PE Prognosis Predictor Manual

Introduction

The web application provides predictive insights into the time-to-delivery of patients diagnosed with preeclampsia. It uses pre-trained models to provide accurate predictions and visualizations, making it a user-friendly tool for healthcare professionals.

Step 1: Choose the panel

The web application has two main panels: the individual prediction panel and the group prediction panel.

• Individual prediction panel: This panel calculates the prognosis score (PI) of time to delivery for a single new patient using pre-trained models. The input should be a text file with one line of values from a single patient.
• Group prediction panel: This panel takes a group of new patients as input and returns the predicted PIs and their percentiles in a table format.

Step 2: Choose pre-trained model

The user needs to choose the desired pre-trained model based on their patient characteristics and data availability. The available pre-trained models are:

Pre-trained Model	Description	Required Input
Baseline Model	Applicable to all PE patients	Baseline variables
Full Model	Applicable to all PE patients	Baseline variables, lab results and vital measuements
EOPE Baseline Model	Applicable to only early onset PE patients	Baseline variables
EOPE Full Model	Applicable to only early onset PE patients	Baseline variables, lab results and vital measuementS

Step 3: Upload new patient data

Both the individual prediction panel and the group prediction panel require a single delimited file as input. The users have the option to choose between three delimiters: “;”, “,”, or “:”. The input file must have the same columns in the same order as specified. The individual prediction panel requires only a single row of data in the input file, whereas the group prediction panel allows for multiple rows.

Selected Model	Required Variables	Variable Description
Baseline Model	diag_GA	PE diagnoses gestational age(days)
	SeverePE	1: the preeclampsia is severe at the first diagnosis; 0: mild preeclampsia at the first diagnosis
	past_pe	1: the patient has previous preeclampsia history; 0:Otherwise
	age	Maternal age at the beginning of current pregnancy (Years)
	EPIS_PARA_COUNT	Previous parity counts
	DiabetesUncomplicated	1: The patient has uncomplicated diabetes comorbidity; 0:Otherwise
	ValvularDisease	1: The patient has valvular disease comorbidity; 0:Otherwise
Full Model	diag_GA	PE diagnoses gestational age(days)
	SeverePE	1: the preeclampsia is severe at the first diagnosis; 0: mild preeclampsia at the first diagnosis
	AST	Aspartate aminotransferase measurement at diagnosis (u/L)
	RRSD	Standard deviation of respirtory rate within 5 days before diagnosis
	EPIS_PARA_COUNT	Previous parity counts
	creatinine_value	Creatinine value measurement at diagnosis
	BPDiaSD	Standard deviation of diastolic blood pressure within 5 days before diagnosis
	age	Maternal age at the beginning of current pregnancy (Years)
	White_Blood_Cell_Count	White blood cell count at diganosis
	BPDiaMean	Average of diastolic blood pressure within 5 days before diagnosis
	Platelet_Count	Platelet count at diagnosis
	White_Blood_Cell_Count	White blood cell count at diagnosis
	past_pe	1: the patient has previous preeclampsia history; 0:Otherwise
EOPE Baseline Model	diag_GA	PE diagnoses gestational age(days)
	past_pe	1: the patient has previous preeclampsia history; 0:Otherwise
	EPIS_PARA_COUNT	Previous parity counts
	Coagulopathy	1: The patient has Coagulopathy comorbidity; 0:Otherwise
	PulmonaryCirculationDisorders	1: The patient has pulmonary circulation disorder; 0:Otherwise
	SeverePE	1: the preeclampsia is severe at the first diagnosis; 0: mild preeclampsia at the first diagnosis
EOPE Full Model	EPIS_PARA_COUNT	Previous parity counts
	NUMBER_OF_FETUSES	Baseline variables, lab results and vital measuementS
	diag_GA	PE diagnoses gestational age(days)
	Coagulopathy	1: The patient has Coagulopathy comorbidity; 0:Otherwise
	PulmonaryCirculationDisorders	1: The patient has pulmonary circulation disorder; 0:Otherwise
	SeverePE	1: the preeclampsia is severe at the first diagnosis; 0: mild preeclampsia at the first diagnosis
	creatinine_value	Creatinine value measurement at diagnosis
	Platelet_Count	Platelet count at diagnosis
	AST	Aspartate aminotransferase measurement at diagnosis (u/L)
	BPDiaMean	Average of diastolic blood pressure within 5 days before diagnosis
	BPSysMean	Average of systolic blood pressure within 5 days before diagnosis
	RRSD	Standard deviation of respirtory rate within 5 days before diagnosis
	past_pe	1: the patient has previous preeclampsia history; 0:Otherwise

Step 4: Choose whether to scale new data

The pre-trained model mandates the scaling of all numeric variables by dividing them by the square root of their mean value. The variables “creatinine_value,” “AST,” and “RRSD” must be log-transformed prior to scaling. If the option “Does the data need to be processed?” is set to “Yes,” the app will automatically scale the input data. However, if set to “No,” it will be the responsibility of the user to scale the data, which is not recommended.