This is a multisite, single-blind, parallel-arm, community-based cluster randomized controlled trial with 2 arms. The trial has been registered with Clinicaltrials.gov (NCT03916926) and is currently in the recruitment phase. This protocol report follows the SPIRIT [28] and CONSORT (Consolidated Standards of Reporting Trials) guidance [29].

The primary aim is to compare 2 evidence-based strategies in low-income older adults aged 62 or older followed for 52 weeks (1 year): a “simple medical” strategy of topical application of SDF (Arm 1) versus a “typical dental” strategy consisting of ART and topical application of FV (Arm 2). The rationale for adding FV to the ART arm is that ART has a direct effect on the treated tooth and perhaps on the adjoining tooth surface while FV has whole mouth benefits, as does SDF. Our primary hypothesis is that simple medical treatment (Arm 1) is not inferior to typical dental treatment (Arm 2) for clinical (caries lesion arrest) and patient-reported (tooth pain/hypersensitivity) outcomes at 52 weeks after treatment. Our secondary hypothesis is that the simple medical treatment (Arm 1) is not inferior to typical dental treatment (Arm 2) for clinical (new caries lesions) and patient-reported (oral health quality of life) outcomes at 52 weeks after treatment.

Study sites are 22 housing facilities for adults of low income in Northeast Ohio. In these facilities, residents live independently in individual apartments. Older adult tenants aged 62 or older will be approached for participation. Those who self-report that they are without teeth will not be recruited. The recruitment will follow a 2-stage process: (1) all participants consented will participate in the dental screening; (2) participants found to have nonurgent cavities (ie, those without irreversible pulpitis, periapical abscess or cellulitis or facial swelling) at the screening and who fulfill the inclusion criteria (see below) will be enrolled into the clinical trial.

Service coordinators at the facilities will serve as site liaisons. The study principal investigator/project manager will provide the coordinators with an introductory letter/flyer (Multimedia Appendix 1) containing study information and ask that it be given to tenants and posted in public areas. The coordinators will arrange an informational meeting (study dentist will give a talk on the interventions as suggested by our stakeholders), and study staff will present information regarding participation at scheduled group events such as tenant meetings or health fairs. For planning purposes, coordinators will have a sign-up sheet for those interested in the sessions. Study staff will schedule potential participants for one-on-one sessions at the housing facility to obtain informed consent and collect a baseline survey. A dental examination and treatment appointments will then be scheduled at each facility according to a designated roll-out schedule for each facility, approximately 1-2 weeks following consent and baseline data collection.

According to the 2-stage process for enrollment, the initial inclusion criteria of participants are as follows: provide signed and dated consent form, willing to comply with all study procedures, being available for the duration of the study (1 year), male or female, aged 62 or older, and living in a participating facility. In the second stage of enrollment, additional inclusion criteria for continuation and participation in the randomized controlled trial are based on the dental screening. A participant must have at least one untreated root surface or coronal caries lesion on any permanent tooth with an International Caries Detection and Assessment System (ICDAS) II [30] active lesion score of 3 or greater (localized enamel cavity to extensive cavity). At this stage, further exclusion criteria are sensitivity to silver or other heavy-metal ions or oral ulcerative gingivitis or stomatitis, which prevents the potential participant from receiving study treatment. The study has a 26-month recruitment time frame that started in October 2019.

Participants are followed at 26 weeks (6 months) and 52 weeks (12 months) that includes dental examinations/treatment and survey completion. This follow-up period is consistent with another caries arrest trial [21]. The schematic of the study design is presented in Figure 1.

All visits will occur at the housing facility where the individual participant resides. Several strategies will be used to retain the participants: promotional items (ie, pens, magnets) with the study logo and contact phone number at recruitment (alternate contact information for family/friends that may allow to reach the participant if primary contact information becomes invalid will be obtained at recruitment); annual birthday/holiday cards will be sent to participants to maintain contact; and newsletters will be sent twice a year with updates on the study’s progress including reminders to update their contact information (by phone or postal mail). Assistance from service coordinators or other facility staff will be sought to maintain contact with hard-to-reach/contact participants. Cash incentives will be given to the participants at the baseline screening and examination (US $25); those enrolled in the clinical trial will receive additional incentives at 26-week follow-up visit (US $40), and at the 52-week visit (US $40) for the completion of all study procedures.

The Case Western Reserve University Institutional Review Board approved all study procedures (STUDY20190481) and written informed consent is obtained from the participant.

Participants’ schedule for data collection is given in Table 1. A summary of the data collected is as follows.

