This cross-sectional study, approved by the Ethics Committee of the Affiliated Stomatological Hospital of Kunming Medical University, was performed from November 2018 to May 2019.
Sample size calculation
According to the previous study, the caries prevalence was set at 69% [17]. The sample size was 913, with a confidence interval width of 6% (n = 4 × 1.962 × p ×[1-p]/L2; [n: no. of elderly, p: prevalence, and L: width of the 95% confidence interval]). We estimated a 90% response rate, and the participants needed to be at least 1014.
Study design and participant selection
Yunnan province is ethnically heterogeneous, with Han people as the predominant population. It is situated in the south-western frontiers of China, bordering Myanmar, Laos, and Vietnam. The gross domestic product of Yunnan province was ranked 20th among the 31 provinces in China [20]. Overall, it is an undeveloped province. Based on the latest national population census, the total population in Yunnan is 46 million, with over 3 million elderly residents. Around half of the population live in rural areas [21].
The participant selection followed a multistage and stratified strategy. All rural districts in Yunnan province were invited and divided into two strata by the average per capita disposable income in Yunnan province in 2017 (US$1400) [22]. In the first stage, two districts, Jianchuan district and Ninger district, were randomly selected to represent the two strata. In 2017, the average per capita disposable income in the rural areas of the two districts was US$1407 and US$1154, respectively [23]. The dentist to population ratio of Jianchuan district was 1:21,000 and 1:24,000 in Ninger district [24]. Then, all the rural villages in the selected districts were numbered. The lottery method was used to randomly select three villages from each district. In the same way, three communities from the lists of each village were randomly selected. However, The dentist to population ratio was not available for each selected village or community. All local residents aged between 65 and 74 years in these communities were invited. Participants excluded were those that were not able to comply with the clinical exam and completion of the questionnaire.
Clinical examination
Two trained and calibrated dentists, wearing headlights, carried out the clinical oral health status assessment with plane dental mirrors and ball-ended probes in a community centre in each village. Dental caries status was evaluated with the DMFT and DF-Root indices following the instructions of the WHO [25]. All permanent teeth, including the wisdom teeth, were evaluated. Coronal and root caries were recorded separately for each decayed tooth (DT). If an unmistakable cavity was present on the crown surface, it was coded as coronal caries. A filled crown with caries or a tooth with a temporary filling was recorded as coronal caries. We recorded root caries as present when a lesion felt soft or leathery on probing with the CPI probe. If any doubt existed, caries was not recorded. The categorization of missing tooth (MT) was used for teeth that were extracted due to caries or other reasons. A filled crown or filled root (FT) was recorded as tooth with filling, without caries on the crown or on the root. The intra- and inter-examiner reliability were evaluated throughout the study by re-examining approximately 10% randomly selected participants. The kappa values of the intra- and inter-examiner reliability were higher than 0.90.
Questionnaire survey
A structured questionnaire was developed based on the previous study conducted in China [26, 27]. A pilot test was conducted before the main study. After the clinical examination, Trained interviewers assisted the participants to fill in the questionnaire during the face-to-face interviews due to the low education level of the elderly. The questionnaire consisted of three parts (Appendix 1):
(i) the elderly’s socioeconomic and demographic information, specifically gender, ethnicity, education level, place of residence, marital status, and annual family income;
(ii) the elderly’s oral health-related behaviours, specifically dental visit history and tooth-brushing habits; and
(iii) the elderly’s oral health-related knowledge about the perceived causes and prevention of dental caries and periodontal diseases. In total, there were four multiple choice questions. For each question, all the alternatives were considered correct answers, excluding “I don’t know” and “no answer". A maximum of three alternatives could be chosen for each question (one point for each correct answer). The total dental knowledge score was calculated, which ranged from 0 to 12 [27].
Statistical analysis
Before the statistical analysis, proofreading and a logic check were conducted. One research assistant conducted the data analysis using IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, New York, United States). Descriptive analysis was conducted to describe the overall measurements of the dental caries, including their distribution, central tendency and dispersion. Chi-square tests were employed to compare the differences in the percentages between groups. Independent t-tests (two groups) and one-way ANOVA (more than two groups) were used to analyse the differences in dental caries experience (mean DMFT scores) between groups. When a variable with more than two subgroups was statistically associated with the caries experience (mean DMFT scores), a Bonferroni test was performed to detect the differences within the subgroups. A multifactor ANCOVA test was used to identify the factors influencing caries status. The dependent variable was the mean DMFT scores. All independent variables, such as the socioeconomic and demographic determinants and oral health-related knowledge scores, were entered into the model. An alpha of 0.05 was used as the cut-off for statistical significance.