Selected studies and its characteristics
A total of n=6,323 hits were identified from the database searches (initial n=4,416; update n=1,907). The total number of relevant articles identified was n=233 (n=230 journal articles, two institutional reports [38, 39], and one master thesis [40]) including n=42 qualitative, n=177 quantitative [n=8 RCTs, n=9 non-RCTs, and n=160 descriptive studies (n=156 cross-sectional, n=4 longitudinal)], and n=14 mixed-methods studies. An overview and the selection process of identified literature in accordance with PRISMA criteria [41] is given in the flow chart in Figure 1. In addition, search strings and hits in databases are listed as additional file (A5. Search strings for databases including hits).
Fig. 1 Flow chart of literature search in databases and screening process; TiAb – title / abstract, E1 – no full-text, E2 – no full-text in English or German, E3 – no quantitative or qualitative secondary study, E4 – patient/s with special needs, E5 – no dental treatment/s, E6 – no patient/s preferences, E7 – multiple publication without relevant additional information
We identified studies in n=49 different countries, with most studies conducted in the UK (n=28), Saudi Arabia (n=23), and the USA (n=22). Only one study per country was conducted in n=19 countries, e.g., Poland [42], Greece [43], or Mexico [44]. A few studies (n=5) included participants from multiple countries such as Bucchi et al. [45] (Portugal, Chile, Spain, France, and Italy) or Ellis et al. [46] (UK and Canada). Some studies could not be clearly assigned to certain countries because those were either (online) studies with an international focus (n=2, [47, 48]), or the place of conduction was not clearly stated (n=9, e.g., [49, 50]). The location and the health sector considered (public: n=29; private: n=16; both: n=28) differed across the studies. We distinguished between dental practice (n=18), (dental) clinic/hospital (n=79) (e.g., in waiting area [51, 52]), academic institution (n=25) (e.g., dental department in university [53] or college [54]), dental service institution (n=7) (e.g., public health service of a city [55]), other non-private areas (n=22) (e.g., school [56, 57], grocery store [40]), and in-private settings (n=39) (e.g., at home [58, 59]). In four cases, focused medical facility was an emergency department [60–63]. Due to the different study designs, various data collection instruments were used, including in-depth and focus group interviews; questionnaires (self-, interviewer-, dentist-administered, etc.) including surveys with scales [e.g., Visual Analogue Scale (VS scale) [64, 65], but also online searches (e.g., blogs [47] and forum posts [66]), or combinations [48, 55]].
The study population ranged from few participants, e.g., n=2 in a case study [44], to n=40,305 participants in a comprehensive questionnaire survey [67]. Most patients were interviewed directly (n=214), but in some cases parents or caregivers on behalf of their children (n=24) or elderly persons (n=2) were interviewed [68, 69]. Accordingly, age varied between young infancy (e.g., 2 years of age [70–72] up to high age groups (max. 101 years of age [73]). In most studies, the proportion of female participants predominated (n=137). In some studies, it was as high as 100% (n=5), also due to the research questions (e.g., investigation of maternal beliefs and motivations for childrens’ first dental visit [74]). In terms of further sociodemographic characteristics, income (i.e., "low" to "high") and education levels of adult respondents (e.g., elementary school education [75] to university degree and above [76]) varied, while study participants with low and middle (household) income predominated (n=21, 37%; and n=22, 39%). Overall, in less than half of the studies, data on participant income (n=74, 32% of articles) and education (n=107, 46% of articles) was reported. The studies focused on various aspects of dental care. In line with our research focus, studies were sorted according to the factors of choice of dental treatments in defined categories: (i) dental treatment (n=138), describing a single treatment, e.g., implant overdenture [77], or teeth whitening [78]; (ii) dental care in general (n=68) including the entirety of dental treatments, e.g., oral health care of patients [44], or utilization of dental services [14]; and (iii) dentist & dental practice (n=27) including, e.g., appearance of the dentist [56], or organization [79] and (technical) equipment of dental practice [80].
