VALIDATION OF THE CONCEPTUAL FRAMEWORK OF THE DIVA QUESTIONNAIRE

Purpose:

DIVA is a questionnaire developed to assess general practitioners’ (GPs) beliefs and engagement toward vaccination. The hypothetical conceptual framework comprised seven domains (see table). An online survey among 9000 GPs registered in the mailing list of the French Society of General Medicine was conducted to validate the hypothetical conceptual framework of DIVA.  The Principal Component Analysis performed on the 1069 returned questionnaires revealed that the empirical conceptual framework comprised eleven domains (see table).This study aimed to validate the final conceptual framework of the DIVA questionnaire.

Method:

The solidity of the empirical conceptual framework was assessed by a Confirmatory Factor Analysis (CFA). Internal consistency reliability of both conceptual frameworks was assessed with the Cronbach’s alpha statistic. Lastly, we compared the two frameworks by conducting two separate multiple linear regressions with the engagement score to vaccinate as the explained variable and the other domain scores as the explanative variables.

Result:

In the CFA, goodness-of-fit indices were satisfactory and suggested that the empirical conceptual framework fitted to data adequately. Both frameworks showed good internal consistency reliability (on average, Cronbach’s alpha > 0.60) for all domains. Compared to the hypothetical scores, the empirical scores did not show better fit to the data (R²: 0.44 and 0.43 respectively).

Table : Effects of domain scores on GPs engagement to vaccinate (N=1069)

Hypothetical conceptual framework (R²=0.44)			Empirical conceptual framework (R²=0.43)
Variable	Coefficient (ß)	P value	Variable	Coefficient (ß)	P value
Intercept	6.33	0.082	Intercept	6.41	0.016
Disease characteristics and expected benefits	0.24	<0.001	Global evaluation of the vaccine	0.45	<0.001
Vaccine properties	0.27	<0.001	Safety of the vaccine	-	NS
Information about vaccination	-	NS	Information about vaccination	-	NS
Practical and organizational aspects	0,16	<0.001	Organizational and regulatory aspects	0.14	0.002
Adaptation to consultation context	-	NS	Context and organization of the  consultation	0.11	0.017
GPs’ individual experience	0.32	<0.001	GPs’ experience	0.19	<0.001
			Physician-patient communication	0.08	0.002
			Vulnerability of the patient	-	NS
			Epidemiological context	-	NS
			Uncovered costs for the patient	-	NS

NS : non statistically significant in univariate analysis

Conclusion:

The empirical conceptual framework confirmed its construct validity. Nonetheless, the hypothetical conceptual framework fitted well the data and was more parsimonious.  Therefore the latest was the final conceptual framework of the DIVA questionnaire.