The study staff members who will facilitate recruitment, scheduling, and data collection will attend a 2-day in-person training on the conduct of the study protocol and logistics. Specialized training in interviewing older adults will also be included. In-class training incorporates the topics of human subject protection, good clinical practice, and the study protocol. Study staff members will be monitored through periodic data audits and through direct observation of calls and recruitment activities. Staff members will receive feedback on their performance and conduct including but not limited to these specific areas, adherence to protocol and good clinical practice. Corrective training will be provided as required.

This study tests the hypothesis that simple medical treatment (Arm 1) is noninferior to typical dental treatment (Arm 2) for primary outcomes (caries arrest, tooth pain/hypersensitivity) and secondary outcomes (new decay, oral health quality of life) at 52 weeks after treatment. For continuous outcomes power was computed using a variance correction (ie, variance inflation factor) to take into account possible correlations of outcomes within cluster. All computations of required effective samples sizes (or corresponding power) were done using the PASS 2005 software. For binary outcomes, a specialized program for cluster randomization (a noninferiority test comparing two proportions) was used.

The first primary outcome tooth pain was defined as change in pain (based on a 100-mm visual analog pain scale, where a higher score indicates worse pain) from baseline to 1 year. A mean difference of 0 between the SDF and ART + FV arms, and noninferiority to be within a margin (mean difference) of 8 between the SDF and ART + FV arms were considered, and based on prior literature we also assumed a common standard deviation of 25 [45]. Further, an average of 25 patients recruited per site, an intraclass correlation of 0.01 (based on prior studies and literature for similar populations) [46,47], and a conservative 15% dropout rate (based on prior studies) were considered including the possibility of deaths among participants. The use of a .025 α-level one-sided t test to test the null hypothesis of inferiority versus the alternative hypothesis of noninferiority (as defined above) in 11 sites per treatment group (corresponding to 275 subjects per treatment group or 550 in total) will provide an estimated 89% power to conclude noninferiority.

A second primary outcome is arrest rate. Previous data [19,48] show a high arrest rate (of around 90%) for ART + FV. A similar arrest rate for biannual application of SDF is expected [23]. For the power calculation, the unit of analysis was assumed to be the person, and considered the binary outcome of arrest, which is defined as all lesions for an individual being arrested. An equal (person-level) arrest rate of 90% for the ART + FV and SDF groups, and noninferiority to be within a margin (difference in arrest proportions for ART + FV versus SDF) of 0.09 were considered. With the same assumptions as before using a .025 α-level one-sided z test, the targeted sample size of 275 subjects in 11 sites per treatment group will provide an estimated 85% power to conclude noninferiority.

Randomization is at the level of the cluster (housing facility) for logistical efficiency, that is, we require supplies and personnel only for one intervention (rather than both) at each site. In addition, having only one intervention at each site will greatly reduce the potential for error (mistakenly giving the wrong treatment) that could otherwise occur with people at the same site assigned to different treatments. Furthermore, keeping the same treatment at each site reduces chances of contamination (ie, participant discussing their treatment with others). Particularly as one of the primary endpoints (self-reported pain/sensitivity) is subjective, it is important, in a context where blinding is not possible, to minimize the possibility that participants may perceive, through communication with others, that the treatment they are receiving is inferior or superior.

Additionally, stratified cluster randomization will be used, that is, a block (constrained) randomization approach in which balance over treatments is assured for 2 key cluster-level (stratification) variables, namely, facility size (>100 versus ≤100 residents) and geographic location (Cuyahoga County vs other). While the randomization is at the housing level, the study objectives, interventions, and primary outcomes all pertain to the individual level.

Study participants are blinded to the study group. Study staff will be blinded at the time of recruitment and obtaining consent.

Each primary outcome will be compared between the SDF and ART + FV groups. For tooth pain, a 95% CI based on a t test for the difference in mean responses (SDF minus ART + FV) will be computed. If the CI lies within the interval (–∞, 8), we may conclude noninferiority of SDF relative to ART + FV treatment. The CI may secondarily be examined to assess possible superiority of one intervention over the other. For arrest rate, a 95% CI for the difference in rates (based on a z statistic) will be computed. If the CI lies within the interval (–0.09, 1), we may conclude noninferiority of the SDF relative to ART + FV treatment. As above, possible superiority of one intervention over the other may also be assessed. For other outcomes, we will also compute 95% CI for differences in means (or proportions for binary outcomes). These secondary outcomes will be assessed in an exploratory manner for possible superiority or inferiority based on appropriate margins.

We will seek to corroborate initial results using a generalized estimating equation (GEE) approach. For each outcome, a GEE (marginal) model will be fitted that includes a treatment indicator and prognostic variables (including sociodemographic variables, medical conditions, and oral health behaviors). Appropriate link functions (eg, logit link for binary outcomes and identity link for continuous outcomes) will be specified and an exchangeable working correlation matrix will be used to allow for correlations within site. The arrest outcome will be analyzed as a binary outcome (as described in the “Sample Size and Power Estimates” section), and secondarily as the number of arrested lesions assuming an appropriate distribution (eg, negative binomial) and link function (eg, log link). Robust t tests with correction for a small number of clusters [49] will be used to test for treatment effects and corresponding 95% CIs computed.

