What is the clinical evidence from this study?

Study design: Cohort. Population: End-stage renal disease on peritoneal dialysis (n=1730). Intervention: High modified Charlson Comorbidity Index (>4) and age ≥ 70.5 years vs. Overall study population. Primary outcome: Mortality (HR 4.61).

May 4, 2020Open Access

Prediction of the Mortality Risk in Peritoneal Dialysis Patients using Machine Learning Models: A Nation-wide Prospective Cohort in Korea

JNJunhyug NohEwha Womans University KYKyung Don YooUlsan College WBWonho BaeSeoul National University

Key Result

A survival-tree algorithm identified that peritoneal dialysis patients with a high modified Charlson Comorbidity Index (>4) and age ≥ 70.5 years had a 4.61-fold higher hazard ratio for mortality compared to the overall study population.

Study Design

Type

Cohort (n=1,730)

Multicenter

Yes

Structured PICO

Do machine-learning algorithms improve mortality risk prediction compared to conventional algorithms in peritoneal dialysis patients?

Population

1,730 adult peritoneal dialysis patients from a nationwide prospective cohort in South Korea, followed for a mean of 30.18 months to predict mortality risk using machine learning models.

Exposure

Machine-learning algorithms (including survival-tree and deep neural network) for mortality risk prediction

Comparator

Conventional algorithms (Cox regression)

Outcome

N-year mortality prediction accuracy (measured by concordance index and AUC)hard clinical

Machine learning models, particularly survival-tree and deep neural networks, outperform conventional Cox regression in predicting mortality risk among peritoneal dialysis patients.

Main Result

Hazard Ratio: 4.61

Limitations

Observational nature of the cohort data with potential unmeasured confounding
Presence of missing data requiring imputation via autoencoder
Original CCI was designed for specific ICD coding which may differ from clinical interviews used in the study

Abstract

Herein, we aim to assess mortality risk prediction in peritoneal dialysis patients using machine-learning algorithms for proper prognosis prediction. A total of 1,730 peritoneal dialysis patients in the CRC for ESRD prospective cohort from 2008 to 2014 were enrolled in this study. Classification algorithms were used for prediction of N-year mortality including neural network. The survival hazard ratio was presented by machine-learning algorithms using survival statistics and was compared to conventional algorithms. A survival-tree algorithm presented the most accurate prediction model and outperformed a conventional method such as Cox regression (concordance index 0.769 vs 0.745). Among various survival decision-tree models, the modified Charlson Comorbidity index (mCCI) was selected as the best predictor of mortality. If peritoneal dialysis patients with high mCCI (>4) were aged ≥70.5 years old, the survival hazard ratio was predicted as 4.61 compared to the overall study population. Among the various algorithm using longitudinal data, the AUC value of logistic regression was augmented at 0.804. In addition, the deep neural network significantly improved performance to 0.841. We propose machine learning-based final model, mCCI and age were interrelated as notable risk factors for mortality in Korean peritoneal dialysis patients.

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper