The Psychological Disturbance Associated with Tooth Loss Questionnaire Revisited: Validation on a National Greek Sample

Featured Application

The PDATL questionnaire is an appropriate remote scale that does not require chairside evaluation for outcome measures in cases where tooth loss is restored with partial or complete dentures.

Abstract

Background. Edentulism is considered a chronic disability since tooth loss, and the required prosthodontic rehabilitation with removable prostheses might irreversibly affect function and aesthetics and significantly impact patients’ psychological well-being and social interactions. The present study aimed to translate, validate, and adapt in the native population the recently developed nine-item self-reporting psychological disturbance associated with tooth loss questionnaire (PDATL). The well-established Depression, Anxiety, and Stress Scale (DASS-21) questionnaire was also employed to identify potential emotional states of depression, anxiety, and stress caused by tooth loss and denture-wearing. Methods. For the present analysis, nationwide baseline data from two samples were used. The participants were using clinically acceptable complete or partial dentures for at least 1-year and were interviewed from May 2021 to October 2021. The dentures were fabricated in two different Dental School settings, the National Kapodistrian University of Athens (NKUA) and the Aristotle University of Thessaloniki (AUTH). Data from both clinical studies were analyzed similarly, in a parallel fashion, to enhance the generalizability of the results. Cronbach’s alpha coefficient measured internal consistency reliability, whereas Principal Component analysis using a Varimax with Kaiser Normalization rotation was subsequently performed to test the questionnaire’s internal structure regarding its pattern and structure matrix. All analyses were performed at α = 0.05 level of significance with statistical software (SPSS Statistics for Windows, v20.0; SPSS Inc. (Chicago, IL, USA)). Results. In both data sets, all items except for “discomfort/pain” and “difficult to relax” contributed significantly to the scale structure. The two items that presented weaker contributions indicated that they were redundant. DASS-21 statements revealed no severe negative emotional symptoms in both data settings. However, tooth loss was acknowledged by the participants as a critical factor of distress in their life. Conclusion. The PDATL questionnaire has several advantages since it allows for summary score calculation, and due to its shortened content, it can be easily distributed during a telephone interview. This is important in older tooth loss sufferers where digital technology has an additional advantage as it allows for remote denture fabrication through digital workflow.

1. Introduction

Extensive edentulism and total tooth loss are considered irreversible oral diseases. Consequently, these permanent conditions due to the individualized disturbance of the oral anatomy implicate functional health and aesthetics and cause considerable psychological discomfort and satisfaction with self-image impairment [1]. Prosthodontic rehabilitation treatment options should aim to provide good quality dentures by considering the pathology-driven outcomes and the patient’s perceptions and preferences based on their individual needs during clinical decision making [2].

Implementing CAD/CAM systems in the denture workflow is gaining popularity and gradually replacing conventional procedures in fabricating removable prostheses. It presents multiple advantages, such as reduced clinical and post-insertion time, shorter appointments, better fit with more precision, and reproducibility with fewer clinical steps compared to conventional dentures [3]. Despite the conventional or digital workflow, a complete or partial denture should promote functional health and improve self-image since self-esteem weakening could influence an individual’s behavior, social interaction, and relationships [4]. Since the 1990s, international research has focused on documenting the effectiveness of denture treatment, aiming to provide valuable clinical screening tools to assess the treatment efficacy based not only on pathology-driven conditions but also on patients’ perceptions [5]. More than ever, short and easy-to-use practical diagnostic tools are essential to highlight changes in functional health, psychological well-being, and positive body image; to evaluate whether digital workflow presents comparable and practically efficient outcomes to the conventional ones [3].

Recently, Kudsi et al. [6] introduced a brief questionnaire, the psychological disturbance associated with tooth loss questionnaire (PDATL), that included nine items. Four items aimed to assess functional health, such as pain and problems in speaking and/or eating, and five were related to psychological discomfort, body image impairment, and social interactions. Kudsi et al.’s nine-item questionnaire was then related to the well-established Depression, Anxiety, and Stress Scale (DASS-21) questionnaire, further validating the impact of tooth loss and denture treatment [6]. Finally, the authors suggested the nine-item questionnaire as a disease-specific measure of denture treatment effectiveness for assessing the psychological disturbance of tooth loss based on the emotional impact and personality traits to provide replacement options for patient-centered care [7]. The present study aimed to (a) Evaluate the PDATL item inter-relations to identify redundant questions and clarify their conceptual validity. The PDATL was adopted in the native language with baseline data derived from two clinical trials that covered a nationwide sample representative of cultural and environmental influences responsible for psychological disturbance and impaired body image. (b) Provide evidence of the conceptual validity of the analysis of significant components to explore if further item removal ends up with a clear-cut scale appropriate to a broader sample profile and various treatment workflows.