In the additional files [A6. Characteristics, factors of choice, and references of included articles (n=233)], extracted study characteristics [i.e., information on study setting (e.g., country) and population (e.g., age, income), treatment by category], as well as the factors of choice and the article references are listed by study design. Furthermore, methodological details of the study can be found as additional files [A7. Methodological characteristics of included articles (n=233), and search details, e.g., recruitment, in-/exclusion criteria, incentives, and source of literature].
Fig. 2 Factors of choice and descriptive results of identified articles; diff. – different; MMAT – Mixed-methods-appraisal tool [I – qualitative studies; II quantitative randomized controlled trials (RCTs); III – quantitative non-randomized controlled trials (non-RCTs); IV – quantitative descriptive studies: a – cross-sectional ~, b – longitudinal ~;
V – mixed-methods studies]; ns – not stated or unclear; w/ – with; 1 – incl. military hospitals; 2 – e.g., parents, relatives; 3 – e.g., implant, veneers; 4 – e.g., dental visits, toothache pain; 5 – articles with consideration of several countries are counted 1x per country; 6 – majority of or all study population, mixed-methods studies counted multiple if necessary; 7 – self-perceived or dentist diagnosed; 8 – in brackets: number of articles in which factors of choice were mentioned w/ percentage of these articles compared to the articles of all countries, multiple factor mentions in articles possible; 9 – one article considering several countries excluded from analysis since separation of factors to countries impossible (the UK: n=27, and Canada: n=8 articles); * Bosnia and Herzegovina, Finland, France, Greece, Hong, Kong, Kuwait, Lebanon, Mexico, Philippines, Poland, Portugal, Russia, South Korea, Sudan, Syria, Taiwan, Tanzania, Trinidad, Spain; ** Austria, Belgium, Bulgaria, Chile, Pakistan, Singapore; *** China, Ireland, Jordan, Netherlands, Switzerland, Thailand
Findings of studies
Factors of choice
Several different factors of choice could be identified. Overall, some of it occur only once [e.g., dentist smell [81], (patient’s) forgetfulness [15]], or a few times (e.g., n=7, chewing ability / function [82, 83]), while others were identified more frequently. Reported in n=132 studies, "out-of-pocket payment" is the most frequently identified factor (e.g., [60, 84, 85]) followed by "fear (of treatment)" (n=53, [61, 86]), and "aesthetics" (n=29, e.g., [87, 88]).
The codebook contained n=176 codes as a framework for the qualitative analysis. There was a total of two coding runs (i, ii), because after the first coding run, the target value of ICA was not reached with α=0.683. In the second run the value was an acceptable α=0.865. After the first coding run, some codes were excluded if merged with other codes that describe the content as well. In addition, some codes were split, to describe the factor of choice in more detail. This was followed by a final consensus discussion which resulted in n=177 codes representing factors of choice. Codes with the same content were summarized thematically and a suitable term was assigned by the reviewers in a discussion. This resulted in n=101 factors assigned to three categories (I) "Dentist & dental institution", (II) "Patient" and (III) "Treatment", developed by the reviewers. Furthermore, factors were grouped into subcategories.
The category "Dentist & dental institution" comprises a total of n=44 factors divided in four subcategories (I.1-4). (I.1) "Access to care" includes factors such as access barriers in general, but also coverage by insurance (on individual request), as well as physical accessibility, such as location, and transportation. The subcategory (I.2) "Communication" includes factors of the relationship between dentist / staff and patient, such as altruism (of dentists), transparency, and understandable information. The subcategory (I.3) "Qualification" includes factors that describe education and training of dentists and staff (e.g., academic institution) while the subcategory (I.4) "Organization" focuses on equipment and processes (e.g., customer service, facilities) in a dental practice.
The category "Patient" comprises three subcategories with a total of n=37 factors representing patient characteristics: (II.1) "Medical characteristics" include factors such as emergency and prevention, (II.2) "Social characteristics" compraise factors such as religion and social environment, and "Individual characteristics" include factors such as aesthetics and pain.
The "Treatment" category with n=20 factors is subdivided in two subcategories, which describe the characteristics of dental treatments on the one hand and highlights the financial aspect from patient’s perspective on the other hand. The subcategory (III.1) "Treatment characteristics" includes factors such as complexity of treatment and tooth saving. The subcategory (III.2) "Cost" includes the three factors out-of-pocket payment, second opinion, and installment.
