Does the availability of a South Asian language in practices improve reports of doctor-patient communication from South Asian patients? Cross sectional analysis of a national patient survey in English general practices

Table 3 Effect of concordance on the average doctor-patient communication difference for South Asians, when compared to White British respondents (single-handed practices)

Ethnic group	Model 1:		Model 2:
	Mean Score (0-100)	Average difference compared to White British respondents*	Average difference compared to White British respondents when a concordant language is:
	Mean Score (0-100)	Average difference compared to White British respondents*	Available at practice**	Not available at practice ^‡
Indian	88.5	-1.9 (-2.5, -1.4)	-1.5 (-2.5, -0.5)	-2.1 (-2.7, -1.5)
Pakistani	88.5	-1.9 (-2.6, -1.2)	-0.2 (-1.5, 1.0)	-2.7 (-3.6, -1.9)
Bangladeshi	87.5	-2.9 (-4.2, -1.6)	4.5 (-1.0, 10.1)	-3.3 (-4.6, -2.0)
	p < 0.0001 ^Ɨ		Likelihood-ratio test: p = 0.0109 ^ƗƗ

Both models were adjusted for age, gender, deprivation, self-rated health status, presence of a mental health condition, and a random effect for practice.
^ƗJoint test of the differences of South Asians from White British.
^ƗƗP-values relates to the Likelihood-ratio test (omnibus test) for whether the effect of ethnicity varies with language concordance (comparing model 2 with a model where the language-ethnicity concordance effect was constrained to be constant across the three ethnic groups).
*There was no evidence (p = 0.19) to suggest that these adjusted mean scores varied across the three ethnic groups (post-hoc Wald test).
**There was evidence (p = 0.0402) to suggest that these adjusted mean scores varied across the three ethnic groups. In particular, the adjusted mean scores varied between Bangladeshi and Indian ethnic group (p = 0.0368). However, adjusted differences between Pakistani and Indian and between Pakistani and Bangladeshi ethnic groups were not significant (p = 0.10 for both) (post-hoc Wald tests).
^‡ There was no evidence (p = 0.19) to suggest that these adjusted mean scores varied across the three ethnic groups (post-hoc Wald test).

ISSN: 2731-4553