Quantitative data collection approaches in subject-reported oral health research: a scoping review

Abstract

Background

A fundamentally different approach is currently needed to address the oral health of populations worldwide namely by considering the perspective of patients or populations and not only dental professionals' views [1]. It seems increasingly necessary to integrate the self-reported perceptions of oral health, as they can complete or even replace clinical measures of dental status in surveys of populations. Indeed, such subjective measures are easy to use in large-scale populations and can provide a broader health perspective as compared to clinically determined measures of dental status alone [2, 3]. Since the topic is broad, this scoping review sets out to identify methods employed in population surveys that discussed self-reported perceptions of oral health, and the extent to which new computer-oriented technological approaches are being incorporated in the research methods.

The literature on oral health and dental-related scoping and systematic reviews includes studies that use specific populations in terms of disease or clinical conditions, treatments, political or social status and typically do not explore oral health status outcome measures [4–15]. These studies only occasionally provide perspectives on general populations. A review by Mittal et al. identifies dental Patient-Reported Outcomes (dPROs), and dental Patient Reported Outcome Measures (dPROMs) related to oral function, oral-facial pain, orofacial pain and psychosocial impact [16]. The study affords a valuable and extensive review of self-reported oral health and quality of life measures, many of which are found in this paper. This scoping review, then, seeks approaches used in subject-reported surveys, including those with general populations, which may broaden the perspective on oral health outcome measures.

The objective of this review is to categorize measures used to evaluate self-reported oral health status and oral health quality of life used in surveys of general populations.

Methods

This work is implemented following the framework of scoping reviews [17–19] and is presented according to the recommendations of the Preferred Reporting of Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR), as listed in Additional file 1: Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) Checklist [20]. Additional file 5: Glossary of Terms provides definitions for the important terms used across the paper.

Selection strategies

Figure 1 outlines the review process utilizing the PRISMA-ScR framework. The title-and-abstract screening was completed by a researcher (D.X.) against the inclusion criteria using a screening tool (Additional file 3: Search Tool). To check for reliability and consistency, one of the researchers (L.Z.) randomly screened 10% of articles independently and compared the inclusion decisions. Given the result of title-and-abstract screening, two researchers (L.Z. and Y.H.) verified the eligibility of the remaining articles independently through full-text review. Inclusion discrepancies were resolved by an additional researcher (D.X.).

Open in a separate window

Fig. 1

PRISMA framework with additional examples

Results

Discussion

This scoping review provides an overview of data collection methods used for subject-reported surveys to measure oral health outcomes. Studies are characterized by four survey methods (in-person, mail-in, Internet-based, and telephone) and by summarized dimensions and attributes of data collection for each method, such as technological approaches, survey population or sampling methods. Studies typically employ in-person surveys and more studies were conducted in Asia-Pacific and Middle East countries than in any other world region. Most studies recruit participants directly from study sites. Both probability and non-probability sampling methods employ typically convenience sampling, simple random sampling, and stratified sampling. Studies that achieve the highest response rate on average use in-person surveys, while the lowest rate occurs in Internet-based surveys. Telephone surveys are used to cover a wider population compared to other data collection methods. Many studies, especially those using in-person and mail-in data collection methods, incorporate supplemental data types and technological approaches. Outcome measures are frequently used to evaluate impacts caused by functional limitations related to physical, psychological, and social factors.

Frequently used self-reported oral health status and OHRQoL measures are OHIP-14, an impact measure, and the GOHAI, a functional measure. Children’s oral health outcomes measures form a considerable portion of the self-reported oral health outcome research papers. Although OHIP-14 is the most utilized single measure, many other papers use only portions of this measure, while adding other outcome measures, such as dental care needs, satisfaction, oral health status, and so on. The validity of these measures is therefore compromised and could not provide insight into the degree that the studies are measuring self-reported oral health status or quality of life [4]. Other measures rate an individual’s oral health status using a simple self-rating scale, from very poor to excellent. This approach is more directly related to a person’s oral conditions and therefore their perceptions and behavior tend to be more consistent with this rating [56, 57]. These self-rating measures focus on the overall dimension of perceived oral health status. Unlike the measures previously discussed, these simple ratings do not delineate the psychological, social and physical dimensions of oral health. Nevertheless, such measures can enable researchers to identify hidden dimensions by analyzing independent variables that account for the respondent’s perception.

This review identifies research that employs more conventional methods. The face-to-face interview and the pencil and paper format are conventionally used in many studies along with a clinical dental exam. While offering unique flexibility and easier administration, in-person approaches are more labor-intensive and normally take more time compared to other methods. Countries, such as Brazil, rely for years on these techniques to develop national epidemiological oral health surveys [28, 58, 59]. Although these surveys are very well-organized and established throughout the country, this review does not find that newer technological approaches are introduced into their conventional approach. In this case, there may be little incentive to change their approach because their methods are well understood and employing more technological approaches may be costly.

