Depression Screening Tool Shows Promise for Teenage Mothers in Uganda, Study Finds

In Uganda, one in four pregnancies occurs among teenagers, and an estimated 20.7 percent of adolescent mothers experience postpartum depression, a condition that can quietly undermine both a young mother’s recovery and her child’s early development. Yet in many low- and middle-income countries, postpartum depression in teenage mothers remains chronically underdiagnosed, largely because mental health resources are scarce and few screening instruments have been formally validated for this specific, vulnerable population. A new cross-sectional study published in BMC Psychiatry by researchers from the Medical College of Wisconsin, Makerere University College of Health Sciences, the University of Ottawa, and collaborating institutions set out to answer a deceptively simple question: can a widely used, decades-old depression questionnaire reliably detect probable depression in postpartum adolescents attending an urban clinic in central Uganda?

The instrument under scrutiny was the Center for Epidemiological Studies Depression Scale, known universally in psychiatric research as the CES-D. Developed originally for community and primary care settings, the 20-item scale asks respondents to rate how often they experienced symptoms such as depressed mood, feelings of guilt or worthlessness, sleep disturbance, appetite changes, and psychomotor slowing during the past week. Each item is scored on a frequency scale, and total scores range from zero to sixty, with higher totals indicating more depressive symptom burden. The CES-D is attractive for resource-constrained settings because it is brief, inexpensive, self-report based, and does not require a clinician to administer. But its performance has never been rigorously tested in postpartum adolescents in Uganda, a group whose symptom presentation may differ from the adult populations in which the tool was originally calibrated.

The research team enrolled 200 adolescent mothers aged 11 to 19 years who were attending postnatal care at an urban-based health center in Kawempe, in central Uganda. The participants had a mean age of 16.0 years, with a standard deviation of 3.4 years, underscoring how young this patient population is. Every participant completed the CES-D using a cutoff score of 26 or higher to flag probable depression, and then underwent a structured diagnostic interview with the Mini International Neuropsychiatric Interview for Children and Adolescents, abbreviated as MINI-KID. The MINI-KID served as the diagnostic reference standard, generating depression classifications aligned with formal psychiatric criteria. This two-step design, screening tool followed by diagnostic interview, is the classic architecture of a validation study, allowing researchers to calculate how often the questionnaire agrees with the gold-standard assessment.

The headline numbers reveal a striking gap between symptom screening and formal diagnosis. On the CES-D, 31.5 percent of participants screened positive for depression, with a 95 percent confidence interval of 25.1 to 37.9 percent. Yet only 8.4 percent of the same sample, with a 95 percent confidence interval of 4.6 to 14.0 percent, met the MINI-KID criteria for depression. In other words, nearly a third of these young mothers reported depressive symptom levels above the screening threshold, but fewer than one in ten received a diagnostic classification of depression from the structured interview. This pattern of over-detection is common when community screening instruments are applied to populations with lower baseline prevalence, and it frames the study’s central analytical challenge: determining whether the CES-D’s cutoff is appropriately calibrated for adolescent mothers in this setting.

The psychometric analysis produced a nuanced performance profile. Against the MINI-KID reference standard, the CES-D at the prespecified cutoff demonstrated a sensitivity of 76.9 percent, with a 95 percent confidence interval spanning 46.2 to 95.0 percent, meaning it correctly identified roughly three-quarters of the adolescents who truly met diagnostic criteria. Specificity was 66.0 percent, with a confidence interval of 57.5 to 73.7 percent, indicating that about two-thirds of those without depression were correctly classified as negative. The most clinically reassuring figure was the negative predictive value: at 95.9 percent, with a 95 percent confidence interval of 91.3 to 98.5 percent, a negative CES-D screen was highly trustworthy, effectively ruling out depression for the overwhelming majority of adolescents who scored below the cutoff. The positive predictive value, by contrast, was just 17.2 percent, with a confidence interval of 8.6 to 29.4 percent, meaning that fewer than one in five adolescents who screened positive actually met diagnostic criteria on follow-up assessment.

Two additional statistical measures rounded out the evaluation. Internal consistency, assessed with Cronbach’s alpha, came in at 0.86, a value comfortably above the conventional 0.70 threshold that signals the 20 items are measuring a coherent underlying construct rather than a scatter of unrelated complaints. The area under the receiver operating characteristic curve, or AUROC, was 0.76, which the authors characterize as acceptable diagnostic accuracy. The AUROC summarizes the tool’s ability to discriminate between depressed and non-depressed individuals across all possible cutoff scores; a value of 0.5 would be no better than chance, while 1.0 would be perfect. At 0.76, the CES-D sits in the moderate-to-good discrimination range, suggesting the instrument captures genuine signal in this population even if the specific cutoff of 26 generates a substantial number of false positives.

