Evaluate Five Different Diagnostic Tests for Dry Mouth Assessment in Geriatric Residents in Long-term Institutions in Taiwan

Yao-Ming Cheng; Shao-Huan Lan; Yen-Ping Hsieh; Shou-Jen Lan; Shang-Wei Hsu


BMC Oral Health. 2019;19(106) 

In This Article


Study Design and Participants

The study design is cross-sectional. We explained our research to 27 appropriated registered LTC institutions in central Taiwan, fifteen of which are recruited in this study intern recruited their residents as research participants. This study recruited reviewers who received training regarding the purpose of the study and content of the questionnaire as well as techniques to measure saliva, the required interview skills, and necessary precautions. These reviewers conducted face-to-face interviews with LTC residents and collected questionnaire data.

The inclusion criteria were residents who had (1) excluded those who are diagnosed as neurocognitive disorder and were willing to have written informed consent; (2) lived in LTC institutions in central Taiwan for at least 1 month and were aged ≥50 years, and (3) could provide clear responses. A total of 577 eligible residents were identified, 568 of them were willing to participate. Of the 568 copies of the questionnaire distributed, 560 valid responses were received.

The study was presented to the Institutional Review Board of the Cheng Ching General Hospital, and was approved (HP140026). The data were collected from August 2014 to March 2015.


The research instrument was a self-developed questionnaire created on the basis of previous studies and scales regarding dry mouth.[7,33,35–37] The content validity was evaluated by three experts with experience of performing oral care services in LTC institutions (two nurses in LTC institutions and one dentist). The reliability was assessed using the Cronbach's α coefficient or the Kuder–Richardson formula 20.

The information obtained from the questionnaire survey comprised the residents' demographic attributes, such as gender, educational attainment, age, type of LTC institutions lived in, marital status, average length of LTC stay (months), daily water intake (mL), number of diseases experienced in the last 6 months, and state of oral health care, (i.e., frequency of tooth brushing and gargling per day, and whether the residents had received teeth cleaning in the previous 6 months).

Questionnaire responses confirmed that the residents' textures of food intake in the LTC institutions were divided into four categories: (1) a soft foods (for older adults with poor chewing ability but normal swallowing function); (2) a finely chopped foods (solid food is processed through methods such as mincing and grinding to produce food that can be swallowed without chewing); (3) a semi-liquid foods (solid food is processed through methods such as mincing and grinding, after which it is added to porridge, soup, and drink to that can be swallowed with or without a little chewing); and (4) a fully liquid foods (semiliquid food is completely liquidized in a juicer). The Barthel Index was used to verify LTC residents' 10 activities of daily living (ADL); their scores ranged from 0 to 100, with a lower score indicating higher reliance on others' assistance in daily life.[38]

The residents' body mass index (BMI) was calculated after their height in centimeter (cm) and weight in kilogram (kg) were measured. The height of bedridden residents or those unable to stand was measured using their knee height (cm).[39] Weight measurements were performed using a weighing scale borrowed from the LTC institution. The approximate amount of water in milliliter (mL) drunk by the residents per day was measured using a 350 mL cup (excluding the amount of water in meals).

Taiwanese short-form of the Oral health impact profile (OHIP-7 T).[40] The Oral Health Impact Profile (OHIP-7 T) was used to measure the residents' oral health status by asking residents about the problems they had experienced in the past year regarding their mouths, teeth, or dentures. Specifically, this instrument comprised the following seven questions: "have you ever experienced problems related to your teeth or dentures?"; "have you ever been interrupted in meal because of problems related to your teeth or dentures?"; "have you ever experienced discomfort because of problems related to your teeth or dentures?"; "have you ever had difficulties of concentrating because of problems related to your teeth or dentures?"; "have you ever experienced difficulties in pronunciation because of problems related to your teeth or dentures?"; "have you ever encountered difficulties in daily life because of problems related to your teeth or dentures?" and "have your sense of taste deteriorated because of problems related to your teeth or dentures?". The residents replied these questions on a 4-point Likert scale (0 = never and 4 = often). The total possible score was 28 points. A lower score indicate more favorable oral health status. The Cronbach's α of this study was 0.95.

Self-perceived ability to chew food.[41] A scale for self-perceived ability to chew food was used to assess the residents' self-evaluated chewing ability. This scale incorporated 24 items: eight fruits were used to test the self-perceived ability of chewing foods with hardness; four fresh foods and meats were used to test the self-perceived ability of chewing foods with toughness; eight cooked vegetables were used to test the self-perceived ability of chewing foods with fracturability; and four viscous foods were used to test the self-perceived ability of chewing foods with viscosity. Table 2 presents the detailed food names. The foods in each category were ranked according to their difficulty to chew.[41] Those listed at the top were the most difficult to chew, and those at the bottom were the easiest to chew. The residents were asked about the difficulty of chewing each category of food, and 1 points were assigned if their response was "easy to chew," whereas 0 points were assigned if their response was "difficult or unable to chew." The highest score for each category is 8 points, yielding a total score of 32 points (total scores of fresh foods, meats, and viscous foods multiplied by 2). A higher score suggested more satisfactory chewing ability. The value of the Kuder–Richardson formula 20 was 0.946, indicating moderate reliability.

