Geriatric Resources for Assessment and Care of Elders (GRACE): A New Model of Primary Care for Low-Income Seniors

Steven R. Counsell, MD; Christopher M. Callahan, MD; Amna B. Buttar, MD, MS; Daniel O. Clark, PhD; Kathryn I. Frank, RN, DNS


J Am Geriatr Soc. 2006;54(7):1136-1141. 

In This Article

Description of the Program

The GRACE intervention includes a nurse practitioner and a social worker who care for low-income seniors in collaboration with the patient's PCP and a geriatrics interdisciplinary team. The foundational principles for the GRACE intervention mirror recommendations in recent reviews of best practices for care of chronic conditions, including specific targeting of older people at risk, availability of collaborative expertise in geriatrics, integration of the program into primary care, coordination of care across all sites of care, the use of an electronic medical record to support physician practices and facilitate monitoring of clinical parameters, and institutionally endorsed clinical practice guidelines.[4,5,23,24]

The catalyst for the GRACE intervention is the GRACE support team, consisting of a certified nurse practitioner and a licensed clinical social worker with clinical experience in geriatrics and employed by the primary care practice. The team completes special training in implementing the GRACE protocols and working as an interdisciplinary team during 12 weekly small group seminars.

The GRACE support team meets with the patient (and their family when possible) in the patient's home. The assessment includes a medical and psychosocial history, medication review, functional assessment, and review of social supports and advance directives. The examination gives special attention to orthostatic vital signs, vision, hearing, gait and balance, affect, and mental status. A home safety evaluation is also performed. Following this in-home assessment, the support team meets with the larger GRACE interdisciplinary team to develop an individualized care plan (including activation of GRACE protocols) and a priority for sequencing of team suggestions. The GRACE protocols were developed with input from local opinion leaders. The GRACE interdisciplinary team includes a geriatrician, pharmacist, physical therapist, mental health social worker, and community-based services liaison from the same primary care practice and affiliated health system.

Prepared with this care plan, the GRACE support team meets face to face with the patient's PCP to review a computer-generated summary of the patient assessment and team suggestions. Typically one to three patients are discussed for about 5 minutes each to modify, prioritize, and gain PCP approval of the plan of care. Although not directly compensated for this time, physicians participating in the study are credited for these meetings in their annual performance review. Collaborating with the PCP, the GRACE support team then implements the care plan consistent with the patient's healthcare goals and starting with an in-home follow-up visit. With the support of an electronic medical record and longitudinal tracking system that monitors episodes and outcomes of care, the GRACE team provides ongoing care management and coordination of care across multiple geriatric syndromes, providers, and sites of care.

The nurse practitioner and social worker implement the care plan and coordinate care between providers and settings through face-to-face and telephone contacts with patients, family members/caregivers, and providers. All patients receive an annual reassessment and follow-up visit. The number, timing, and content of additional patient contacts occurs as appropriate to implement the care plan. Each patient receives a minimum of one phone contact per month. In addition, a face-to-face home visit occurs after any emergency department visit or hospitalization. Monthly contacts provide a mechanism for proactive follow-up and patient assistance and a check for new problems, including medication changes or changes in social supports or living arrangements. Patients also contact their GRACE team as needed via a dedicated phone line. Teams encourage goal setting and self-care, teach problem-solving skills, provide education using low-health-literacy materials corresponding to each GRACE protocol, prepare patients and physicians to address problems and team suggestions during office visits, and assist with transportation arrangements.

The GRACE support team also makes contacts with other providers for coordination of care. For example, the GRACE team facilitates consultation by the ACE inpatient team whenever a patient is hospitalized. Providers in the primary care and specialty clinics, emergency department, and hospital receive automated prompts via the electronic medical record system to contact the GRACE support team for information and assistance with follow-up and coordination of care.

Regularly scheduled GRACE interdisciplinary team reviews are held for each patient at 3 and 6 weeks and 3, 6, and 9 months after the initial and annual care-planning meetings. Additional team reviews are held for a particular patient following a major change in status, emergency department visit, or hospitalization. A key feature of these meetings is to review the GRACE support team's success in implementing the care protocols and in improving care (e.g., depression severity scores).

An expert panel has identified a set of geriatric conditions representing optimal targets for quality improvement.[25] In partnership with several PCPs and opinion leaders working within the public health system, 12 of these conditions were chosen to target in the GRACE intervention: advance care planning, health maintenance, medication management, difficulty walking/falls, chronic pain, urinary incontinence, depression, malnutrition/weight loss, visual impairment, hearing loss, dementia, and caregiver burden. For each condition, GRACE protocols specify recommendations for evaluation and management. These recommendations are based on published practice guidelines and input from local primary care providers. Each patient enrolled in the GRACE program is proactively assessed for these 12 conditions, and when present, the corresponding GRACE protocol is activated along with specific team suggestions. The Advance Care Planning and Health Maintenance protocols are implemented in all patients, whereas the other protocols are only implemented when appropriate. Although disease management for specific comorbidities such as diabetes mellitus and congestive heart failure is not the primary focus of the GRACE intervention, the GRACE support team notes these conditions on the health maintenance protocol, and team suggestions are made when appropriate.

GRACE protocols are available upon request from the corresponding author. As an example, the difficulty walking/falls protocol contains 31 total suggestions for further evaluation, management, consultation, and patient education, including 16 intended for review with PCP before implementation and 15 routine team interventions. Review with PCP suggestions for difficulty walking/falls include confirm diagnosis and contributing causes and update medical record; evaluate and treat causes; order laboratory evaluation, including complete blood count, comprehensive metabolic panel, thyroid-stimulating hormone, and B12 levels; optimize pain management; and consider physical therapy consultation. Routine team interventions include monitor orthostatic vital signs, encourage increased fluid intake, recommend exercise program for balance and strengthening, recommend walking program, and provide patient education on falls prevention.

A fundamental component of the GRACE intervention is improving access and care coordination for the patient's entire constellation of problems rather than targeting patients with a single disease. In addition, each of the targeted areas above requires attention to patient education, help in improving patient self-management, and aid in linking the patient with existing health-system and community-based services. The innovation of GRACE is the integration of care across multiple chronic conditions and across the continuum of care.

Patients were recruited from a large primary care practice affiliated with Wishard Health Services, a university-affiliated urban healthcare system serving medically indigent patients in Indianapolis, Indiana. This practice includes seven community-based health centers providing primary care to approximately 6,000 adults aged 65 and older. Faculty from the Indiana University School of Medicine staff each of these health centers. All PCPs working in this health system use the Regenstrief electronic medical record system to record diagnoses, order diagnostic tests and medications, review diagnoses and orders of other physicians involved in the patient's care, review hospital discharge summaries, review results of diagnostic tests, schedule new appointments, and assign billing codes. Thus, PCPs interact with the electronic medical record whenever they are involved in patient care in the hospital, emergency department, or outpatient clinics.

At the time of implementation of the GRACE intervention, the following geriatric clinical services existed at Wishard in support of the primary care practice: outpatient geriatric assessment and multispecialty clinic, inpatient ACE unit, skilled nursing facility, and physician house calls program. Psychiatric care was available through the health systems' community mental health center.

Research assistants screened patients with scheduled appointments at the targeted primary care practices to determine eligibility. Patients were eligible for enrollment if they were aged 65 and older, had an annual household income below 200% of the federal poverty level (e.g., <$25,000 for family unit of two persons), had one or more primary care visits in the past 12 months, resided in the community, and had access to a telephone. Excluded from the study were patients unable to understand English and those who had severe hearing impairment, were on dialysis or enrolled in another research study, or had severe cognitive impairment without an available caregiver to consent to participate. The Indiana University institutional review board approved the study, and participants provided written informed consent.

Of 332 subjects enrolled in the GRACE intervention from January 2002 through June 2003; 254 have been followed for 2 years and are described in this article. The 78 intervention subjects who did not complete 2 years included 28 withdrawals (most of whom refused home visits or telephone interviews), 28 deaths, nine who moved out of the metropolitan area, eight who were placed in nursing homes, and five lost to follow-up.

Characteristics of the 254 GRACE intervention patients are provided in Table 1 . The mean age was 72 (range 65–92). Approximately 95% reported having Medicare, and 37% reported having Medicare and Medicaid (dually eligible). The mean number of physician visits in the 12 months before enrollment was 4.4 (range 1–31).

