Predicting neurodevelopmental outcomes in children with perinatal HIV using a novel machine learning algorithm
ABSTRACT Background A subset of children with perinatal HIV (pHIV) experience long-term neurocognitive symptoms despite treatment with antiretroviral therapy. However, predictors of neurocognitive outcomes remain elusive, particularly for children with pHIV residing in low-to-middle income countries. The present study utilized a novel data analytic approach to identify clinically-relevant predictors of neurocognitive development in children with pHIV.Methods Analyses were conducted on a large repository of longitudinal data from 285 children with pHIV in Thailand (n=170) and Cambodia (n=115). Participants were designated as neurocognitively resilient (i.e., positive slope; n=143) or at risk (i.e., negative slope; n=142) according to annual performances on the Beery-Buktenica Developmental Test of Visual-Motor Integration over an average of 5.4 years. Gradient-boosted multivariate regression (GBM) with 5-fold cross validation was utilized to identify the optimal combination of demographic, HIV disease, blood markers, and emotional health indices that predicted classification into the two neurocognitive subgroups. Model performance was assessed using Receiver Operator Curves and sensitivity/specificity.Results The analytic approach distinguished neurocognitive subgroups with high accuracy (93%; sensitivity and specificity each > 90%). Dynamic change indices and interactions between mental health and biological indices emerged as key predictors.Conclusion Machine learning-based regression defined a unique explanatory model of neurocognitive outcomes among children with pHIV. The predictive algorithm included a combination of HIV, physical health, and mental health indices extracted from readily available clinical measures. Studies are needed to explore the clinical relevance of the data-driven explanatory model, including potential to inform targeted interventions aimed at modifiable risk factors..
| Medienart: |
|
|---|
Preprint| Erscheinungsjahr: |
2022 |
|---|---|
| Erschienen: |
2022 |
| Enthalten in: |
bioRxiv.org - (2022) vom: 17. Sept. Zur Gesamtaufnahme - year:2022 |
|---|
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Paul, Robert [VerfasserIn] |
|---|
| Links: |
Volltext [kostenfrei] |
|---|
| Themen: |
|---|
| doi: |
10.1101/632273 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
XBI017907969 |
|---|
| LEADER | 01000caa a22002652 4500 | ||
|---|---|---|---|
| 001 | XBI017907969 | ||
| 003 | DE-627 | ||
| 005 | 20230429081958.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 200517s2022 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1101/632273 |2 doi | |
| 035 | |a (DE-627)XBI017907969 | ||
| 035 | |a (biorXiv)10.1101/632273 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Paul, Robert |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Predicting neurodevelopmental outcomes in children with perinatal HIV using a novel machine learning algorithm |
| 264 | 1 | |c 2022 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 520 | |a ABSTRACT Background A subset of children with perinatal HIV (pHIV) experience long-term neurocognitive symptoms despite treatment with antiretroviral therapy. However, predictors of neurocognitive outcomes remain elusive, particularly for children with pHIV residing in low-to-middle income countries. The present study utilized a novel data analytic approach to identify clinically-relevant predictors of neurocognitive development in children with pHIV.Methods Analyses were conducted on a large repository of longitudinal data from 285 children with pHIV in Thailand (n=170) and Cambodia (n=115). Participants were designated as neurocognitively resilient (i.e., positive slope; n=143) or at risk (i.e., negative slope; n=142) according to annual performances on the Beery-Buktenica Developmental Test of Visual-Motor Integration over an average of 5.4 years. Gradient-boosted multivariate regression (GBM) with 5-fold cross validation was utilized to identify the optimal combination of demographic, HIV disease, blood markers, and emotional health indices that predicted classification into the two neurocognitive subgroups. Model performance was assessed using Receiver Operator Curves and sensitivity/specificity.Results The analytic approach distinguished neurocognitive subgroups with high accuracy (93%; sensitivity and specificity each > 90%). Dynamic change indices and interactions between mental health and biological indices emerged as key predictors.Conclusion Machine learning-based regression defined a unique explanatory model of neurocognitive outcomes among children with pHIV. The predictive algorithm included a combination of HIV, physical health, and mental health indices extracted from readily available clinical measures. Studies are needed to explore the clinical relevance of the data-driven explanatory model, including potential to inform targeted interventions aimed at modifiable risk factors. | ||
| 650 | 4 | |a Biology |7 (dpeaa)DE-84 | |
| 650 | 4 | |a 570 |7 (dpeaa)DE-84 | |
| 700 | 1 | |a Cho, Kyu |e verfasserin |4 aut | |
| 700 | 1 | |a Mellins, Claude |e verfasserin |4 aut | |
| 700 | 1 | |a Malee, Kathleen |e verfasserin |4 aut | |
| 700 | 1 | |a Robbins, Reuben |e verfasserin |4 aut | |
| 700 | 1 | |a Kerr, Stephen |e verfasserin |4 aut | |
| 700 | 1 | |a Sophonphan, Jiratchaya |e verfasserin |4 aut | |
| 700 | 1 | |a Jahanshad, Neda |e verfasserin |4 aut | |
| 700 | 1 | |a Aurpibul, Linda |e verfasserin |4 aut | |
| 700 | 1 | |a Thongpibul, Kulvadee |e verfasserin |4 aut | |
| 700 | 1 | |a Kosalaraksa, Pope |e verfasserin |4 aut | |
| 700 | 1 | |a Kanjanavanit, Suparat |e verfasserin |4 aut | |
| 700 | 1 | |a Ngampiyaskul, Chaiwat |e verfasserin |4 aut | |
| 700 | 1 | |a Luesomboon, Wicharn |e verfasserin |4 aut | |
| 700 | 1 | |a Wongsawat, Jurai |e verfasserin |4 aut | |
| 700 | 1 | |a Vonthanak, Saphonn |e verfasserin |4 aut | |
| 700 | 1 | |a Chettra, Kea |e verfasserin |4 aut | |
| 700 | 1 | |a Suwanlerk, Tulathip |e verfasserin |4 aut | |
| 700 | 1 | |a Valcour, Victor |e verfasserin |4 aut | |
| 700 | 1 | |a Balla, Lila |e verfasserin |4 aut | |
| 700 | 1 | |a Garcia-Egan, Paola M. |e verfasserin |4 aut | |
| 700 | 1 | |a Preston, Rebecca |e verfasserin |4 aut | |
| 700 | 1 | |a Ananworanich, Jintanat |e verfasserin |4 aut | |
| 700 | 1 | |a Puthanakit, Thanyawee |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t bioRxiv.org |g (2022) vom: 17. Sept. |
| 773 | 1 | 8 | |g year:2022 |g day:17 |g month:09 |
| 856 | 4 | 0 | |u http://dx.doi.org/10.1101/632273 |z kostenfrei |3 Volltext |
| 912 | |a GBV_XBI | ||
| 912 | |a SSG-OLC-PHA | ||
| 951 | |a AR | ||
| 952 | |j 2022 |b 17 |c 09 | ||