According to the World Health Organization (WHO), every day, worldwide, about 1,000 women die due to causes related to pregnancy or childbirth and, every year, more than eight million children in low and middle income countries die before reaching five years of age. The WHO was clear: maternal and child health is a topic of enormous medical importance and requires investments, projects, energy and commitment; it is an essential part of the public health of human populations. Improving the approach and access to health care, making qualified assistance, drug treatment and training of the operators more available, but also elementary preventive interventions during pregnancy, childbirth and the early years of a child's life, can prevent avoidable deaths and reduce several neonatal outcomes. Given the complexity of all the issues and problems concerning births and maternal and child health, through this thesis I propose a path divided into several stages which covers various topics starting from the socio-economic profile of the mother, moving to the pharmacological profile of pregnancy, up to the prevention of stillbirths. Several statistical methods were implemented to answer the different questions depending on the aim of each study. Log-binomial regression was used for estimating the association between the mother’s exposure during pregnancy and the selected neonatal outcomes. The fully conditional specification (FCS) model was performed to generate appropriate values of missing data for those women with missing covariates. The rule-out approach described by Schneeweiss was implemented to make our estimates, which might be affected by unmeasured confounder, more robust. The mediation analysis described by VanderWeele and Vansteelandt was used to assess the role that some adverse neonatal events at presentation (mediator) play in the relationship between the mother’s exposure during pregnancy (exposure) and adverse neonatal events later in life (outcome). Lastly, the Propensity Score Stratification derived from the predicted probability of treatment estimated in a logistic-regression model, as well as the high-dimensional propensity score algorithm to evaluate hundreds of inpatient diagnosis, procedures, and pharmacy claims, were completed to account for all potential confounders. The aim of my thesis is to identify factors to develop and improve the health care related to maternal- fetal and maternal-child world (before and after birth, respectively) from a sociodemographic, farmacoepidemiology, and clinical point of view. The layout of the thesis has been divided into different sections. I will proceed in the first instance by giving an overview of the methods used in the various studies carried out during my PhD, proceeding with a detailed description of the latter.

Secondo l'Organizzazione Mondiale della Sanità (OMS), ogni giorno, in tutto il mondo, circa 1.000 donne muoiono per cause legate alla gravidanza o al parto e, ogni anno, più di otto milioni di bambini nei paesi a basso e medio reddito muoiono prima di raggiungere i cinque anni di vita. L'OMS è stata chiara: la salute materno-infantile è un tema di enorme importanza sanitaria e richiede investimenti, progetti, energie e impegno, rappresentando una componente fondamentale della salute pubblica delle popolazioni umane. Migliorando l'approccio e l'accesso alle cure sanitarie, rendendo più disponibili assistenza qualificata, trattamenti farmacologici e la formazione degli operatori, ma anche semplici interventi preventivi durante la gravidanza, nel corso della gravidanza, del parto e dei primi anni di vita del bambino è possibile, infatti, prevenire le morti evitabili. A fronte della complessità dell’insieme dei temi e delle problematiche relative al percorso nascita e alla salute materno-infantile, ho ritenuto di suddividere in più fasi il mio progetto sviluppando diversi profili, come quello socio-economico, farmacologico, e clinico. Sono state utilizzate diverse metodologie statistiche a seconda dell’obiettivo dello studio. Le associazioni tra le diverse esposizioni della madre, prima e/o durante la gravidanza, e gli esiti neonatali selezionati, sono state valutate attraverso modelli di regressione logistica. In alcuni studi c’è stata la necessità di imputare i dati mancanti. Vista la natura di questi ultimi, arbitrary missing data pattern, è stato utilizzato il modello “fully conditional specification (FCS)”, ipotizzando l'esistenza di una distribuzione congiunta per le variabili mancanti. Per rendere le stime più robuste, a seguito della presenza di confondenti non misurati, è stato utilizzato l'approccio rule-out descritto da Schneeweiss. Per valutare il ruolo che gli eventi avversi neonatali alla nascita svolgono nel rapporto tra l’esposizione della madre durante la gravidanza e gli esiti neonatali in questione, ho utilizzato la “mediation analysis” descritta da VanderWeele e Vansteelandt. Infine, a seguito dell’utilizzo di dati osservazionali, le caratteristiche basali di esposti e non esposti potrebbero essere sbilanciate. Ho quindi utilizzato la tecnica del “propensity score startification” che permette di creare gruppi di pazienti con simile probabilità di ricevere il trattamento. Il propensity score è stato stimato sia attraverso il metodo classico della regressione logistica, sia utilizzando l’algoritmo “high-dimensionale propensity score” per valutare le centinaia di diagnosi, procedure, e prescrizioni più significative. Lo scopo della mia tesi è quello di identificare i possibili fattori per sviluppare e migliorare la salute materno-infantile da un aspetto socio-demograficho, farmacologico, e clinico. Ho strutturato la mia tesi in diverse sezioni. Procederò in prima istanza, dando una panoramica dei metodi utilizzati nei vari studi effettuati durante il mio dottorato, procedendo con una descrizione dettagliata di questi ultimi.

