Effectiveness of the Oral Care Program in Type 2 Diabetic Elderly Patients with Periodontitis: A Clustered Randomised Controlled Study With 6-month Follow-Up

Background: Diabetes mellitus and periodontal disease have a bidirectional relationship. This study aimed to assess the effectiveness of the oral care program in reducing glycaemic status and improving oral health behaviours and oral hygiene among type 2 diabetes mellitus (T2DM) older patients with periodontitis. Methods: A clustered randomised controlled study was conducted in Nakhon Ratchasima province between July 2019 and January 2020. The two health centres were randomly assigned to control and intervention groups, which thirty-ve T2DM older patients with chronic periodontitis were recruited in each group. The intervention group received the program consisted of oral health education based on Health Belief Model (HBM), individual oral hygiene instruction and scaling and root planning at one month and followed by individual oral hygiene instruction at three months. The patients in the control group received the routine program provided by the health centre. Outcomes were measured using the interviewed questionnaire, simplied oral hygiene index, and the glycaemic status (HbA1c) at baseline, 3 months and 6 months, respectively. Data were analysed using a descriptive statistic, Chi-square test, t-test, repeated measure ANOVA at p-value = 0.05. Results: After the 3 and 6 months, the results showed that the intervention group signicantly improved in the HBM scores, oral health behaviours scores, and decreased in OHI-S scores and HbA1c level (p < 0.05), while there was no signicant difference in the control group. Conclusion: The ndings suggested that the proposed program was effective among diabetic older people with periodontitis to improve oral health perception, behaviours, oral hygiene, and decreased glycaemic status at 3- and 6-month evaluation. 0.05, respectively. A repeated measure ANOVA showed a statistically signicant difference of perceived self-ecacy scores within the DOC group between baseline, 3- and 6- month follow-ups. Bonferroni post-hoc analysis revealed that the baseline score (10.74 ± 1.42) was signicantly lower than 3- and 6-month follow-ups (11.80 ± 0.47, p < 0.001 and 11.89 ± 0.40, p < 0.001). There were no statistically signicant differences between 3- and 6- month follow up scores (p = 1.00). For the patients within the control group, there were no statistically signicant differences of perceived self-ecacy within the group between baseline, 3- and 6- month follow-ups, (p = 0.26).


Background
Diabetes is one of the fastest-growing global health problems that cause morbidity and mortality due to long-term complications 1 . In 2019, the International Diabetes Federation estimated that the number of people with diabetes was 463 million and will reach 700 million by 2045 2 . Since the diabetes prevalence increases with age, the highest estimated prevalence is in people older than 65, accounting about one-fth of the population 2 . The prevalence of diabetes in the old age group varies signi cantly between regions driven by population ageing, economic development and increasing urbanisation, which leads to more Type 2 diabetes mellitus (T2DM) 3 . The vast majority (87%) of diabetes-related deaths occur in low-and middle-income countries 2 , and more than 60% of diabetes live in Asia 4 .
Periodontitis was rated to be the sixth complication of diabetes 5 . A recent systematic review con rmed that periodontitis has a signi cant impact on diabetes control 6 . The risk of periodontitis increases by approximately three times in diabetic patients 7 . The elevation of pro-in ammatory factors in the gingiva of patients with poorly controlled diabetes suggests a biological pathway that aggravates periodontitis 8 . Diabetes increases the risk of periodontal disease and periodontal in ammation also negatively affects glycemic control indicated its two-way relationship 9,10 . The previous study demonstrated that patients achieved glycemic controlled through prevention and management of periodontitis 11 . A metaanalysis showed a signi cant reduction of glycated haemoglobin on T2DM patients with periodontitis after non-surgical periodontal therapy using scaling and root planing (SRP) with a minimum of 3-months follow-up. A previous study in Japan found a signi cant improvement of bleeding on probing after 4-time, every 6-8 months dental education program for oral self-care, but did not change in glycated haemoglobin among T2DM patients with periodontitis 12 . Thus, with the current evidence, professional periodontal treatment is required to improve oral health and glycemic control in diabetic patients with periodontitis, and further study needs more follow-up periods 13 .
Thailand is among the countries that have a high prevalence of diabetes 14 . The prevalence of diabetic patients in Thailand increased from 6.9% in 2009 to 8.8% in 2014 15 . The prevalence was higher in urban areas than in rural areas, higher in women than in men, and highest at aged 60-69 accounted for 17% of total diabetic patients 16 . The Thai National Health Examination Survey also found that only 20% of the patients could effectively control their blood sugar, and other 30% although treated could not control their blood sugar 17 . In Nakhon Ratchasima, the largest province in Thailand, the number of older people with T2DM has been rising rapidly according to the Health Data Centre of the Ministry of Public Health (HDC program), and the number of patients is highest in the population over aged 60 18 . Currently, multiple sectors attempt to educate the public about a healthy diet and physical activity to reduce the risk of T2DM. However, the relationship between periodontal diseases and diabetes are less likely to address. Thus, the aim of this study was to propose the newly develop oral health program and evaluate its effectiveness on improving the glycemic status and oral hygiene status among older uncontrolled diabetic patients with chronic periodontitis.

