Unmet Needs of Caregivers of Patients Referred to a Dementia Care Program

Abstract

Background/Objectives

Caregiver strain and low self-efficacy for managing dementia-related problems are common among those caring for patients with dementia, but the level of unmet need and relation to provider type has not been well characterized.

Design

Cross-sectional observational cohort

Setting

Urban academic medical center

Participants

Caregivers of community-dwelling adults with dementia referred to the program

Measurements

Caregivers were surveyed and completed the Patient Health Questionnaire (PHQ-9) about themselves, the Modified Caregiver Strain Index, the Neuropsychiatric Inventory Questionnaire, which measures patient symptom severity and related caregiver distress, and a 9-item caregiver self-efficacy scale developed for the study.

Results

Of 307 patient and caregiver dyads surveyed over a one year period, 32% of caregivers reported confidence in managing dementia-related problems, 19% knew how to access community services to help provide care, and 28% agreed the patient’s provider helped them work through dementia care problems. Thirty-eight percent reported high levels of caregiver strain, and 15% reported moderate to severe depressive symptoms. Caregivers of patients referred by geriatricians more often reported having a healthcare professional to help work through dementia care problems than those referred by internists, family physicians, or other specialists, but self-efficacy did not differ. Low caregiver self-efficacy was associated with higher caregiver strain, more caregiver depressive symptoms, and caring for a patient with more severe behavioral symptoms.

Conclusion

Most caregivers perceived inadequate support from the patient’s provider in managing dementia-related problems, reported strain, and had low confidence in managing caregiving. New models of care are needed to address the complex care needs of patients with dementia and their caregivers.

INTRODUCTION

In 2012, 5.2 million Americans were estimated to have Alzheimer’s disease, and 15.4 million Americans provided an estimated 17.5 billion hours of unpaid care for people living with dementia.¹ Caring for a person with dementia poses several challenges. Caregivers must provide increasing supervision and personal care as the person with dementia becomes more functionally dependent. Caregivers must learn to manage behavioral disturbances, personality changes, and the loss of the ability to communicate effectively with their loved one. As dementia progresses, caregivers experience increasing strain and burnout, depression, disruptions in employment, and depleted finances.¹

High levels of caregiver strain are associated with worse outcomes for patients with dementia, including higher rates of nursing home placement.^2–4 Moreover, high levels of strain also result in poor health outcomes for caregivers, including depression,⁵ increased risk of cardiovascular disease,^6,7 and perhaps increased risk of mortality among those caregivers with the highest levels of stress.^8,9 Low self-efficacy for dementia caregiving, specifically low perceived ability to manage behavioral symptoms or access community resources, has been associated with more depressive symptoms and lower health-related quality of life among family caregivers.^10–12

Interventions developed to decrease the negative aspects of caregiving have focused on training caregivers on strategies to manage behavioral symptoms, enhancing caregiver social support, and providing respite. Multi-component interventions that longitudinally engage multiple family members to provide ongoing assistance as the disease progresses have been most effective at reducing caregiver strain.^13–16 However, busy physicians have neither the time nor, in some cases, the skills to adequately implement interventions to reduce caregiver strain and improve self-efficacy. In primary care settings, fewer than half of dementia caregivers receive counseling about managing behavioral symptoms or accessing caregiver support,^17–19 and caregiver mental health is often insufficiently addressed.⁵ Dementia care is also provided by specialists, including geriatricians, neurologists, and psychiatrists, and less is known about how caregiver strain, depressive symptoms, and self-efficacy may differ by provider type. Community resources (e.g., the Alzheimer’s Association), can improve the quality of dementia care by providing education and support for caregivers; however, these organizations are underutilized.¹⁹

In response to these needs, in July 2012, UCLA launched the Alzheimer’s and Dementia Care Program (UCLA ADC), a dementia management program that provides comprehensive, coordinated, patient-centered care for patients with dementia and their families. The program uses a co-management model with a nurse practitioner dementia care manager partnering with primary care physicians and five local community-based organizations.²⁰

In order to better understand the needs of caregivers on entry into the program, we conducted a baseline evaluation of caregiver reports of their perceived ability to manage dementia-related problems, perceptions of the patient’s dementia care, caregiver mental health and strain, and the relationship of these to the type of referring provider. In this paper, we describe the unmet caregiver needs for the first 307 patients enrolled in the program.

METHODS

We evaluated all patients enrolled in the UCLA ADC, which required that patients be community-dwelling, have a diagnosis of dementia of any type, and be referred by a UCLA provider. UCLA is an urban, academic health system in west Los Angeles. Patients’ primary caregivers were surveyed as part of the clinical data collection of the UCLA ADC Program. The UCLA Office of the Human Research Protection Program determined that this quality improvement evaluation did not require IRB review.

The primary caregivers of the first 307 patients enrolled in the Program completed the 9-item Patient Health Questionnaire (PHQ-9)²¹ and the 13-item Modified Caregiver Strain Index (MCSI)²² about themselves, and the 12-item self-administration version of the Neuropsychiatric Inventory Questionnaire (NPI-Q)²³ about the patient. Prior to the intake visit, 227 caregivers also completed a 9-item survey of self-efficacy for managing dementia-related problems and accessing help. This survey (developed for the UCLA ADC evaluation) was added to the pre-visit assessment three months after the start of the program, and thus, only 254 caregivers were eligible for the survey.

The PHQ-9 is a validated tool used to assess depressive symptoms using the DSM-IV criteria for major depression and is scored from 0–27 with scores greater than 10 indicating moderate depressive symptoms and scores greater than 20 indicating severe depressive symptoms.²¹ The MCSI is a validated tool used to assess severity of caregiver strain. The index targets financial, physical, psychological, and social aspects of strain and is scored from 0 to 26 with higher scores indicating greater levels of strain.²² The NPI-Q is a validated survey that assesses the caregiver’s perception of the severity of 12 dementia-related psychiatric and behavioral symptoms and the level of distress experienced by the caregiver in response to these symptoms. NPI-Q severity score ranges from 0–36, and NPI-Q distress score ranges from 0–60 with higher scores indicating more severe symptoms and distress, respectively.²³

Patient cognitive status was measured using the Mini Mental Status Exam (MMSE),²⁴ which ranges from 0 to 30 with lower scores indicating greater impairment, and patient functional status was measured using the Functional Activities Questionnaire (FAQ),²⁵ which ranges from 0 to 30 with higher scores indicating more functional dependence.

We developed a 9-item pre-visit caregiver survey to assess three domains: 1) caregiver perception of the primary care provider’s recognition of the patient’s cognitive impairment and its impact on health conditions, 2) whether the caregiver had received advice on dementia-related topics, and 3) the caregiver’s self-efficacy for caring for the patient with dementia and for accessing help, including community resources. Response options were on a five-point rating scale with 1=strongly disagree and 5=strongly agree. Six items were adapted from prior surveys,^10,26,27 including 3 items from the Fortinsky scale for dementia-caregiving self-efficacy.¹⁰ Cognitive interviews were performed with ten caregivers of patients with dementia to evaluate the comprehension, clarity, and timing of the questions.

Statistical Analysis

Frequencies or means and standard deviations were used to examine demographic characteristics among dementia patients and their caregivers, including age, gender, race and ethnicity, insurance status, type of dementia, functional and cognitive status, whether the patient lived with his or her caregiver, the caregiver’s relationship to the patient, and referring provider.

Cronbach’s alpha was calculated for the responses to the four survey items addressing caregiver self-efficacy for managing dementia-related problems and accessing services (items 4–7 in Table 2). A self-efficacy score was calculated for each caregiver as the average of the responses to these four survey items within the self-efficacy domain. The self-efficacy score ranged from 1 to 5 with higher scores indicating greater self-efficacy. Responses to the pre-visit caregiver survey were then collapsed into 3 categories (strongly agree or agree, neutral, or disagree or strongly disagree) for ease of presentation.

Table 2.

