herdiantrisufriyana/prom

Prognostication for prelabor rupture of membranes and the time of delivery in nationwide insured women: development, validation, and deployment

Herdiantri Sufriyana, MD, MSc; a, b Yu-Wei Wu, PhD; a, c Emily Chia-Yu Su, PhD a, c, d, *

a Graduate Institute of Biomedical Informatics, College of Medical Science and Technology, Taipei Medical University, 250 Wu-Xing Street, Taipei 11031, Taiwan. b Department of Medical Physiology, Faculty of Medicine, Universitas Nahdlatul Ulama Surabaya, 57 Raya Jemursari Street, Surabaya 60237, Indonesia. c Clinical Big Data Research Center, Taipei Medical University Hospital, 250 Wu-Xing Street, Taipei 11031, Taiwan. d Research Center for Artificial Intelligence in Medicine, Taipei Medical University, 250 Wu-Xing Street, Taipei 11031, Taiwan. * Corresponding author at: Clinical Big Data Research Centre, Taipei Medical University Hospital, 250 Wu-Xing Street, Taipei 11031, Taiwan. Phone: [phone] ext. 1515.

The preprint can be found here: https://www.medrxiv.org/content/10.1101/2021.06.16.21258884v1

The journal article will be published soon.

Supplement and other files can be found in any of above publications.

System requirements

We used R 4.0.2 programming language (R Foundation, Vienna, Austria) to conduct most steps of the data analysis. For steps related to the deep-insight visible neural network (DI-VNN), we also used Python 3.6.3 programming language (Anaconda Inc., Austin, TX, USA). The integrated development environment software was RStudio 1.3.959 (RStudio PBC, Boston, MA, USA). To ensure reproducibility, we used Bioconductor 3.11;61 thus, versions of the included R packages were all in sync according to versions in this Bioconductor version. For all models except the DI-VNN, we used an R package of caret 6.0.86 that wraps R packages for the modeling algorithms, which were glmnet 4.1, Rborist 0.2.3, and gbm 2.1.8. For the DI-VNN, we used keras 2.3.0 and tensorflow 2.0.0 Python libraries via R packages of reticulate 1.16, keras 2.3.0.0, and tensorflow 2.0.0. We also created R packages and a Python library for many steps in the data analysis, including the DI-VNN, which are: medhist 0.1.0, gmethods 0.1.0, rsdr 0.1.0, clixo 0.1.1, and divnn 0.1.3 (both an R package and Python library) For model deployment, we used Shiny Server 1.4.16.958 and Node.js 12.20.0. Details on other R package versions and all of the source codes (vignette) for the data analysis are available (eTable 6 in the Supplement).

To reproduce our work, a set of hardware requirements may be needed. We used a single machine for all models, except the DI-VNN, with 16 logical processors for the 2.10 GHz central processing unit (CPU) (Xeon® E5-2620 v4, Intel®, Santa Clara, CA, USA), 128 GB RAM, and 11 GB graphics processing unit (GPU) memory (GeForce GTX 1080 Ti, NVIDIA, Santa Clara, CA, USA). Parallelism was applied for CPU computing. Meanwhile, the DI-VNN required a higher GPU capability than that provided by the previous specification. For hyperparameter tuning and training, we used multiple machines in a cloud with 90 GB RAM and 32 GB GPU memory (Tesla V100-SXM2, NVIDIA, Santa Clara, CA, USA). For predictions, the DI-VNN only needed a CPU in a local machine, or that in a cloud machine for the web application.

Installation guide

Please follow through the R Markdown (prom.Rmd) or R Script (prom.R). Installation approximately requires ~5 minutes.

Demo

All codes require ~30 minutes to complete for non-expensive computation. We provided pre-existing files to substitute the expensive computation. One can set a variable that defines whether the expensive parts will be conducted. This may take weeks to complete.

We also provide small datasets to demo the 5 packages we made for this study. Please follow through the vignettes. These show simple examples to demo the packages.

Instructions for use

Briefly, all system requirements, installation guide, demo, and instructions for use are available in R Markdown (prom.Rmd) and R Script (prom.R), and other files in this repository.

The R Markdown (.Rmd) contains the same texts with this document but including the programming codes for the data analysis in-between as shown in the Supplement. An exception is subsection of "Comparison to previous studies". The codes for core steps in the analysis pipeline are also provided exclusively in an R Script (.R). The codes beyond the core steps were used for analytic decision or creating tables or figures. These are shown to provide details on how data are processed to construct all tables and figures in both the main text and this Supplement, including those in Appendix of this Supplement (eTables 1 to 5, and 13) and those shown only in the R Markdown due to the complexity (eTables 6 to 12 and 14 to 21). We also provided a 5-minute Video 1 to briefly explain technical details on deep-insight visible neural network (DI-VNN) pipeline.

Nevertheless, one cannot run the markdown or script unless having the raw dataset. To get raw data, one need to request an access from the BPJS Kesehatan for their sample dataset published in August 2019. Up to this date, there are three sample datasets they published in February 2019, August 2019, and December 2020. For the first and second versions, a request is applied via https://e-ppid.bpjs-kesehatan.go.id/, while the third is applied via https://data.bpjs-kesehatan.go.id.

To preprocess the raw data into the input dataset of this study, follow the codes of the R Markdown in data.Rmd: https://github.com/herdiantrisufriyana/medhist/tree/main/preprocessing.