Introduction

Oral health is fundamental to overall health and well-being, but the separation of medical and dental care in terms of funding, education and care delivery has resulted in persistent oral health inequities and challenges in achieving oral health for millions in the USA.1 Dental diseases including periodontal disease and dental caries are associated with worse chronic disease control,2 worse nutrition3 and social isolation,4 with unplanned dental visits resulting in an estimated 92 million lost work or school hours annually.5 6 Dental disease and subsequent tooth loss are more common in individuals from historically under-resourced groups, including people who identify as racial and ethnic minorities and low-income people.7 8 This is compounded by the structure of dental insurance in the USA, which leaves more than a third of the population without coverage. The largest public health insurance programmes in the USA both offer limited to no dental benefit. Medicare, which provides health insurance coverage to adults over age 65 and adults with disabilities, does not offer a dental benefit. Medicaid, which provides health insurance coverage to low-income adults and children, does not offer a universal adult dental benefit, and services covered vary substantially by state of residence.9 Even in states that offer a dental benefit through the Medicaid programme, only 33% of dentists nationally choose to participate, further reducing access to dental care.10

Insufficient access to dental care leads to negative downstream consequences for individuals and the health system. There are nearly 2.5 million visits to hospital emergency departments (EDs) for dental problems each year,11 though definitive dental treatment is almost never available12; in one study, only half of patients with an ED visit for a dental problem successfully visited a dentist within 6 months, and many return to the ED for the same issue.13 14

Ideally, patients with unmet oral health needs would be supported in accessing oral healthcare before they required an ED visit. Previous work has demonstrated that even when patients gain dental insurance, additional individual (eg, transportation barriers) and community (eg, low density of dentists) factors continue to play an important role in dental care access.15 16 Adverse social determinants of health (aSDoH), such as food or housing insecurity, are associated with both ED utilisation and poor oral health.17 18 Efforts to screen ED patients for aSDoH are nascent but may be successful in connecting patients with resources and improving health.19 20 It is unknown how screening for aSDoH can affect dental care access for those patients with unmet oral health needs.

To better understand the interaction between aSDoH and unmet oral health need, and to develop best practices in supporting patients in meeting both oral health and social resource needs, we present the protocol for a randomised controlled trial (RCT) of ED-based screening and care referral. Reporting in this manuscript is in accordance with the Standard Protocol Items: Recommendations for Interventional Trials guidelines.21 Our study’s primary aims are to: (1) determine if addressing identified aSDoH can improve linkage to oral healthcare for ED patients, (2) evaluate the impact of addressing aSDoH and unmet oral health need on subsequent ED visits for dental problems and (3) identify the association between neighbourhood dentist and aSDoH resource density on intervention effectiveness.

Methods and analysisStudy design and setting

The overall design is an RCT that will test intervention efficacy, while gathering information on implementation. The study population is ED patients with unmet oral health needs because untreated oral disease is both a marker of challenges accessing care and a cause of worsening health outcomes.22 This is a multicentre trial across three EDs (two academic, one community) affiliated with the same health system, enrolling patients with residence within the geographical catchment area for any of the three hospitals.

This study is a randomised, unblinded, parallel-group trial with three arms. Patients in Arm A will receive standard care information with general oral health and aSDoH resources. Patients in Arm B will receive geographically proximate aSDoH resources and a search link to identify geographically proximate oral healthcare. Patients in Arm C will receive information with geographically proximate aSDoH resources information, a search link to identify geographically proximate oral healthcare and phone-delivered active navigational assistance. The randomisation will use a central stratified permuted block randomisation procedure by clinical site (proportionally allocated based on usual volume) and age group (approximately 90% adult and 10% paediatric) to ensure balance of these factors across the three arms. The randomisation sequence will be a priori generated by the study statistician and concealed from all other study staff until assignment to a participant. Randomisation will be performed using the Research Electronic Data Capture (REDCap) randomisation module, which is programmed to prevent access by study staff and will only display the next appropriate allocation once a participant is deemed eligible and provides consent. The study statistician will have no other role in participant enrolment and assignment. There will be no blinding of study group assignment due to the nature of the study procedures. Study design is portrayed in figure 1.

Figure 1

Study design with points of study team contact. aSDoH, adverse social determinants of health; ED, emergency department; OHIP, Oral Health Impact Profile.