Caregiver Dementia Care and Self-efficacy Survey Responses, N=227

Survey item	Agree or strongly agree N (%)	Neutral N (%)	Disagree or strongly disagree N (%)
Domain: Experience with advice on dementia-related topics
1. I have received advice about what problems to expect in the future related to Alzheimer's or dementia.27	79 (35%)	50 (22%)	96 (43%)
2. I have received advice about handling problems like the patient's memory loss, wandering or behavior problems.27	87 (39%)	52 (23%)	83 (38%)
3. I am aware of services available to me to help me provide care.	57 (25%)	74 (33%)	95 (42%)
Domain: Self-efficacy for caring for patient with dementia and accessing help
4. I know how to get community services that will help me provide care.	43 (19%)	54 (24%)	130 (57%)
5. I feel confident that I can handle problems like the patient's memory loss, wandering or behavior problems.10	73 (32%)	68 (30%)	86 (38%)
6. I feel confident that I can deal with the frustrations of caregiving.10	78 (35%)	60 (27%)	88 (39%)
7. I know where to turn to get answers to questions about problems like the patient's memory loss, wandering, or behavior problems.10	82 (36%)	51 (23%)	92 (41%)
Domain: Perception of patient’s dementia care prior to program entry
8. The patient's regular doctor understands how memory or behavior problems complicate other health conditions.26	178	36 (16%)	12 (5%)
9. I have a healthcare professional who helps me work through dementia care problems.	63 (28%)	47 (21%)	112 (51%)

Caregiver self-efficacy survey responses by type of referring provider were compared using chi-square tests. One-way ANOVA or Kruskal Wallis tests for variables with non-normal distributions were used to compare self-efficacy score, MCSI, PHQ-9, and NPI-Q severity score by type of referring provider.

Two-sample t-tests with equal variances or Wilcoxon-Mann-Whitney tests (for continuous variables) or chi-square tests or Fisher exact tests (for categorical variables) were used to compare patient and caregiver characteristics between caregivers with high and low self-efficacy. Low self-efficacy was defined as a self-efficacy score of less than three. Two-sample t-tests or chi-square tests were also used to compare patient and caregiver characteristics between caregivers with high and low levels of strain (MCSI<13). The midpoint of the MCSI was chosen as a cut point to define high and low levels of strain based on the definition of a positive screen for the original Caregiver Strain Index²⁸ and use of the scale in prior studies.^29,30 All analyses were cross-sectional and were performed using STATA IC version 13.³¹

RESULTS

Over twelve months, 307 patient and caregiver dyads were evaluated as part of the baseline assessment of the dementia care program. The mean age of patients in the study was 82 years; 63% were female; 25% were racial or ethnic minorities; and 10% were Medicare and Medicaid dually insured. The majority of patients (90%) had Alzheimer’s, mixed vascular and Alzheimer’s dementia, or an unspecified type of dementia; the mean MMSE score was 16.2; and the mean FAQ score was 22 on a scale where 0 is complete functionality and 30 indicates maximum functional dependence. Fifty-seven percent of patients were referred to the program by a geriatrician, 23% by an internist or family medicine physician, and 19% by neurologists. (Table 1)

Table 1.

Patient and Caregiver Characteristics, N=307

Patients	N (%) or M (SD)
Patient age, years	81.8 (8.6)
Patient female	193 (63%)
Patient race/ethnicity
White, non-Hispanic	230 (75%)
Black, non-Hispanic	36 (12%)
Hispanic	22 (7%)
Asian or Pacific Islander	19 (6%)
Medicare/Medicaid dually insured	31 (10%)
Type of dementia
Alzheimer’s	138 (45%)
Mixed or unspecified	138 (45%)
Vascular	13 (4%)
Lewy Body	11 (4%)
Frontotemporal	6 (2%)
Parkinson’s	1 (0.3%)
Mini Mental State Examination (MMSE) score, N=297 (range 0–30)	16.2 (7.7)
Functional Assessment Questionnaire score, N=305 (range 0–30)	22 (7.5)
Neuropsychiatric Inventory Questionnaire (NPI-Q) symptom severity score, N=301 (range 0–36)	11.1 (7.1)
Neuropsychiatric Inventory Questionnaire (NPI-Q) caregiver distress score, N=301 (range 0–60)	13.9 (10.1)
Referring provider
Internal Medicine or Family Practice	71 (23%)
Geriatrics	174 (57%)
Neurology	59 (19%)
Psychiatry	3 (1%)
Caregivers
Caregiver female	206 (67%)
Caregiver lives with patient	198 (65%)
Caregiver relationship to patient
Spouse/partner	125 (41%)
Child/child-in-law	149 (49%)
Other	33 (11%)
Modified Caregiver Strain Index score, N=299 (range 0–26)	11.5 (7.0)
--Score >13	115 (38%)
Caregiver PHQ-9 score, N=298 (range 0–27)	4.7 (4.7)
--Score>9	44 (15%)
Self-efficacy score, N=227 (range 0–5)	2.8 (0.98)
--Scores 1.0–2.9	115 (38%)

