Who Chooses Prepaid Dental Care?
All data sets included clinically recorded measures of oral health, together with a questionnaire focusing on health and health beliefs. Patients were followed at 12 or 18-month intervals, and after 3 and 6 years, respectively, they were again given the opportunity to change or to stick with their present payment scheme; prepayment or fee-for-service payment.
The questionnaire obtained information about demographics, self-reported oral and general health, lifestyle, dental care habits, preventive measures, experience of dental care, attitudes and beliefs towards health and disease. The procedure was standardized through formalized instructions to the dental staff to ensure that the questionnaire was completed by the patient him/herself before the clinical examination. All completed questionnaires were stored at the respective clinics until the data collection was completed and the data were transferred to a computer file. The Regional Ethical Review Board in Gothenburg has approved the study (No. 323–07).
The dependent variable indicated the patient's choice of either of the two payment schemes:
The independent variables included the answers to the following questions in the questionnaire.
The response options to some questions were trichotomized for the multiple regression analysis merging low value options, as described in Table 1.
Analyses were performed using the SPSS, version 20.0. The Mann–Whitney U test and the chi-square analysis were used to detect statistically significant differences in the distribution of questionnaire responses between the two payment schemes, for continuous and ordinal-scale independent variables, respectively.
A logistic regression model was developed using a stepwise forward strategy. Independent variables were categorized as covariates or confounders, according to our graphically outlined understanding of their relationship. All the available independent variables were considered for the final logistic regression model after they were determined not to be correlating hazardously with each other, based on the correlation analysis (Spearman's ρ ≤ 0.348) or when cross-tabulated, variable by variable. Each independent variable was then independently included in the final regression model if it exerted a statistically significant influence on the value of the dependent variable in the bivariable analysis (p < 0.25 in the log likelihood test), together with a similarly statistically significant influence in the multivariable analysis (p < 0.25 in the Wald test). A number of interaction terms; for instance, the product of the variables age and assessment of own oral health were considered but rejected, as they failed to show statistical significance as described above. The final full model aimed at predicting the choice of payment scheme and to explain the amount of variability assessed by the independent variables. Results were expressed as odds ratios (OR) with 95% confidence intervals. The traditional payment scheme was used as the reference category of the dependent variable payment system.
The Method
All data sets included clinically recorded measures of oral health, together with a questionnaire focusing on health and health beliefs. Patients were followed at 12 or 18-month intervals, and after 3 and 6 years, respectively, they were again given the opportunity to change or to stick with their present payment scheme; prepayment or fee-for-service payment.
The questionnaire obtained information about demographics, self-reported oral and general health, lifestyle, dental care habits, preventive measures, experience of dental care, attitudes and beliefs towards health and disease. The procedure was standardized through formalized instructions to the dental staff to ensure that the questionnaire was completed by the patient him/herself before the clinical examination. All completed questionnaires were stored at the respective clinics until the data collection was completed and the data were transferred to a computer file. The Regional Ethical Review Board in Gothenburg has approved the study (No. 323–07).
Included Variables
The dependent variable indicated the patient's choice of either of the two payment schemes:
0 = The traditional fee-for-service scheme
1 = The prepayment scheme; i.e., the new dental insurance policy Frisktandvård ('Dental care for health').
The independent variables included the answers to the following questions in the questionnaire.
The response options to some questions were trichotomized for the multiple regression analysis merging low value options, as described in Table 1.
Data Analysis
Analyses were performed using the SPSS, version 20.0. The Mann–Whitney U test and the chi-square analysis were used to detect statistically significant differences in the distribution of questionnaire responses between the two payment schemes, for continuous and ordinal-scale independent variables, respectively.
A logistic regression model was developed using a stepwise forward strategy. Independent variables were categorized as covariates or confounders, according to our graphically outlined understanding of their relationship. All the available independent variables were considered for the final logistic regression model after they were determined not to be correlating hazardously with each other, based on the correlation analysis (Spearman's ρ ≤ 0.348) or when cross-tabulated, variable by variable. Each independent variable was then independently included in the final regression model if it exerted a statistically significant influence on the value of the dependent variable in the bivariable analysis (p < 0.25 in the log likelihood test), together with a similarly statistically significant influence in the multivariable analysis (p < 0.25 in the Wald test). A number of interaction terms; for instance, the product of the variables age and assessment of own oral health were considered but rejected, as they failed to show statistical significance as described above. The final full model aimed at predicting the choice of payment scheme and to explain the amount of variability assessed by the independent variables. Results were expressed as odds ratios (OR) with 95% confidence intervals. The traditional payment scheme was used as the reference category of the dependent variable payment system.
Source...