Validity of an automated algorithm using diagnosis and procedure codes to identify decompensated cirrhosis using electronic health records

Figures & data

Table 1 ICD-9/10 and CPT codes comprising the 10 cluster variables (C1–C10)

Condition	Associated ICD-9 and CPT codes	Associated ICD-9/10 and CPT codes
C1: Liver transplant	V42.7, 996.82, 050.5, 050.51, 050.59, 47135, 47136	V42.7, 996.82, 050.5, 050.51, 050.59, 47135, 47136,T86.40, T86.41, T86.42, 0FY00Z0, 0FY00Z1, 0FY00Z2
C2: Liver cancer	155.0, 155.1, 155.2	155.0, 155.1, 155.2, C22.0, C22.2, C22.7, C22.8, C22.1, C22.9
C3: Hepatorenal syndrome	572.4	572.4, K76.7
C4: Hepatic encephalopathy	572.2	572.2, K72.09, K72.91
C5: Portal hypertension/portal decompression procedures	572.3, 37140, 37160, 37180, 37181 37182, 37183	572.3, 37140, 37160, 37180, 37181, 37182, 37183, K76.6
C6: Esophageal varices complications (bleeding) and procedures	456.0, 456.20, 43204, 43205, 43243, 43244, 43400, 43401, 42.91, 44.91, 96.06	456.0, 456.20, 42.91, 44.91, 96.06, 43204, 43205, 43243, 43244, 43400, 43401, I85.01, I85.11, 06L30CZ, 06L30DZ, 06L30ZZ, 06L33CZ, 06L33DZ, 06L33ZZ, 06L34CZ, 06L34DZ, 06L34ZZ, 06L20ZZ, 06L23ZZ, 06L24ZZ, 0DL57DZ, 0DL58DZ
C7: Other gastrointestinal hemorrhage	530.7, 530.82, 578.0, 578.1, 578.9	530.7, 530.82, 578.0, 578.1, 578.9, K22.6, K22.8, K92.0, K92.1, K92.2
C8: Ascites/paracentesis procedures	789.5, 789.59, 49080, 49081, 54.91	789.5, 789.59, 54.91, 49080, 49081, R18.8, 0D9S30Z, 0D9S3ZZ, 0D9S40Z, 0D9S4ZZ, 0D9T30Z, 0D9T3ZZ, 0D9T40Z, 0D9T4ZZ, 0D9V30Z, 0D9V3ZZ, 0D9V40Z, 0D9V4ZZ, 0D9W30Z, 0D9W3ZZ, 0D9W40Z, 0D9W4ZZ, 0W9F30Z, 0W9F3ZZ, 0W9F40Z, 0W9F4ZZ, 0W9G30Z, 0W9G3ZZ, 0W9G40Z, 0W9G4ZZ, 0W9J30Z, 0W9J3ZZ
C9: Other sequelae of chronic liver disease	572.8	572.8, K72.10, K72.90
C10: Cirrhosis	571.2, 571.5	571.2, 571.5, K70.30, K74.0, K74.60, K74.69

Abbreviations: CPT, Current Procedural Terminology; ICD, International Classification of Diseases.

Table 2 Study population characteristics

Characteristic	N=296
Age (mean, standard deviation)	58.6 (8.3)
Age category, (n)
30≤40 years	4 (1%)
41≤50 years	27 (9%)
51≤60 years	161 (54%)
61≤70 years	75 (25%)
>70 years	29 (10%)
Male sex, (n)	175 (59%)
Race, (n)
Asian/Pacific Islander	16 (5%)
White	162 (55%)
Black	108 (37%)
Native American	4 (1%)
Unknown	6 (2%)
Hispanic ethnicity, (n)	9 (3%)
Median annual household income,a (n)
<$15,000	6 (2%)
$15,000<$30,000	49 (17%)
$30,000<$50,000	106 (36%)
$50,000<$75,000	79 (27%)
≥$75,000	47 (16%)
Missing	9 (3%)
Insurance status, (n)
Medicaid	20 (7%)
Medicare	133 (45%)
Private	139 (47%)
None	4 (1%)
Unknown	334 (3%)
Liver conditions, (n)
C1: Liver transplant	54 (18%)
C2: Liver cancer	94 (32%)
C3: Hepatorenal syndrome	11 (4%)
C4: Hepatic encephalopathy	30 (10%)
C5: Portal hypertension/portal decompression procedures	25 (8%)
C6: Esophageal varices complications (bleeding) and procedures	24 (8%)
C7: Other gastrointestinal hemorrhage	53 (18%)
C8: Ascites/paracentesis procedures	65 (22%)
C9: Other sequelae of chronic liver disease	21 (7%)
C10: Cirrhosis	203 (67%)

Notes:

a Currency is US$.

Table 3 Classification models for identification of DC

		(A) Optimized LASSO			(B) Five-cluster LASSO			(C) Three-cluster CART			(D) Five-cluster Hybrid
	Model	−1.52+0.98×C1+1.17×C2+0.98×C6+1.25×C8+0.48×C10			C1+C2+C6+C8+C10 (with all coefficients=1)			C2, C6, C8 (Figure 1)			Combined LASSO/CART ≥2 of C1, C2, C6, C8, and C10 (Figure 2)
	Cutoff	2.28			1.77						2
	Actual class	Non-DC	DC	Correct	Non-DC	DC	Correct	Non-DC	DC	Correct	Non-DC	DC	Correct
Testing data/model validation
Non-DC	169	148	21	88%	145	24	86%	144	25	85%	145	24	86%
DC	127	21	106	83%	23	104	82%	17	110	87%	19	108	85%
Total	296	169	127		168	128		161	135		164	132
AUROC		89%			90%			86%			86%
Specificity		88%			86%			85%			86%
Sensitivity		83%			82%			87%			85%
PPV		83%			81%			81%			82%
Learning Data/Model Development
Non-DC	169	151	18	89%	145	24	86%	143	26	85%	145	24	86%
DC	127	21	106	83%	19	108	85%	17	110	87%	19	108	85%
Total	296	172	124		164	132		160	136		164	132
AUROC		92%			90%			87%			85%
Specificity		89%			86%			85%			86%
Sensitivity		83%			85%			87%			85%
PPV		85%			82%			81%			82%

Abbreviations: C1, liver transplant; C2, liver cancer; C6, esophageal varices complications (bleeding) and procedures; C8, ascites/paracentesis procedures; C10, cirrhosis; AUROC, area under receiver operating characteristic; CART, Classification and Regression Trees; DC, decompensated cirrhosis; LASSO, Least Absolute Shrinkage and Selection Operator; PPV, positive predictive value.

Figure 1 The three cluster-variable CART model.

Abbreviations: C1, liver transplant; C2, liver cancer; C6, esophageal varices complications (bleeding) and procedures; C8, ascites/paracentesis procedures; C10, cirrhosis; CART, Classification and Regression Trees; DC, decompensated cirrhosis; TN, terminal node.

Figure 2 The hybrid CART model.

Abbreviations: C1, liver transplant; C2, liver cancer; C6, esophageal varices, complications (bleeding), and procedures; C8, ascites/paracentesis procedures; CART, Classification and Regression Trees; DC, decompensated cirrhosis; TN, terminal node.