Most caregivers were children (49%) or spouses (41%), female (67%), and lived with the patient (65%). The mean caregiver PHQ-9 score was 4.7, reflecting mild depressive symptoms among caregivers on average; however, 44 caregivers (15%) had PHQ-9 scores of 10 or greater consistent with moderate or severe depressive symptoms. The mean score on the MCSI was 11.5, reflecting moderate strain, with 115 caregivers (38%) reporting high levels of strain with scores greater than 13 (index midpoint). (Table 1)

On the NPI-Q, caregivers reported a mean patient symptom severity score of 11.1 on a scale of 0–36 and a mean distress score of 13.9 on a scale of 0–60. Agitation or aggression was the most commonly reported behavioral symptom with 70% of caregivers reporting it as present in the last month. Over 60% of caregivers reported depression/dysphoria, anxiety, apathy/indifference, irritability/lability, or nighttime behaviors as present in the last month. Apathy (17%) and nighttime behaviors (22%) were the two behaviors most commonly rated as severe by caregivers, and agitation or aggression (18%) and nighttime behaviors (16%) were the behaviors most commonly reported as very or extremely distressing to caregivers, causing difficulty with coping or an inability to cope.

Internal consistency in this sample, measured using Cronbach’s alpha, was high for all scales (PHQ-9 0.87, MCSI 0.90, NPI-Q severity scale 0.80, NPI-Q distress scale 0.82, and FAQ 0.72).

Caregiver Self Efficacy and Perception of Dementia Care

Eighty-nine percent of caregivers (N=227) completed the pre-visit caregiver survey. Patient and caregiver characteristics, including MCSI score, caregiver PHQ-9 score, NPI-Q severity and distress scores, MMSE score, and FAQ score, did not differ between caregivers who completed the pre-visit survey and those who did not (N=80). Thirty-five percent agreed that they had received advice about problems to expect in the future related to dementia, 39% had received advice about managing dementia-related problems, and 25% were aware of services available to help them provide care. Concerning self efficacy, 19% of caregiver respondents knew how to get community services to help them provide care, 32% felt confident handling problems like the patient’s memory loss, wandering, or behavioral problems, and 35% felt confident dealing with the frustrations of caregiving. Although the majority of respondents (79%) agreed that the patient's regular doctor understands how memory or behavior problems complicate other health conditions, only 28% agreed that they have a healthcare professional who helps them work through dementia care problems. (Table 2)

The internal consistency of the four survey questions assessing the caregiver’s self-efficacy for managing dementia-related problems and accessing help (items 4–7 in Table 2) was high (Cronbach’s = 0.84). The mean caregiver self-efficacy score was 2.8 with a range of 1 to 5 with higher scores indicating greater self-efficacy. Low caregiver self-efficacy, defined as a self-efficacy score of <3.0, was associated with higher caregiver strain, more caregiver depressive symptoms, and more severe behavioral symptoms on the NPI-Q. (Table 3)

Table 3.

Caregiver and Patient Characteristics by Level of Caregiver Self-Efficacy

	Caregivers
	Low Self-Efficacy (self-efficacy score <3.0), N=115 N (%) or M (SD)	High Self-Efficacy, N=112 N (%) or M (SD)	p-value
Modified Caregiver Strain Index Score	13.9 (7.2)	9.1 (6.1)	<0.001
Caregiver PHQ-9 score	5.9 (5.2)	3.5 (3.6)	<0.001
NPI-Q severity score	12.4 (7.2)	10.2 (6.7)	0.02
NPI-Q distress score	16.5 (10.9)	11.9 (8.8)	<0.001
Caregiver female	84 (73%)	72 (64%)	0.16
Caregiver lives with patient	76 (66%)	67 (60%)	0.33
Caregiver relationship
Spouse	50 (43%)	39 (35%)	0.29
Child	56 (49%)	59 (53%)
Other	9 (8%)	14 (12%)
Patient age, years	82.4 (7.4)	80.6 (10.3)	0.14
Patient female	67 (58%)	78 (70%)	0.07
Patient racial or ethnic minority	29 (25%)	27 (24%)	0.85
Patient Medicare and Medicaid dually insured	13 (11%)	11 (10%)	0.72
MMSE	18.1 (7.0)	15.1 (7.7)	0.002
Functional Assessment Questionnaire score	22.0 (7.5)	21.3 (8.0)	0.66
Dementia Type
Alzheimers	52 (45%)	60 (54%)	0.22
Mixed or Unspecified	49 (43%)	45 (40%)
Other	14 (12%)	7 (6%)
Referring Provider
Internal Med/Family Practice	30 (26%)	23 (21%)	0.06
Geriatrics	50 (44%)	66 (59%)
Neurology/Psychiatry	35 (30%)	23 (21%)

Continuous variables were compared using two-sample t-tests or Wilcoxon-Mann-Whitney tests for variables with non-normal distributions (PHQ-9 and FAQ). Categorical variables were compared using chi-square tests or Fisher exact tests for variables with small sample size (any cell with N<10).

Relationship of Unmet Caregiver Needs and Referring Provider

There was no difference in MCSI score (F=0.40, p=0.67), caregiver PHQ-9 score (chi2=0.64, p=0.73), caregiver self-efficacy score (F=2.51, p=0.08), or NPI-Q severity score (F=1.21, p=0.30) by type of referring provider. Similarly, there were no differences in responses to the four survey items (items 4–7 in Table 4) addressing caregiver self-efficacy for managing dementia-related problems and accessing services by specialty of referring provider. However, caregivers were more likely to report that they had a healthcare professional who helped them work through dementia care problems if they were referred to the program by a geriatrician (42%) as compared to those referred by an internist or family practitioner (14%) or neurologist or psychiatrist (16%). Caregivers of patients referred by neurologists or psychiatrists also less frequently agreed that: the patient’s regular doctor understood how dementia complicates other health conditions, they had received advice about handling dementia-related problems or about what to expect in the future, and they were aware of services to help provide care. (Table 4)

Table 4.

Caregiver Dementia Care and Self-efficacy Survey Response by Referring Provider, N (%) Agree or Strongly Agree

Survey Item	All Caregivers, N=227	Referred by Internal Medicine or Family Practice, N=53	Referred by Geriatrics, N=116	Referred by Neurology or Psychiatry, N=58	p-value
Domain: Experience with advice on dementia-related topics
1. I have received advice about what problems to expect in the future related to Alzheimer's or dementia.27	79 (35%)	20 (38%)	49 (42%)	10 (17%)	0.005
2. I have received advice about handling problems like the patient's memory loss, wandering or behavior problems.27	87 (39%)	25 (47%)	51 (44%)	11 (19%)	0.002
3. I am aware of services available to me to help me provide care.	57 (25%)	13 (25%)	36 (31%)	8 (14%)	0.043
Domain: Self-efficacy for caring for patient with dementia and accessing help
4. I know how to get community services that will help me provide care.	43 (19%)	8 (15%)	27 (23%)	8 (14%)	0.23
5. I feel confident that I can handle problems like the patient's memory loss, wandering or behavior problems.10	73 (32%)	17 (32%)	40 (34%)	16 (28%)	0.72
6. I feel confident that I can deal with the frustrations of caregiving.10	78 (35%)	18 (34%)	41 (35%)	19 (33%)	0.93
7. I know where to turn to get answers to questions about problems like the patient's memory loss, wandering, or behavior problems.10	82 (36%)	20 (38%)	47 (41%)	15 (26%)	0.15
Domain: Perception of patient’s dementia care prior to program entry
8. The patient's regular doctor understands how memory or behavior problems complicate other health conditions.26	178 (79%)	46 (87%)	99 (85%)	33 (57%)	<0.001
9. I have a healthcare professional who helps me work through dementia care problems.	63 (28%)	7 (13%)	47 (41%)	9 (16%)	<0.001

Pearson chi-square tests were used to compare groups.