Participants
The population for this study was the uncontrolled T2DM patients who received services at health centres, Muang district, Nakhon Ratchasima province, Thailand. Based on the previous study 19 , the sample size was 35 for each group calculated from Cochran's Formula 20 . Inclusion criteria were the uncontrolled T2DM diagnosed patients aged over 60 years, who had at least 10 natural teeth, and the periodontal screening and recording (PSR) more than code 3 showing that is periodontitis 21 . The patients with serious systemic disease including stroke, severe heart disease, hyperosmolar hyperglycaemic nonketotic syndrome (HHNS), severe neuropathy, end-stage renal failure, and mental disorder were excluded. Patients who had manual dexterity problems, depended on the caretaker, deaf or blind, was unable to speak Thai, were also excluded. Cluster random sampling was used for choosing two health centres. Yangyai and Khok Kruat health centres were randomly assigned by drawing the envelope. to the intervention and control groups. Yangyai Health Centre received the intervention, and Khok Kruat Health Centre received routine program (control group). Two hundred and fty-seven older diabetic patients in Yangyai Health Centre and 253 older diabetic patients in Khok Kruat Health Centre were eligible based on the inclusion and exclusion criteria. Then, 35 patients were randomly selected using simple random sampling technique into the intervention and the control groups (Fig. 1). All participants were blinded for their group status. The research proposal was approved and reviewed by the ethical committee of the Faculty of Dentistry/Faculty of Pharmacy, Mahidol University, Institutional Review Board (No.MU-DT/PY-IRB 2019/042.0307; 3 July 2019).

Examiner calibration
The interviewers were two health o cers who did not work in both health centres and blinded to the group assignment. The interviewers attended a one-day training program before collecting data. The examiners were two dentists who did not work in both health centres and did not know whether patients belonged to the intervention or the control groups. Intra-examiner reliability of two dentists using Cronbach's alpha coe cient was 0.99 and 0.99. The inter-examiner reliability between two dentists using the Cronbach's alpha coe cient was 0.98.

Program description
The intervention group received four weeks of the Diabetic and Oral Care program (DOC) comprised of a group oral health education, individual oral hygiene instruction and scaling and root planing (SRP). The rst week, a 1-hour oral health education program was provided in a group session for T2DM patients by the interdisciplinary team. This educational program was based on the Health Belief Model (HBM) 22 using slide presentations, including the relationship and complications between T2DM and oral health, oral care and the healthy diet for diabetic patients. During the second week, patients practised oral cleaning and self-oral examination conducted by a dentist and dental hygienist. During the last two weeks, patients were scheduled for SRP (5 patients per day) performed by a dentist. Before the treatment, an individual oral hygiene instruction on tooth brushing, using an interdental brush, cleaning dentures, and how to self-check oral health was provided by a dental hygienist. At 3-month recall, an individual oral hygiene instruction was followed up and emphasised by a dentist and dental hygienist again.
The control group received a routine program that included seeing the doctor once a month, blood testing, and collecting pharmaceutical items from a nurse every 3-month and undergoing oral examination twice a year.

Measurement of outcomes
The patient's interview, oral examination and blood sample testing were obtained at baseline, 3-month and, 6-month follow up. The patients did not know to which group they belonged to, and the interviewers were blinded to the intervention assignment. The T2DM patients in both groups were interviewed about general characteristics, the oral health perception based on HBM, and oral health behaviours. The structured questionnaires were validated by 3 experts in dentistry consisted of experts in periodontology, community dentistry and, advanced dentistry. The Item-Objective Congruence Index (IOC) was 0.95. A pilot study was conducted with 30 diabetic patients receiving the service in another health centre before data collection in order to test the reliability of the questionnaires. Cronbach's alpha coe cient was 0.73-0.80, indicating good internal consistency reliability. An oral hygiene examination was performed by the two calibrated dentists using the simpli ed oral hygiene index (OHI-S) 23 . Blood samples were tested for HbA1c by nurses every 3-months, and the results of the blood samples were retrieved from the patients' medical records.

