.024), body image dissatisfaction (p = 0.009) and self-esteem (0.035) (Table 5).

According to the hierarchical regression analysis correlating the PSQI scores to the mental distress variables, after checking the key assumptions of the multiple regression (the linear relationship between independent variables and the dependent variable, the normality of residuals, the uncorrelated residuals, as well as the homoscedasticity) and after controlling for socio-demographic variables (age and medical coverage system), greater dissatisfaction with body image and lower self-esteem contributed significantly to the quality of sleep (Table 6).