Caregiver Strain and its Correlates

Caregivers reporting higher levels of strain (>13 on the MCSI) were more likely to be female (77% vs. 61%, p=0.004), reported more depressive symptoms (PHQ-9 score of 6.7 (SD 5.1) vs. 3.5 (SD 4.0), p<0.001) and cared for more functionally impaired patients (FAQ score of 25.0 (SD 4.7) vs. 20.1 (SD 8.2), p<0.001) with more behavioral symptoms (NPI-Q severity score of 14.0 (SD 6.8) vs. 9.1 (SD 6.3), p<0.001) as compared to caregivers reporting lower levels of strain. (Appendix 5)

Appendix.

Patient and Caregiver Characteristics by Level of Caregiver Strain

	Caregivers reporting lower levels of strain (mCSI ≤ 13), N=184 N (%) or M (SD)	Caregivers reporting higher levels of strain (mCSI>13), N=115 N (%) or M (SD)	Significance
Patients
Patient age, years	80.7 (9.1)	83.0 (7.4)	p=0.02
Patient female	114 (62%)	74 (64%)	p=0.68
Patient race/ethnicity			p=0.33
White, non-Hispanic	140 (76%)	82 (71%)
Black, non-Hispanic	24 (13%)	12 (10%)
Hispanic	11 (6%)	11 (10%)
Asian or Pacific Islander	9 (5%)	10 (9%)
Medicare/Medicaid dually insured	16 (9%)	14 (12%)	p=0.33
Type of dementia			p=0.66
Alzheimers	86 (47%)	48 (42%)
Mixed or unspecified	81 (44%)	54 (47%)
Other	17 (9%)	13 (11%)
Mini Mental State Examination (MMSE) score, N=289 (range 0–30)	16.9 (7.8)	15.2 (7.4)	p=0.06
Functional Assessment Questionnaire score, N=298 (range 0–30)	20.1 (8.2)	25.0 (4.7)	p<0.001
Neuropsychiatric Inventory Questionnaire (NPI-Q) symptom severity score, N=292 (range 0–36)	9.1 (6.3)	14.0 (6.8)	p<0.001
Referring provider			p=0.24
Internal Medicine/Family	36 (20%)	32 (28%)
Practice	108 (59%)	62 (54%)
Geriatrics	40 (22%)	21 (18%)
Neurology or Psychiatry
Caregivers
Caregiver female	113 (61%)	89 (77%)	p=0.004
Caregiver lives with patient	115 (63%)	81 (70%)	p=0.16
Caregiver relationship to patient			p=0.09
Spouse/partner	80 (43%)	44 (38%)
Child/child-in-law	80 (43%)	63 (55%)
Other	24 (13%)	8 (7%)
Caregiver PHQ-9 score, N=292 (range 0–27)	3.5 (4.0)	6.7 (5.1)	p<0.001
Self-efficacy score, N=224 (range 0–5)	3.0 (0.93)	2.5 (0.99)	p<0.001

DISCUSSION

Caregivers of patients referred to a dementia care program experienced high levels of strain, had low confidence in their ability to manage caregiving and access help, and frequently felt that they did not have a healthcare professional to help them with dementia-related problems. One in seven caregivers had symptoms consistent with moderate to severe depression. Caregivers in this cohort who reported low self-efficacy for dementia-caregiving also reported higher levels of strain, more depressive symptoms, and cared for a patient with more severe behavioral symptoms.

In prior cross-sectional studies of dementia caregivers, low self-efficacy for managing dementia-related behaviors was similarly associated with greater caregiver burden, more depressive symptoms, and poor self-rated health.^{10–12,32–35} Self-efficacy for dementia symptom management has also been shown to mediate the relationship between patient neuropsychiatric symptoms and caregiver burden and depression in those caring for patients with greater cognitive impairment.³³ Assessment of self-efficacy for managing dementia-related symptoms and accessing help may be a good way to identify caregivers who are most vulnerable to strain and depression. Furthermore, if self-efficacy can be modified by an intervention (such as a dementia care program), it may have implications for caregiver strain and depression.

The Caregiver Strain Index (CSI) has been widely used to evaluate strain among those caring for patients with a variety of chronic illnesses, ^36–39 including dementia,^29,30,40,41 and high CSI scores have been linked to poor mental health outcomes for caregivers.^36–39 This is consistent with the strong correlation between MCSI and PHQ-9 scores among caregivers in this cohort.

The need for additional caregiver support is underscored by the finding that most caregivers felt they did not have a healthcare professional to help them with dementia-related problems. This is particularly striking considering that more than half of referring providers were geriatricians, who are trained in the management of dementia and often have panels with large numbers of dementia patients. Although patients and caregivers referred to the program by geriatricians, internists, and family practitioners were more likely to report they had received advice on dementia-related topics as compared to patients referred by other subspecialists, the percentage of caregivers that had received advice was low overall, and there was no difference in self-efficacy for managing dementia-related problems or accessing help by type of referring provider. Similarly, neither caregiver strain or depressive symptoms nor the severity of patient behavioral symptoms differed by type of referring provider.

These findings suggest that the unmet need is less due to deficits in physician knowledge or ability, but rather that the current model of primary care does not facilitate adequately addressing the complexities of dementia care. For example, busy primary care physicians often do not have the time or resources available to adequately counsel patients and families and may not be familiar with community organizations to which they can direct caregivers. Moreover, caregivers may not have access to the patient’s primary care physician after-hours for advice to help manage dementia-related behaviors and avert crises.

To improve health outcomes for patients with dementia and their caregivers, a more comprehensive model of dementia care is needed. This model should routinely provide: individualized care plans that address all aspects of the disease; caregiver counseling and anticipatory guidance about the expected progression and complications of dementia; formal linkage to community organizations for respite care, caregiver training and education programs, and support groups; guidance on advance care planning; and 24/7 access to a dementia care provider to manage problems in real time and prevent unnecessary emergency department visits and hospitalizations. These enhanced services cannot be realistically provided by an individual primary care provider alone, but could be accomplished through other novel care delivery models, such as a co-management model with a nurse practitioner dementia-care specialist^20,42,43 or an interdisciplinary team approach with collaboration among the primary care provider and other health professionals including psychologists, pharmacists, social worker or nurse care managers, and occupational, recreational, and physical therapists.

There is also an increased national awareness of the current gaps in care for adults living with dementia and their caregivers, further supporting the implementation of new models of comprehensive dementia care. In January 2014, President Obama signed into law a $122 million increase for Alzheimer's research, education, outreach and caregiver support to enable the continued implementation of the National Plan to Address Alzheimer's Disease.^{44, 45} The 2013 update to the National Alzheimer’s Plan includes new provisions to improve the quality of dementia care provided by primary care practitioners; better address the needs of those in the late stages of the disease; and connect those living with dementia and their caregivers to available resources.⁴⁶

Our study has several limitations. The study cohort included the caregivers of the first 307 patients referred to a new dementia care program. Patients that were first referred to the program may have more severe behavioral symptoms or may have caregivers that are more strained than the general population of caregivers of patients with dementia. Also, our dementia care program serves an insured population at a large, metropolitan academic medical center that may not be generalizable to other populations. As this is a baseline evaluation of a quality improvement program, we have little sociodemographic information and no health status data for caregivers, and these caregiver characteristics likely influence patient and caregiver health outcomes. Lastly, this is a cross-sectional observational study using data collected at the time of program enrollment; thus, we cannot determine causality among severity of patient dementia-related symptoms, caregiver strain, caregiver self-efficacy, or caregiver depressive symptoms.

Longitudinal studies will help better identify those groups with the highest strain and lowest self-efficacy and allow implementation of changes to the program to better meet caregiver needs. Future research should also focus on further validation of task-specific caregiver self-efficacy scales, including the one used by our program. A better understanding of the specific areas in which caregivers feel least confident may allow the development of more tailored interventions to improve self-efficacy and ultimately caregiver health outcomes.

In conclusion, the findings of this study illustrate the substantial gap between caregiver needs and the current care provided for dementia. These needs include low self-efficacy for dementia-caregiving, caregiver strain and, in some cases, substantial depressive symptoms. These findings provide a strong rationale for implementing comprehensive care management approaches for patients with Alzheimer’s disease and other dementias.