Reliability of newborn and child quality of care indicators in health facilities receiving support from the USAID ASSIST Project in Antigua

The USAID Applying Science to Strengthen and Improve Systems (ASSIST) Project has implemented Zika-related health systems strengthening efforts in Antigua since July 2018. To assess the validity of the quality improvement data being collected by ASSIST-support quality improvement teams, understand the perceptions and practices of teams with respect to data, and to recommend potential corrective actions, we carried out a validation study in six primary health care facilities.  Data were extracted for deliveries (n=47) at the public hospital and well-baby visits (n=306) for seven indicators being collected at six primary health care facilities and compared to data reported in collection forms submitted to the project’s Improvement Indicator Database. Eight interviews were conducted with key informants and health facility personnel.

The study found that the difference between main indicators as measured by the validation study as compared to the reported data was 4 to 47 percentage points, with the validation study finding lower values for six out of seven indicators. Qualitative results pointed to training, staff workload, tools and equipment, and the perceived benefits of data as influencers of data quality.

Key recommendations for newborn indicators are to use a single form for indicators to be kept in the maternal chart along with more precise guidance on indicator definitions. With regard to well-baby indicators, it is recommended that clinical providers conduct and monitor clinical encounters completely, that data entry personnel have a clear understanding of indicator definitions, and that internal audits of data quality be conducted.


Report Author(s): 
Chris Andersen | Astou Coly | Tamar Chitashvili | Salwan Hager | Shivon Belle
USAID Applying Science to Strengthen and Improve Systems (ASSIST) Project/URC
ASSIST publication: 
ASSIST publication
Facebook icon
Twitter icon
LinkedIn icon
e-mail icon