Inaccuracy of ICD-9 Codes for Chronic Kidney Disease: A Study from Two Practice-based Research Networks (PBRNs)

Article Figures & Data

Table 1. Demographics of Chronic Kidney Disease (CKD) Stage 3 Patients at Three Residency Teaching Sites

Characteristic	Total Sample (n = 325)	Buffalo Practice 1 (n = 109)	Buffalo Practice 2 (n = 95)	Kansas City Practice (n = 121)
Female sex	56 (182)	61 (67)	44 (42)	60 (73)
Age (years)
Mean (SD)	66.8 (13.5)	67.8 (13.4)	63 (12.8)	69.1 (13.5)
Range	32–96	32–96	35–93	34–95^*
Race/ethnicity
Hispanic (any race)	3 (11)	5 (5)	0 (0)	5 (6)
Black	68 (222)	83 (90)	86 (82)	41 (50)
White	25 (82)	10 (11)	14 (13)	56 (58)
American Indian/Alaska Native	0 (1)	0 (0)	0 (0)	0.8 (1)
Asian/unknown/other/decline to answer	5 (15)	3 (3)	0 (0)	10 (12)

Table 2. Chart Review Protocol

ICD-9 Code	585.3
Race/ethnicity	White, African American, Hispanic, or Native American
Age	Numeric value
Body mass index	Numeric value
Two most recent GFR values, with dates	Numeric value^*
Two most recent ACR values, with dates	Numeric value^*
Diagnosis based on GFR correct?	Yes/No
Diagnosis based on ACR correct?	Yes/No
ACR date	Date
ACR verified? (2 successive ACRs at least 90 days apart)	Dates and values of the 2 ACRs
Comorbidities
Diabetes mellitus	Yes/no
Hypertension	Yes/no
Congestive heart failure	Yes/no
Sleep apnea	Yes/no
Acute kidney injury	Yes/no
Coronary artery disease	Yes/no

↵* Values were recorded only at the Buffalo sites.
ACR, albumin-to-creatinine ratio; GFR, glomerular filtration rate; ICD-9, International Classification of Diseases, Ninth Revision.

Table 4. Breakdown of Misdiagnosed Patients^*