Self-perceived levels of dry mouth. Another scale was conceived to measure residents' self-perceived levels of dry mouth. A total of nine items were designed, which were evaluated using a 5-point Likert scale (1 = never and 5 = frequently). The total score is 45 points; a higher score indicated higher levels of self-perceived dry mouth. This scale examine residents self-perceived symptoms of dry mouth in the last month with the following statements: "I feel dryness in my mouth"; "I feel dryness in my lips"; "I feel that my gums are swollen, hurt, or have a burning sensation"; "I need fluids (e.g., soup or water) to help me swallow my food"; "I need to get up at night to drink water"; "I often feel like my mouth is dry after having a meal;" "I have difficulty eating food without water content"; "I feel dryness in my nose"; and "I feel dryness in my eyes." The Cronbach's α was 0.80.

Oral moisture checking. Researches have indicated that decreased oral moisture levels correlate with increased oral dryness.[42–44] Therefore, we measured the actual moisture levels in the mouth of the residents to determine their levels of dry mouth. An oral moisture checking device (Moisture®, approval number: 22200BZX00640000, Life Co., Ltd., Saitama, Japan)[45] was used for the measurement. As shown in Figure 1, the sensor on the left was used to detect the moisture of the oral cavity and the green button was for the measurement; the right side of the figure shows the data displayed after the measurement. Because this device is easy and quick in measurement, it is suitable for use among residents in LTC institutions.[43,46] Measurement was conducted according to the following steps:

Figure 1.

Oral moisture checking device (Moisture®, approval number: 22200BZX00640000, Life Co., Ltd., Saitama, Japan)

  1. The surface of the tongue 10 mm from the apex linguae was measured. The measurement for each resident was conducted using a new sense cover. For each measurement attempt, approximately 200 g of pressure was applied using the device, and a value was then provided after 2 s. Three consecutive readings were taken, and the median of these is the final measurement. Figure 2 presents the measurement using the oral moisture checking device.

  2. The residents were requested to sit and rest for 5 min before the measurement. With reference to the suggestion provided by Saito et al. (2008), the measurement was conducted between two meals, ideally from 10:00 AM to 11:00 AM or from 2:30 PM to 4:30 PM.[47]

  3. Definitions of the measurement values were as follows: ≥29.6 was defined as normal, ≤27.9 was defined as dry mouth, and 28.0–29.5 was defined as borderline dry mouth.[43]

Figure 2.

Measurement of oral moisture degree

Repetitive saliva swallowing test (RSST). The repetitive saliva swallowing test (RSST) was employed to determine the residents' swallowing movements. Oguchi et al. (2000) suggested that RSST is a simple and non-invasive examination.[48] In this test, research personnel assessed the residents' frequency of saliva swallowing by touching their prominentia laryngea through palpation and counting the frequency of swallowing over the span of 30 s. Residents who swallowed saliva three times or more within 30 s were considered to have normal swallowing ability. Those who swallowed saliva only one or two times were considered to have a moderate risk of dysphagia, and those who could not successfully swallow saliva or choked during the process were considered to have a high risk of dysphagia.[49,50]

Statistical Analyses

Statistical analysis was performed by using SPSS version 19. Descriptive statistics including frequencies, percentages, mean and 95% confidence interval (CI) were reported. The correlation coefficient was used to confirm the correlation level between the oral moisture measurement and other variables.

Chi-squared automatic interaction detection (CHAID) was employed to create a decision tree. Chi-square automatic interaction detection (CHAID) is a nonmathematical decision tree that apply the stepwise classification method (tree classification method) to detect interactions. A CHAID technique involves a tree analysis between one dependent variable and several independent variables. This method enables classifying both continuous and categorical variables. CHAID operates by examining dependent variables and independent variable through a chi-square test to identify significant variables. If no significant difference exists between the sequential groups of selected independent variables, these variables are combined as one group. That is, each interval of that variable is automatically separated as being mutually independent from other variables through category merging.[51]

The dependent variable of this study was oral moisture (interval variable). The independent variable consisted of the residents' demographic characteristics, oral health status, OHIP-7 T results, self-perceived ability to chew food, self-perceived level of dry mouth, and RSST results. We input all independent variables in a CHAID model and did not conduct a hierarchical analysis. To determine the selection sequence of independent variables, this study adopted the automatic model selection method of CHAID. SPSS was used to identify the most significantly lowest probability value in order to divide variables into groups. Likelihood ratio was used in the chi-squared tests. To prevent overfitting, the minimum parent node size was set at 100 and the minimum child node size was set as 50. The eligibility level and merge level were set to a p-value threshold of .05.[51,52]