Strengthening capacity of health workers to diagnose birth defects in Ugandan hospitals from 2015 to 2021.

BACKGROUND: Limited diagnostic capabilities, resources and health worker skills have deterred the advancement of birth defects surveillance systems in most low- and middle-income countries (LMICs). Empowering health workers to identify and diagnose major external birth defects (BDs) is crucial to establishing effective hospital-based BD surveillance. Makerere University-Johns Hopkins University (MU-JHU) Research Collaboration BD Surveillance System consists of three diagnostic levels: (1) surveillance midwives, (2) MU-JHU clinical team, and (3) U.S. Centers for Disease Control and Prevention (CDC) birth defects subject matter experts (SMEs) who provide confirmatory diagnosis. The diagnostic concordance of major external BDs by surveillance midwives or MU-JHU clinical team with CDC birth defects SMEs were estimated. METHODS: Study staff went through a series of trainings, including birth defects identification and confirmation, before surveillance activities were implemented. To assess the diagnostic concordance, we analyzed surveillance data from 2015 to 2021 for major external BDs: anencephaly, iniencephaly, encephalocele, spina bifida, craniorachischisis, microcephaly, anophthalmia/microphthalmia, anotia/microtia, cleft palate alone, cleft lip alone, cleft lip with cleft palate, imperforate anus, hypospadias, talipes equinovarus, limb reduction, gastroschisis, and omphalocele. Positive predictive value (PPV) as the proportion of BDs diagnosed by surveillance midwives or MU-JHU clinical team that were confirmed by CDC birth defects SMEs was computed. PPVs between 2015 and 2018 and 2019-2021 were compared to assess the accuracy of case diagnosis over time. RESULTS: Of the 204,332 infants examined during 2015-2021, 870 infants had a BD. Among the 1,245 BDs identified, 1,232 (99.0%) were confirmed by CDC birth defects SMEs. For surveillance midwives, PPV for 7 of 17 BDs was > 80%. For the MU-JHU clinical team, PPV for 13 of 17 BDs was > 80%. Among surveillance midwives, PPV improved significantly from 2015 to 2018 to 2019-2021, for microcephaly (+ 50.0%), cleft lip with cleft palate (+ 17.0%), imperforate anus (+ 30.0%), and talipes equinovarus (+ 10.8%). Improvements in PPV were also observed among MU-JHU clinical team; however, none were significant. CONCLUSION: The diagnostic accuracy of the midwives and clinical team increased, highlighting that BD surveillance, by front-line health care workers (midwives) in LMICs is possible when midwives receive comprehensive training, technical support, funding and continuous professional development.

Mobile tablets for real-time data collection for hospital-based birth defects surveillance in Kampala, Uganda: Lessons learned.

Sustainable birth defects surveillance systems provide countries with estimates of the prevalence of birth defects to guide prevention, care activities, and evaluate interventions. We used free and open-source software (Open Data Kit) to implement an electronic system to collect data for a hospital-based birth defects surveillance system at four major hospitals in Kampala, Uganda. We describe the establishment, successes, challenges, and lessons learned from using mobile tablets to capture data and photographs. After intensive training, surveillance midwives collected data using Android tablets with inbuilt logic checks; another surveillance midwife checked the quality of the data in real-time before data were securely uploaded onto a local server. Paper forms were used when needed as a backup for the electronic system. We experienced several challenges implementing the surveillance system, including forgotten passwords, unstable network, reduced tablet speed and freezing, loss of touch-screen sensitivity, decreased battery strength, and repetitive extensive retraining. We addressed these challenges by backing up and removing all photos from the tablet, uninstalling irrelevant applications to the study to increase storage space and speed, and monitoring and updating the system based mainly on feedback from the midwives. From August 2015 to December 2018, surveillance midwives documented information on 110,752 births at the participating hospitals. Of these, 110,573 (99.8%) were directly entered into the electronic data system and 179 (0.2%) were captured on paper forms. The use of mobile tablets for real-time data collection was successful in a hospital-based birth defects surveillance system in a resource-limited setting. Extensive training and follow-up can overcome challenges and are key to preparing staff for a successful data collection system.