Cross-sectional clinical validation of the Periodontal Disease Self-Report measure

Abstract

Background:

The Periodontal Disease Self Report (PDSR) measure was originally created and psychometrically validated using a nationwide sample via online data collection. No clinical parameters were included in the prior validation of the PDSR. Thus, this study evaluated potential evidence for clinical validity of the measure by examining associations between the PDSR scores and various clinical parameters obtained from a new sample of participants in which full-mouth periodontal examinations were conducted.

Methods:

Adults from a community sample (n = 114) provided demographic information, responded to the PDSR measure, and received a full-mouth clinical periodontal examination. Individual self-report items, subscale scores, and total scores obtained from the PDSR were evaluated against clinical parameters of periodontitis. Regression models and receiver operating characteristic statistics were also utilized to test the ability of the PDSR to predict clinical outcomes.

Results:

PDSR total scores were positively correlated with mean probing depth (r = 0.50, p < 0.01) and mean clinical attachment loss (r = 0.52, p < 0.01). After accounting for common risk factors in periodontal disease, the PDSR predicted mean probing depth (β = 0.45, 95% CI: 0.02–0.04; (ΔR² = 0.19). The area under the curve for the PDSR scores distinguishing between CDC/AAP no/mild periodontitis and moderate/severe periodontitis categories was 0.71 (95% CI: 0.62–0.81).

Conclusions:

Clinical data support the use of the PDSR measure as a screening tool for periodontal disease. Additionally, the PDSR may offer added utility compared to other measures due to less reliance on information obtained via clinical encounters.

Introduction

Self-reported oral health status remains a necessary approach for capturing population oral health outcomes, including in periodontology, given the low cost and ease of conduct relative to clinical examination (Abood et al. 2016; Blicher et al. 2005). Periodontitis continues to be exceedingly common, particularly among older adults (Eke et al., 2016; Eke et al., 2020) and self-report measures that aim to capture the complex nature of the condition are needed. Individual self-report items and composite scales (i.e., measures that combine multiple items to create a total score) have been tested and compared to various clinical parameters and in various languages and contexts (Eke et al 2018; Carra et al 2018; Deng et al. 2021). The most used self-report measures of periodontitis in population surveillance and oral health research, however, are limited in scope and rely heavily on knowledge obtained through clinical encounters that involve diagnosis and treatment (Eke et al. 2013). Considering the documented lack of access to oral health care and periodontal treatment and maintenance regimens, particularly for low-income and marginalized populations (Jin et al. 2011), along with the estimated population prevalence of periodontitis exceeding 40% among adults aged 30 years and older in the United States (Eke et al. 2018), continued development of self-report measures in combination with clinical validation is warranted (Eke et al. 2020; Eke et al. 2013; Heaton et al. 2017).

Self-report measures of periodontitis often are written in a way that expects patients to have prior clinical experiences or interactions with providers and vary in their relative sensitivity or specificity in discriminating between various disease phenotypes. That is, other extant self-report measures of periodontitis lack the ability to assess multi-dimensional phenotypes or clinical nuances of the disease (e.g., physiological versus functional limitations). This is particularly important given it is often the functional limitations of periodontitis that impact oral health-related quality of life (Tonetti et al. 2017). Thus, additional work is needed to improve the ability to detect meaningful clinical correlates of periodontitis using self-report measures (Carra et al. 2018). Such an approach may be particularly advantageous for populations in which access to oral health care services is limited, such as for rural and/or remote populations, and where methods to support teledentistry, including the triaging of signs and symptoms of oral disease and dental care needs, is of interest (Reiniger et al. 2020).

The Periodontal Disease Self Report (PDSR) measure is a 17-item scale that was partially based on previously published and validated items that were further adapted by including clinical signs and symptoms (Wright et al. 2021). The PDSR was initially validated using a latent-variable and psychometric approach in a general nationwide USA sample using online survey responses (Wright et al. 2021). The measure was demonstrated to be internally consistent and moderately sensitive to a single item self-report question about periodontal disease (Wright et al. 2021). PDSR scores were congruent with prior literature in terms of associations with periodontitis risk factors. Given the online nature of data collection, however, clinical data were not available at the time to provide evidence of clinical validity.

From a psychometric or scale development perspective, exploring or testing whether PDSR scores correlate or are associated with various clinical parameters can help to provide evidence for content or convergent validity of the self-report measure (DeVellis 2017). That is, comparing clinical parameters to the self-report measure allows us to understand the extent to which there is congruence between the two ways of assessing the underlying construct, in this case, periodontitis. Examining clinical data in comparison to the PDSR is an important step in the validation process and evaluating the validity of using such a measure in periodontology.

