Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.780462
Title: Clinical prediction modelling to guide decision making for essential neonatal services in Kenyan hospitals
Author: Aluvaala, Martin Jalemba
ISNI:       0000 0004 7966 1046
Awarding Body: University of Oxford
Current Institution: University of Oxford
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
Introduction: More than 98% of neonatal deaths occur in low and middle income countries (LMICs) and hospitals in these regions are expected to play a key role toward achievement of SDG 3.2 "reduce neonatal mortality to 12 per 1000 live births and below". One approach in efforts to achieve this is to use data for decision making to improve quality of hospital care for neonates. Methods: Using the largest neonatal unit data set in Kenya (11,850 observations): (1) I conducted the first analyses of in-hospital mortality in Kenya using Kaplan-Meier and Competing Risks analysis, (2) developed, validated and explored the potential utility of two novel and simple prognostic models using the Prognosis Research Strategy (PROGRESS) recommendations. Model development was informed by a systematic review guided by the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) recommendations. Results: A tenth of the neonates studied died (839/9115, 9.2% [95% C. I 8.7 to 9.8%]). The overall median length of stay was 2 days (range of 0-98 days, IQR 1-5days). Competing risk analysis showed different cumulative incidence of death for the 1.5 to 1.99kg and 2 to 2.49kg weight categories that are often grouped together (1.5 to 2.49). In addition, the competing risk method provides a more clinically meaningful estimate of in-hospital death compared to Kaplan-Meier. Two simple prognostic models that use treatments (NETS) and signs and symptoms (SENSS) as predictors provide accurate predictions of in-hospital death. Discrimination by c-statistic is 0.89 for each model with calibration slope 0.90 for SENSS and 0.76 for NETS. Conclusion: Prognostic data on in-hospital mortality may inform decisions on treatment, demand and capacity planning as well as balancing of groups in clinical trials. Scale up of such data collection and analyses is required in LMICs to support efforts to improve neonatal survival.
Supervisor: English, Mike ; Collins, Gary S. ; Berkley, Jay Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.780462  DOI: Not available
Share: