Research Article

Topological network analysis of patient similarity for precision management of acute blood pressure in spinal cord injury

Weill Institute for Neurosciences; Brain and Spinal Injury Center (BASIC), Department of Neurological Surgery, University of California, San Francisco; Zuckerberg San Francisco General Hospital and Trauma Center, United States
Division of Physical Medicine and Rehabilitation, Department of Orthopaedics and Rehabilitation, University of New Mexico School of Medicine, United States
San Francisco Veterans Affairs Healthcare System, United States
Computational Research Division, Lawrence Berkeley National Laboratory, United States
Lawrence Berkeley National Laboratory, United States
Rehabilitation Research Center, Santa Clara Valley Medical Center, United States
Department of Psychiatry and Behavioral Science, and University of Minnesota, United States
Institute for Health Informatics, University of Minnesota, United States
Department of Radiology and Biomedical Imaging, University of California, San Francisco, United States
Department of Emergency Medicine, University of California, San Francisco; Zuckerberg San Francisco General Hospital and Trauma Center, United States
Department of Physical Medicine and Rehabilitation, Santa Clara Valley Medical Center, United States
Department of Neurosurgery, Stanford University, United States
Department of Anesthesia and Perioperative Care, University of California, San Francisco; Zuckerberg San Francisco General Hospital and Trauma Center, United States

Nov 16, 2021

https://doi.org/10.7554/eLife.68015

Open access
Copyright information

Version of Record: December 3, 2021
Version of Record: December 2, 2021
Accepted Manuscript: November 17, 2021
Accepted Manuscript: November 16, 2021

Download
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
- Article PDF
- Figures PDF
Open citations (links to open the citations from this article in various online reference manager services)
- Mendeley
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Abel Torres-Espín

Jenny Haefeli

Reza Ehsanian

Dolores Torres

Carlos A Almeida

J Russell Huie

Austin Chou

Dmitriy Morozov

Nicole Sanderson

Benjamin Dirlikov

Catherine G Suen

Jessica L Nielson

Nikos Kyritsis

Debra D Hemmerle

Jason F Talbott

Geoffrey T Manley

Sanjay S Dhall

William D Whetstone

Jacqueline C Bresnahan

Michael S Beattie

Stephen L McKenna

Jonathan Z Pan

Adam R Ferguson

The TRACK-SCI Investigators

(2021)
Topological network analysis of patient similarity for precision management of acute blood pressure in spinal cord injury

eLife 10:e68015.

https://doi.org/10.7554/eLife.68015
- Download BibTeX
- Download .RIS
Cite
Share
CommentOpen annotations (there are currently 0 annotations on this page).

Figures
Tables
Additional files

5 figures, 12 tables and 3 additional files

Figures

Figure 1

Download asset Open asset

High-frequency monitoring operating room (OR) data.

Flowchart of retrospective study and cohort selection criteria (a). A final cohort of 118 patients were identified and values of mean arterial pressure (MAP) (b) and heart rate (HR), (c) by time …

Figure 2 with 2 supplements

Download asset Open asset

Topological network analysis of intra-operative monitoring.

Intra-operative mean arterial pressure (MAP) and heart rate (HR) sampled every 5 min (Q5) were curated, processed, and formatted in a unique data matrix (a) (Figure 2—figure supplement 1). The …

Figure 2—figure supplement 1

Download asset Open asset

Data preprocessing, dimensionality reduction, and clustering optimization.

The time-series signal between patients is not aligned (changes at specific time point might not reflect the same for each patient) and of different length due to the different duration of surgery (a…

Figure 2—figure supplement 2

Download asset Open asset

Exploratory network analysis of AIS (American Spinal Injury Association [ASIA] Impairment Scale) at discharge.

The proportion of AIS grades at discharge for each cluster was mapped into the network and its association with the mean average mean arterial pressure (aMAP) per cluster explored. Clusters with …

Figure 3

Download asset Open asset

Non-linear relationship of average mean arterial pressure (aMAP) with the probability of improving at least one AIS grade.

Logistic regression models were fitted to study the potential non-linearity of the aMAP predictor as suggested by the exploratory analysis. Six different models were studied: naïve, linear, …

Figure 4 with 3 supplements

Download asset Open asset

Range of mean arterial pressure (MAP).

To find the optimal MAP range, a moving MAP range was computed and the time of MAP outside range calculated (a and d, example of the same patient for symmetric and asymmetric map range, …

Figure 4—figure supplement 1

Download asset Open asset

Exploratory network analysis of mean arterial pressure (MAP) out of range.

We further explored the relationship of the distribution of patients in the network and the amount of time their MAP was out of a specific range (Figure 4). For each MAP range, the assortativity …

Figure 4—figure supplement 2

Download asset Open asset

Logistic least absolute shrinkage and selection operator (LASSO) regression with leave-one-out cross-validation (LOOCV) of symmetric range.

(a) The deviance of the binomial model used as cross-validation error (mean and standard deviation) in the LOOCV (see Methods) against the values of l (in log form). The two vertical doted lines …

Figure 4—figure supplement 3

Download asset Open asset

Logistic least absolute shrinkage and selection operator (LASSO) regression with leave-one-out cross-validation (LOOCV) of asymmetric range.

Figure 5

Download asset Open asset

Leave-one-out cross-validation (LOOCV) performance of prediction models.

We built three prediction models, one to predict American Spinal Injury Association (ASIA) Impairment Scale (AIS) improvement of at least one grade at discharge (AIS impro., a), one to predict AIS A …

Tables

Table 1

Cohort demographics split by AIS (American Spinal Injury Association [ASIA] Impairment Scale) improvement.

	AIS improve. N/A(n = 15)	AIS improve. NO(n = 61)	AIS improve. YES(n = 42)	Univariate p-value
Age (years)				0.12
Mean (SD)	46.0 (17.6)	45.3 (19.1)	51.4 (19.7)
Median [min, max]	45.5 [19.0, 87.0]	47.0 [18.0, 82.0]	55.0 [18.0, 86.0]
Missing	1 (6.7%)	2 (3.3%)	1 (2.4%)
AIS admission				0.013
A	1 (6.7%)	33 (54.1%)	18 (42.9%)
B	0 (0%)	5 (8.2%)	8 (19.0%)
C	0 (0%)	5 (8.2%)	11 (26.2%)
D	0 (0%)	14 (23.0%)	5 (11.9%)
E	0 (0%)	4 (6.6%)	0 (0%)
Missing	14 (93.3%)	0 (0%)	0 (0%)
AIS discharge				<0.0001
A	0 (0%)	35 (57.4%)	0 (0%)
B	0 (0%)	5 (8.2%)	5 (11.9%)
C	1 (6.7%)	4 (6.6%)	15 (35.7%)
D	0 (0%)	14 (23.0%)	17 (40.5%)
E	1 (6.7%)	2 (3.3%)	5 (11.9%)
Missing	13 (86.7%)	1 (1.6%)	0 (0%)
Surgery duration (min)				0.66
Mean (SD)	433 (167)	392 (146)	407 (181)
Median [min, max]	432 [121, 725]	389 [120, 728]	343 [151, 950]
Missing	1 (6.7%)	2 (3.3%)	1 (2.4%)
Surgery to discharge (days)				0.33
Mean (SD)	9.50 (2.12)	18.8 (20.6)	23.4 (23.8)
Median [min, max]	9.50 [8.00, 11.0]	11.0 [1.00, 128]	14.5 [4.00, 120]
Missing	13 (86.7%)	4 (6.6%)	2 (4.8%)
Dichotomized neurological level of injury at admission				0.054
Cervical	2.00 (13.3%)	36 (59%)	33 (78.6%)
Non-cervical	13.00 (86.7%)	25 (41%)	9 (21.4%)

Table 2

Logistic regression likelihood ratio test and leave-one-out cross-validation (LOOCV) error (n = 103 patients).

Model	AIC	Residual df	Residualdeviance	Deviance	p-Value	LOOCV error
Null model $(l = β_{0})$	141.26	102	139.26			0.246
Linear model $(l = β_{0} + β_{1} x)$	134.8	101	130.80	8.46(vs. null model)	0.0036**(vs. null model)	0.231
Quadratic model $(l = β_{0} + β_{1} x + β_{2} x^{2})$	128.48	100	122.48	8.32(vs. linear model)	0.0039**(vs. linear model)	0.210
Cubic model $(l = β_{0} + β_{1} x + β_{2} x^{2} + β_{3} x^{3})$	126.97	99	118.97	3.50(vs. quadratic model)	0.061(vs. quadratic model)	0.213
Natural Spline model (df = 2) $(l = β_{0} + f_{1} (x))$	128.29	100	122.29	8.50(vs. linear model)	0.0035**(vs. linear model)	0.210
Natural Spline model (df = 3) $(l = β_{0} + f_{1} (x))$	127.13	99	119.13	3.34(vs. quadratic model)	0.067(vs. quadratic model)	0.213

** p < 0.01.

Table 3

Evaluation of logistic regression (Wald test) and leave-one-out cross-validation (LOOCV) error.

Model: $l = β_{0} + β_{1} x_{1} + β_{2} x_{1}^{2}$ where $x_{1} :$ average MAP (n = 103 patients)
LOOCV: average observed accuracy = 0.66; average kappa statistic = 0.334
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = –0.55	0.242	–2.293	0.02183*
Average MAP ( $x_{1}$ )	$β_{1}$ = 8.62	2.944	2.931	0.00338**
Average MAP ( $x_{1}^{2}$ )	$β_{2}$ = –7.601	3.039	–2.501	0.0123*

*p < 0.05; **p < 0.01.

Table 4

Evaluation of logistic regression with covariates (Wald test) and leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{11} x_{1} + β_{12} x_{1}^{2} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5} + β_{6} x_{6} + β_{7} x_{7} + β_{8} x_{8} + β_{9} x_{9}$ , where $x_{1} :$ average MAP; $x_{2}$ : average HR; $x_{3} :$ length of surgery (min); $x_{4} :$ days to AIS discharge (days); $x_{5} :$ age; $x_{6} :$ AIS admission D (‘yes’,’no’); $x_{7} :$ AIS admission C (‘yes’,’no’); $x_{8} :$ AIS admission B (‘yes’,’no’); $x_{9} :$ AIS admission A (‘yes’,’no’); (AIS admission E was set as the reference level for AIS admission variable and is part of the intercept) (final n = 93)
LOOCV: average observed accuracy = 0.688; average kappa statistic = 0.362
Predictor		Coef. estimate (logit)		Std. error		z-Value		p-Value
Intercept		$β_{0}$ = –1.530		121.8		–0.013		0.99
Average MAP ( $x_{1}$ )		$β_{11}$ = 7.398		3.112		2.377		0.017*
Average MAP ( $x_{1}^{2}$ )		$β_{12}$ = –8.053		3.530		–2.281		0.022*
Average HR ( $x_{2}$ )		$β_{2}$ = –2.087		0.0245		–0.851		0.394
Length of surgery ( $x_{3}$ )		$β_{3}$ = 0.0011		0.0015		0.728		0.466
Days to AIS discharge ( $x_{4}$ )		$β_{4}$ = 0.0037		0.0109		0.344		0.730
Age ( $x_{5}$ )		$β_{5}$ = 0.0082		0.013		0.634		0.526
AIS admission D ( $x_{6}$ )		$β_{6}$ = 1.454		1.218		0.012		0.990
AIS admission C ( $x_{7}$ )		$β_{7}$ = 1.645		1.218		0.014		0.989
AIS admission B ( $x_{8}$ )		$β_{8}$ = 1.585		1.218		0.013		0.989
AIS admission A ( $x_{9}$ )		$β_{9}$ = 1.527		1.218		0.013		0.990
Correlation matrix (Spearman)
	Average MAP	AverageHR	Length of surgery		Days to AIS discharge		Age		AIS admission
Average MAP	1
Average HR	–0.126	1
Length of surgery	–0.152	0.101	1
Days to AIS discharge	0.088	–0.059	0.165		1
Age	0.006	–0.245	0.011		0.022		1
AIS admission	0.024	0.003	–0.01		0.258		–0.13		1

*p < 0.05.

Table 5

Evaluation of logistic regression covariates only (Wald test) and leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5} + β_{6} x_{6} + β_{7} x_{7} + β_{8} x_{8} + β_{9} x_{9}$ , where $x_{2}$ : average HR; $x_{3} :$ length of surgery (min); $x_{4} :$ days to AIS discharge (days); $x_{5} :$ age; $x_{6} :$ AIS admission D (‘yes’,’no’); $x_{7} :$ AIS admission C (‘yes’,’no’); $x_{8} :$ AIS admission B (‘yes’,’no’); $x_{9} :$ AIS admission A (‘yes’,’no’); (AIS admission E was set as the reference level for AIS admission variable and is part of the intercept) (final n = 93)
LOOCV: average observed accuracy = 0.612; average kappa statistic = 0.17
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = –1.585	138.2	–0.011	0.991
Average HR ( $x_{2}$ )	$β_{2}$ = –0.0209	0.029	–0.911	0.362
Length of surgery ( $x_{3}$ )	$β_{3}$ = 0.0016	0.00141	1.151	0.250
Days to AIS discharge ( $x_{4}$ )	$β_{4}$ = 0.0105	0.0106	0.993	0.320
Age ( $x_{5}$ )	$β_{5}$ = 0.0052	0.012	0.424	0.672
AIS admission D ( $x_{6}$ )	$β_{6}$ = 1.511	1.382	0.011	0.991
AIS admission C ( $x_{7}$ )	$β_{7}$ = 1.715	1.382	0.012	0.991
AIS admission B ( $x_{8}$ )	$β_{8}$ = 1.643	1.382	0.012	0.990
AIS admission A ( $x_{9}$ )	$β_{9}$ = 1.574	1.382	0.011	0.991

Table 6

Evaluation of logistic regression in American Spinal Injury Association (ASIA) Impairment Scale (AIS) A at admission cohort (Wald test) and leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{11} x_{1} + β_{12} x_{1}^{2} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5}$ , where $x_{1} :$ average MAP; $x_{2}$ : average HR; $x_{3} :$ length of surgery (min); $x_{4} :$ days to AIS discharge (days); $x_{5} :$ age (final n = 51)
LOOCV: average observed accuracy = 0.63; average kappa statistic = 0.197
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = –0.931	3.433	–0.271	0.786
Average MAP ( $x_{1}$ )	$β_{11}$ = 10.79	5.014	2.153	0.031*
Average MAP ( $x_{1}^{2}$ )	$β_{12}$ = –6.73	4.591	–1.468	0.142
Average HR ( $x_{2}$ )	$β_{2}$ = –0.016	0.035	–0.468	0.639
Length of surgery ( $x_{3}$ )	$β_{3}$ = 0.0039	0.0026	1.504	0.132
Days to AIS discharge ( $x_{4}$ )	$β_{4}$ = 0.0067	0.014	0.477	0.633
Age ( $x_{5}$ )	$β_{5}$ = –0.012	0.020	–0.599	0.549

*p < 0.05.

Table 7

Neurological level of injury cases.

	Cervical (n = 71)							Non-cervical (n = 32)
NLI	C2	C3	C4	C5	C6	C7	C8	T2	T3	T4	T5	T6	T7	T8	T9	T10	T11	T12	S1	S5
Cases	3	3	24	28	4	8	1	1	3	3	1	2	1	3	1	3	2	4	6	2

Table 8

Evaluation of logistic regression in Cervical cohort (Wald test) and leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{11} x_{1} + β_{12} x_{1}^{2} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5} + β_{6} x_{6} + β_{7} x_{7} + β_{8} x_{8}$ , where $x_{1} :$ average MAP; $x_{2}$ : average HR; $x_{3} :$ length of surgery (min); $x_{4} :$ days to AIS discharge (days); $x_{5} :$ age; $x_{6} :$ AIS admission D (‘yes’,’no’); $x_{7} :$ AIS admission C (‘yes’,’no’); $x_{8} :$ AIS admission B (‘yes’,’no’); (AIS admission A was set as the reference level for AIS admission variable and is part of the intercept, no AIS admission E was present in this cohort) (final n = 93)
LOOCV: average observed accuracy = 0.688; average kappa statistic = 0.362
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = 2.747	3.018	0.91	0.363
Average MAP ( $x_{1}$ )	$β_{11}$ = 7.594	3.056	2.485	0.013*
Average MAP ( $x_{1}^{2}$ )	$β_{12}$ = –7.528	3.358	–2.242	0.025*
Average HR ( $x_{2}$ )	$β_{2}$ = –0.055	0.034	–1.608	0.108
Length of surgery ( $x_{3}$ )	$β_{3}$ = 0.0014	0.0019	0.720	0.472
Days to AIS discharge ( $x_{4}$ )	$β_{4}$ = 0.0022	0.012	0.182	0.855
Age ( $x_{5}$ )	$β_{5}$ = 0.0079	0.016	0.482	0.630
AIS admission D ( $x_{6}$ )	$β_{6}$ = –0.747	0.87	–0.840	0.730
AIS admission C ( $x_{7}$ )	$β_{7}$ = 0.745	0.80	0.925	0.355
AIS admission B ( $x_{8}$ )	$β_{8}$ = 0.301	0.88	0.346	0.401

*p < 0.05.

Table 9

Evaluation of logistic regression in non-cervical cohort only (Wald test) and leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{11} x_{1} + β_{12} x_{1}^{2} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5} + β_{6} x_{6} + β_{7} x_{7} + β_{8} x_{8} + β_{9} x_{9}$ , where $x_{1} :$ average MAP; $x_{2}$ : average HR; $x_{3} :$ length of surgery (min); $x_{4} :$ days to AIS discharge (days); $x_{5} :$ age; $x_{6} :$ AIS admission D (‘yes’,’no’); $x_{7} :$ AIS admission C (‘yes’,’no’); $x_{8} :$ AIS admission B (‘yes’,’no’); $x_{9} :$ AIS admission A (‘yes’,’no’); (AIS admission E was set as the reference level for AIS admission variable and is part of the intercept) (final n = 93)
LOOCV: average observed accuracy = 0.688; average kappa statistic = 0.362
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = –1.883	352.4	–0.005	0.996
Average MAP ( $x_{1}$ )	$β_{11}$ = –0.206	4.713	–0.044	0.965
Average MAP ( $x_{1}^{2}$ )	$β_{12}$ = –8.064	7.643	–1.055	0.291
Average HR ( $x_{2}$ )	$β_{2}$ = –0.0002	0.0649	0.004	0.997
Length of surgery ( $x_{3}$ )	$β_{3}$ = 0.0018	0.0054	0.336	0.737
Days to AIS discharge ( $x_{4}$ )	$β_{4}$ = 0.076	0.0613	1.240	0.215
Age ( $x_{5}$ )	$β_{5}$ = –0.0047	0.051	–0.921	0.357
AIS admission D ( $x_{6}$ )	$β_{6}$ = 1.727	3.524	0.005	0.996
AIS admission C ( $x_{7}$ )	$β_{7}$ = 3.557	5.782	0.005	0.996
AIS admission B ( $x_{8}$ )	$β_{8}$ = 1.738	3.524	0.005	0.995
AIS admission A ( $x_{9}$ )	$β_{9}$ = 1.686	3.524	0.005	0.996

Table 10

Least absolute shrinkage and selection operator (LASSO) solution and logistic regression of most predictive symmetric range with leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{1} x_{1}$ , where $x_{1} :$ time of MAP outside range 76–104 mmHg (n = 103 patients)
LOOCV: average observed accuracy = 0.61; average kappa statistic = 0.158
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = 0.368	0.333	1.103	0.269
Time MAP out 76–104 ( $x_{1}$ )	$β_{1}$ = –0.006	0.0026	–2.566	0.0103*

*p < 0.05.

Table 11

Least absolute shrinkage and selection operator (LASSO) solution and logistic regression of most predictive asymmetric range with leave-one-out cross-validation (LOOCV).

Model: $l = β_{0} + β_{1} x_{1}$ , where $x_{1} :$ time of MAP outside range 76–117 mmHg (n = 103 patients)
LOOCV: average observed accuracy = 0.5728; average kappa statistic = 0.102
Predictor	Coef. estimate (logit)	Std. error	z-Value	p-Value
Intercept	$β_{0}$ = 0.2881	0.287	1.002	0.316
Time MAP out 76–117 ( $x_{1}$ )	$β_{1}$ = –0.00788	0.0027	–2.828	0.00468**

**p < 0.01.

Table 12

Best prediction models of outcome.

Model predicting AIS improvement: $l = β_{0} + β_{11} x_{1} + β_{12} x_{1}^{2} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5}$ Model predicting AIS A: $l = β_{0} + β_{11} x_{1} + β_{12} x_{1}^{2} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5} + β_{6} x_{6} + β_{7} x_{7}$ Model predicting AIS D: $l = β_{0} + β_{2} x_{2} + β_{3} x_{3} + β_{4} x_{4} + β_{5} x_{5} + β_{8} x_{8} + β_{9} x_{9}$ where $x_{1} :$ average MAP; $x_{2}$ : AIS admission A (‘yes’, ‘no’); $x_{3}$ : AIS admission B (‘yes’, ‘no’); $x_{4}$ : AIS admission C (‘yes’, ‘no’); $x_{5}$ : AIS admission D (‘yes’, ‘no’); $x_{6}$ : NLI non-cervical; $x_{7}$ : Time MAP out 76–117; $x_{8}$ : Length of surgery; $x_{9}$ : Age; (AIS admission E and NLI cervical were set as the reference levels for the corresponding variable and are part of the intercept). All metrics are on LOOCV prediction (n = 93)

	Model AIS improv.	Model AIS A	Model AIS D
Predictor	Coef. estimate (logit)	Coef. estimate (logit)	Coef. estimate (logit)
Intercept	$β_{0}$ = –16.24	$β_{0}$ = 20.466	$β_{0}$ = 1.558
Average MAP ( $x_{1}$ )	$β_{11}$ = 7.374	$β_{11}$ = 27.031
Average MAP (Cohn et al., 2010) ( $x_{1}$ )	$β_{12}$ = –8.215	$β_{12}$ = –17.138
AIS admission A ( $x_{2}$ )	$β_{2}$ = 15.54	$β_{2}$ = –22.814	$β_{2}$ = 2.324
AIS admission B ( $x_{3}$ )	$β_{3}$ = 16.1818	$β_{3}$ = –20.38	$β_{3}$ = 0.41
AIS admission C ( $x_{4}$ )	$β_{4}$ = 16.752	$β_{4}$ = –19.01	$β_{4}$ = –2.591
AIS admission D ( $x_{5}$ )	$β_{5}$ = 14.828	$β_{5}$ = 0.217	$β_{5}$ = –2.624
NLI non-Cervical ( $x_{6}$ )		$β_{6}$ = –1.228
Time MAP out 76–117 ( $x_{7}$ )		$β_{7}$ = 0.017
Length of Surgery ( $x_{8}$ )			$β_{8}$ = –0.0044
Age ( $x_{9}$ )			$β_{9}$ = 0.03
Model performance metric	Metric value	Metric value	Metric value
Accuracy (95% CI)	0.73 (0.629, 0.818)	0.82 (0.735, 0.898)	0.806 (0.71, 0.881)
AUC	0.743	0.88	0.87
Kappa	0.45	0.629	0.573
Sensitivity	0.71	0.812	0.793
Specificity	0.74	0.836	0.812
Positive predicted value	0.658	0.72	0.657
Negative predicted value	0.788	0.89	0.896

Additional files

Supplementary file 1 Code.output.html: output html file results of running the source code.: https://cdn.elifesciences.org/articles/68015/elife-68015-supp1-v3.zip
Download elife-68015-supp1-v3.zip
Supplementary file 2 Source.code.Rmd: Rmarkdown file containing the source code that reproduced the paper.: https://cdn.elifesciences.org/articles/68015/elife-68015-supp2-v3.zip
Download elife-68015-supp2-v3.zip
Transparent reporting form: https://cdn.elifesciences.org/articles/68015/elife-68015-transrepform1-v3.docx
Download elife-68015-transrepform1-v3.docx

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Figures

High-frequency monitoring operating room (OR) data.

Topological network analysis of intra-operative monitoring.

Data preprocessing, dimensionality reduction, and clustering optimization.

Exploratory network analysis of AIS (American Spinal Injury Association [ASIA] Impairment Scale) at discharge.

Non-linear relationship of average mean arterial pressure (aMAP) with the probability of improving at least one AIS grade.

Range of mean arterial pressure (MAP).

Exploratory network analysis of mean arterial pressure (MAP) out of range.

Logistic least absolute shrinkage and selection operator (LASSO) regression with leave-one-out cross-validation (LOOCV) of symmetric range.

Logistic least absolute shrinkage and selection operator (LASSO) regression with leave-one-out cross-validation (LOOCV) of asymmetric range.

Leave-one-out cross-validation (LOOCV) performance of prediction models.

Tables

Cohort demographics split by AIS (American Spinal Injury Association [ASIA] Impairment Scale) improvement.

Logistic regression likelihood ratio test and leave-one-out cross-validation (LOOCV) error (n = 103 patients).

Evaluation of logistic regression (Wald test) and leave-one-out cross-validation (LOOCV) error.

Evaluation of logistic regression with covariates (Wald test) and leave-one-out cross-validation (LOOCV).

Evaluation of logistic regression covariates only (Wald test) and leave-one-out cross-validation (LOOCV).

Evaluation of logistic regression in American Spinal Injury Association (ASIA) Impairment Scale (AIS) A at admission cohort (Wald test) and leave-one-out cross-validation (LOOCV).

Neurological level of injury cases.

Evaluation of logistic regression in Cervical cohort (Wald test) and leave-one-out cross-validation (LOOCV).

Evaluation of logistic regression in non-cervical cohort only (Wald test) and leave-one-out cross-validation (LOOCV).

Least absolute shrinkage and selection operator (LASSO) solution and logistic regression of most predictive symmetric range with leave-one-out cross-validation (LOOCV).

Least absolute shrinkage and selection operator (LASSO) solution and logistic regression of most predictive asymmetric range with leave-one-out cross-validation (LOOCV).

Best prediction models of outcome.

Additional files

Supplementary file 1

Supplementary file 2

Transparent reporting form

Download links

Be the first to read new articles from eLife

Share this article

Cite this article

High-frequency monitoring operating room (OR) data.

Topological network analysis of intra-operative monitoring.

Data preprocessing, dimensionality reduction, and clustering optimization.

Exploratory network analysis of AIS (American Spinal Injury Association [ASIA] Impairment Scale) at discharge.

Non-linear relationship of average mean arterial pressure (aMAP) with the probability of improving at least one AIS grade.

Range of mean arterial pressure (MAP).

Exploratory network analysis of mean arterial pressure (MAP) out of range.

Logistic least absolute shrinkage and selection operator (LASSO) regression with leave-one-out cross-validation (LOOCV) of symmetric range.

Logistic least absolute shrinkage and selection operator (LASSO) regression with leave-one-out cross-validation (LOOCV) of asymmetric range.

Leave-one-out cross-validation (LOOCV) performance of prediction models.

Cohort demographics split by AIS (American Spinal Injury Association [ASIA] Impairment Scale) improvement.

Logistic regression likelihood ratio test and leave-one-out cross-validation (LOOCV) error (n = 103 patients).

Evaluation of logistic regression (Wald test) and leave-one-out cross-validation (LOOCV) error.

Evaluation of logistic regression with covariates (Wald test) and leave-one-out cross-validation (LOOCV).

Evaluation of logistic regression covariates only (Wald test) and leave-one-out cross-validation (LOOCV).

Evaluation of logistic regression in American Spinal Injury Association (ASIA) Impairment Scale (AIS) A at admission cohort (Wald test) and leave-one-out cross-validation (LOOCV).

Neurological level of injury cases.

Evaluation of logistic regression in Cervical cohort (Wald test) and leave-one-out cross-validation (LOOCV).

Evaluation of logistic regression in non-cervical cohort only (Wald test) and leave-one-out cross-validation (LOOCV).

Least absolute shrinkage and selection operator (LASSO) solution and logistic regression of most predictive symmetric range with leave-one-out cross-validation (LOOCV).

Least absolute shrinkage and selection operator (LASSO) solution and logistic regression of most predictive asymmetric range with leave-one-out cross-validation (LOOCV).

Best prediction models of outcome.

Supplementary file 1

Supplementary file 2

Transparent reporting form