Study interventions were developed on the basis of the investigators’ prior work in care navigation for aSDoH,19 20 23 as well as qualitative interviews with ED patients, ED clinicians and community dentists about barriers and facilitators to meeting resource and dental needs.24

Participants and recruitment

All subjects will be enrolled during an ED visit. Eligible patients are those who reside within the geographical catchment area for the three hospitals, who have capacity to consent, are able to communicate in English or Spanish, are medically stable for enrolment (have an Emergency Severity Index score of 2 or greater) and have a phone number for follow-up. Speakers of other languages will not be included because of the need to fully validate all navigational materials in subjects’ preferred language. Patients on involuntary holds, presenting from carceral facilities or presenting for acute mental healthcare or assistance with intimate partner violence or care following sexual assault will not be eligible to participate. We will not enrol these patients due to potential inability to consent and to maximise privacy and reduce contact with non-critical personnel for vulnerable patients.

ED research assistants (RAs) will screen the electronic medical record (EMR) track board for eligible patients. In this study, we define an ‘unmet oral health need’ to be a positive screen for patient-reported lack of or challenge with access to dental care using a validated screening tool. For adults, we will use the Hope Home25 screening tool which contains three questions about not seeing a dentist in the last 5 years, dental pain and/or inability to see a dentist in the last 6 months. For children, we will use the Gazzaz26 questions on delayed or forgoing any dental visit and no preventive dental care visit in the past 12 months. A positive answer to any of the questions will be considered an unmet oral health need for the purpose of this study. Enrolment and randomisation are portrayed in figure 2.

Figure 2

Subject enrolment and randomisation. aSDoH, adverse social determinants of health; RA, research assistant.

Data collection

Enrolment will be conducted by trained English and/or Spanish-speaking RAs who are certified as Qualified Bilingual Staff or use the clinical interpreter service if non-bilingual. We will use time-shift sampling to balance recruitment across areas of the EDs, time of day and day of week to limit bias in recruitment. RAs will complete study forms in REDCap on iPads or laptops depending on site resources. Screening of the EMR for eligible participants and entry of patient medical record number (MRN), age and demographic information occurs prior to consent and assent. Participants who are able to complete the study questionnaires on their own device (eg, smartphone) will then receive a text message link to respond to study surveys electronically. The remainder of participants will have the form read to them by the RA who will enter the answers into REDCap via iPad or computer. Consent will be obtained verbally by RAs prior to initial screening and enrolment. In addition to parent/guardian consent, verbal assent will be obtained from paediatric subjects aged 7 and older.

On enrolment, participants will complete the Baseline Demographics and aSDoH Screen to assess social needs and risks, as well as the five-question Oral Health Impact Profile (OHIP-5), a validated assessment of oral health-related quality of life.27 The demographic information includes racial/ethnic identity, preferred language, level of education, income, home address (for geographical resource linkage) and measures of literacy and numerical literacy. The aSDoH screening tool includes previously validated and published social risk and social needs questions about food, housing, transportation and utilities.19 The tool features three questions, respectively, on housing and food security, and two each for utilities and transport support. A participant will be noted to have an aSDoH need if any of them respond affirmatively to any of the questions.

Participants will be randomised to one of three arms. Participants in Arm A will receive information with general oral health and aSDoH resources. The oral health resources are identical to those used in the EMR discharge instructions for patients presenting with oral health complaints. The aSDoH resources are identical to those provided in the EMR as a ‘tip sheet’ for patients identified as having a specific aSDoH concern. Neither are geographically specific.

Participants in Arm B will receive information with geographically proximate aSDoH resources. For dental resources, those completing the study on their own device will receive the link to the state’s public insurance (Medicaid) website to search for dental providers in their neighbourhood. Those completing the study with RA assistance will receive a paper printout of search results. For aSDoH resources, they will receive geographically targeted resources based on a directory created by the study team, which will be offered the choice of completing the survey electronically on their own device or on paper.

Participants in Arm C will receive information with geographically proximate aSDoH resources and oral health search information (Arm B information) in addition to active navigational assistance. Active navigational assistance will be provided by Mapping Oral health and Local Area Resources (MOLAR) Study Navigators by phone 2–7 days following the ED visit or discharge date if participants are admitted and not discharged until a later date. The MOLAR Study Navigator will use a structured checklist to review the patient’s social risks and needs, specifically around food, housing, transportation and utilities, confirming the answers given in the earlier screening process. This is critical because participants may not want navigational assistance with all identified needs. Based on the patient’s identified needs, the Study Navigator will then provide active support in connecting participants to existing programmes and resources that match these unmet social needs. This can include providing personalised resources for food, housing, transportation and utility needs. The Study Navigator will work to identify a dental clinic with available appointments and support the participant in booking and attending their appointment. Depending on participant needs, this may include calling with the participant to directly book appointments or register with organisations or dental offices. Patients will be contacted eight times before being considered a non-respondent for navigational assistance. Non-respondents will still be contacted for survey completion and will be included in chart review.

