Combining Inpatient and Outpatient Data for Diagnosis of Non-Valvular Atrial Fibrillation Using Electronic Health Records: A Validation Study

Figures & data

Table 1 Source of Information for the Three AF Algorithms

Algorithm Data Source	1	2	3
Community: Supporting free text	✓
Community: ICD-9 codes	✓	✓
Hospital: ICD-9 codes	✓	✓	✓

Note: Check marks represent the specific source of information that was used for each algorithm.

Abbreviations: AF, atrial fibrillation; ICD-9, International Classification of Diseases, ninth revision.

Table 2 Characteristics of Individuals with Atrial Fibrillation^a

	Individuals Identified as Having AF Based on Open-Chart Review (n=41)	Clalit Members Aged >40 with Documented AF (n=174,188)
Age (years)
40–44, n (%)	0	1555 (0.9)
45–54, n (%)	0	8957 (5.1)
55–64, n (%)	8 (4.0)	22,342 (12.8)
65–74, n (%)	76 (38.0)	43,679 (25.1)
75–84, n (%)	91 (45.5)	62,541 (35.9)
85+, n (%)	25 (12.5)	35,114 (20.2)
Mean (SD)	76.7 (6.5)	74.7 (11.5)
Median (IQR)	76 (72–80)	76 (68–83)
Sex, n (%)
Male	131 (65.5)	82,671 (47.5)
Female	69 (34.5)	91,517 (52.5)
Comorbidity, n (%)
Hypertension	41 (100)	134,627 (77.3)
Ischemic stroke	3 (7.3)	18,112 (10.4)
Ischemic heart disease	100 (50.0)	87,406 (50.2)
Diabetes	59 (29.5)	57,581 (33.1)

Note: ^aAs of date of AF diagnosis.

Abbreviations: Clalit, Clalit Health Services; AF, atrial fibrillation; SD, standard deviation; IQR, interquartile range.

Table 3 Validity and Accuracy Measurements of the Three Different Algorithms Based on Different Data Sources

	Algorithm
	1	2	3
Data source	Inpatient: ICD-9 codes Outpatient: ICD-9 codes and supporting free text	Inpatient: ICD-9 codes Outpatient: ICD-9 codes	Inpatient: ICD-9 codes
Sensitivity	85.4%	82.9%	70.7%
Specificity	95.0%	95.0%	96.9%
PPV	81.4%	81.0%	85.3%
NPV	96.2%	95.6%	92.8%

Abbreviations: PPV, positive predictive value; NPV, negative predictive value.