A comprehensive assessment for community-based, person-centered care for older adults

Serving Seniors’ Gary and Mary West Senior Wellness Center (SWC) and the Gary and Mary West Senior Dental Center (SDC) are non-profit CBOs co-located in downtown San Diego, California. Serving over 900 clients daily, the SWC provides physical activity, meals, and social, legal, care-coordination, and mental health services to low-income older adults. The SDC, launched in October 2016, is a four-chair dental clinic designed to focus on collaborative care management for underserved and vulnerable older adults, situated on the second floor of the SWC [29]. This study reports on a convenience sample of 996 new and existing SWC clients who consented for assessment and referral to services between October 1, 2016 and January 1, 2018. Age under 60 was the only exclusion criterion.

This Continuous Quality Improvement (CQI) project employed Community-Based Participatory Research (CBPR) [30‐33] and human-centered design approaches [34, 35] to co-develop and pilot a person-centered care model linking health and wellness, dental, and social services by real-time CGA and metrics-based referral in a community-based setting. A multidisciplinary team of social workers, nurses, physicians, dentists, and care navigators met regularly over two years, with input from clients, and co-developed the CGA, with the new oral health assessment, prior to the SDC’s opening.

The clinicians and staff sought to extend, enhance and improve care quality and efficiency at the SWC. The team created and introduced a person-centered model of care to better identify and improve triage processes and care referrals. The addition of dental providers and services altered client care teams at this location and required shifting workflows and adding a pathway to dental care. This team proposed and refined workflows and informatics tools that provided preliminary metrics and rules for referral to support real-time decision-making for referrals and follow-ups. Not all clients that sought dental care were referred directly to dental services. By co-addressing all health and social needs with the CGA, the appropriate first referral could be made.

The CGA captured clients’ demographics, social factors, healthcare utilization (including recent primary care visits, ED visits and hospitalizations), fall history, nutrition (hunger), health status (including medical conditions, and oral health and mental health status), medications, functional status, and quality of life (Table 1). In addition to standard medical history (conditions, medications, allergies), we included validated instruments for oral health, functional status, mental health and social determinants.

We developed a new Oral Health Assessment (OHA) by drawing from a repository of existing oral health measures from the National Health and Nutrition Examination Survey [11, 36‐38]. We queried a range of multiple dental symptoms, treatment needs, hygiene behaviors, self-reported oral health status, access barriers, and utilization patterns. There were 14 symptoms on a checklist, that ranged from mild to major concerns. These symptoms included experiencing dry mouth when eating or sleeping, issues with too much or too little saliva, bad breath or taste in mouth, sores in the mouth, bleeding gums, pain and difficulties with tasting, swallowing or chewing. Pain and functional limitations from chewing difficulty were examined separately, as they were considered more severe and urgent problems needing attention. The number of positive responses were summed, and a total score tabulated, with a possible range of 0–14. The symptom variable reflected a count of the total number of symptoms currently experienced by the client, per self-report. We explored different ways to summarize the information from the symptom checklist as there is no standard scoring, and conducted sensitivity analyses and consulted with the dental providers on the care team. Clients in this cohort scored between 0 and 9. We ultimately found that four or more reported symptoms appeared to be the most sensitive threshold, and these individuals were categorized as experiencing many dental symptoms and having greater needs.

Validated instruments for quality of life (Brief Older People’s Quality of Life Questionnaire, OPQoL-Brief) [39], depressive symptoms (Patient Health Questionnaire, PHQ-9) [40], and functional status (Vulnerable Elders Survey, VES-13) [41] served as a backbone to query psychosocial status and function. The PHQ-9 and VES-13 are widely validated measures, scored based on standard cutoff values accepted in the literature [39‐42]. PHQ-9 scores greater than 5 were categorized to indicate presence of moderate to high risk for depressive symptoms (possible PHQ-9 full scale range: 0–27). VES-13 scores of 3 or higher were categorized as moderate functional limitation (possible VES-13 full scale range: 0–10). OPQOL-B did not have standard cutoff scoring, and the full scale ranges 13–65. For this study, after sensitivity analyses, we categorized the lowest quartile (those in the sample scoring 26–47) as having low QoL.

Cohort characteristics are shown in Table 2. Clients’ average age was 72, and 51% were men. Most clients experienced general pain (77%), 51% reported on average 2 or more medical conditions (range = 0–7), 5 prescribed medications (range = 1–20), 2 or more medical symptoms (range = 0–11) and 43% reported 4 or more dental symptoms (range = 0–9). About half (54%) were medically complex (by triage metrics) and reported more than five medications and conditions combined (including diabetes, hypertension, arthritis, and hyperlipidemia). More than two-thirds (68%) had either acute or complex dental disease, and 39% reported toothache. Half (50%) had difficulty chewing, and 14% had difficulty swallowing. Most reported good QoL (OPQoL-B = 50.2, range = 26–65), moderate functional limitation (VES-13 = 1.7, range = 0–10), and low to moderate depressive symptom scores (PHQ-9 = 3.3, range = 0–24). A significant proportion of clients were at high risk for malnutrition and food insecurity (46%). Almost half (49%) reported vision and memory problems (46%). Of those with a significant mental health diagnosis (25%), 41% reported no active treatment. About one-third (36%) reported isolation or loneliness.

We captured all CGA results in a prototype, shared health record which allowed data intake, metrics conversion and sharing of data with all members of the care teams. The care navigator recorded intake data and reviewed results and metrics for errors. Twenty-two metrics covering health and wellness, dental, and psychosocial needs were displayed as color-coded lights based on high (red), moderate (yellow), and low (green) risk categories (Fig. 1a). On the first pass, the care navigator could view a client list with a summary of four metrics corresponding to services referrals – medical complexity (care coordination), oral health status (dental care), depression risk/suicidality (mental health), and case management (social work) – to quickly locate key patient needs and match with available care (Fig. 1b). The care navigator then made a second pass to consider other pertinent data and metrics in the record and synthesize these for referral decisions.

Four primary referral metrics were calculated based on composite scores of multiple measures: (a) case management included “homelessness,” “lack of insurance,” “lack of Primary Care Provider (PCP),” and being “low-income” (which was defined as earning ≤199% of the U.S. federal poverty level [FPL]); (b) medical complexity included “number of medical conditions” and “number of medications,” with both defined continuously as counts, with medical conditions categorized as two or more, which was indicative of being a higher risk client with care coordination needs; (c) oral health status was comprised of: “self-rated oral health status,” “missing teeth”, experiencing more than four current dental symptoms (from a list of 14 queried), and functional impairment, including difficulty chewing and food limitation; and (d) depressive symptomology was determined by PHQ-9, with additional weight placed on suicidality (Item 9) for triage purposes.

Clients were referred to services based on the care navigator’s review of extracted metrics and key data such as symptoms, medications, and medical conditions. We designed the metrics to accelerate data review only, and during this pilot phase, referral was ultimately based on the care navigator’s judgement and client’s preference.

Staff used an Acute Clinical Screen (ACS) to assess clients with flagged medical and dental symptoms (e.g., chest pain, jaw pain, shortness of breath, bleeding, fever), and triaged them to 911, urgent medical care, or immediate dental services. Care navigators gave non-acute/well patients a brief take-home questionnaire (Pre-CGA) that included contact and basic medical information (e.g., allergies, medications, problem list, insurance, and medical contacts). Patients completed the Pre-CGA paper forms prior to the first appointment. The care navigator collected and entered clients’ Pre-CGA information, completed the assessment, and referred them to services in the same appointment. Clients with one yellow or red light were typically referred only to that service unless the client requested other services. Clients with 2 or more red lights were referred for simultaneous care, based on their most urgent need. Clients with 1 or more yellow lights were referred based on urgency and navigator’s judgement, and clients with 4 green lights seeking dental care were assigned to dental “usual-care” (described in the next section).

Clients referred to care coordination were assessed by a registered nurse for active medical symptoms, conditions, and establishment of an appropriate medical home. The nurse also provided clients with health education, nursing care, and helped them create personal health action-plans. For clients referred to case management, social workers provided assistance with housing, legal matters, insurance eligibility, and individual and group counseling. Clients referred to mental health received acute assessments (e.g., for suicidality), acute therapy, and medication reconciliation from either the center’s psychiatric nurse or an outside provider.

Non-acute clients referred to dental care participated in a mandatory hour-long oral health education (OHE) course, (developed by Oral Health America; which has now ceased operations, but oral health resources for older adults are still available at https://​www.​toothwisdom.​org/​) followed by a Basic Screening Survey (BSS) both administered by a registered dental hygienist (RDH) prior to the client’s first dental appointment. The OHE covered topics on older adults’ oral health and common chronic conditions and provided hygiene demonstrations. We used the BSS, a brief visual exam conducted by RDH, to determine dental urgency and needs as “ground-truth,” compared to self-reported data, and to prioritize scheduling clients for dental care. We followed the standard BSS procedures for screening older adults, using the standard American Association of State and Territorial Dental Directors (ASTDD) Toolkit for Older Adults [43]. Based on the BSS, the RDH determines if a client should be seen urgently (immediately/as soon as possible, or within the next 2 weeks), soon (within weeks), or no obvious problem or dental need, and the client can continue their regular care pattern for their next check-up.

Of 1335 clients engaged, 1012 completed a pre-CGA and 996 completed a CGA without missing responses. Of 1335 clients, 22 reported acute conditions (e.g. active chest pain, fever, severe jaw pain) and completed the ACS. Of these, 9 were referred to PCP/911 and 13 directly to urgent dental care (Table 3). CGA average time to completion was 52 min (range = 34–68). There were no significant differences in response or completion time between English, Spanish, and Mandarin speakers or written versus digital format. Of the clients completing the CGA, 158 declined services or were lost to follow-up and 52 new clients required only nutrition services.

Care navigators made 1210 referrals for 996 clients (dental = 905, care coordination =172, social work = 100, mental health = 33). Of those, 701 needed only a single referral, and 225 received multiple referrals (maximum 4 referrals). The majority of single referrals (683; 98%) were for dental services. Nine clients were referred only to care coordination, 6 to case management and 2 for mental health services. Two hundred twenty clients were referred to dental and at least one more service.

There were no instances of acute medical or psychiatric issues during dental care. Almost all clients referred to dental services completed the OHE and BSS segments, with no-show rates of < 1%. Total no-show rate for dental care was 5.7%. Table 3 summarizes the distribution of services received for each referral (and see Supplemental Figure 1 for a diagram of these referral pathways and outcomes). For dental care, the most common services were preventative (598), periodontics (440) and prosthetics/dentures (217). The most common care coordination services were for medical complexity or management of specific symptoms (59), connection to medical home (46) and enrollment in external services such as specialty dental or medical care (41). Case management services were primarily for health insurance enrollment (32), housing (30) and income assistance (9). Mental health services included linkage to care (11) or counseling or support at Serving Seniors (9). Eight clients were referred to mental health for behavioral or non-compliance issues with the dental center.

Correlation analysis of CGA data (Fig. 2) showed significant relationships between the CGA dimensions (all p-values < 0.05). Pain was strongly negatively correlated with general health (− 0.49) and QoL (− 0.32), and positively correlated with depression risk (0.41) and lower functional status (0.43). Self-reported general health was the strongest positive correlate of QoL (0.47) and lower general health was strongly correlated with lower functional status (.53). Number of medical conditions was significantly associated with lower self-reported health (− 0.34) and QoL (− 0.18), correlated with depression (0.26), and strongly correlated with functional limitation (0.35). Medical symptoms were negatively correlated with general heath (− 0.41) and QoL (− 0.26) and associated with depressive symptoms (0.54), low functional status (0.43), and fall risk (0.30). Dental symptoms were negatively correlated with self-rated health (− 0.30) and QoL (− 0.24) and strongly correlated with depressive symptoms (0.43). Dental symptoms were significantly correlated with dental pain (0.54), food limitation (0.21), and difficulty chewing (0.59). Finally, Isolation was associated with medical (0.32) and dental (0.23) symptoms and functional limitation (0.19), strongly correlated with poor general health (− 0.29), pain (0.35) and depressive symptoms (0.43), and was the strongest correlate of low QoL (− 0.50).

Table 4 summarizes the adjusted odds ratios (AOR) and 95% confidence intervals (CI) for recent ED visits, hospitalizations, and dental urgency. ED visits were most strongly associated with homelessness (AOR = 3.18, 95% CI [1.59, 6.39]), lack of PCP (AOR = 2.43, 95% CI [1.00, 5.90]), recent falls (AOR = 2.88, 95% CI [2.01, 4.15]) and key specific medical conditions such as chronic obstructive pulmonary disease (COPD) (AOR = 3.05, 95% CI [1.45, 6.44]) and congestive heart failure (CHF) (AOR = 8.86, 95% CI [2.39, 32.80]). Those with moderate to severe depressive symptoms (AOR = 2.36, 95% CI [1.59, 3.50]) or lower QoL (AOR = 1.96, 95% CI [1.32, 2.88]) were approximately 2 times more likely to have had an ED visit within 6 months.

For recent hospitalizations, the most strongly associated with homelessness (AOR = 2.61, 95% CI [1.48, 4.62]), lack of PCP (AOR = 3.04, 95% CI [1.09, 8.51]), poor self-rated general health (AOR = 2.52, 95% CI [1.76, 3.59]), hypertension (AOR = 2.04, 95% CI [1.43, 2.90]), COPD (AOR = 5.05, 95% CI [2.70, 9.42]), CHF (AOR = 6.94, 95% CI [2.61, 18.46]), low functional status (AOR = 2.21, 95% CI [1.53, 3.21]), active depression (AOR = 1.97, 95% CI [1.34, 2.89]), and low QoL (AOR = 2.42, 95% CI [1.67, 3.51]).

Dental urgency was most strongly associated with self-rated general health (AOR = 1.78, 95% CI [1.31, 2.43]), the presence of 4 or more dental symptoms (AOR = 2.39, 95% CI [1.78, 3.20]), the presence of dental pain (AOR = 2.06, 95% CI [1.55–2.74]), and difficulty chewing (AOR = 2.80, 95% CI [2.09–3.76]). Those with more than 4 dental symptoms were also more likely to have had recent ED visits (AOR 1.61, 95% CI [1.12–2.30]) or hospitalizations (AOR 1.47, 95% CI [1.04–2.1]).