2. Materials and Methods

2.1. Data Source

The PDATL questionnaire consists of nine questions assessing functional health (items one to four) and emotional/body image (items five to nine). The questionnaire uses a 5-point Likert scale to rate the items representing the frequency of the disturbance, “never”, “rarely”, “sometimes”, “often”, and “very often” corresponding to the numeric values 5, 4, 3, 2, and 1, respectively. It also includes the “don’t know” option, which corresponds to the numeric value 6 [6]. The original English version of the questionnaire was adopted to the Greek language using the forward and backward translation methodology to utilize it as a documentation tool and further validate it as a summated rating scale to formulate a valuable tool in clinical settings [8]. The DASS-21 questionnaire was also employed to identify potential emotional states of depression, anxiety, and stress caused by tooth loss and denture-wearing. DASS-21 has been successfully adopted in the Greek language in a previous study [9]. Furthermore, the Distress Thermometer (DT), a visual analog scale used by the National Comprehensive Cancer Network to screen distress in cancer patients, was applied to measure the overall impact of tooth loss and denture-wearing. DT measures distress directly on a 10-point scale ranging from “no distress” to “severe distress” and is available in Greek [10].

Two nationwide samples were collected at baseline, and data were organized in two studies according to the study design followed in a previous research report [11]. The first sample was acquired from the Dental School at the National and Kapodistrian University of Athens (NKUA) (study 1-NKUA), and the second sample was acquired from the Dental School at the Aristotle University of Thessaloniki (AUTH) (study 2-AUTH). Calibrated interviewers (one interviewer for study 1-NKUA and two for study 2-AUTH) conducted telephone interviews following the questionnaire items. Since two calibrated examiners interviewed the participants in the AUTH data set, the Mann–Whitney non-parametric test was performed at a statistical level of α = 0.05 to explore statistically significant differences between the two interviewers. The results showed that the distribution of items 1, 2, 3, 5, and 6, 7, 8 was not statistically significantly different between the two examiners, with corresponding p-values 0.81, 0.61, 0.90, 0.80, 0.40, 0.19, and 0.19, respectively. However, for items 4 and 9, the corresponding p-values were statistically significant (p < 0.01 and 0.04, respectively), indicating that the participants presented with a wide variation in “discomfort/pain” threshold levels, as well as in the “difficult to relaxing” perception when interviewed by the two different interviewers.

Adult patients who responded positively to the study call from May 2021 to October 2021 and were successfully wearing dentures for at least one year were recruited for this study. The dentures were fabricated at the corresponding University clinical settings following the same fabrication procedures and construction protocol. All participants were ambulatory, living in the community, presented no mental illness, and came to the university clinical setting for rehabilitation with complete or removable dentures. NKUA and AUTH Schools of Dentistry Committees for Research Ethics approved the protocols of these studies (IRB# 455/25.01.2021 and 13/14.04.2021, respectively). All participants signed an informed consent according to the general recommendations of the Declaration of Helsinki.

2.2. Participant Profile

The study 1-NKUA recruited 160 participants with a mean age of 64 ± 10 years; sixty-two participants were female, of which 28 belonged to the older adult group (age 65+, within the range 65 to 94 years old) and 34 in the younger group (age ≤ 64, within the range 38 to 64 years old). In contrast, 98 participants were male, 54 belonging to the older adult group and 43 to the younger group. Ninety-two (58%) participants had mixed dentitions wearing a complete denture on one jaw, a partial denture on the opposing jaw, and 68 (42%) wearing a pair of complete dentures. All participants with the mixed dentition had artificial functional tooth units. One hundred and seventeen (73%) participants, in a ratio of 2:1, wore their dentures from one to three years.

The reason for recruiting adults who, although suffering from extensive tooth loss, still retain some natural teeth was to explore how this limited presence contributed to better functional health and body image. The study 2-AUTH recruited 105 complete denture wearers as participants with a mean age of 69 ± 8 years to further explore functional and emotional disturbances from edentulism, potentially leading to a lower rating of body image. Forty-four participants were female, of which 34 belonged to the older adult group (age 65+, within the range from 65 to 83 years old) and 10 in the younger group (age ≤ 64, within the range 52 to 64 years old). In contrast, 61 participants were male, 42 belonging to the older adult group (age 65+, within a range from 65 years to 90 years), and 19 to the younger group (age ≤ 64, within the range from 55 years to 64 years). All participants wore complete dentures from one to three years. The participants’ distribution in gender and age is illustrated in

Table 1. Participants’ profiles in gender and age distribution of both settings.

