Factors influencing private dentists’ attitudes and practice participation in dental insurance: a validation study among Kenyan practitioners

Abstract

Introduction: The Kenya National Oral Health Survey (2015) reported a high incidence of untreated oral disease. The Social Health Authority (SHA) and Private Dental Insurance provide minimal dental benefits. Dental care has been identified to have the most self-reported financial barriers compared to other types of healthcare. Dental insurance has been shown to improve access to dental services, hence reducing the hazard of all-cause mortality of oral disease. However, dentists’ participation in dental insurance plans has been shown to be significantly associated with their attitudes. Problem Statement: In Kenya, only 12.8 % of Kenyans have dental insurance. Most dental plans are run by non-professionals, and usually set prices that are lower than that recommended by the Kenya Medical Practitioners and Dentists’ Council. Problems such as delayed payment, preauthorization time and coverage limits have been identified as reasons for dissatisfaction by dentists. However, little is known in Kenya about how dental insurance plans may influence a dental practitioner’s attitudes and dental practice participation in dental insurance networks. This knowledge gap hinders development of evidence-based strategies to enhance dental insurance plans and improve the access between dentists and patients. An effective research tool is therefore needed to conduct periodic surveys to collect feedback so as to improve the industry. Aim: The present study aimed at constructing and validating a research tool that investigates the factors influencing private dentists’ attitudes and their dental practices’ participation in dental insurance in Kenya. Methods: The study was an analytical cross-sectional study conducted among private dentists practicing in Kenya who were retained in the register in the year 2025. The study was carried out between May and August 2025. A convenience sampling technique was used. A study questionnaire was constructed based on literature review and guided by the Theory of Planned Behaviour by Ajzen (1991). The study was carried out in four phases; an expert review, a pilot study, a validation study and a test-retest study. A non-probabilistic convenience sampling method was employed on the calculated sample sizes of both the pilot and validation study. The questionnaire was subjected to the different phases of the study. Data was collected via a self- administered questionnaire provided on the online REDCap software platform. Data analysis was carried out using Microsoft Excel and Jamovi version 2.6.44. Content validity was assessed after the expert review by the computation of content validity indices I-CVI, S-CVI (Average) and S-CVI (Universal Agreement). Face validity, was examined using both quantitative and qualitative methods, during the pilot study and the results were represented descriptively and in bar graphs. Reliability and internal consistency was tested after the pilot study by calculating Cronbach’s Alpha for each of the different sections of the questionnaire. Construct Validity was evaluated after the validation study by performing exploratory factor analysis (EFA) in order to determine the underlying factor structure of the questionnaire. Test-retest reliability was established after the test-retest study by computation of the inter-class correlation coefficient (ICC 2,1) to establish whether the responses were reproducible. The data was presented in bar graphs, scree plots and tables. Results: For content validity, all the items in the questionnaire except one, had an I-CVI ≥ 0.83, and hence were retained. The S-CVI (Average) was 0.95 while the S-CVI (Universal Agreement) was 0.80. Face Validity was ascertained in the pilot study where majority of the participants (88.8%) thought the tool was very relevant or extremely relevant to the topic being investigated. Similarly, 88.6% of the pilot sample thought that the questionnaire was “very clear” or “extremely clear” in its intended purpose, while 88.7% felt that it had clarity of comprehension, meaning that it was either “easy” or “very easy” to understand. Overall, 69.8% of the respondents felt that the tool measured the factors influencing dental insurance “very well” or “extremely well”. Reliability and internal consistency was established of the Likert Items. The overall Cronbach’s Alpha for the 21 Likert items was 0.80 demonstrating good reliability. The Cronbach’s Alpha for Part 2 (items measuring attitude) was 0.745 indicating acceptable reliability, while that of Part 3 (Items measuring factors influencing Participation) was 0.871 signifying good reliability. Part 4 (items measuring non participation) were analyzed with frequency and tests of association and no significant statistical associations were found. Construct validity was established using EFA and this showed that Part 2 (items measuring attitude) was supported by a three-factor structure with an excellent model fit, while Part 3 (items measuring factors influencing participation) was supported by a five-factor structure with an excellent model fit. Test-Retest reliability was demonstrated in the items with Likert scales in Part 2 (items measuring attitude) and Part 3 (items measuring factors influencing participation) with the items ICC (2,1) of both parts being 0.750 and 0.796 respectively.