Code frequency showed that out-of-pocket payment (n=148) ranked first, followed by dental fear (n=73), aesthetics (n=64), and pain (n=58). N=23 codes were identified only once, e.g., quality assessment culture, social class, or willingness-to-travel (WTT). In seven out of eight countries with most identified articles [UK (n=28 articles), Saudi Arabia (n=23), USA (n=22), India (n=19), Brazil (n=14), Turkey (n=11), Germany (n=9) and Canada (n=9)], the factor out-of-pocket payment is most prevalent, e.g., Canada [n=6 articles (67%) [77, 85, 89–92]], India [n=13 articles (68%) [50, 54, 66, 93–102]], Germany [n=6 articles (67%) [68, 103–107]], the UK (53.6%) [n=15 articles [13, 63, 84, 88, 107–117]], and the USA [n=12 articles (55%) [39, 53, 58, 60, 118–125]]. Frequencies of factors dental fear, aesthetics, and pain vary among the countries. For example, dental fear is a frequently mentioned factor in studies from India [n=8 articles (42%) [50, 56, 59, 93, 94, 98, 99, 101]], Saudi Arabia [n=9 articles (39%) [15, 61, 126–132]], and Brazil [n=5 articles (36%) [133–137]], but it has less weight in studies from Germany [n=1 article (11%) [138]], and Canada (n=0). Aesthetics is mentioned most frequently in Brazilian studies [n=8 articles (57%) [55, 133–136, 139–141]] followed by Saudi Arabia [n=7 articles (30%) [72, 78, 83, 127, 129, 142, 143]]. It is less often mentioned in studies of other countries, e.g., India [n=2 articles (11%) [66, 144]], and Turkey [n=1 article (9%) [145]). Pain is a frequently mentioned factor of choice in Brazilian studies [n=6 articles (43%) [134–136, 141, 146, 147]], followed by Canada [n=2 articles (25%) [92, 148]], and the USA [n=5 articles (23%) [60, 74, 118, 120, 121]]. It was not identified for the German context (n=0). Figure 2 includes a bar graph showing the frequency of the factors mentioned and the number of articles of the four most frequently mentioned factors of choice. In this analysis, we only considered those articles in which the identified factors could be clearly assigned to one country, e.g., for Ellis et al. 2011 [46] this was not possible. In addition, Figure 2 provides an overview of characteristics of the identified articles, e.g., study types, health sectors in which the studies where conducted, and number of articles per country.
Figure 3 shows the factors subdivided into (sub-)categories. The additional files contain the coding scheme and the codebook, including the framework for coding (A8. Coding scheme, codebook, and framework), and the final codes definitions (A9. Code definitions) representing the factors of choice.
Fig. 3 Factors of choice in dental care from patient perspective in (sub-)categories, N=101
Willingness-to-pay in dental care
As "out-of-pocket payment" is the most frequently identified factor, we additionally give an overview on n=37 articles analyzing the willingness-to-pay (WTP) of patients (n=3 RCTs, n=6 non-RCTs, n=26 descriptive studies (cross-sectional), and n=2 mixed-methods studies). A large proportion of the WTP articles examined WTP for a particular dental treatment (n=29), e.g., dental implant [142], filling replacement [149], and fluoride varnish [115]. N=6 articles examined WTP for a "dental service in general" (e.g., dental tourism [103], dental services [150]), and n=2 articles for the category "dentist & dental office" (e.g., osteoporosis risk assessment in primary dental care [151]). Different methodological approaches were used to determine WTP, e.g., bidding game, and payment card method, with closed [e.g., Christell et al. (2019) [151]], and open [e.g., Tianviwat et al. (2008) [152]] response options.
Re et al. [153–155], reported patients would be willing to pay an additional fee to receive a particular treatment. According to McKenna et al. (2016) [156], there is a strong WTP regarding basic and functional treatments. For more aesthetical treatments, there is even a higher WTP [88, 140]. Individual (sociodemographic) characteristics of patients, e.g., age [140], gender [142], income [142, 149, 157], education level, and number of missing teeth [158], have an impact on WTP. Widström et al. (2012) [149] reported, that individuals with higher incomes are willing to pay higher prices than individuals with lower incomes. Articles also reported that patients are not willing to pay any price for a dental treatment [159, 160].