Data analysis
Statistical analysis was performed using the SPSS software (IBM SPSS statistics version 23; SPSS Inc). Data were analysed using descriptive statistics, including frequency distribution and percentages, to describe general characteristics. Mean, and the standard deviation was used to describe a patient's HbA1c, oral hygiene index, the HBM and oral health behaviour scores. Chi-square and independent sample t-tests were used to compare the difference between the two groups at baseline, 3-and 6-months. Repeated measure ANOVA and Bonferroni post-hoc analysis was used to compare the innergroup differences between baseline, 3-and, 6-months. All analyses used statistically signi cant at p-value = 0.05.

Results
The study was conducted during July 2019 and January 2020. All 70 patients in the intervention and control group completed the study protocol with 100% response rate and no loss of follow-up. At baseline, the results showed no statistically signi cant differences in general characteristics of patients between the two groups in terms of age, gender, body mass index, duration of being diabetic, systemic disease, history of smoking, occupation, education level, income, and health insurance (Table 1).

Perceived susceptibility
The mean scores of perceived susceptibility showed no statistically signi cant differences between the two groups at baseline (p = 0.19). There were statistically signi cant differences between the two groups at 3-and 6-month follow-ups, the mean differences were 0.54 ± 0.21, p < 0.05 and 0.54 ± 0.22, p < 0.05, respectively. A repeated measure ANOVA showed statistically signi cant differences of perceived susceptibility scores within the DOC group between baseline, 3-and 6month follow-ups, (p < 0.05). Bonferroni post-hoc analysis revealed that the baseline (10.94 ± 1.59) was lower than that of the 3-and 6-month follow-ups with statistically signi cant differences (11.97 ± 0.17, p < 0.001 and 11.98 ± 0.16, p < 0.001) whereas there were no statistically signi cant differences between 3-and 6-month follow-ups (p = 1.0). For the patients within the control group, there was no statistically signi cant differences of perceived susceptibility scores within-group between baseline, 3-and 6-month follow-ups, (p = 0.95).

Perceived severity
For perceived severity mean scores, there were no statistically signi cant differences between the two groups at baseline (p = 0.33). However, there was a statistically signi cant difference between the two groups at 3-and 6-month follow-ups, the mean differences were 0.51 ± 0.19, p < 0.05 and 0.09 ± 0.08, p < 0.05, respectively. A repeated measure ANOVA showed statistically signi cant differences of perceived severity scores within the DOC group between baseline, 3-and 6-month follow-ups, (p < 0.05). Bonferroni post-hoc analysis revealed that the baseline score (11.31 ± 1.13) was lower than at the 3and 6-month follow-ups with statistically signi cant differences (11.93 ± 0.36, p < 0.001 and 11.94 ± 0.34, p < 0.05), whereas there were no statistically signi cant differences between 3-and 6-month follow-ups (p = 0.97). For the patients in the control group, it showed no statistically signi cant differences of perceived severity scores within-group between baseline, 3and 6-month follow-ups, (p = 0.11).
Perceived bene t For perceived bene t mean scores, there were no statistically signi cant differences between the two groups at baseline (p = 0.33). However, there was a statistically signi cant difference between the two groups at 3-and 6-month follow-ups. The mean differences were 0.63 ± 0.27, p < 0.05 and 0.49 ± 0.21, p < 0.05, respectively. A repeated measure ANOVA showed statistically signi cant differences of perceived bene t scores within the DOC group between baseline, 3-and 6-month follow-ups: F (1.099,37.354) = 55.881, p < 0.001. Bonferroni post-hoc analysis revealed that the baseline score (11.20 ± 1.32) was signi cantly lower than 3-and 6-month follow-ups (11.94 ± 0.24, p < 0.001 and 11.95 ± 0.20, p < 0.001), whereas no statistically signi cant differences between 3-and 6-month follow-ups (p = 1.00). For the patients within the control group, there were no statistically signi cant differences of perceived bene ts scores within-group between baseline, 3-and 6month follow-ups, (p = 0.66).

