VALIDATION OF A DOMAIN ASSESSING THE ENGAGEMENT OF GENERAL PRACTITIONERS TOWARD VACCINATION USING THE RASCH MODEL

Purpose:

The DIVA© questionnaire was developed to assess general practitioners' (GPs) beliefs and behaviors toward vaccination. This study presents the Rasch validation of one DIVA domain, assessing GPs' engagement toward vaccination.

Methods:

The Rasch model (RM), using a rating scale model, was applied to the six items from the ‘Engagement in vaccination' domain in a cross-sectional study conducted in France to finalize the scoring and validate the DIVA questionnaire. The items were formatted in a four-level Likert scale.

Results:

Table 1 represents items' location on the latent trait. Higher location values were associated with stronger engagement in vaccination. The weakest sign of engagement was given by item 6 (location=-0.93) while the strongest by item 4 (location=0.88). The order of DIVA item location followed a theoretical model of GP's engagement process with vaccination: First, belief and values (item 6), then communication (item 1), action (item 2), repetition of action (item 3) and control (item 4).  Item 5 was neutral in this process (location=0.10).

Table 1: Location of DIVA items from the ‘Engagement in vaccination' domain on the latent trait from Rasch analysis (N=1069)

Item label	Location
6. My attitude towards prescribing the vaccine is in agreement with my convictions	-0.93
1. I raise the subject of vaccination	-0.54
2. I am used to prescribe vaccination	-0.07
5. Vaccination is a subject that interests me	0.10
3. I insist on vaccination if the patient is reticent	0.56
4. I make sure that my prescription for vaccination has been followed	0.88

The person-item threshold distribution (Figure 1) represents the distribution of the study population and item thresholds on the same metric. All six items were able to distinguish GPs according to their level of engagement and GPs in the study were fairly engaged toward vaccination.

Figure 1: Person-item threshold distribution of DIVA items from the ‘Engagement in vaccination' domain (N=1069)

Conclusions:

The RM validated the number and content of items and modalities of the engagement domain. The level of engagement of GPs in the vaccination against a specific disease can be measured accurately with the six items of the engagement domain of the DIVA questionnaire.