Table 4.

The impact of provider use of Joint Longitudinal Viewer during outpatient primary care visits on provider ordering of duplicate images (Stage 2 full output).^a

Duplicate imaging (second stage)				Odds ratio (95% CI)		P value
Joint Longitudinal Viewer encounter				0.08 (0.01-0.81)		.03
Anscombe residual				3.16 (1.29-7.79)		.01
Provider characteristics
	Provider history of ordering imaging studies (quartiles)
		1	Refb		Ref
		2	0.83 (0.23-3.07)		.78
		3	3.11 (1.18-8.22)		.02
		4	3.93 (1.42-10.94)		.009
	Provider type
		Physician	Ref		Ref
		Physician assistant/Nurse practitioner	1.24 (0.61-2.51)		.56
Patient characteristics
	Gender
		Female	Ref		Ref
		Male	1.49 (0.72-3.08)		.28
	Age (years)
		<30	Ref		Ref
		30-39	2.28 (0.98-5.34)		.06
		40-49	0.87 (0.32-2.42)		.79
		≥50	1.50 (0.38-5.93)		.57
	Race
		White	Ref		Ref
		Black or African American	1.03 (0.48-2.18)		.95
		Other	1.57 (0.65-3.80)		.32
	Elixhauser comorbidity score
		0	Ref		Ref
		1	0.43 (0.20-0.91)		.03
		2	0.43 (0.15-1.21)		.11
		3 and above	0.16 (0.04-0.57)		.005
	Fiscal month
		1	Ref		Ref
		2	1.41 (0.00-602.30)		.91
		3	1.31 (0.00-542.59)		.93
		4	1.43 (0.00-639.45)		.91
		5	0.75 (0.00-290.28)		.93
		6	1.10 (0.00-523.39)		.98
		7	3.15 (0.01-1539.80)		.72
		8	1.77 (0.01-819.76)		.86
		9	4.11 (0.01-1699.58)		.65
		10	4.41 (0.01-2011.70)		.64
		11	1.96 (0.01-762.17)		.83
		12	5.02 (0.01-2177.41)		.60
Consc				0.05 (0.00-22.75)		.34

^aAverage incremental effects are estimated from the 2-stage residual inclusion logistic regression controlling for all variables shown in the table.

^bRef indicates baseline in the analysis.

^cCons: Constant term in the regression.