	Study 1-NKUA			Study 2-AUTH
Participants	Gender
Age	Female	Male	Total	Female	Male	Total
Older (65+)	28 (45%)	54 (56%)	82 (52%)	34 (77%)	42 (69%)	76 (72%)
Younger (≤64)	34 (55%)	43 (44%)	77 (48%)	10 (23%)	19 (31%)	29 (28%)
Total	62 (100%)	98 (100%)	160 (100%)	44 (100%)	61 (100%)	105 (100%)

.

2.3. Statistical Analysis

Data from both studies were analyzed similarly and in parallel to enhance the generalizability of the results. All analyses were performed at α = 0.05 level of significance with statistical software (IBM SPSS Statistics for Windows, v26.0; IBM Corp, Armonk, NY, USA).

3. Results

3.1. Descriptive Statistics of PDATL Questionnaire

Descriptive statistics of the two data sets for the nine-item PDATL, including the mean, standard deviation (SD), and variance (Var), are presented in

Table 2. Descriptive statistics of the PDATL-1 questionnaire in both data sets.

Domains	Items (I)		Study 1-NKUA				Study 2-AUTH
			Mean	SD	Var	Mean	SD	Var
Functional Health	I1	Trouble speaking	3.98	1.18	1.40	3.50	1.08	1.16
	I2	Trouble eating	3.64	1.24	1.54	3.45	1.19	1.40
	I3	Change diet	3.61	1.23	1.51	3.60	1.17	1.36
	I4	Discomfort/pain	3.94	1.28	1.63	3.85	0.85	0.73
Body Image	I5	Uncomfortable/impact on the appearance	2.79	1.44	2.08	3.51	1.09	1.20
	I6	Think a lot about tooth loss	2.68	1.41	1.99	3.61	1.14	1.30
	I7	Avoid social situations	4.29	1.14	1.30	4.09	0.98	0.96
	I8	Stress in relationship/marriage	4.24	1.14	1.29	4.50	0.90	0.81
	I9	Difficult to relax	4.46	1.02	1.04	3.74	1.01	1.02

. The distribution of the participants’ responses within the Likert scale was comparable in both data sets, as more than 50% of the participants presented rare or no functional disturbance (

Table 3. The distribution of the answers within the Likert scale for both data sets.

		I1	I2	I3	I4	I5	I6	I7	I8	I9
very often	Study 1-NKUA	8 (5%)	14 (9%)	12 (8%)	12 (7%)	44 (27%)	48 (30%)	8 (5%)	7 (4%)	7 (4%)
	Study 2-AUTH	5 (5%)	9 (9%)	7 (7%)	1 (1%)	8 (8%)	5 (5%)	3 (3%)	3 (3%)	4 (4%)
often	Study 1-NKUA	10 (6%)	13 (8%)	16 (10%)	4 (4%)	28 (17%)	29 (18%)	4 (2%)	7 (4%)	1 (1%)
	Study 2-AUTH	11 (10%)	14 (13%)	13 (12%)	4 (4%)	5 (5%)	14 (13%)	3 (1%)	1 (1%)	7 (7%)
sometimes	Study 1-NKUA	35 (22%)	39 (24%)	43 (27%)	29 18%)	30 (19%)	30 (19%)	27 (17%)	25 (16%)	18 (11%)
	Study 2-AUTH	37 (35%)	22 (21%)	19 (18%)	29 (27%)	37 (35%)	24 (23%)	8 (8%)	8 (8%)	25 (24%)
rarely	Study 1-NKUA	31 (19%)	45 (28%)	40 (25%)	30 (19%)	33 (21%)	33 (21%)	16 (10%)	23 (14%)	19 (12%)
	Study 2-AUTH	31 (30%)	41 (39%)	42 (40%)	47 (45%)	35 (33%)	36 (34%)	21 (20%)	21 (20%)	45 (43%)
never	Study 1-NKUA	76 (48%)	49 (31%)	49 (30%)	78 (49%)	25 (16%)	20 (12%)	105 (66%)	98 (62%)	115 (72%)
	Study 2-AUTH	21 (20%)	19 (18%)	24 (23%)	24 (23%)	20 (19%)	26 (25%)	72 (68%)	72 (68%)	24 (23%)

). In the study1-NKUA, the participants showed a higher percentage range of rare or no functional problems (between 56% and 67%). Moreover, as expected, the denture wearers of Study-2-AUTH presented a lower percentage range of rare or no disturbance (between 50% and 57%) in items of functional health. Notably, all participants provided complete answers to all questions, except that no participants from both studies used the response “don’t know” in any of the questions. The remaining percentage of complete denture wearers who presented with functional disturbances was still substantial, potentially attributed to the quality of the complete dentures or anatomical conditions that contributed to the observed functional disorders.

3.2. Defining Internal Consistency and Construct Validity of the PDATL Scale

3.2.1. The Study 1-NKUA

Reliability analysis was performed at a significance level of α = 0.05 for the study1-NKUA data set. Cronbach’s α coefficient and corrected item-total correlations were used to explore the inter-relationship of all nine items and identify redundant questions. Cronbach’s α coefficient was 0.92, and all items were found to contribute to the scale significantly for this data set. Thus, all items were considered unidimensional, strongly associated with each other, and represented a single concept [12].

Principal Component analysis using a Varimax with Kaiser Normalization rotation, at a significant level of 0.05, was subsequently performed to test the questionnaire’s internal structure regarding its pattern and structure matrix; thus, appropriate for data reduction, enabling the removal of redundant questions [2]. For each factor, for the loading value (the percentage of variance in an original variable explained by a factor) to be considered significant based on the sample size of this data set, the factor loading coefficient was set at 0.45 for the study 1-NKUA. In this data set, three principal components explained the total variance of 82% (Kaiser–Meyer–Olkin Measure of Sampling Adequacy (KMO) was 0.86, and Bartlett’s Test of Sphericity (BTS) Sig = 0.00). The KMO results indicated that the identified three underlying grouping factors (components) appropriately explained the correlations between variables. The grouping factors were component (1): items 1,2,3, component (2): items 4,5, and 6, and component (3): items 7, 8, and 9. Moreover, the BTS results indicated that the formulated null hypothesis assuming that the existing correlation matrix is not different from an identity matrix and therefore questions would be unrelated, was rejected (p < 0.05) and confirmed the offered correlations.

In light of establishing higher standards for the formulated scale, new tests were performed by splitting the samples by gender and age using all nine items [2]. It was proved that the loading pattern was similar for both age groups, females, participants with mixed dentitions consisting of functional tooth units (FTUs) on artificial teeth, and participants having their dentures in use between 2 and 3 years. The complete denture wearer group was potentially small for item 4 to contribute to the measurement model. Moreover, item 4 did not contribute to factor loading in recent both partial and complete denture wearers. Additionally, although males were a large group of participants, item 9 did not contribute to the factor model in the male group. Reliability analysis was rerun after removing items 4 and 9 at a significant level of 0.05. Cronbach’s α coefficient and corrected item-total correlations were again used to reveal the potential existence of low contribution items in the scale and remove them. Cronbach’s α coefficient was 0.89, slightly lower, but all seven items excellently contributed to the scale. Principal Component analysis using a Varimax with Kaiser Normalization rotation, at a significant level of 0.05, was again performed to test the 7-item scale’s internal structure regarding its pattern and structure matrix. In this 7-item scale, three principal components were highlighted that explained the total variance of 87% (Kaiser–Meyer–Olkin Measure of Sampling Adequacy KMO = 0.80 and Bartlett’s Test of Sphericity (BTS) Sig = 0.00) (

Table 4. The rotated component matrix of the 7-items of the PDATL-1 for the study 1-NKUA.

	Components
	1	2	3
Item 1	0.904
Item 2	0.872
Item 3	0.812
Item 5		0.885
Item 6		0.918
Item 7			0.765
Item 8			0.890

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 5 iterations.

).

3.2.2. The Study 2-AUTH

Reliability analysis was also performed at a significant level of 0.05 for all nine items in the study 2-AUTH data set. Cronbach’s α coefficient and corrected item-total correlations were used to reveal the potential of low contribution items in the scale and remove them. Cronbach’s α coefficient was 0.90, and seven items significantly contributed to the scale. Items 4 and 9 presented weaker contributions, indicating that items 4 and 9 were redundant. Thus, they were removed and not included in the following Principal Components analysis. For each factor, for the loading value to be considered significant based on the sample size of the study 2-AUTH data set, the factor loading coefficient was set at 0.55. That data set, which was further validated by age and gender, also confirmed the three principal components, as in study 1-NKUA, which explained the total variance of 83% (KMO = 0.80 and BTS Sig = 0.00) (

Table 5. The rotated component matrix of the 7-items of the PDATL-1 for the study 2-AUTH.

	Components
	1	2	3
Item 1	0.780
Item 2	0.904
Item 3	0.792
Item 5		0.553
Item 6		0.817
Item 7			0.817
Item 8			0.887

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 5 iterations.

).