The GRACE interdisciplinary team meeting occurred within 30 days of enrollment in 85% of patients (mean 24.3 days; range 4–162 days). Of the 12 GRACE protocols, a mean of 5.3 were activated per patient in Year 1 (range 2–10) and Year 2 (range 2–11). The activation frequency of each GRACE protocol is presented in Table 2 . Most of the protocols used for an individual patient in Year 1 were selected again in Year 2.

Of 275 total possible team suggestions contained in the 12 protocols, a mean of 62.6 (range 33–131) were selected per patient in Year 1 and 33.6 (range 7–84) in Year 2. Adherence to GRACE interdisciplinary team suggestions was high in both years (81% in Year 1 and 79% in Year 2). Of the suggestions selected but not completed, 73% were not completed because of patient disagreement and 4.7% because of physician disagreement, as judged by the GRACE support team. The remainder of selected suggestions were not completed for logistical reasons (e.g., missed appointment), because the service was not available, or because the suggestion became "no longer applicable." The mean number of physician visits reported by GRACE patients in Year 1 was 4.8 (median 4; range 0–59) and in Year 2 was 4.6 (median 4; range 0–25). Table 3 shows the number of additional patient contacts made by the GRACE support team. Team face-to-face contacts were usually home visits and only occasionally in the office, hospital, or nursing home. Two-thirds of the team's contacts with the patient occurred in the first half of each year. There were 22 combined physician visits and team patient contacts in Year 1 and 21 in Year 2. Support team contacts with healthcare providers (per patient) for coordination of care in Year 1 and Year 2, respectively, were as follows: primary care, 2.83 (range 0–14), 2.72 (range 0–11); specialty care, 1.13 (range 0–18), 0.84 (range 0–21); emergency department, 0.21 (range 0–5), 0.13 (range 0–2); hospital, 0.28 (range 0–6), 0.16 (range 0–10); nursing facility, 0.10 (range 0–11), 0.04 (range 0–5); home health care, 0.28 (range 0–11), 0.15 (range 0–9); and total, 7.92 (range 0–68), 6.63 (range 0–79). The mean number of extra GRACE interdisciplinary team reviews per patient in Year 1 and Year 2 was 0.51 (range 0–5) and 0.22 (range 0–4), respectively.

A brief survey was developed to determine physician satisfaction with the GRACE intervention. Surveys were sent to PCPs asking them to compare their satisfaction with the resources available in their practice to treat elderly patients who are enrolled in the GRACE intervention with their satisfaction with the resources available to treat patients who are not enrolled in the intervention. Providers were surveyed using standard mail survey techniques with three follow-up mailings. Likert scales were used to rate satisfaction (range 1–5; 5=very satisfied), helpfulness (range 1–5; 5=very helpful), and amount of care (range 1–3; 1=not enough, 2=too much, and 3=just right).

Survey response rate was 85%, with 100% of 21 faculty and 79% of 58 residents responding. Physicians were much more satisfied with the resources available to treat patients in the GRACE program (mean rating 4.28 vs 2.98; P <.001 by paired t test; n=49). There were no significant differences in this outcome by physician sex, faculty versus resident, or number of patients seen. In addition, the GRACE intervention was rated as somewhat to very helpful in providing care to older patients (mean 4.52). The amount of care provided by the GRACE nurse practitioner and social worker was rated as just right (mean 2.96), implying that the supplemental care provided by the GRACE support team was neither insufficient to meet patient needs nor excessive or duplicative.

Implementation costs included patient recruitment and training of the nurse practitioners, social workers, support staff, PCPs, and health center staff. Estimating an eventual caseload of 125 patients per GRACE support team, two nurse practitioner/social worker teams were hired for start-up, and approximately five patients were enrolled per week. A third team was hired after approximately 200 patients were enrolled in the intervention. In steady-state and to maintain a census of approximately 250 patients, the program would need to employ two full-time equivalent (FTE) nurse practitioners, two social workers, one administrative assistant, and 0.2 FTEs each of a geriatrician, pharmacist, physical therapist, mental health social worker, community-based services liaison, and practice manager. Including salary and benefits for personnel listed above, mileage reimbursement, pager and cell phone costs, home visit bags, and office supplies, the estimated direct cost of providing the GRACE intervention was $1,000 per patient per year. Approximately 10% of the total cost of the intervention was covered under current Medicare reimbursement. GRACE intervention costs per patient are likely to be less without the added costs attributable to the clinical trial requirements.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.