Skip to main content

Table 1 Distribution of Instrumental Variable Causal Forest Algorithm Early Surgery Absolute Effects on the Benefit Outcomea for Medicare Patients with Proximal Humerus Fractures in 2011 by Number of Trees in IV-CFA Forest and Minimum Leaf Node Population Size in Each Tree

From: Assessing the ability of an instrumental variable causal forest algorithm to personalize treatment evidence using observational data: the case of early surgery for shoulder fracture

Trees in IV-CFA Forest

Minimum Leaf Node Population Size

Mean

St Dev

Percent of Patients with Positive Effect

Min

10th

25th

50th

Median

75th

90th

Max

3000

50

.198

.327

75%

-1.448

-.193

.0001

.203

.411

.586

1.368

100

.196

.236

81%

-.729

-.088

.056

.204

.350

.481

.908

200

.197

.175

88%

-.449

-.033

.100

.210

.311

.418

.675

300

.198

.147

90%

-.284

-.005

.110

.213

.299

.373

.587

400

.198

.131

91%

-.220

.013

.118

.214

.290

.358

.518

4000

50

.197

.331

75%

-1.457

-.208

-.004

.200

.413

.593

1.366

100

.197

.237

81%

-.727

-.094

.058

.204

.353

.480

.906

200

.196

.177

87%

-.434

-.038

.096

.208

.312

.421

.659

300

.197

.148

90%

-.283

-.008

.109

.213

.299

.374

.565

400

.198

.132

91%

-.206

.014

.116

.215

.290

.362

.512

5000

50

.198

.327

75%

-1.389

-.201

.003

.202

.413

.590

1.375

100

.197

.234

81%

-.721

-.087

.060

.201

.352

.477

.882

200

.197

.172

87%

-.417

-.032

.100

.209

.308

.414

.663

300

.198

.145

90%

-.278

-.004

.112

.214

.296

.369

.560

400

.198

.129

91%

-.189

.018

.117

.217

.286

.354

.505

  1. a1 if patient survives 61–365 days after index proximal humerus fracture with less than $300 of shoulder-related healthcare costs, 0 otherwise