Secondarily, the above GEE approach will be extended to analyze the repeated (baseline and 26 and 52 week) measures for each outcome. The models for each outcome will include the same prognostic variables as before, as well as time and a time × treatment interaction, and will allow for correlations among the repeated measures (eg, using a first-order autocorrelation structure). If a substantial within-facility correlation is found, the need to incorporate facility as a second cluster level (within which a person [ie, the first cluster level] is nested) will be assessed. To compare trends over time for the 2 interventions, estimation and testing (via a robust t test) for the interaction term will be done. If the use of 2 cluster levels is not found to be feasible in the GEE approach, a generalized mixed effects model approach will be considered.

The study results will be disseminated through local and national conferences, scientific publications, and channels established by a 11-member stakeholder engagement group established for this project. For local and city levels, dissemination will occur through presentations, workshops, and educational programs conducted in community organizations. At the state level, results will be shared with the Ohio Department of Health, Medicaid policymakers, and third-party payers who can revise reimbursement policies. Nationally, Housing and Urban Development (HUD) will create flyers to be distributed to their housing facilities and share results with other states. All stakeholder partners have suggested sharing the results on their organization’s respective webpages and newsletters, which reach large national, state, and local audiences.

The nonidentified data files, data dictionary, and supporting documentation will be available without restriction by application to the project biostatistician.

Dental practitioners lack robust guidance and standards of care for the treatment of older adults. The ability to provide treatment and the treatments themselves are impacted by the consequences of multiple medical problems, physical frailty, and dementia. Access is limited by the ability to pay. Dentists typically apply treatments developed for younger, healthier individuals, with predictably poor results. An article on the dental woes of an aging population [50] best summarizes this dilemma. This report relates an anecdote in which a public health dentist recalls a lecture given 20 years earlier on geriatric treatment during which the speaker presented slides of a patient with substantial evidence of previous treatment with excellent crowns and fillings. However, with gum recession, inflammation, and caries they had led to failure of the fillings and crowns. The patient was a retired dentist who was knowledgeable and had taken care of his teeth throughout his life, but medical problems, side effects of medications, and frailty took their toll. Conventional treatment with a drill was not an option in this case nor is it for many other older adults. The article concludes that with a growing geriatric population, a new standard of dental care was urgent. Nonsurgical treatments such as SDF and ART and FV are effective caries management strategies for older adults that address many of the limitations of traditional treatments.

Our study is a comparative effectiveness study of 2 nonsurgical treatments for addressing both root and coronal caries outcomes in older adults. Previous studies outside of the United States have mainly focused on the prevention and arrest of root caries lesions in community-dwelling older adults [19,21-23]. Both SDF and ART are two relatively inexpensive, evidence-based nonsurgical treatments that can be provided by dental hygienists (dental paraprofessionals) in most states outside of traditional dental clinics to address the unique oral health needs of community-dwelling older adults.

This study adds further to existing knowledge by addressing patient-reported outcomes such as pain and hypersensitivity that are of importance to older adults, as they affect their overall oral health quality of life. In our prior study of older adults living independently in 16 subsidized housing facilities in Northeast Ohio, 62% of the participants reported that their perceived tooth condition was fair/poor [6] with 40% reporting tooth pain and 54% tooth sensitivity. Both treatment strategies have the ability to kill plaque bacteria and strengthen the tooth surface [24-27] and thus address pain and sensitivity. SDF, in particular, has been cleared by the FDA as a tooth desensitizer.

This comparative trial can also provide needed evidence regarding the effectiveness of these interventions in community-based settings (ie, for public health purposes). Currently, there are no community-based public health interventions for older adults like those for children. For any dental caries preventive/treatment intervention to be useful on a population level, 5 presumed attributes are necessary: pain and infection control, simplicity of use, cost affordability, minimal personnel time and training, and noninvasiveness [51]. Potentially, these nonsurgical interventions have these attributes and can address the limitations of the current clinic-based dental care delivery system which is expensive and ill adapted [52].

The state requirements for supervision vary considerably. The results of this trial may be applicable to settings in which dental hygienists, or perhaps dental therapists, are permitted to provide services under general supervision outside of traditional clinics and practices. The results are limited by the necessity to select particular dental materials based on our current assessment of their practicality and ease of application. Other materials may behave differently as industry is constantly innovating and changing the materials. Nevertheless, evidence for this class of treatment alternatives can result in more robust treatment guidelines and standards that are broadly generalizable.