(2017). Maternal and Child Health. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2017).

Maternal and Child Health

CANTARUTTI, ANNA
2017

Abstract

According to the World Health Organization (WHO), every day, worldwide, about 1,000 women die due to causes related to pregnancy or childbirth and, every year, more than eight million children in low and middle income countries die before reaching five years of age. The WHO was clear: maternal and child health is a topic of enormous medical importance and requires investments, projects, energy and commitment; it is an essential part of the public health of human populations. Improving the approach and access to health care, making qualified assistance, drug treatment and training of the operators more available, but also elementary preventive interventions during pregnancy, childbirth and the early years of a child's life, can prevent avoidable deaths and reduce several neonatal outcomes. Given the complexity of all the issues and problems concerning births and maternal and child health, through this thesis I propose a path divided into several stages which covers various topics starting from the socio-economic profile of the mother, moving to the pharmacological profile of pregnancy, up to the prevention of stillbirths. Several statistical methods were implemented to answer the different questions depending on the aim of each study. Log-binomial regression was used for estimating the association between the mother’s exposure during pregnancy and the selected neonatal outcomes. The fully conditional specification (FCS) model was performed to generate appropriate values of missing data for those women with missing covariates. The rule-out approach described by Schneeweiss was implemented to make our estimates, which might be affected by unmeasured confounder, more robust. The mediation analysis described by VanderWeele and Vansteelandt was used to assess the role that some adverse neonatal events at presentation (mediator) play in the relationship between the mother’s exposure during pregnancy (exposure) and adverse neonatal events later in life (outcome). Lastly, the Propensity Score Stratification derived from the predicted probability of treatment estimated in a logistic-regression model, as well as the high-dimensional propensity score algorithm to evaluate hundreds of inpatient diagnosis, procedures, and pharmacy claims, were completed to account for all potential confounders. The aim of my thesis is to identify factors to develop and improve the health care related to maternal- fetal and maternal-child world (before and after birth, respectively) from a sociodemographic, farmacoepidemiology, and clinical point of view. The layout of the thesis has been divided into different sections. I will proceed in the first instance by giving an overview of the methods used in the various studies carried out during my PhD, proceeding with a detailed description of the latter.
CORRAO, GIOVANNI
SES-disparities; Antidepressant; Stillbirths; neonatal-outcomes; Pregnancy
SES-disparities; Antidepressant; Stillbirths; neonatal-outcomes; Pregnancy
MED/09 - MEDICINA INTERNA
English
10-apr-2017
SANITA' PUBBLICA - 78R
29
2015/2016
open
(2017). Maternal and Child Health. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2017).
File in questo prodotto:
File Dimensione Formato  
phd_unimib_718004.pdf

accesso aperto

Descrizione: tesi di dottorato
Tipologia di allegato: Doctoral thesis
Dimensione 3.01 MB
Formato Adobe PDF
3.01 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/158179
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact