- about DPGP -

Why "Potencial Pregnancy days lost" (DPGP in portuguese)?


Problem: In low-and middle-income countries (LMIC), in addition to the maternal and child health problems associated with poverty and poor access to safe care (“too little, too late”, TLTL), we also have problems associated to the unregulated, inappropriate use of interventions in childbirth (“too much, too soon”, TMTS, usually less visible to information systems). One of the effects of TMTS, especially of elective cesareans, is the shortening of pregnancy. Babies born in the "early term" period (37w 0/7d to 38w 6/7d) may have health outcomes more similar to those born preterm than those born in the "full term" period (> 39 weeks), including increased respiratory and metabolic complications, NICU admission, low birthweight, breastfeeding difficulties, and long-term effects. 
Our idea was to increase the granularity of gestational age (GA) data to days of pregnancy, and develop an innovative measure of GA, called "potential pregnancy days lost" (PPDL=GA-280 days), as both dependent and independent variable, to explore its association with maternal and child attributes, morbidity and mortality. Given the recent increase in infant and maternal mortality, we worked with local partners to optimize knowledge translation from the research and develop visualization alternatives for specific settings.
Methods: in partnership with the division of Live Birth Information System (SINASC) of the City Government of São Paulo (MSP) Health Department (SMS) we analyzed data from 1,525,759 live births in the city (2012-2019), 504,302 with GA in days, and at the national level, 8,854,727 live births, 3.329.339 with GA in days. Both datasets were linkage to the SIM (mortality information system), and in the city level, also with the hospital admission system (SIH).  Records without GA information in days were analyze in weeks of pregnancy, and alternatively, models of imputation of missing data were developed for the MSP base. Descriptive and stratified analysis included socio-demographic, obstetric and Robson groups, public or private sector, spatial analysis, and the use of Equiplots. Hazard ratio (HR) for early neonatal, neonatal and infant mortality were calculated for each day of GA during term period, controlling for maternal education, funding system, mode of delivery, color of skin and parity. We compared the data in days and in weeks, and the trends of GA and cesarean section rate in the period 2012-2019.
Some of the main insights and results are:
- Within the term period, in national and city level, all models show significant differences in HR neonatal mortality by days lost in early term period, confirming that “each day counts”. 
- The cesarean of the rich and of the poor women are different in GA, obstetric and clinical attributes, with the poorer with worse outcomes, reflecting differences in access to technology to compensate effects of PPDL (ex. NICU admission).
- Women with more education, living in higher HDI areas, tend to give birth in early term, losing more days in term period. In all analysis, we find an “inversion of the expected disparity” in GA, since in previous decades we would expect the opposite.
- For women with term pregnancies, the risk of maternal long hospital admission and readmission, is higher in women with shorter pregnancies.  
- The analysis of Robson groups is useful not only to indicate the excess of cesareans and calculate the loss of days by group, but also to study the overuse of induction (“high-intervention / low compassion care”). Quality of information about induction (ex. distinguishing “induction” from augmentation) is variable.
- Data literacy of providers and health managers is variable and tends to be low, so we are producing a website, podcasts and videos about the research, and oganizing a course on Data Literacy to be held in February/2021, in the USP School of Public Health, experimenting data visualization and knowledge translation, including an app about local mortality and GA trends. 
- We propose a small but highly promising change in the fields 31/32 of the SINASC form, to include GA in days, based on (1) the last menstrual period (LMP)  (2) ultrassound (US) with GA 

 

- common questions -

A pesquisa já tem resultados publicados?

Sim, veja a aba
Publicações e Resultados preliminares para saber mais.