The aim of this study was to clinically validate the PDSR by correlating self-report scores with commonly used clinical case definitions of periodontitis including the Centers for Disease Control and Prevention and American Academy of Periodontology (CDC/AAP) status, as well as various clinical parameters of the disease such as probing depth or attachment loss. It was hypothesized that higher PDSR total and subscale scores would be significantly associated with increased mean probing depth and attachment loss, greater number of teeth with deeper probing depth, greater clinical attachment loss, a greater percentage of sites with bleeding on probing, and greater percentage of sites with plaque. It was also expected that higher PDSR scores would be associated with more severe disease states, and lower PDSR scores less severe disease, independent of common risk factors.

Methods

This study was part of a larger project targeted at understanding biobehavioral correlates of oral inflammation. Participants included were from a cross-sectional, community-based sample of 114 adults in the greater Morgantown, West Virginia area in the United States. Participants were recruited through a convenience sampling method resulting from the distribution of advertisements in dental clinics or non-profit organizations, word of mouth, and from the graduate periodontics clinic at the West Virginia University School of Dentistry. Advertisements were distributed to diverse groups to avoid excess homogeneity in the sample. Given the community sample, the desire was to cast a broad net and capture the variability inherent in the community.

Potential participants were scheduled to come into the clinic when they called in response to a flyer or referral from a friend/colleague. Once they arrived at the study site (i.e., the graduate periodontics clinic at West Virginia University), they were met by a study staff member who confirmed eligibility (e.g., 18 years or older, English speaking, no other inflammatory conditions, not pregnant), conducted the informed consent process, and provided them with initial self-administered questionnaires. Following completion, they were seated for the clinical examination, after which they completed other psychosocial surveys, including the PDSR. All procedures were conducted in accordance with West Virginia University’s Institutional Review Board (Protocol #1902473418) requirements and procedures and conforms to STROBE guidelines. REDCap (Harris et al. 2009) was used to facilitate screening, as well as data entry, collection, and management.

The first participant was enrolled on September 18, 2019. After enrolling 51 participants, recruitment was paused from March 14, 2020, until September 16, 2020, and again from December 5, 2020, until February 27, 2021 due to impacts of the COVID-19 pandemic. The last day of data collection was March 19, 2021. Other than additional safety protocols being initiated (e.g., the wearing of additional personal protective equipment, taking temperatures, and screening participants for illness), the procedures for the study remained unchanged.

Periodontal examination and assessment

A board-certified periodontist and a trained and calibrated periodontal resident conducted full-mouth clinical periodontal examinations for all participants. Repeat examinations on approximately 10% of the participants were conducted to ensure appropriate calibration and determine intra- (i.e., compared to self) and inter- (e.g., periodontist repeat of resident) examiner reliability via kappa scores. A substantial majority (85.1%) of the exams were conducted by the board-certified periodontist (AS) and several (13.2%) by the periodontics resident (SP). Two individuals (1.8%) were not examined due to missing teeth at the time of the visit but who had passed initial screening.

All study-related periodontal examinations followed the National Health and Nutrition Examination Survey (NHANES) oral health examination procedures (Centers for Disease Control and Prevention 2013), as follows. The examiner probed six sites per tooth (excluding third molars), using a UNC 15 periodontal probe graduated from 1 to 15 mm. First, the distance between the free gingival margin and the bottom of the sulcus or periodontal pocket (i.e., pocket depth) was measured. Next, the examiner measured the distance between the free gingival margin and the cemento-enamel junction (CEJ). For cases in which the gingival margin was below the CEJ (i.e., measure of gingival recession), recession was documented. Each of these measures was taken from the mesial, mid, and distal sites on both the buccal and lingual sides of each tooth (i.e., six sites per tooth). Each measurement was called out by the examiner to a trained research assistant and recorded on a printed paper chart. Bleeding on probing was also recorded for each site probed, and plaque scores were recorded for the mesial, mid, and distal sites of the buccal side and the mid site on the lingual (i.e., total of four sites per tooth). Each of the four sites were charted by either the presence or absence of plaque (Carvalho et al. 2023; O’Leary et al., 1972). Missing teeth or severely decayed teeth where the periodontist rendered probing or attachment immeasurable were also recorded.

