Table 3. Overall Rating on Quality by Office Size, Ownership, and Health Information Technology (HIT) Implementation

Fixed Effect and Categories	No. in Category	Effects Modeled Separately		Combined Model
Fixed Effect and Categories	No. in Category	Role-Adjusted % Positive Ratings^*	Global P	Adjusted % Positive Ratings^{^†}	Global P
Office size
Small (3–15 staff)	120	75 (71–79)	<.0001	75 (72–79)	<.0001
Medium (16–40 staff)	117	65 (61–69)^{^‡^§}		65 (61–69)^{^‡^§}
Large (41–70 staff)	47	60 (56–65)^{^‡}		60 (56–65)^{^‡}
Very large (>70 staff)	22	55 (48–61)^{^‡^§}		55 (48–61)^{^‡^§}
Ownership
Provider owned	78	72 (67–78)^{^§}	.0077	Not included^{^¶}
Government owned	25	60 (52–67)^{^§}
University/academic	61	62 (56–69)
Hospital/health system	133	69 (64–74)
Other	9	68 (64–74)
HIT implementation level
Low	16	76 (67–85)^{^§}	.0268	74 (67–82)^{^§}	.0102
Partial	86	64 (59–70)^{^§}		64 (60–68)^{^§}
High	109	69 (64–74)		70 (66–73)^{^§}
Full	95	67 (62–72)		68 (64–72)

Data provided as mean (95% confidence interval).
↵* Estimated through mixed models that included respondent role and the specified office characteristic as fixed effects. Models treated the 4 scores due-role as repeated measures for the office and included health system group membership as a random effect. The main effect of role was significant at P < .0001 in all models.
↵† Estimated from the combined model with role, office size, and HIT implementation level as fixed effects.
↵‡ Post hoc pair-wise comparison with small (3–15 staff) significant at P < .0001 after Tukey adjustment.
↵§ Post hoc pair-wise comparison significant at P < .05 after Tukey adjustment.
↵¶ If ownership is added-the combined model P = .4610.