Perceived barriers
For perceived barriers mean scores, there were no statistically signi cant differences between the two groups at the baseline (p = 0.50). However, there was a statistically signi cant difference between the two groups at 3-and 6-month follow-ups. The mean differences were 8.06 ± 0.38, p < 0.001 and 8.74 ± 0.34, p < 0.001, respectively. A repeated measure ANOVA showed statistically signi cant differences of perceived barriers scores within the DOC group between baseline, 3-and 6month follow-ups, (p < 0.001). Bonferroni post-hoc analysis revealed that the baseline score (10.46 ± 0.18) was signi cantly higher than 3-and 6-month follow-ups (2.40 ± 0.35, p < 0.001 and 1.69 ± 0.25, p < 0.001). No statistically signi cant differences were recorded between 3-and 6-month follow-ups (p = 0.29). While, the patients within the control group, there were no statistically signi cant differences of perceived barriers scores within-group between baseline, 3-and 6-month follow-ups, (p = 0.06).

Perceived self-e cacy
In the case of the mean scores for the last component, perceived self-e cacy, there were no signi cant differences between the two groups at baseline (p = 0.06). There were statistically signi cant differences between the two groups at 3-and 6month follow-ups. The mean differences were 0.89 ± 0.31, p < 0.05 and 0.49 ± 0.21, p < 0.05, respectively. A repeated measure ANOVA showed a statistically signi cant difference of perceived self-e cacy scores within the DOC group between baseline, 3-and 6-month follow-ups. Bonferroni post-hoc analysis revealed that the baseline score (10.74 ± 1.42) was signi cantly lower than 3-and 6-month follow-ups (11.80 ± 0.47, p < 0.001 and 11.89 ± 0.40, p < 0.001). There were no statistically signi cant differences between 3-and 6-month follow up scores (p = 1.00). For the patients within the control group, there were no statistically signi cant differences of perceived self-e cacy within the group between baseline, 3-and 6-month follow-ups, (p = 0.26).
Oral health behaviour scores At baseline, the result showed no statistically signi cant differences in oral health behaviour scores between the two groups (p = 0.07). The oral health behaviour scores between the two groups demonstrated statistically signi cant differences at 3and 6-month follow-ups (p < 0.001, and p < 0.001). The mean difference was highest (3.94 ± 0.45) at 6-month follow-ups.
A repeated measure ANOVA showed statistically signi cant differences in oral health behaviour scores within the DOC group between baseline, 3-and 6-month follow-ups. Bonferroni post-hoc analysis revealed that the baseline score (18.43 ± 3.37) was the lowest (p < 0.001). For the patients within the control group, there were no statistically signi cant differences in oral health behaviour scores within the group between baseline, 3-and 6-month follow-ups, (p = 0.08).

Oral hygiene status
Comparing the oral hygiene status between the two groups was no statistically signi cant difference in the mean for OHI-S at baseline (p = 0.56). The mean difference of OHI-S was signi cantly highest (2.53 ± 0.14) at 3-and 6-month follow-ups (p < 0.001, and p < 0.001). Moreover, a repeated measure ANOVA showed statistically signi cant differences of OHI-S scores within the DOC group between baseline, 3-and 6-month follow-ups. Bonferroni post-hoc analysis revealed that the baseline score (3.31 ± 1.04) was signi cantly higher than that at the 3-and 6-month follow-ups (0.88 ± 0.46, p < 0.001 and 0.68 ± 0.53, p < 0.001). There were no statistically signi cant differences between 3-and 6-month follow-ups (p = 0.09). No statistically signi cant differences of OHI-S scores were found within the control group between baseline, 3-and 6-month follow-ups, (p = 0.18).

Glycaemic status
At baseline, the result showed no statistically signi cant differences for HbA1c values between the two groups (p = 0.99). A signi cant difference in glycaemic status between the two groups was found at 3-and 6-month follow-ups (p = 0.04, and p = 0.02, respectively). The mean difference of HbA1c value was signi cantly highest (0.97 ± 0.39) at 6-month. A repeated measure ANOVA showed statistically signi cant differences of HbA1c values within the intervention group between baseline, 3-and 6-month follow-ups, (p < 0.05). Bonferroni post-hoc analysis revealed that the baseline score (8.94 ± 1.95%) was signi cantly higher than the 3-and 6-month follow-ups (8.08 ± 1.66%, p < 0.001 and 7.83 ± 1.81%, p < 0.05). There were no statistically signi cant differences between 3-and 6-month (p = 1.00). For the patients within the control group, there were no statistically signi cant differences of HbA1c value within-group between baseline, 3-and 6-month follow-ups, (p = 0.10).

