Oral Health Intervention in Adult Primary Care

Overview

This study is a Stage III cRCT to test the efficacy of multi-level interventions at the practice- and provider-level to address low dental utilization (attendance) among Medicaid-enrolled older adults 55 years or older attending non-urgent primary care visits (PCV) in MetroHealth practice settings. Twelve practices will be randomized into two arms: A) Intervention arm will receive the multi-level intervention that includes: 1. Practice-level: EHR changes to include: ask, advise, assess, and connect (AAAC) strategies; 2. Provider-level: Medical staff (MA, nurse): Training in the AAAC process and complete AAAC for enrolled older adults; Clinicians (physician/nurse practitioner): CSM-based education (didactic), skills training (video training with standardized patients), and view completed AAAC in EHR to deliver core oral health (OH) facts to older adults, reinforce importance of dental visits, and document in EHR that OH facts were delivered. B) Control arm will receive, at the provider-level only (clinicians), non-theory-based information about retaining a healthy mouth using the ADA Mouth Healthy Series and deliver standard OH care for patients. Older Adults will be followed at 12 months and 24 months to determine if the participant had any dental attendance. The primary objective is to test the efficacy of the practice level EHR strategy to ask \[OH risk assessment\], advise \[going to dentist\], assess \[willingness for referral\], and connect \[eReferral, resources\] together with clinician theory-based education to communicate OH facts versus clinician alone (standard oral health care) in increasing dental attendance in primary care settings. The secondary objectives are to assess oral hygiene behavior, Geriatric Oral Health Quality of life, biometric measures (BP, serum cholesterol, blood glucose, hbA1c) abstracted from EHR data, potential mediators and moderators to investigate pathways that affect the primary and secondary outcomes, and assess implementation strategies: adoption, reach, fidelity, and maintenance of providers and practices that affect older adult primary and secondary outcomes. The hypothesis is that medical staff completing the AAAC strategy and clinicians with improved OH knowledge (chronicity, systemic effects) will deliver consistent oral health messaging to older adults at PCVs that will result in increased preventive and restorative dental utilization compared to those providers delivering standard care.

Study Design: This study will utilize a multi-level cluster randomized clinical trial design (Phase III) to assess interventions intended to address poor dental utilization among Medicaid-enrolled older adults aged 55 years or older attending non-urgent primary care visits (PCV) in primary care settings. The focus is on addressing factors (determinants) at the three socio-ecological levels: the practice/organization, provider (medical staff (MA, nurse) and clinicians (physician, nurse practitioner), and the older adult. A total of 12 practices will be randomized to 2 arms as follows: Intervention Arm (Arm A) includes practice-level changes in Epic EHR for the provider-level medical staff to deliver the ask, advise, assess, connect (AAAC) strategies to screen, refer, and provide resources for oral health + clinician CSM theory-based training to deliver oral health facts to older adults and acknowledge giving facts in EHR; Control Arm (Arm B) will receive provider-level standard non-theory based oral health training for the clinician and no changes to the EHR system or completion of AAAC by medical staff. Arm B participants will receive standard of care which is to address any oral health issues if brought to the attention of the provider. Each arm will consist of 6 practices (n= 12), = 47 clinicians (n= 95), and 400 older adults (n= 800). Only the 6 practices in Arm A will enroll = 95 medical staff. Older adults will be recruited at the first PCV and then be followed for two consecutive PCVs (for approximately 24 months). Providers will complete the training dependent upon arm (EHR, didactic + skills or standard didactic) prior to enrolling any patient and will participate in the study for a total of 24-36 months duration (includes initial and booster training). Booster training for clinicians will be completed prior to PCV #2. Immediately after randomization of practices, recruitment of older adult participants will be rolled out during a two-year period, i.e. 6 practices with 400 participants in each of the two years. The primary outcome will be dental attendance (receipt of dental care) assessed through CareSource Medicaid claims data. The dental attendance from Medicaid claims will be validated with a sub-set of clinical data from the dental EHR. Clinical data (i.e. encounter date, tooth level caries, periodontal data) will be abstracted for those patients who requested an eReferral to Metro dental. Secondary outcomes from participant questionnaires will be assessed at 24 months post 1st PCV as follows: oral hygiene behavior and oral health quality of life. Biometric measures (i.e. BP, weight, height, hbA1c) will be abstracted from EHR data. Participants: Study participants include medical staff (intervention arm), clinicians, and older adult patients from primary care practices within Cuyahoga County. Medical staff must be solely at the participating practice (not floating between practices) and not planning to leave the practice within a year. Clinicians must be solely in the participating practice (not floating between different practices), have a minimum of 2 patient-care days per week, and not planning to leave the practice within a year. Older adult patient participants are those 55 years or older, attending a wellness or a non-urgent PCV, and enrolled in CareSource Medicaid. All participants meeting the eligibility criteria will be enrolled in the study upon signing the consent form. Procedures: Pre-PCV #1 Visit: Providers: Reviews the written/electronic consent form and signs the form. Medical staff (Arm A) attends EHR training and clinician attends OH didactic education and skills training (Arm A) to communicate core OH facts (dental and systemic links, chronicity of dental disease, dental visits) to older adults or clinician (Arm B) attends standard ADA based didactic training. Providers complete administered pre- and post-tests (clinicians only), socio-demographics (before only), and self-efficacy (clinicians only) to providers before and after the training session. PCV #1 (Baseline): Providers: Medical staff in Arm A asks oral health questions and provides an eReferral/structured referral. Clinician delivers oral health facts and documents OH in EHR, based on study arm. Participants (before clinician encounter): Research staff will review inclusion/exclusion criteria, the written/electronic consent form, and patient signs the form. Medical staff then proceed with the AAAC process (Arm A) in EHR and provide an eReferral to Metro dental or structured referral to Medicaid-accepting dentists with resources. In the exam room, participants will complete the following baseline questionnaires: Brief Illness Perception Questionnaire-Revised for Dental Use in Older/Elder Adults (Brief IPQ-RDE) and the Older Adult Questionnaire. Participants (during clinician encounter): Older adults will receive the core oral health facts from the clinician in Arm A and reinforcement for dental visits. Arm B participants will receive the clinician's standard oral health care. Older adults attending PCVs may have research staff directly observe the encounter to evaluate the clinician's delivery of OH communication for a sample of visits. Participants (after clinician encounter): Older adults will provide feedback about the OH information given to them during the medical visit with a short exit questionnaire. Older adults will be given the B1: Follow-up Brief IPQ-RDE to be completed and returned. Participants after PCV #1: Older adults will receive telephone calls and text messages from research staff to remind them to complete and/or return the Follow-up Brief IPQ-RDE if it has not already been returned. Older adults will receive telephone calls and text messages from research staff 8 weeks after the PCV to complete the Dental Attendance Questionnaire. Before PCV #2: Providers: Provide OH didactic education booster session for clinicians based on study arm. PCV #2: Providers: Same assessments as PCV #1 Participants: Same assessments as PCV #1 PCV #3: Providers: Same assessments as PCV #1 (no booster training). Participants: Same assessments as PCV #1 except follow-up Brief IPQ-RDE and Dental Attendance Questionnaire. Analysis Plan: Primary Statistical Analysis: For the primary outcome of dental attendance (preventive, restorative) by 12 months (similarly for attendance between 12 and 24 months) the investigators will use a generalized estimating equations (GEE) approach with practices as clusters. This approach will be based on a marginal model with a logit link and include as covariates an intervention indicator variable and selected patient baseline variables considered as potential confounders. A corrected standard error, adjusting for the small number of clusters, and a corresponding score test will be used to test for an intervention effect, and 95% confidence intervals for the intervention effect (on an odds ratio scale) will be computed. To assess possible effect modification, interaction terms involving the intervention indicator and each moderator (for example, sociodemographic and social support variables) will be added to the above model and tested. The investigators will consider alternative within-cluster association models, for example, by allowing for a second level of clustering (of providers nested within practices), and through an alternating logistic regression approach (using odds ratios as measures of association among binary outcomes in a cluster or sub cluster). As a supplementary analysis, the investigators will analyze dental attendance in the two time periods (from baseline to 12 months and from 12 to 24 months) via a longitudinal model. Specifically, the investigators will extend the above GEE approach by accounting for clustering of the repeated measures within individual as well as within practice (and possibly provider). The investigators will test for an (assumed common) effect of the intervention on dental attendance each year, as well as test for varying effects over time by adding an intervention by time interaction term to the model. The initial analysis will assume that missing data are missing completely at random (MCAR). To account for missing data, the investigators will conduct sensitivity analyses in which responses are imputed under conservative assumptions. Additionally, sensitivity analysis will be conducted for those without a dental home vs. those with a dental home at the start of the study. For the longitudinal analysis, the MCAR assumption will be assessed by modeling missingness of the dental attendance outcome, and testing whether missingness in the second year is related to dental attendance in the first year. If a violation of MCAR is found, alternative approaches (such as use of a GLIMMIX model) will be considered. This model will use a logit link, include the same covariates as above, and incorporate random effects for practices, as well as second-level random effects for individuals in the longitudinal model. Secondary Statistical Analysis: Secondary outcomes (OH related quality of life, oral hygiene behavior, biometric measures collected at all three PCVs) will be analyzed using a GEE/marginal model approach similarly to that described for dental utilization, but using the appropriate link function depending on the variable; for e.g., an identity link (linear model) may be appropriate for continuous outcomes such as the quality of life measure and biometric measures. Log transformations will be considered for the sake of improving approximations to a normal distribution. Data Management: The research staff or older adult will enter answers either electronically or on paper and the research staff will review the questionnaires after completed and prior to the older adult leaving the PCV. The older adult will have a copy of the questionnaire to reference during the surveys as needed. Each paper form and questionnaire will be entered by one individual and verified by a different person. All study forms and questionnaires collected on paper will be double-checked and kept secure at all times. Further, study forms will be entered into the database and checked for inconsistencies and range and assessed for missing data. Any inconsistencies, outliers, or missing data observed will be compared to the paper document and appropriate corrective actions carried out. REDCap's native data resolution workflow will be used to document and fix any data anomalies. The Data Manager will also respond to data queries generated by the PI, Study Coordinator, or other research staff. Following collection of all data from the project, additional data processing will be required by the Data Manager, e.g., longitudinal coding of medical illness and dental claims data, creation of psychosocial scale scores. Following PI concurrence, the database will be locked. Data for this study will include: (1) oral health screening questions, (2) study questionnaires: older adults, provider, medical staff, and practice, (3) abstracted medical and Dentrix dental data, (4) abstracted Medicaid dental claims data, (5) data from observation of providers and medical staff, and (6) EHR audit data. Form revisions should be minimal; however, should they occur, changes will be submitted to the Data Manager for updating and dissemination to research staff. Quality control is primarily conducted at the study team level through internal processes of data review/data monitoring using periodic custom reports. The Data Manger will run regular validation reports to detect data anomalies and will work with the research staff to resolve any data anomalies that arise during data entry. REDCap's native data resolution workflow will be used to document and fix any data anomalies. The Data Manager will also respond to data queries generated by the PI, Study Coordinator, or other research staff.