Participants will be reached for follow-up at 3, 6, 9 and 12 months by text message or phone based on their preference. No additional needs will be assessed or resources provided during the follow-up assessments. At each follow-up timepoint, participants will complete the OHIP-5 as well as a linkage assessment and a healthcare utilisation questionnaire. The linkage assessment will ask if respondents were able to see a dentist since their last follow-up as well as the date of their appointment, and if they were able to make contact with given aSDoH resources, whether they enrolled in or received new resources and how helpful these resources were. Participants will be asked if they visited a primary care provider, paediatrician, urgent care or ED and if so, how many times they were seen at each one. All participants will report their satisfaction with the intervention at the first follow-up contact. If randomised to receive care navigation, participants will also be asked about their ability to connect with the ED navigator and their level of assistance using a validated Likert scale.28 At 12 months, a retrospective chart review will be performed by the RAs to ascertain healthcare utilisation as documented in the EMR. The chart review will assess both general and oral health-specific utilisation at any time during the defined study period. Subjects who are not able to be reached for follow-up will still be included in the chart review.

Data storage

All data will be stored in REDCap,29 an encrypted, password-protected database. Enrolment forms completed on paper will be entered into REDCap and the original forms will be stored in a locked file cabinet in the principal investigator’s office.

Sample size

There are little data on linkage to care for patients with unmet oral health needs in the ED setting to power this study. The best information that we have is that a pilot evaluation of patients admitted to one of the three hospitals for non-dental concerns, completed by our team, found that 68% of subjects had at least moderate periodontal disease.30 Regarding follow-up and outcomes, our pilot work on text messaging for aSDoH in the ED showed 74% of patients completed the text message follow-up. In that study, none of the control participants used the resources, compared with 9% of those in the intervention arm. Assuming a control group linkage to care rate of either 1% or 10% (based on the baseline and intervention states from our pilot) and that the intervention group response is 5–10 percentage points higher (similar to our pilot data), a range of 255–2049 patients would be required for analysis conservatively accounting for multiple comparisons between arms with a Bonferroni correction.31

Because of the limited data to power the study, we have planned for a conservative enrolment size of 2049 (683 per arm). Based on a sample of 683 patients in usual care and 683 patients in oral health information and aSDoH groups, we will have over 80% power to see a 6 percentage point difference in linkage to care at 3 months. Based on the demographics of included hospitals, we estimate that 60% of patients will present from a low dental density neighbourhood. It is possible that the intervention will be only effective in high dental density areas (ie, with a 10% increase in completion of dental appointments); in contrast, a previous paediatric study suggested that neighbourhood factors explain only 2% of the difference in dental attendance yearly.32 With the 1366 patients receiving the dental intervention, with the recognition that robust SEs will be used to account for clustering, we would have between 80% and 90% power under the assumptions above.

For an analysis sample of 2049, assuming 70% will follow-up at 1 month, we will need to enrol 2927 people; if 60% of those approached consent and enrol, we would need to approach 4879 individuals. We therefore budgeted to approach 4900 individuals.

Data analysis

All primary analyses will be based on the intention-to-treat population. The intention-to-treat population will include all participants who were randomised and completed the baseline aSDoH screening. Our primary outcome is to test the efficacy of two forms of customised oral health and aSDoH navigation assistance intervention to improve linkage to oral healthcare for individual ED patients. Our secondary outcomes include testing the impact of the intervention on repeat ED visits for oral health problems; and identifying the association between geographical distribution of resources and intervention effectiveness at community level.

Data will be analysed in STATA (StataCorp, College Station, Texas). Significance will be set at p<0.05. Trial data will be reported in accordance with Consolidated Standards of Reporting Trials guidelines. The final analysis will be performed after all subjects have completed the 12-month visit or dropped out prior to 12 months. No interim analyses are planned.

