Dataset Information

Dealing with missing data in the Center for Epidemiologic Studies Depression self-report scale: a study based on the French E3N cohort.

ABSTRACT:

Background

The Center for Epidemiologic Studies - Depression scale (CES-D) is a validated tool commonly used to screen depressive symptoms. As with any self-administered questionnaire, missing data are frequently observed and can strongly bias any inference. The objective of this study was to investigate the best approach for handling missing data in the CES-D scale.

Methods

Among the 71,412 women from the French E3N prospective cohort (Etude Epidémiologique auprès des femmes de la Mutuelle Générale de l'Education Nationale) who returned the questionnaire comprising the CES-D scale in 2005, 45% had missing values in the scale. The reasons for failure to complete certain items were investigated by semi-directive interviews on a random sample of 204 participants. The prevalence of high depressive symptoms (score ? 16, hDS) was estimated after applying various methods for ignorable missing data including multiple imputation using imputation models with CES-D items with or without covariates. The accuracy of imputation models was investigated. Various scenarios of nonignorable missing data mechanisms were investigated by a sensitivity analysis based on the mixture modelling approach.

Results

The interviews showed that participants were not reluctant to answer the CES-D scale. Possible reasons for nonresponse were identified. The prevalence of hDS among complete responders was 26.1%. After multiple imputation, the prevalence was 28.6%, 29.8% and 31.7% for women presenting up to 4, 10 and 20 missing values, respectively. The estimates were robust to the various imputation models investigated and to the scenarios of nonignorable missing data.

Conclusions

The CES-D scale can easily be used in large cohorts even in the presence of missing data. Based on the results from both a qualitative study and a sensitivity analysis under various scenarios of missing data mechanism in a population of women, missing data mechanism does not appear to be nonignorable and estimates are robust to departures from ignorability. Multiple imputation is recommended to reliably handle missing data in the CES-D scale.

SUBMITTER: Resseguier N

PROVIDER: S-EPMC3602286 | biostudies-literature | 2013 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Dealing with missing data in the Center for Epidemiologic Studies Depression self-report scale: a study based on the French E3N cohort.

Resseguier Noémie N Verdoux Hélène H Giorgi Roch R Clavel-Chapelon Françoise F Paoletti Xavier X

BMC medical research methodology 20130221

<h4>Background</h4>The Center for Epidemiologic Studies - Depression scale (CES-D) is a validated tool commonly used to screen depressive symptoms. As with any self-administered questionnaire, missing data are frequently observed and can strongly bias any inference. The objective of this study was to investigate the best approach for handling missing data in the CES-D scale.<h4>Methods</h4>Among the 71,412 women from the French E3N prospective cohort (Etude Epidémiologique auprès des femmes de l ...[more]

PMID: 23433105

Similar Datasets

Project description:BackgroundThe aims of this study were to conduct a cross-cultural validation of the Panic Disorder Severity Scale - Self-Report (PDSS-SR) and to examine psychometric properties of the French-Canadian version.MethodsA sample of 256 adults were included in the validation study based on data from the baseline interview of a clinical trial on transdiagnostic cognitive-behavioral therapy for mixed anxiety disorders. Participants completed the Anxiety and Related Disorders Interview Schedule (ADIS-5), and self-report instruments including the PDSS-SR, Beck Anxiety Inventory (BAI), Mobility Inventory for Agoraphobia (MIA), Sheehan Disability Scale (SDS), Patient Health Questionnaire (PHQ-9), Social Phobia Inventory (SPIN), Insomnia Severity Index (ISI) and Penn State Worry Questionnaire (PSWQ). The cross-cultural adaptation in French of the PDSS-SR included a rigorous back-translation process, with an expert committee review. Sensitivity to change was also examined with a subgroup of patients (n = 72) enrolled in the trial.ResultsThe French version of the PDSS-SR demonstrated good psychometric properties. The exploratory factor analysis supported a one factor structure with an eigenvalue > 1 that explained 64.9% of the total variability. The confirmatory factor analysis (CFA) corroborated a one-factor model with a good model fit. Internal consistency analysis showed a .91 Cronbach's alpha. The convergent validity was adequate with the ADIS-5 clinical severity ratings for panic disorder (r = .56) and agoraphobia (r = .39), as well as for self-report instruments [BAI (r = .63), MIA (accompanied: r = .50; alone: r = .47) and SDS (r = .37)]. With respect to discriminant validity, lower correlations were found with the SPIN (r = .17), PSWQ (r = .11), ISI (r = .19) and PHQ-9 (r = .28). The optimal threshold for probable diagnosis was 9 for the PDSS-SR and 4 for the very brief 2-item version. The French version showed good sensitivity to change.ConclusionsThe French version of the PDSS-SR has psychometric properties consistent with the original version and constitutes a valid brief scale to assess the severity of panic disorder and change in severity over time, both in research and clinical practice.