The confidence intervals around some of these estimates deserve careful attention. The sensitivity estimate, in particular, carries a wide interval from 46.2 to 95.0 percent, a consequence of the small number of true depression cases in the sample. Because only about 8 percent of the 200 participants met MINI-KID criteria, the analysis rested on roughly 17 diagnosed individuals, and each additional or missing case shifts the sensitivity estimate considerably. This is a structural feature of validation studies in low-prevalence settings: the rarer the condition, the larger the sample needed to pin down sensitivity and positive predictive value with precision. The authors’ use of exact binomial, or Clopper-Pearson, confidence intervals reflects an appropriately conservative approach to quantifying this uncertainty rather than overstating the precision of their estimates.

The practical implications for clinical care flow directly from the asymmetry in the predictive values. Because the negative predictive value is so high, the CES-D functions well as a first-line triage instrument in this setting: a young mother who scores below the cutoff can be reasonably reassured and does not need immediate further psychiatric evaluation, which conserves scarce specialist resources. Conversely, because the positive predictive value is low, a positive screen should be treated as a flag for secondary diagnostic confirmation rather than a diagnosis in itself. The study’s conclusion states this explicitly: the CES-D demonstrated acceptable screening performance and high internal consistency in this urban sample of adolescent mothers, but given its low positive predictive value, positive screens require secondary diagnostic confirmation. In a health system where mental health professionals are few, routing every positive screen to a structured diagnostic interview such as the MINI-KID represents a feasible two-stage pathway that balances sensitivity at the population level with diagnostic precision at the individual level.

Beyond its immediate clinical message, the study addresses a broader equity issue in global mental health research. Screening tools are frequently exported across languages, cultures, and age groups on the assumption that a questionnaire validated in one population will perform equivalently in another, yet symptom expression, idioms of distress, and the meaning of items about fatigue or appetite can shift substantially across contexts, particularly for postpartum adolescents whose somatic complaints may overlap with normal postnatal recovery. By formally evaluating sensitivity, specificity, predictive values, internal consistency, and discriminant accuracy in the exact population where the tool would be deployed, the researchers provide an evidence base that most screening programs in similar settings lack. The work also carries ethical significance: the participants were adolescent mothers considered emancipated minors under Uganda National Council for Science and Technology guidelines, and written informed consent was obtained from all participants, with the study approved by The AIDS Support Organization Institutional Review Board and the Medical College of Wisconsin Institutional Review Board and conducted in accordance with the Declaration of Helsinki.

For the roughly one in four Ugandan pregnancies that occur among teenagers, and for the estimated one in five adolescent mothers experiencing postpartum depression, this study offers a measured but meaningful step forward. It does not declare the CES-D a perfect instrument; the wide confidence intervals, the modest specificity, and the low positive predictive value all counsel caution and confirmatory testing. What it does establish is that a brief, low-cost, 20-item questionnaire can serve as a defensible first filter in an urban Ugandan postnatal clinic, reliably ruling out depression for most young mothers while identifying a smaller group who warrant deeper evaluation. As maternal mental health gains recognition as a cornerstone of child health and development worldwide, validation studies of this kind, grounded in the realities of adolescent motherhood in low-resource settings, provide the technical foundation on which scalable, equitable screening programs can finally be built.

Subject of Research: Validation of the CES-D depression screening scale among postpartum adolescent mothers in urban Uganda

Article Title: Validity of the center for epidemiological studies depression (CES-D) scale for depression screening in postpartum adolescents in Uganda

Article References: Bollepalli, G., Abbo, C., Babikako, H. M., Loewus, J., Fahhoum, M., Annyapu, E., Racheal, B., James, M., Dickson-Gomez, J., Cassidy, L. D., & Anguzu, R. (2026). Validity of the center for epidemiological studies depression (CES-D) scale for depression screening in postpartum adolescents in Uganda. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08672-z

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08672-z

Keywords: CES-D, postpartum depression, adolescent mothers, Uganda, depression screening, psychometrics, MINI-KID, sensitivity, specificity, predictive values, global mental health, BMC Psychiatry