Due to the three-arm design of this study, analyses will include multiple comparisons with pairwise comparisons between arms (AB, AC, BC). Because of the planned pairwise comparisons, we have elected to use a false discovery rate approach using the Benjamini-Hochberg procedure to control the inflated type I error.33 To evaluate outcomes of our customised oral health and aSDoH navigation assistance intervention, we will compare linkage to oral healthcare between the three study groups using a generalised linear mixed model accounting for clustering by study site to calculate risk ratio and risk difference. To test the impact of the intervention on repeat visits for oral health problems, we will assess the association between the (a) baseline aSDoH and (b) study group (usual care, oral health information alone, oral health+aSDoH) and any return ED visit using a logistic regression model. To identify the association between geographical distribution of resources and intervention effectiveness at community level, we will generate separate models to determine association between density of resources and linkage to care at the neighbourhood and individual levels. For all outcomes, because this is an RCT, we will use unadjusted models for the primary analysis, with subsequent sensitivity analyses adjusting for potential confounders.

A sensitivity analysis is planned using the per protocol population, defined as participants who completed the intervention as originally allocated. We will also complete a priori planned subgroup analyses to compare the primary outcome by participant sex, gender identity, race/ethnicity and language using the intention-to-treat and per protocol populations.

Patient and public involvement

As part of initial study development, we conducted a preliminary series of qualitative interviews with 25 ED patients. These interviews were used to guide how screening and resource dissemination will be conducted in the trial. We do not currently have plans to involve patients in dissemination.

Timeline

We anticipate enrolment to take 3.5 years, and each subject will be followed for up to 12 months. Human subjects review board approval has been obtained for all three sites and the Manual of Procedure and all trial documents were finalised in July 2023. Enrolment began in September 2023; as of November 2023, two hundred and twenty participants have been enrolled and randomised.

Ethics and dissemination

All three sites are affiliated with a single healthcare system (Mass General Brigham), which allows for centralised human subjects review board approval (Mass General Brigham Protocol 2022P003224) that incorporates all eligibility screening, recruitment, enrolment and follow-up data collection.

The primary risk to participants in the study is privacy breach which will be minimised by the use of coded data and a secure database (REDCap). A secondary risk is potential distress completing the screening tool and reporting aSDoH or unmet oral health needs. This will be minimised by providing community resources to reduce unmet oral health needs and social needs, and through involvement of the subject’s treatment team (if during enrolment) and social work as needed.

In addition to the actions of the study team, study oversight will be under the direction of a Data Safety and Monitoring Board (DSMB) composed of members with expertise in clinical, statistical, scientific and ethical disciplines. The Board is comprised of four members with overlapping expertise in public policy, health services research, social emergency medicine, statistics, and public health. This DSMB has been appointed by the National Institute of Dental and Craniofacial Research (NIDCR) and is independent of the study team and grantee institution. NIDCR has a process in place to review conflict of interest disclosures provided by the members of the oversight board. Clinical site monitoring will be performed by the NIDCR’s Clinical Research Operations and Management Support contractor in collaboration with study staff through a Clinical Monitoring Plan and will also be notified of any protocol modifications.

Our study has several limitations. First, our intervention is restricted to three EDs affiliated with a large medical system; while we include both academic EDs and a community hospital ED, we hope to pilot the programme in alternative settings (eg, rural community EDs) on the basis of this trial. Because of the intensive contact required for the trial, we restricted enrolment to English and Spanish-speaking ED patients. While this includes most ED patients at these sites, future work must ensure efficacy for linguistically diverse populations. Lastly, our state offers uniquely generous Medicaid dental benefits. The absence of a dental benefit for many adults in other US states may impact the efficacy of navigational support for dental care.

The final results of the trial will be shared in a peer-reviewed publication, with ongoing participation in conferences relating to dental health, emergency medicine and health policy. Findings will be provided directly to the NIDCR and health system leadership for consideration of further scaling and evaluation in other clinical settings. After deidentification and removal of information that may be used to infer the identity of individuals, data will be shared under controlled access conditions as required by institutional policy.

On completion of our study, we will have established whether addressing aSDoH improves access to dental care for patients with unmet oral health needs, and the value of care navigation compared with usual care. We will also understand how the interaction between aSDoH and unmet oral health need affects oral health-related quality of life; and we will evaluate the role of the density of social and dental resources within subjects’ communities on successfully addressing aSDoH and unmet oral health need. Moreover, we will understand the impact of linkage to care and aSDoH on subsequent ED utilisation for dental problems. Taken together, these outcomes will allow us to contextualise individual access to oral health and social resources within health systems and communities. We believe this will demonstrate actionable recommendations for health system administrators and policy makers to address upstream determinants of health and improve care for patients with unmet oral health needs.