The three-factor solutions corresponded to the construction of sub-scales with identifiable dimensionality related to the following conceptual definition. The first construct consisted of items 1, 2, and 3 communicating functional speaking and eating difficulties. The second construct consisted of items 5 and 6 linked with how the participants perceived self-image impairment. Finally, items 7 and 8 corresponded to how self-image influenced the individual’s behaviors, social interaction, and relationships. The participants of both these studies failed to describe how they perceived discomfort or pain from tooth loss and whether it was difficult for them to relax because of their tooth loss.

3.3. Defining DASS Statements and the Impact of Tooth Loss on Life

Frequencies were provided on a visual analog scale rating the effect of tooth loss in life and the DASS questionnaire in both data settings. The participants rated the tooth loss impact from 0 (no distress) to 5 (mild distress) at 29% for study 1-NUKA and 39% for study 2-AUTH. Moreover, the rating ranged from 6 to 10 (severe distress), 71% for study 1-NKUA and 61% for the study2-AUTH.

Frequencies were also provided on the DASS questionnaire’s responses in both data settings. No significant differences were found in the rating of the answers. All participants rated DASS statements mostly with 0 and 1, between 74.4% and 99%. Those ratings showed that tooth loss and the treatment of the participants of both studies did not challenge severe negative emotional symptoms regarding depression, anxiety, and stress; however, they pointed out a diverse impact on their life.

4. Discussion

The PDATL questionnaire has been recently introduced in two studies [6,7], where functional health and body image impairment related to negative depression, anxiety, and stress emotional states described in the well-established DASS questionnaire and to personal behavioral traits described in Eysenck short-form questionnaire. They aimed at providing replacement options for tooth loss through a patient-centered treatment. The present study adds significant additional validation by confirming the inter-relationship between the variables and highlighting the PDATL scale’s good properties as an outcome measure for prosthetic interventions where tooth loss has occurred.

A three-factor solution was revealed where the first factor was linked to functional health, including trouble speaking, trouble eating, and dietary changes, which are the primary operational concerns related to denture wearing and chewing on artificial tooth units. The second and third factors were related to body image; however, the second factor referred to an inner sense of body dissatisfaction, while the third factor to an outer sense of negative body image, which entailed an emotional-behavioural component that also affected relationships with others.

The correlation matrices indicated removing pain/discomfort because of tooth loss. Discomfort/pain appeared not to be critical for identifying functional problems related to conventional partial and complete dentures for both data sets. Thus, we pulled out that question since it was not loaded by most participants in study 1-NKUA and all participants in study 2-AUTH. Likewise, we removed the question on the difficulty of relaxing as the participants of both studies at most significant percentage did not worry about that perception. Both questions were discarded by PCA when the 9-item questionnaire and additional analyses after the removal of these items led to a 7-item, more concise version with a similar excellent internal consistency (i.e., 0.89 to 0.92 for the study 1-NKUA and 0.90 for study 2-AUTH). The shortened version of the scale highlighted the participants’ lower disturbance of tooth loss, as displayed by the variability of the responses rating.

The nationwide sample from both studies constitutes an asset as the same cultural and environmental factors influence them. Those common factors indirectly reinforced existing behavioral attitudes, personality traits, and eating habits. They were also influenced comparably by receiving treatment with removable partial and complete dentures under identical clinical protocols in university settings. Additional data quality was that all participants chewed on artificial tooth units; therefore, the samples of the two studies were considered parallel in function and mainly regarding eating and dietary habits. The higher percentage range of rare or no functional problems observed in the study 1-NKUA could be attributed to the lower total number of complete denture wearers and the inclusion of partially dentate participants who retained some of their natural teeth. In addition, since the participants were motivated patients who sought treatment for their tooth loss, the DASS statements revealed no severe negative emotional symptoms, and the results were similar in both data settings. However, tooth loss was acknowledged as a critical factor of distress in their life. That fact seemed to be compensated by the good quality dentures people received at university clinics, where the protocols were well established within the student curricula.

The present theoretical model offers ground for calculating summary scores for each item subset, differentiating amongst different groups, and confirming longitudinal response changes in future studies. An additional advantage raised was its shortened content which can readily be displayed by telephone interview to assess functional disturbances and how they affect denture wearers’ behaviors, interactions, and relationships within the close or broader environmental settings. That possibility is an excellent asset in recent years, where clinicians are led to remote digital workflows in denture fabrication, which are gaining popularity due to ease of reproducibility without patient involvement. The interview potentially provides the remote subjective perception of denture treatment options to add to the pathology-driven procedures in decision-making, which is paramount for patient-centered care. Limitations of the present study include the addition of a third group of complete or partially edentulous patients rehabilitated with implant-supported prostheses.