Impact of Covid-19 behavior
Our update search of databases during the Covid-19 pandemic showed that Covid-19 influences patients’ choice regarding dental treatments reported in n=1 of the articles identified. Papautsky et al. (2021) [161] used a convenience non-representative sample of 2,570 US patients in their mixed-methods study to investigate reasons for cancelling or postponing (dental) medical treatments at the onset of the Covid-19 pandemic via an online survey. They found that fear of SARS-CoV-2 infection led patients to cancel and postpone dental treatments. Dental care respondents were most likely to report delays (38.1%), compared to other medical areas [e.g., preventive care (29.2%), and diagnostic services (16.4%)]. Age, gender, education, and self-reported concerns for overall health were significantly associated with health care postponement.
Quality of studies
ICR coefficient for the initial search had a moderate value of k=0.41 (vs. substantial k>=0.61 [34]). Therefore, a cross-check quality assessment of a minimum sample size was conducted for the update review. A minimum sample of n=47 articles was randomly selected [using the software Rstudio (version 1.2.5)] from the literature hit pool of the update search [34]. This sample of articles was assessed independently by the two reviewers SF and JFH, using the MMAT. Subsequently, the results of the assessment, including questions S1., S2. and the five questions regarding the study design, were compared, and ICR calculated. Since a substantial value of k=0.697 was obtained, SF finalized quality assessment, after discussion and consensus of the previous results. Results of ICR and ICA, and references of the articles included in the cross-check can be found as additional files [A10. Calculation of ICR and ICA, and A7. Methodological characteristics of included articles (n=233), and search details].
Quality of the studies, assessed by MMAT, varied. The score points ranged between 0 (n=1) and 5 (n=56) with most articles scoring 4 (n=86) (1: n=10, 2: n=25, and 3: n=40). N=11 articles failed question S1 and n=5 articles failed question S2. These were excluded from score evaluation. Overall, qualitative studies and RCTs (modus=5) were assessed best while quantitative descriptive studies (modus=4) were rated worst. Nevertheless, in all study designs, articles could be identified that scored highest (score 5). The lowest score of 0 was given to an RCT [77]. Poor scores of 1 were found for quantitative descriptive studies and once for qualitative studies.
The studys of the different designs were scored as follows: for qualitative studies, for example, points were deducted for question 1.5 if the research question was not clearly answered in the results (e.g., Johannsen et al. (2012) [162]), or if no direct relationship between generated citations and derived results was apparent (e.g., Serban et al. (2019) [110]). For RCTs, one point was deducted if randomized group assignment was not performed appropriately (Question 2.1), among others (e.g., [163]). Here, only flipping a coin was applied during randomization. In addition, one point was deducted if the intervention and comparison groups were not comparable at baseline (Question 2.2), e.g., in terms of group size (e.g., [77]). An example of point deductions for non-RCTs is Eyuboglu et al. (2020) [65], who did not state where the study participants were recruited (Question 3.1), nor did they provide information on follow-up (Question 3.3) and confounders (Question 3.4). A common reason for point deductions in quantitative descriptive studies was the non-conduction of a pre-test in the case of using non-validated questionnaires or guides (Question 4.3) (e.g., [164], [165]). Other examples of point deduction are Aldaij et al. (2018) [78] and Wall et al. (2015) [39]. In both articles, it is not reported how many participants were considered in the analysis. In Dalanon et al. (2018) [166], no statistical significance of the results is reported (Question 4.5). Reasons for deducting points for "mixed-methods studies" included the fact that it remained unclear how the differend study designs fit together or were built on each other (Question 5.1) (e.g., Clarkson et al. (2020) [116]). Further points were deducted when considering quality of the individual studies (Question 5.5); in the case of quantitative descriptive studies, for example, when information on pre-tests was missing (e.g., Paisi et al. (2020) [117], Papautsky et al. (2021) [161]). An overview of the MMAT assessment of all studies is given as additional files (A11.–A15. Quality assessment by MMAT: study design I–V).