Project description:BackgroundDepression is among the top mental health problems with a major contribution to the global burden of disease. This study aimed at identifying the latent factor structure and construct validity of the Center for Epidemiologic Studies Depression (CES-D) Scale.Participants and settingA cross-sectional survey of 562 adults aged 18 years and above who were randomly selected from the Eritrean refugee community living in the Mai-Aini refugee camp, Ethiopia.MeasuresThe CES-D Scale, Primary Care PTSD (PC-PTSD) screener, premigration and postmigration living difficulties checklist, Oslo Social Support Scale (OSS-3), Sense of Coherence Scale (SoC-13), Coping Style Scale and fast alcohol screening test (FAST) were administered concurrently. Confirmatory factor analysis was employed to test prespecified factor structures of CES-D.ResultFirst-order two factors with second-order common factor structure of CES-D (correlated error terms) yielded the best fit to the data (Comparative Fit Index =0.975; root mean square error of approximation=0.040 [90% CI 0.032 to 0.047]). The 16 items defining depressive affect were internally consistent (Cronbach's ?=0.932) and internal consistency of the 4 items defining positive affect was relatively weak (Cronbach's ?=0.703). These two latent factors have a weaker standardised covariance estimate of 33% (24% for women and 40% for men), demonstrating evidence of discriminant validity. CES-D is significantly associated with measures of adversities, specifically, premigration living difficulties (r=0.545, p<0.001) and postmigration living difficulties (r=0.47, p<0.001), PC-PTSD (r=0.538, p<0.001), FAST (r=0.197, p<0.001) and emotion-oriented coping (r=0.096, p?0.05) providing evidence of its convergent validity. It also demonstrated inverse association with measures of resilience factors, specifically, SoC-13 (r=-0.597, p<0.001) and OSS-3 (r=-0.319, p<0.001). The two correlated factors model of CES-D demonstrated configural, metric, scalar, error variance and structural covariance invariances (p>0.05) for both men and women.ConclusionsUnlike previous findings among Eritreans living in USA, second-order two factors structure of CES-D best fitted the data for Eritrean refugees living in Ethiopia; this implies that it is important to address culture for the assessment and intervention of depression.

Project description:Aims: Depression is prevalent among university students worldwide, and the prevalence appears to be increasing. As an intermediate stage between being healthy and having depression, students with subthreshold depression could develop worsening depression or recover with intervention to prevent depression. The Center for Epidemiologic Studies Depression Scale (CES-D) is a useful tool to assess subthreshold depression. The primary purpose of the current study was to evaluate the psychometric characteristics of CES-D in Chinese university students. Secondly, we aimed to describe the prevalence of subthreshold depression among the student sample and examine its demographic correlates. Methods: A total of 2,068 university students participated in the study, and they were asked to respond to the Chinese CES-D, Beck Depression Inventory-II (BDI-II), and Positive and Negative Affect Schedule (PANAS). The factor structure was evaluated by conducting exploratory (EFA) and confirmatory factor analysis (CFA) using a structural equation modeling approach. The reliability was assessed by calculating Cronbach's alpha, inter-item correlation, and item-total correlation coefficients. The prevalence of subthreshold depression was calculated and demographic correlates of gender, grade, and major were examined by multiple regression. Results: The final sample included 1,920 participants. The EFA results suggested extraction of three factors (somatic symptoms, negative affect, and anhedonia) that account for 52.68% of total variance. The CFA results suggested that the newly derived model with 14 items was the best fit for our data. Six items were removed from the original scale (item 9, 10, 13, 15, 17, and 19). The Cronbach's alpha of the 14-item CES-D was 0.87. The prevalence of subthreshold depression among university students reached 32.7% for the 20-item CES-D and 31% for the 14-item CES-D, although there was no significant difference of prevalence in gender, grade, and major. Conclusions: The CES-D has good reliability and validity for assessing subthreshold depression in Chinese university students.

Project description:BACKGROUND:Depression is a prevalent and serious mood disorder and a major source of disability adjusted life years (DALY) in Uganda. Furthermore, evidence from Uganda and other countries throughout sub-Saharan Africa suggests that nearly a third of persons living with human immunodeficiency virus (HIV) suffer from depression and it adversely affects healthcare seeking behavior. The high burden of disease attributable to depression makes data on the prevalence of depression in Uganda, a country with a generalized HIV epidemic, a public health priority. This paper describes the psychometric properties of the Center for Epidemiologic Studies-Depression (CES-D) measure when administered to men and women residing in three fishing communities along the shore of Lake Victoria. METHODS:We applied methods based on item response theory and classical test theory approaches to assess individual item characteristics, conducted exploratory factor analysis and assessed internal reliability, and construct and content validity of the measure. All analyses were performed in R Studio. RESULTS:The study sample consisted of 300 residents of fishing communities in Wakiso District, Uganda. Fifty-six percent of the sample was female and 19.7% reported being HIV positive. Seven items of the measure that did not perform well, either because they could not differentiate between levels of the latent trait or because they did not map onto the primary factor, were removed from the scale. A single factor structure best fit our final set of 13-items and we found an overall coefficient alpha of 0.89, indicating high internal consistency in this population. CONCLUSIONS:Based on our findings, we recommend that future use of the CES-D in this population utilize our revised scale with the final set of 13-items. The addition of other measures that can improve the rigor of CES-D validation efforts, such as inclusion of a clinical depression measure and administration in both a clinical and a general population sample in this setting are needed.

Dataset Information

Dealing with missing data in the Center for Epidemiologic Studies Depression self-report scale: a study based on the French E3N cohort.

Background

Methods

Results

Conclusions

Publications

Dealing with missing data in the Center for Epidemiologic Studies Depression self-report scale: a study based on the French E3N cohort.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure