A systematic review investigating patient knowledge and awareness on the association between oral health and their systemic condition

The objective of this review is to identify inequalities in the dissemination of information regarding the oral-systemic link, by investigating the awareness of patients with major systemic conditions, regarding the link between oral disease and their condition.

Studies were evaluated using an analytical approach, quantifying associations between participant factors and knowledge and awareness outcomes. The following PICOS framework was proposed, according to Li et al. [23]:

Participants: patients with major systemic conditions (DM, respiratory disease, CVD, pregnancy and bone disease).

Intervention: explore knowledge and awareness of participants regarding the association between oral health and their condition.

Comparison: not applicable.

Outcome: assessment of knowledge and awareness.

Studies: observational study design.

Study selection was based on the following inclusion criteria: (1) observational studies; (2) published in English; (3) adult participants; (4) patients with major systemic conditions (diabetes mellitus (DM), cardiovascular disease (CVD), respiratory disease, bone disease and pregnancy); (5) publications within the time frame of 2011–2020 to ensure an up-to-date measure of knowledge; (6) quantitative, questionnaire-based studies. Excluded studies involved: (1) reviews, case reports, case studies, opinions or commentary and/or editorials on searched topics; (2) studies involving health professionals and healthcare students which may contribute to knowledge bias; (3) unpublished studies.

An extensive literature search was conducted by the primary reviewer from six databases: Medline (Ovid), CINAHL, The Cochrane Library, Web of Science, Informit Health Databases and Scopus, using keywords and Medical Subject Headings (MeSH) term headings. Boolean phrases such as “AND” or “OR” were included. Individual search strings were adapted for each database. A complete electronic search strategy for Medline (Ovid) is attached as a supplementary file (see Additional file 2). International studies were searched, without limitations, according to the following research question: ‘are patients with major systemic conditions aware and knowledgeable of the oral health associations to their condition?’. A final search was completed on 3 August 2020, to ensure the most recent literature. A grey literature search was also conducted on Google Scholar for unpublished studies, although no studies satisfying the inclusion criteria were found. The reference lists of included full-text articles, were manually searched for studies that were not identified through the electronic search. Screening and removal of duplicates were completed using the Endnote program (X9.3.3, Clarivate Analytics, Philadelphia, PA, United States of America).

Data selection was performed independently by two reviewers (SA and MR). Throughout the screening process, any conflicts or uncertainty regarding inclusion or exclusion of the articles, were resolved by discussion between the primary and secondary reviewer, or consultation with a third reviewer (SL). The first stage involved the primary reviewer (SA) screening for all relevant titles and abstracts, complying with the inclusion and exclusion criteria. The selected articles were verified by a second reviewer (MR). If a title or abstract provided insufficient information for exclusion, it was included for a full-text review. In the second stage, full-text articles were screened and analysed independently, by two reviewers (SA, MR). Corresponding authors of the included studies were contacted for unavailable studies, or additional studies complying with the review aim.

Two reviewers (SA, MR) independently completed data extraction using a pilot-tested, standardised spread-sheet on Excel (Microsoft Corp, Redmond, Washington). Conflicts were resolved via consensus between the two reviewers, or via consultation with the third reviewer (SL). Data extraction included information regarding author, year of publication, study population characteristics (population size, age, gender, type of systemic condition), study location, study design, study setting, knowledge and awareness outcomes, a summary of major findings, ethical approval, statistical analysis and quality assessment (see Additional file 3).

A final search strategy from all six electronic databases resulted in 6878 total articles. Thirty additional articles from manual searching were also selected for screening. Removal of duplicates resulted in 4780 articles eligible for screening of relevant titles and abstracts. Ninety-four articles were admitted for full-text article screening. Following full text screening, 24 articles were accepted for inclusion in the systematic review, each satisfying the inclusion criteria. A total of 4756 articles were excluded in the study selection process. The search strategy followed the PRISMA guidelines and a checklist flowchart is provided in Fig. 1 [22].

Inter-rater reliability between the two reviewers was 98.8% (83 of 84), with a Kappa score of 0.99 for screening of titles and abstracts and 100% (20 of 20), with a Kappa score of 1.0 for the included full-text articles. Notably, a Kappa score of excellent reliability ranges from 0.81-1.0. Any conflicts were resolved via discussion to arrive at a consensus, or via consultation with a third reviewer. Each section of data extraction demonstrated a very high inter-agreement reliability, between reviewer one and two.

Two reviewers (SA and SL) independently assessed the quality of the included full-text articles, at study level (n = 24). Risk of bias was evaluated using the Joanna Briggs Institute (JBI) checklist for analytical cross-sectional studies, which is an 8-item scale including the options: Yes, No, Unclear or Not applicable [24]. Each paper was rated with high (score 80–100%), fair (50–79%), or low (< 50%) quality. The intention of quality assessment was to influence the interpretation of study findings, to support reliable and accurate generalisations.

A Synthesis Without Meta-Analysis (SWiM) was conducted on the included studies, due to heterogeneity of the population and outcome measures [25]. Studies were grouped based on systemic condition. A total measure of knowledge on the oral-systemic link was determined from the main findings of each study, describing either: poor (< 50%), average (50%), or good (> 50%) knowledge. Table 1 presents the main study characteristics, including study design, screening, interventions and outcomes. This enabled informal investigation of heterogeneity.

Data included studies that originated in 14 different countries and India was the most prominent location of the included studies (n = 6). The overall age range across all studies was 18–99 years. Studies investigating patients with diabetes, heart disease, bone disease or pregnancy were included. There were no studies investigating patients with respiratory disease, applicable to the inclusion criteria. Self-administered questionnaires were the most prominent form of data collection (n = 14). Only three studies reported a face-to-face questionnaire design [26‐28]. The majority of studies collected data from university clinics or teaching hospitals (n = 12). Public health facilities were more common than private; out-patient settings were more common than in- patient settings. One study in mainland China was distributed nation-wide [27]. Another study was internet-based [29]. Few studies involved rural populations [26, 30‐32].

The majority of the included studies received a fair quality rating (n = 16) according to the JBI checklist (see Additional file 4). Several studies were unclear, or did not provide information on the measurement of validity and reliability. There was also a generalised lack of identification and adjustment for confounding factors, in the majority of studies (n = 15), which is important to consider when interpreting study findings. No study was excluded due to the assessment of bias alone. Only two of the included studies fulfilled the complete checklist, demonstrating high internal validity [33, 34]. The inter-rater reliability between the two critical appraisers was 100%, corresponding with a Kappa score of 1.0.

Studies investigating participants with the following major systemic conditions were included in this systematic review: DM, heart disease, bone disease and pregnancy. Overall, the included studies assessed patient knowledge regarding the oral manifestations of their systemic condition, the impact of relevant medications on their oral health and the effect of poor oral health on their systemic condition. Pregnant participants were the most consistent patient group to demonstrate good knowledge (scoring > 50%), across multiple studies. The included studies were based in various countries, strengthening global data. Compared to male participants, females generally demonstrated greater knowledge regarding the oral implications of their systemic condition, which was not determined by level of education [30, 31, 35, 36]. Overall, approximately 70.8% of patients with major systemic conditions, had poor knowledge and awareness (scoring < 50%), regarding the relationship between oral health and their systemic condition.

Five studies investigated the knowledge and awareness of adult pregnant patients [32, 34, 36‐38]. The influence of poor oral health on pregnancy, the effect of pregnancy hormones on oral health and the importance of dental visits during pregnancy, was assessed. The majority of these studies demonstrated an adequate level of knowledge, regarding the link between pregnancy and oral health [29, 31, 36]. However, Payal et al [37], reported limited knowledge of pregnant participants. A significant finding was that only 19.38% of participants were aware that oral hygiene can affect their growing baby, and none sought a routine checkup during pregnancy [35]. Abiola et al [36], conducted a study in Nigeria, identifying that 14.8% of patients agreed pregnancy caused gum problems [36]. Despite this poor awareness (< 50%), it was concluded that this was an acceptable level of knowledge, which may be true for this population demographic.

Three studies assessed the knowledge of adult patients with heart disease, regarding the correlation between oral health and heart disease. The majority of studies demonstrated a lack of awareness and limited knowledge [33, 39]. For instance, Hollatz et al [33], conducted a study in Germany indicating that approximately 73% of patients with adult congenital heart disease (ACHD), had inadequate or non-existent knowledge regarding the interrelation between oral health and heart disease [33]. Alternatively, a study conducted in Saudi Arabia by Rasouli-Ghahroudi et al [31], indicated that 72% of participants scored moderate or good knowledge, which was attributed to repeated health education programs in the community.

Adult diabetic patients were the most studied population group and accounted for 15 of the included articles. It was summarized that the majority of diabetic patients have inadequate knowledge and awareness (scored < 50%) [13, 27, 30, 40‐45]. Few studies demonstrated adequate knowledge (scored > 50%) regarding the relationship between oral health and diabetes, including Bangash et al [28] (64%), Al Amassi et al [29] (81%), Mian et al [35] (76.4%), Wang et al [26] (81.1%) and Ummadisetty et al [46] (61.7%). Exclusion of participants with type 1 diabetes mellitus (T1DM) was apparent in several studies [27, 30, 40, 44]. A study by Weinspach et al [12], demonstrated that participants with T1DM were more aware of the bi-directional relationship between diabetes and periodontitis, than those with type 2 diabetes mellitus (T2DM), potentially due to earlier age of onset. Additionally, several studies concluded that diabetic patients were more knowledgeable of associated systemic complications, rather than oral complications [41, 47]. Few studies including healthy (non-diabetic) participants, revealed that diabetic participants demonstrated higher oral health knowledge [13, 26, 35]. Wang et al [26], demonstrated a 5.5% difference in knowledge between rural diabetics and healthy participants.

Rotman-Pikielny et al [48], investigated patients with bone disease. Participants were assessed on the relationship and influence of osteoporosis on oral health, in addition to the associations between oral health and osteoporosis treatment. The study findings reported low knowledge of the oral health associations to osteoporosis and osteopenia [48]. Further research is recommended to support this finding.

The majority of patients had not received adequate information about the oral health implications of their systemic condition, suggestive of a lack of health practitioner-patient communication. Information was sourced most commonly from dentists, other health professionals and the media [13, 34, 35, 41, 46‐48]. Source of knowledge was not reported in several studies (n = 10). A study on Australian cardiac patients, by Sanchez et al [39], indicated that patients with valvular conditions (40.6%) received more information about oral health, than those with cardiovascular conditions (7.4%). This suggests an inequality in the dissemination of oral health information amongst at-risk groups.

Various measures are required to address the poor awareness of the oral associations relevant to patients with heart disease, diabetes and bone disease. This must be directed at both dental and non-dental health practitioners, depending on access to health care services, globally. Patient-practitioner communication of the oral-systemic link, is currently undermined as a routine practice. Therefore, improving communication and education programs globally, whilst accounting for language, cultural differences and access barriers in remote locations, is necessary to addressing inequalities in the dissemination of information on oral-systemic complications [26, 32, 41, 42, 47]. Considering that the oral-systemic link is a constantly evolving health topic, it is essential to ensure that health practitioners are trained with current and evidence-based oral health knowledge, to encourage prompt action and referral. Policymakers can integrate basic education guidelines into clinical settings, regarding the oral-systemic link, to enforce routine patient-practitioner discussion. Mass media health promotion is also recommended, considering this was a common information source. Overall, these implications in clinical practice can address disparities in the dissemination of oral-systemic education, to address the rate of potentially preventable chronic conditions and related morbidity.

The findings from this systematic review recommend that future research be conducted on more diverse populations to increase applicability (external validity), to the global population of patients with major systemic conditions. Investigations on the knowledge of patients with respiratory disease, regarding relevant oral implications, is also recommended. Additionally, to measure the effectiveness of educational programs and policy changes on patient knowledge and chronic disease burden, follow-up studies are advised. This would particularly benefit the systemically compromised patient groups (bone disease, heart disease and diabetes), that demonstrated mostly poor awareness on the oral-systemic link in the current review.

The strengths of this review include the range of countries involved, the socio-demographic characteristics of study participants and the consistency of study design and data collection methods. Eleven studies identified non-significant results, which suggests low reporting bias.

Several limitations were noted in the included studies. As the majority of studies utilised a self- administered questionnaire for the assessment of knowledge and awareness, results are prone to measurement bias. This includes an increased prevalence of recall bias due to lack of feedback during the intervention, or bias towards social desirability and over- reporting [26, 33, 34, 37, 42]. Few studies involved interview administration of questionnaires, to minimise the risk of incompleteness and allow for on-going feedback [26‐28]. Despite these benefits, one study argued that the absence of an interviewer could encourage the patient’s own opinions and knowledge, when responding [26]. The measure of knowledge summarized from each study, is dependent on the specific questions involved in the intervention, which differs according to questionnaire design, contributing to some heterogeneity in outcome measure. Additionally, the majority of studies failed to identify and adjust for confounding factors, despite measuring socio-demographic variables such as age, gender, occupation, income, educational status and co-morbidities. Few studies mentioned the use of regression analysis to adjust for confounding factors, influencing oral health knowledge [12, 33, 34, 36, 39]. Participant selection was mostly via convenience sampling, from single- centre sites, contributing to selection bias and low generalisability. Few studies investigating diabetic patients excluded T1DM patients, without reason, which may also contribute to selection bias [30, 35, 40]. The good knowledge and awareness demonstrated by the majority of pregnant participants, may be attributed to greater support and responsibility for the naturally occurring condition. Reporting bias was apparent in some publications that did not provide tabulated data [37, 47]. Additionally, Abiola et al [36]. demonstrated conflicting analyses of significance in a Chi-square and ANOVA test, respectively reporting insignificance and significance between educational status and oral health knowledge. This reporting ambiguity, is reflected in the JBI appraisal which reported unclear methodological quality in several areas (see Additional file 4). Although grey literature was searched, there were no relevant unpublished studies that would influence the overall findings of the included published studies.