Table 3.

Association Between EHR and Time to HbA_1c and LDL-C Follow-up Retesting, by HbA_1c and LDL-C Level

Index Test Value	Predicted Probability of Retest, %*						Overall Time to Retest:HR (95% CI)†
	Within 1 y			Within 90 d
	No EHR	EHR	OR (95% CI)	No EHR	EHR	OR (95% CI)
HbA1c
<7%	85.2	86.7	1.18 (1.05 to 1.34)	16.0	14.6	0.90 (0.85 to 0.90)	1.01 (0.97 to 1.04)
7%–8.9%	87.0	88.9	1.34 (1.17 to 1.54)	24.6	25.2	1.04 (0.96 to 1.04)	1.08 (1.04 to 1.13)
≥9%	83.5	86.7	1.40 (1.19 to 1.66)	27.5	28.3	1.04 (0.97 to 1.04)	1.13 (1.08 to 1.20)
LDL-C
<2.6 mmol/L (<100 mg/dL)	86.1	87.9	1.14 (1.01 to 1.29)	16.6	15.0	0.88 (0.83 to 0.88)	0.99 (0.95 to 1.03)
2.6–3.3 mmol/L (100–129 mg/dL)	90.6	92.8	1.21 (1.07 to 1.37)	24.9	25.6	1.04 (0.97 to 1.04)	1.07 (1.03 to 1.11)
≥3.4 mmol/L (≥130 mg/dL)	86.2	89.7	1.30 (1.15 to 1.48)	26.0	27.0	1.05 (0.99 to 1.05)	1.10 (1.06 to 1.16)

EHR = electronic health record; HbA_1c = hemoglobin A_1c; LDL-C = low-density lipoprotein cholesterol; OR = odds ratio.

^*Retests within 1 y and 90 d were analyzed using logistic regression with an interaction between EHR status and index HbA_1c or LDL-C level, adjusted for calendar month and year; medical center; and patient characteristics, including treatment intensification, age, sex, neighborhood socioeconomic status, race or ethnicity, and other chronic conditions and with SEs adjusted for clustering at the facility level by using the Stata logistic command with the cluster option.

^†Overall time to retest was analyzed using the time between any given test (index test) and the subsequent test (retest). We used a Cox model stratified by index HbA_1c or LDL-C level, with an interaction between EHR status and index HbA_1c or LDL-C level, adjusted for calendar month and year; medical center; and patient characteristics, including treatment intensification, age, sex, neighborhood socioeconomic status, race or ethnicity, and other chronic conditions and with SEs adjusted for clustering at the facility level.