The use of Internet-based surveys is increasingly common in the medical field. Although these surveys end with potentially lower response rates, this approach is normally more cost-effective [60]. Internet-based surveys have many notable advantages, including easy administration, fast data collection process, lower cost, wider population coverage and better data quality with fewer overall data errors and fewer missing items [61–63]. However, this data collection method is constrained by sample bias, topic salience, data security concerns and low digital literacy that may affect response rates [62]. In settings where Internet-based surveys are not practical, longstanding and effective conventional oral health data collection methods in research will continue. It is evident from this review that the use of computerized technological approaches is limited. While such approaches in survey research improve the quality and quantity of data collection, only twelve studies in this review employ them. The most widely used technical approaches are Computer-Assisted Personal Interviewing (CAPI) and online survey platforms (e.g., Google Forms and SurveyMonkey).

Two noteworthy approaches to survey research methodology emerge from this review, particularly in: (1) sample selection, and (2) mode of interaction with research subjects. North American researchers found different platforms to access subjects for their studies. Canadian studies use random digit dialing to recruit and conduct computer-assisted interviews [54]. In the United States, researchers access existing polling populations or use Amazon’s MTurk platform for “workers” who are paid small amounts for each survey they respond to [64]. The second approach is the use of computers to collect self-reported data. The basic surveying technique is CAPI with interviewers directly entering the data into a database. There is also Computer-Assisted Telephone Interviewing (CATI), a survey technique, where the interviewer follows a scripted interview guided by a questionnaire that appears on the screen. A third Internet-based survey technique, the Computer-Assisted Web Interviewing (CAWI), requires no live interviewer. Instead, the respondent follows a script made in a program for designing web interviews that may include images, audio and video clips, and web-based information.

An innovative technological approach worth noting is the use of OralCam to perform self-examination using a smartphone camera [65]. The study applies research used in medicine to detect liver problems from face photos as well as other diseases [66]. The paper describes the use of a smartphone camera to interact with a computer using diagnostic algorithms, such as the deep convolutional neural network-based multitask learning approach. Based on over three thousand intraoral photos, the system learns to analyze teeth and gingiva. The smartphone camera takes a picture using a mouth opener. The computer’s algorithms analyze the captured picture, along with survey data, to diagnose several dental conditions including caries, chronic gingival inflammation, and dental calculus. This use of multitask learning technology, with the extensive availability of cell phones, may revolutionize oral health research and care.

This scoping review is limited to oral health survey-based studies in peer-reviewed journals and MedRxiv published in English between 2011 to 2021. A further limitation is that many of the reviewed papers do not adequately describe the methods they use to collect data. Publications using secondary data from national studies are excluded, The exclusion is based on the fact that these researchers are not engaged in designing the methods or conducting the data collection. Often, the publications refer to the original study to describe the method used. Also, the original data collection may have occurred before the time frame of this review. The fifteen papers that use secondary data published over this study’s time frame represent only about six percent of the reviewed papers. Thus, the overall impact of this exclusion is minimal on this scoping review’s results.

Conclusions

This scoping review provides an assessment of oral health outcome measures, including subject-reported oral health status, and notes newly emerging computer technological approaches recently used in surveys conducted on general populations. Such technological approaches, although rarely used in the reviewed studies, hold promise for both researchers and the various populations that use or need oral health care. Future studies employing more developed computer applications for survey research to boost recruitment and participation of study subjects with wide and diverse backgrounds from almost unlimited geographic areas can then provide a broader perspective on oral health survey methods and outcomes.

Supplementary Information

Acknowledgements

Abbreviations

Author contributions

C.M., D.X., M.M. and H.L. conceptualized the study and designed the data collection form and established the data analysis plan. A.O. developed search strategies and carried out searching on multiple databases. D.X., Y.L., Y.H., J.S. and Y.L. performed additional searching and tested the data charting form. D.X., Y.L., and Y.H. helped to screen studies for relevance and data charting. C.M. and M.M. reviewed full-text papers and verify the data charting results. C.M., D.X., and M.M. drafted the original manuscript. D.X. and L.Z. prepared Tables ​Tables1,1, ​,22 and Fig. 1. C.M., D.X., M.M., and L.Z prepared Tables ​Tables33 and ​and4.4. All authors read and provided substantial comments/edits on the manuscript and approved the final version. All authors read and approved the final manuscript.

Funding

This research was supported by an NIDCR/NIH grant to the University of California, Los Angeles (UCLA) (U01DE029491). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Availability of data and materials

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Footnotes

References