Risk controlled decision trees and random forests for precision Medicine
© 2021 John Wiley & Sons Ltd..
Statistical methods generating individualized treatment rules (ITRs) often focus on maximizing expected benefit, but these rules may expose patients to excess risk. For instance, aggressive treatment of type 2 diabetes (T2D) with insulin therapies may result in an ITR which controls blood glucose levels but increases rates of hypoglycemia, diminishing the appeal of the ITR. This work proposes two methods to identify risk-controlled ITRs (rcITR), a class of ITR which maximizes a benefit while controlling risk at a prespecified threshold. A novel penalized recursive partitioning algorithm is developed which optimizes an unconstrained, penalized value function. The final rule is a risk-controlled decision tree (rcDT) that is easily interpretable. A natural extension of the rcDT model, risk controlled random forests (rcRF), is also proposed. Simulation studies demonstrate the robustness of rcRF modeling. Three variable importance measures are proposed to further guide clinical decision-making. Both rcDT and rcRF procedures can be applied to data from randomized controlled trials or observational studies. An extensive simulation study interrogates the performance of the proposed methods. A data analysis of the DURABLE diabetes trial in which two therapeutics were compared is additionally presented. An R package implements the proposed methods ( https://github.com/kdoub5ha/rcITR).
| Medienart: |
E-Artikel |
|---|
| Erscheinungsjahr: |
2022 |
|---|---|
| Erschienen: |
2022 |
| Enthalten in: |
Zur Gesamtaufnahme - volume:41 |
|---|---|
| Enthalten in: |
Statistics in medicine - 41(2022), 4 vom: 20. Feb., Seite 719-735 |
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Doubleday, Kevin [VerfasserIn] |
|---|
| Links: |
|---|
| Themen: |
Decision trees |
|---|
| Anmerkungen: |
Date Completed 31.03.2022 Date Revised 01.05.2023 published: Print-Electronic Citation Status MEDLINE |
|---|
| doi: |
10.1002/sim.9253 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
NLM333244133 |
|---|
| LEADER | 01000naa a22002652 4500 | ||
|---|---|---|---|
| 001 | NLM333244133 | ||
| 003 | DE-627 | ||
| 005 | 20231225221247.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 231225s2022 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1002/sim.9253 |2 doi | |
| 028 | 5 | 2 | |a pubmed24n1110.xml |
| 035 | |a (DE-627)NLM333244133 | ||
| 035 | |a (NLM)34786731 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Doubleday, Kevin |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Risk controlled decision trees and random forests for precision Medicine |
| 264 | 1 | |c 2022 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a ƒaComputermedien |b c |2 rdamedia | ||
| 338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Date Completed 31.03.2022 | ||
| 500 | |a Date Revised 01.05.2023 | ||
| 500 | |a published: Print-Electronic | ||
| 500 | |a Citation Status MEDLINE | ||
| 520 | |a © 2021 John Wiley & Sons Ltd. | ||
| 520 | |a Statistical methods generating individualized treatment rules (ITRs) often focus on maximizing expected benefit, but these rules may expose patients to excess risk. For instance, aggressive treatment of type 2 diabetes (T2D) with insulin therapies may result in an ITR which controls blood glucose levels but increases rates of hypoglycemia, diminishing the appeal of the ITR. This work proposes two methods to identify risk-controlled ITRs (rcITR), a class of ITR which maximizes a benefit while controlling risk at a prespecified threshold. A novel penalized recursive partitioning algorithm is developed which optimizes an unconstrained, penalized value function. The final rule is a risk-controlled decision tree (rcDT) that is easily interpretable. A natural extension of the rcDT model, risk controlled random forests (rcRF), is also proposed. Simulation studies demonstrate the robustness of rcRF modeling. Three variable importance measures are proposed to further guide clinical decision-making. Both rcDT and rcRF procedures can be applied to data from randomized controlled trials or observational studies. An extensive simulation study interrogates the performance of the proposed methods. A data analysis of the DURABLE diabetes trial in which two therapeutics were compared is additionally presented. An R package implements the proposed methods ( https://github.com/kdoub5ha/rcITR) | ||
| 650 | 4 | |a Journal Article | |
| 650 | 4 | |a Research Support, N.I.H., Extramural | |
| 650 | 4 | |a decision trees | |
| 650 | 4 | |a precision medicine | |
| 650 | 4 | |a random forests | |
| 650 | 4 | |a risk control | |
| 650 | 4 | |a variable importance | |
| 700 | 1 | |a Zhou, Jin |e verfasserin |4 aut | |
| 700 | 1 | |a Zhou, Hua |e verfasserin |4 aut | |
| 700 | 1 | |a Fu, Haoda |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Statistics in medicine |d 1984 |g 41(2022), 4 vom: 20. Feb., Seite 719-735 |w (DE-627)NLM012664596 |x 1097-0258 |7 nnns |
| 773 | 1 | 8 | |g volume:41 |g year:2022 |g number:4 |g day:20 |g month:02 |g pages:719-735 |
| 856 | 4 | 0 | |u http://dx.doi.org/10.1002/sim.9253 |3 Volltext |
| 912 | |a GBV_USEFLAG_A | ||
| 912 | |a GBV_NLM | ||
| 951 | |a AR | ||
| 952 | |d 41 |j 2022 |e 4 |b 20 |c 02 |h 719-735 | ||