Discussion
The outcomes in this study demonstrate that the proposed intervention program improved the glycaemic and oral hygiene status in T2DM older patients with chronic periodontitis. On completion of the program, the patients' glycaemic values signi cantly decreased when comparing the DOC group and the control group at 3-and 6-months follow-ups. The results showed an improvement in HbA1c value in the intervention group from 8.94-8.08% in 3-month, and 8.08-7.83% in 6-month. Therefore, the program could reduce HbA1c 0.86% in 3-month, and 0.25% in 6-month. However, the reduction of HbA1c level did not meet the target of American Diabetes Association treatment goal that the HbA1c level must be reduced to less than 7.5 in healthy older people 12 . In other previous studies [24][25][26][27] , periodontal therapy in diabetic patients could signi cantly reduce HbA1c levels from 0.29-0.64% in 3-month and 0.2% in 6-month after completion of the therapy 28 . As the combination of lifestyle changes and dental care program also decreased HbA1c (0.29%) after 3-months follow-ups, whereas the HbA1c increased 0.09% in the control group 19 . Simpson et al. 13 found periodontal therapy could reduce the HbA1c values of 0.29% at 3-month and 0.02% at the 6-month period. However, the effect of periodontal treatment alone on uncontrolled type 2 diabetes mellitus in Thai older subjects was not signi cant although HbA1C values for the treatment group dropped by -0.2% three months after completion of the treatment 29 . Similar to full-mouth non-surgical scaling and root planing, it could improve periodontal health, but it had no signi cant effects on glycaemic control based on HbA1c values of -0.04% 30 .
The oral hygiene status of participants, including OHI-S in the intervention group, also decreased after the DOC program was nished, and it showed the improvement of the OHI-S score when compared with the control group. Previous research showed a similar result that non-surgical periodontal therapy and oral hygiene instruction of T2DM subjects with chronic periodontitis could reduce the plaque index by more than 80% after three months 31 . Intensive oral hygiene care for periodontitis in T2DM patients could signi cantly reduce the plaque index after six months 32 . The oral health instruction of periodontitis patients with T2DM also reduced plaque index signi cantly, and HbA1c was reduced by 0.2% within 6-months period, although the difference was not statistically signi cant 33 .
In this study, the component of the Health Belief Model, perceived susceptibility, perceived severity, perceived bene t of diabetic older patients increased in the DOC group. This increase can be attributed to the effect of education, presenting the image on a slide presentation, distributing lea ets, interactions of patients in the group training and individual oral hygiene consulting also could increase their understanding. The perceived barrier of these patients decreased due to the education intervention, oral care practising, available of dental treatment services, distributing a package of oral care tools. Moreover, perceived self-e cacy also increased by empowering the patient's practical training to increase the capability of controlling and managing their oral health care behaviours continuously. Similar to the previous study found that knowledge, behaviour and attitude toward T2DM and oral health were increased after the intervention 19 whereas the education package in T2DM patients did not affect knowledge scores between the two groups after one month 34 due to lack of motivation periods.
Moreover, a similar study based on the Health Belief Model to promote the oral and dental health of patients with type 2 diabetes mellitus, three months after the intervention, awareness of the patients and perceived susceptibility, bene ts, selfe cacy, internal cue to action, and performance in oral and dental hygiene-related behaviours had increased in the intervention group 35 .
The strengths of this study are 100% response rate, the use of biomarkers to examine outcomes, and a double-blinded technique due to our study has gathered data in one place that dentists and interviewers did not know that patients come from interventions or control groups. Moreover, patients did not know which group they belong to, and there is less contamination between the two groups because they lived in a remote village, and we made an appointment at different times. The limitation of this study is to reduce the oral hygiene index from the DOC program, which cannot ensure that the decline came from changes in oral health behaviour or scaling and root planing treatment since our program included of both activities. We suggest that future research should have a longer follow-up period or repeat DOC program at six months per year. For immediate action, with the potential bene t to all diabetic patients, the DOC program could be used routinely in other health centres or hospitals at least every six months. For further action, we recommend that the policy should make fast-track dental services for diabetic patients to receive scaling and root planning every 6 months for preventing periodontal disease and control glycaemic status. Finally, the program needs to be repeated in other contexts such as in different provinces and cross cultures to con rm the positive effects and broad application in various settings.

Conclusion
The diabetic and oral care program in type 2 diabetic elderly patients with periodontitis was effective in reducing glycaemic status and improving oral hygiene and oral health behaviours. Through the oral health education based on Health Belief Model, scaling and root planing and continuous active follow-ups can improve oral health perception, oral health behaviours, oral hygiene, and glycaemic status within six months among diabetic older patients with chronic periodontitis.