Clinical periodontal disease measures

Clinical attachment loss was calculated as the sum of the probing depth and recession in millimeters if the gingival margin was below the CEJ. If above the CEJ, attachment loss was calculated by subtracting the distance from the CEJ to the gingival margin from the probing depth. Additionally, mean probing depth, mean attachment loss, number of missing teeth other than third molars, the percent of sites that bled on probing, the percentage of sites with plaque, and the number of teeth with probing depths greater than or equal to 4 millimeters (mm) or 5 mm, as well as the number of teeth with clinical attachment loss greater than or equal to 3mm, 4mm or 6mm were calculated. Using these parameters, an Excel sheet calculated the data and assigned diagnostic categories according to the 2012 CDC/AAP case definition (e.g., no disease, mild, moderate, severe) (Eke et al. 2012).

Self-report periodontal disease measures

The PDSR (Wright et al. 2021) includes 17 items targeted at various symptoms and experiences related to periodontal disease. Participants were asked to indicate on a three-point scale how true (“Not at all true;” “Somewhat true;” or “Definitely true”) statements were about their gum health. Example items include “I have gum problems” or “My teeth seem longer than they used to.” Items are summed to create a total score. The measure consists of two subscales, one focused on physiological symptoms and one focused on functional symptoms. The subscales were derived from exploratory factor analysis (Wright et al. 2021), and each describe self-reported status related to symptomatology of periodontal disease. The physiological symptoms subscale asks participants to rate items such as bleeding gums, whereas the functional symptoms subscale asks participants to rate items such as difficulty chewing. A copy of the PDSR can be found in the online supplementary materials.

Covariates and Sample Characteristics

Common risk factors of periodontal disease were measured via the self-report questionnaires, including age in years, gender (female, male, or other), body mass index (BMI; assessed continuously), smoking status (‘not at all’, ‘some days’, ‘every day’), and diabetes status. The parent study assessed diabetes as a variable, but the prevalence was so low in the sampled population that it would not be informative. Additional characteristic data were collected including number of years of education and income.

Statistical analyses

To explore the associations between the self-reported PDSR scores and clinical parameters, probing depth, clinical attachment loss, the percentage of sites with bleeding on probing, the percentage of sites with plaque, missing teeth, and severely decayed teeth were evaluated in relation to the PDSR total and subscale scores. This was done using correlations, analysis of variance (ANOVA), and multiple linear regression. ANOVA was used to compare CDC/AAP periodontal status with the PDSR scores as the dependent variable. Additionally, multiple linear regression models were fit to assess the additive explanatory power of the PDSR scores in predicting mean probing depth and mean clinical attachment loss after accounting for common risk factors for periodontal disease such as participant age, gender, smoking status, and BMI. Diabetes status prevalence in the sample was too low to include in the analysis. The first model included only covariates. PDSR scores were then included in the second model. Beta coefficients and the change in r-squared were evaluated. Finally, receiver operating characteristic (ROC) and area under the curve analyses were conducted to evaluate the sensitivity and specificity of the PDSR to discriminate between categories of periodontal disease. Listwise deletion was used to handle missingness resulting in a net n of 98 to be included in the regression models. All statistical analyses were conducted using SPSS (IBM).

Results

Participants (n = 114) were 36.1 (SD = 14.0) years old on average, approximately half (50.9%) of whom identified as female. Approximately three-quarters of the sample was White and 21.1% were Black or African American, which is more diverse than the catchment area used for recruitment¹⁴. Participants had an average of 15.2 years of education (SD = 2.9). In terms of the distribution of periodontal disease, 46.5% of the sample had no disease, 12.3% had mild disease, 34.2% had moderate, and 5.3% had severe disease. Sample characteristics are displayed in Table 1.

Table 1.

Sample characteristics (N = 114).

	Mean / N	SD / %
Age (years)	36.1	14.0
Gender
Female	58	50.9
Male	50	43.9
Other	3	2.6
Missing	3	2.6
Race/Ethnicity
White	86	75.3
Black/African American	24	21.1
Asian	1	0.9
American Indian/Native American	1	0.9
Other	1	0.9
Missing	1	0.9
Education (Years)*	15.2	2.9
Income
Less than $10,000	17	14.9
$10,000-$14,999	20	17.5
$15,000-$24,999	16	14.0
$25,000-$34,999	20	17.5
$35,000-$49,999	8	7.0
$50,000-$74,999	18	15.8
$75,000-$99,999	3	2.6
$100,000-$149,999	4	3.5
$150,000-$199,999	6	5.3
Missing	2	1.8
Body Mass Index (BMI)*	28.2	8.2
Diabetes Status
Yes	4Φ	3.5
No	108	94.7
Missing	2	1.8
Current Smoker
Not at all	88	77.2
Some days	9	7.9
Every day	16	14.0
Missing	1	0.9
Periodontal Disease Self Report (PDSR)
Total Score	7.8	6.9
Physiological Symptoms Subscale	5.8	5.2
Functional Symptoms Subscale	2.0	2.4
CDC/AAP Status
No periodontitis	53	46.5
Mild periodontitis	14	12.3
Moderate periodontitis	39	34.2
Severe periodontitis	6	5.3
Missing (no exam)	2	1.8
Number of missing teeth (not 3rd molar)	1.5	2.5
Number of grossly decayed teeth	0.2	1.0
Mean probing depth	2.3	0.5
Mean attachment loss	0.8	1.1
Percent of sites with BOP	20.9	17.5
Percent of sites with plaque	72.5	27.9
Number of teeth with CAL 3+	4.5	7.2
Number of teeth with CAL 4+	2.3	5.0
Number of teeth with CAL 6+	0.6	2.5
Number of teeth with PD 4+	6.3	6.4
Number of teeth with PD 5+	2.2	4.1

Notes:

^*=7 missing Education; 7 missing Body Mass Index;

^Φ=2 type II and 2 type I diabetes. CDC/AAP = Centers for Disease Control and Prevention/American Academy of Periodontology; BOP = bleeding on probing; CAL = clinical attachment loss; PD = probing depth. Missingness on continuous variables was not greater than n = 2 on any given variable other than plaque scores (missing = 25) due to collection of this variable started later in the study.

Regarding periodontal rater reliability, the board-certified periodontist intra-rater reliability for probing depth was measured using an average measures ICC and was 0.92. Absolute agreement plus or minus 1mm was 99.9%. For the measurement of free gingival margins, the average measures ICC was 0.96 and 99.9% absolute agreement plus/minus 1mm. In terms of interrater reliability between the periodontist and the resident rater, the exact agreement for probing depth and free gingival margins was 64.3% but within plus or minus 1mm, it was 98.2%. The average measures ICC for interrater reliability was 0.63.

PDSR total scores and subscale scores were associated with several of the periodontal clinical examination data parameters. PDSR total scores were moderately correlated (0.36 < r < 0.53) with periodontal disease status, mean probing depth, mean attachment loss, the percentage of sites with bleeding on probing, as well as plaque scores. Correlation coefficients tended to be larger (r > 0.50) between PDSR scores and indicators of less severe or less advanced disease, such as clinical attachment loss of ≥3mm or probing depth of ≥4mm. The magnitude of the correlation coefficients was smaller with greater attachment loss and deeper probing depths. The PDSR Functional Symptoms Subscale was positively correlated with number of missing teeth (r = 0.37, p < 0.01), whereas the Total Score and Physiological Symptoms Subscale were not. All correlations are displayed in the online supplement.

Results of an ANOVA comparing differences in mean PDSR Total Scores between periodontal status classifications were statistically significant (F(3,106) = 11.839, p < 0.001, η² = 0.25, 95% CI: 0.11–0.36). Using a Tukey post-hoc analysis, significant differences were detected between the ‘no disease’ and ‘moderate disease’ groups (p = 0.013), as well as ‘no disease’ and ‘severe disease’ groups (p < 0.001). Similarly, differences were detected between ‘mild’ and ‘severe’ disease (p < 0.001) as well as ‘moderate’ and ‘severe’ disease classifications (p < 0.001). ANOVA results for the PDSR Total Scores and the two subscales are displayed in Figure 1.

Figure 1.

ANOVA results indicating Periodontal Disease Self Report (PDSR) means by Centers for Disease Control and Prevention/American Academy of Periodontology (CDC/AAP) status.

When evaluating the explanatory power of the PDSR Total Score, both models (including only covariates and the second model further including the PDSR) were statistically significant. The change in R² also significantly improved when PDSR scores were added to the model in predicting mean probing depth. That is, when accounting for the covariates of age, gender, current smoking status, and BMI, PDSR scores contributed significantly to the model and accounted for 31% of the variance (β = 0.45, 95% CI: 0.02–0.04; ΔR² = 0.19. Similarly, PDSR scores significantly added to a model predicting mean attachment loss that accounted for 56% of the variance (β = 0.46, 95% CI: 0.06–0.10); ΔR² = 0.20) after accounting for covariates. Regression results are displayed in Table 2.

Table 2.

Stepwise regression results indicating added value of PDSR Total scores to predict mean probing depth and mean clinical attachment loss after accounting for common covariates.

Mean Probing Depth					t	p-value	95% CI
		B	Std. Error	Beta			Lower Bound	Upper Bound
1	(Constant)	1.79	0.19		9.47	< 0.001	1.42	2.17
	Age	0.00	0.00	0.12	1.24	0.22	0.00	0.01
	Gender	0.17	0.09	0.18	1.75	0.08	-0.02	0.35
	Current Smoker	0.21	0.12	0.17	1.76	0.08	-0.03	0.44
	BMI	0.01	0.01	0.12	1.23	0.22	0.00	0.02
2	(Constant)	1.68	0.17		9.91	< 0.001	1.35	2.02
	Age	0.00	0.00	0.10	1.17	0.25	0.00	0.01
	Gender	0.21	0.08	0.22	2.44	0.02	0.04	0.37
	Current Smoker	0.13	0.11	0.11	1.24	0.22	-0.08	0.34
	BMI	0.00	0.01	0.06	0.63	0.53	-0.01	0.01
	PDSR Total	0.03	0.01	0.45	5.05	< 0.001	0.02	0.04
	Mean Clinical Attachment Loss	B	Sth. Error	Beta	t	p-value	Lower Bound	Upper Bound
1	(Constant)	-0.92	0.40		-2.29	0.02	-1.71	-0.12
	Age	0.04	0.01	0.49	5.81	< 0 .001	0.03	0.06
	Gender	0.00	0.20	0.00	0.02	0.99	-0.39	0.40
	Current Smoker	0.99	0.25	0.34	4.00	< 0.001	0.50	1.49
	BMI	0.00	0.01	0.00	0.05	0.96	-0.02	0.02
2	(Constant)	-1.20	0.34		-3.56	<.001	-1.86	-0.53
	Age	0.04	0.01	0.47	6.69	<.001	0.03	0.05
	Gender	0.11	0.17	0.05	0.63	0.53	-0.23	0.44
	Current Smoker	0.80	0.21	0.27	3.84	<.001	0.39	1.22
	BMI	-0.01	0.01	-0.06	-0.90	0.37	-0.03	0.01
	PDSR Total	0.08	0.01	0.46	6.52	< 0.001	0.06	0.10

Note: BMI = body mass index; CI = confidence interval; PDSR = Periodontal Disease Self Report. B = non-standardized coefficient; t = t-score.

The area under the curve for the PDSR scores distinguishing between none/mild and moderate/severe disease was 0.71 (95% CI: 0.62–0.81). The PDSR Functional Symptoms subscale had an area under the curve of 0.72 (95% CI: 0.62–0.82). The PDSR Physiological subscale had an area under the curve value of 0.67 (95% CI: 0.57–0.78). ROC curve and AUC data are displayed in Figure 2.

Figure 2.

ROC Curve for Periodontal Disease Self Report (PDSR) Total as well as Functional Symptoms subscale and Physiological Symptoms subscale scores when comparing CDC/AAP no-to-mild periodontitis and moderate-to-severe periodontitis groups.

Discussion

The present study was designed to compare PDSR scores with various clinical parameters of periodontal disease. More specifically, this study tested associations between self-report data and CDC/AAP status, along with various combinations of probing depth, clinical attachment loss, bleeding on probing, and plaque scores. Overall, PDSR scores including the total and subscale scores displayed statistically significant associations with several clinical indicators of periodontal disease. Additionally, PDSR scores explained a significant portion of the variance in models with mean probing depth and mean clinical attachment loss as outcomes, while also accounting for common risk factors of periodontitis. PDSR scores were associated with moderate sensitivity and specificity in their ability to discriminate between none-to-mild disease and moderate-to-severe disease using the CDC/AAP classification.

Overall, clinical associations with the PDSR measure and subscales provide evidence of clinical validity at levels comparable to that of other self-report measures (Abbood et al. 2016). Our findings also provide additional evidence of “convergent validity”, also referred to as “content” validity. Specifically, higher PDSR scores were seen with moderate and severe disease status when compared with no disease or mild disease and the correlations between PDSR scores and various clinical parameters were positive. Thus, the PDSR scores and clinical examination data “hang together” in a way that suggests the PDSR is measuring something akin to the clinical measures. The extent of the congruence, however, between PDSR scores and clinical data should be understood in light of the other correlation coefficients, the regression models and area under the curve results. The findings around greater correlation coefficient magnitude with those with indicators of less severe disease (e.g., number of teeth with PD of 4mm+ versus 5mm+ were of note. It could be that the PDSR maps more precisely onto a less severe disease phenotype. These results could also be a function of the greater number of individuals with CAL of 4mm+ or PD of 4mm+ where it is rarer to have more severe disease markers. The PDSR measure only accounted for a small portion of the variance of mean probing depth or attachment loss and the levels of sensitivity/specificity indicate that the PDSR measure may or may not be an accurate diagnostic tool for periodontal disease. As a general screening measure or as a research instrument, though, the PDSR measure shows promise for future work. Additional work could compare the predictive utility or added benefits of the PDSR when compared to other self-report measures within the same sample or study.

Interestingly, the correlations between the PDSR Physiological Symptoms subscale and mean probing depth were higher relative to correlations with mean attachment loss. Whereas the inverse was true of the PDSR Functional Symptoms subscale. An increase in probing depth and bleeding on probing at a site, in the absence of attachment loss at that site, is typically described as a milder disease, as indicated by the CDC/AAP categorization. The presence of deeper pocket depth and attachment loss at a site is more indicative of moderate or even severe periodontitis (Kwon et al. 2021). It could be that the Physiological Symptoms subscale reflects symptoms of mild disease, and the Functional Symptoms subscale is reflective of more severe periodontitis. The Physiological Symptoms subscale includes items around bleeding, or “puffy” gums whereas the Functional Symptoms scale includes items around difficulty chewing which is more likely with advanced clinical attachment loss. This also pairs with findings indicating greater correlation between the Functional Subscale scores and the number of missing teeth.

The PDSR measure has previously been demonstrated as a potentially useful self-report correlate of periodontitis (Wright et al. 2021). While no self-report measure is perfect at measuring the complex clinical phenotype of periodontitis, this study provides some evidence for clinical validity and utility given the results. Relative to other self-report measures, the PDSR does not rely on a patient’s prior clinical encounters (e.g., having seen a dentist or periodontist) and the patient does not need to have high clinical literacy to be able to self-rate the status of their periodontal health (Bond et al. 2024). The utility of such a measure could enhance some of the known benefits of using self-report measures in population health studies or where a low-cost and efficient measure of periodontal disease is needed.

In terms of limitations, the sample size and the distribution of periodontal disease being skewed toward healthy participants are important to note. The relatively low number of participants with severe periodontal disease naturally limits the inferences one can make regarding the association of PDSR scores and severe disease, though mean population estimates of severe disease within the population are also lower (i.e., 7.8%; National Institute of Dental and Craniofacial Research 2024). While it appeared that the PDSR was more strongly associated with less severe cases, future studies testing the validity of the PDSR would benefit from a higher prevalence of severe disease to better characterize the true association between PDSR scores and more advanced stages of periodontitis. Such work would be important in establishing additional evidence for clinical validity of the PDSR.

It is also important to note that evidence for both the psychometric and clinical validity of a self-report measure are driven in part by the characteristics of the population under study that are associated with properties of self-report and clinical disease state. Therefore, continued work in population groups that differ from ours is certainly warranted. Additionally, some of the covariates that one would expect to be associated with periodontal disease were not significantly associated with some of the periodontal disease parameters in this population. Likewise, other possible covariates were not included in the regression models due to statistical power considerations. In general, a larger sample offering more statistical power would allow for additional estimates in the predictive power of the PDSR measure while including other common risk factors of periodontal disease. Finally, given the PDSR was administered after the clinical examination, there could be order effects in terms of the sequencing of data collection and future work could address this with counter balancing data collection (e.g., some having the exam first and others the self-report first).

Conclusion

This study provided clinical evidence for the content and criterion-related validity of the PDSR measure. The PDSR is a measure with some evidence for clinical validity, at least in identifying less severe periodontitis, that can be used in community samples. With future validations and resultant refinement in additional samples, the PDSR could help in settings where clinical examinations are more cost-prohibitive or where other resources are limited (e.g., rural settings or integrated settings for screening purposes). Future work will aid in providing additional data to validate the PDSR in a variety of samples and in clinical-specific populations.

Knowledge Transfer Statement:

This study provides evidence for clinical and criterion validity of a self-report measure which could have future utility as a feasible measure of periodontitis for large scale studies or clinical settings. Such a measure can provide a low-cost and easy to implement way to approximating various aspects including physiological and functional factors of periodontal disease.