TY - JOUR
T1 - Joint modelling of competing risks and current status data
T2 - an application to a spontaneous labour study
AU - Lee, Youjin
AU - Wang, Mei Cheng
AU - Grantz, Katherine L.
AU - Sundaram, Rajeshwari
N1 - Funding Information:
The authors thank Dr Guoqing Diao for insightful comments during the Conference on Life-time Data Science. The research was partly supported by the intramural research programme of the Eunice Kennedy Shriver National Institute of Child Health and Human Development. The authors also acknowledge that this study utilized the high performance computational capabilities of the Biowulf Linux cluster at the National Institutes of Health, Bethesda, Maryland (http://biowulf.nih.gov).
Funding Information:
The authors thank Dr Guoqing Diao for insightful comments during the Conference on Lifetime Data Science. The research was partly supported by the intramural research programme of the Eunice Kennedy Shriver National Institute of Child Health and Human Development. The authors also acknowledge that this study utilized the high performance computational capabilities of the Biowulf Linux cluster at the National Institutes of Health, Bethesda, Maryland (http://biowulf.nih.gov).
Publisher Copyright:
© 2019 Royal Statistical Society
PY - 2019/8
Y1 - 2019/8
N2 - The second stage of labour begins when the cervix is fully dilated and pushing begins until the fetus is delivered. A Caesarean delivery (CD) or operative vaginal delivery (OVD) is typically encouraged after the recommended time set by ‘expert consensus’. This recommended time has been set out of concern for an increased chance of maternal and neonatal morbidities due to a prolonged second stage of labour, but without thorough consideration of heterogeneous risks for spontaneous vaginal delivery (SVD) and morbidities among women. To provide quantitative evidence for the recommendation, the first step is to compare the risks for SVD, CD or OVD, and the risks of maternal or neonatal morbidities simultaneously across the duration of the second stage of labour. To address such risk comparisons statistically, one needs to study the joint distribution for the time to delivery due to each mode and time to maternal or neonatal morbidity given information provided for each individual. We introduce a joint model which combines the competing risks data for delivery time and current status data for any type of maternal or neonatal morbidity given each woman's baseline characteristics. These two processes are assumed dependent through individual-specific frailty under the joint model. Our numerical studies include a simulation that reflects the structure of observed real data and a detailed real data analysis based on nearly 12000 spontaneous labours. Our finding indicates the necessity to incorporate maternal characteristic such as age or body mass index in assessing the probability for delivery due to SVD, CD or OVD and the onset of morbidities across the second stage of labour.
AB - The second stage of labour begins when the cervix is fully dilated and pushing begins until the fetus is delivered. A Caesarean delivery (CD) or operative vaginal delivery (OVD) is typically encouraged after the recommended time set by ‘expert consensus’. This recommended time has been set out of concern for an increased chance of maternal and neonatal morbidities due to a prolonged second stage of labour, but without thorough consideration of heterogeneous risks for spontaneous vaginal delivery (SVD) and morbidities among women. To provide quantitative evidence for the recommendation, the first step is to compare the risks for SVD, CD or OVD, and the risks of maternal or neonatal morbidities simultaneously across the duration of the second stage of labour. To address such risk comparisons statistically, one needs to study the joint distribution for the time to delivery due to each mode and time to maternal or neonatal morbidity given information provided for each individual. We introduce a joint model which combines the competing risks data for delivery time and current status data for any type of maternal or neonatal morbidity given each woman's baseline characteristics. These two processes are assumed dependent through individual-specific frailty under the joint model. Our numerical studies include a simulation that reflects the structure of observed real data and a detailed real data analysis based on nearly 12000 spontaneous labours. Our finding indicates the necessity to incorporate maternal characteristic such as age or body mass index in assessing the probability for delivery due to SVD, CD or OVD and the onset of morbidities across the second stage of labour.
KW - Competing risks model
KW - Current status data
KW - Random effect
KW - Shared parameter model
UR - http://www.scopus.com/inward/record.url?scp=85064165551&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85064165551&partnerID=8YFLogxK
U2 - 10.1111/rssc.12351
DO - 10.1111/rssc.12351
M3 - Article
AN - SCOPUS:85064165551
VL - 68
SP - 1167
EP - 1182
JO - Journal of the Royal Statistical Society. Series C: Applied Statistics
JF - Journal of the Royal Statistical Society. Series C: Applied Statistics
SN - 0035-9254
IS - 4
ER -