Causal Effects of ICT Infrastructure on Global Health Outcomes With DML and Causal Forest Approaches
Implications of Information and Communication Technologies for Global Public Health, Gamze Sart,Yılmaz Bayar,Funda Hatice Sezgin, Editör, IGI Global, Pennsylvania, ss.145-173, 2027
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2027
- Doi Numarası: 10.4018/979-8-3373-8909-7.ch006
- Yayınevi: IGI Global
- Basıldığı Şehir: Pennsylvania
- Sayfa Sayıları: ss.145-173
- Editörler: Gamze Sart,Yılmaz Bayar,Funda Hatice Sezgin, Editör
- İstanbul Üniversitesi-Cerrahpaşa Adresli: Evet
Özet
In recent years, the rapid development of Information and Communication Technologies (ICT) has accelerated the momentum of digital health transformation. In this direction, the effect of the ICT infrastructure level on global health outcomes has become a problem investigated in the literature. However, in addition to digital infrastructure, the non-linear relationships caused by country-specific economic and governance-related factors that may potentially affect global health outcomes make it difficult to analyze the individual effect of ICT infrastructure. The existing literature mostly addresses this problem through econometric and parametric approaches and does not sufficiently consider non-linear relationships and cross-country heterogeneity. In this study, a Two-Way Fixed Effects (TWFE) regression model is included as an initial comparison model on a panel dataset consisting of 3,058 observations covering 236 countries between 2010 and 2022; after estimating causal effects using an XGBoost-based Double Machine Learning (DML) approach, the cross-country heterogeneity of these effects is examined using the Causal Forest method. Finally, the risk of reverse causality is tested through lagged treatment analysis. This multi-stage methodological framework aims to provide a comprehensive causal analysis. In this context, the effects of four main ICT infrastructure indicators on five global health outcome parameters are analyzed by considering six control variables. The findings show that ICT infrastructure indicators generate statistically significant causal effects at different levels on public health parameters, while also revealing that actions and policy decisions regarding ICT infrastructure should be planned holistically by taking into account the important effect of ICT on the health system.