Reducing Health Care Costs for Dementia Patients

Abstract

Objective: Estimating cost savings based on limited health care events within a short time span for a dementia care program. Method: Data on health care utilization of persons with dementia (PWDs) and caregivers were gathered in the Dementia Care Services Program in North Dakota from January 2010 to January 2012. Data were aggregated into 3-month intervals and compared to 3 months before program intervention. Paired and cross-time models were used to estimate cost savings. Results: Health care cost savings for PWDs were estimated at US$143,118 to US$180,102 during the first 3 months after intervention, then decreased over time. Only the first 9 months could be used in the paired model due to small N and low power. Discussion: For programs with short time spans and limited health care events, a cross-time model can be used to estimate cost savings while producing results similar to paired models.

Keywords

geriatrics dementia rural aging health services

Introduction

Health and long-term care costs for persons with Alzheimer’s disease (AD) or other dementias are massive and continue to grow at an alarming rate. In 2012, the total amount spent on health care, long-term care, and hospice for persons with AD or other dementias was approximately US$200 billion, which is projected to increase to US$1.1 trillion in 2050 (in 2012 dollars; Alzheimer’s Association, 2010). In acknowledgement of this growing problem, the U.S. government introduced a Healthy People 2020 developmental objective for curbing the proportion of preventable hospitalizations among persons with diagnosed AD and other dementias (U.S. Department of Health and Human Services, 2012).

Equally disquieting, the totality of the adverse health effects among dementia caregivers in the United States is linked to approximately US$8.7 billion in additional health care costs (Alzheimer’s Association, 2012; Shriver, 2010). A recent study found that dementia caregivers, within 18 months of initiating their care provision, had a 25% increase use of all health services and paid an average of US$4,766 more per year in health care costs than noncaregivers (National Alliance for Caregiving, Schulz, & Cook, 2011). Caregiving in rural areas is even more difficult due to social isolation, poorer access to services, and long travel times (Butler, Turner, Kaye, & Downey, 2005). No studies were located that addressed health care cost outcomes of dementia support programs within a rural context.

Studies have shown that persons with dementia (PWDs) experience higher total medical costs, primarily due to hospitalization. Increased numbers of hospitalization and longer lengths of stay may be a result of a higher level of comorbidity in patients with dementia when compared to patients without dementia. Total costs for all-payers adjusted for comorbidity have ranged from 1.25 to 1.6 times higher for patients with AD compared to patients without dementia (Frytak et al., 2008; Gutterman, Markowitz, Lewis, & Fillit, 1999; Hill et al., 2002; Zhao, Huo, Weir, Kramer, & Ash, 2008). These hospitalization costs include preventable admissions and other ambulatory-sensitive costs such as emergency room (ER) and home health care costs. Reasons for these costs may include higher rates of trauma or falls, failure to manage chronic disease, inability of PWDs to coordinate care or report symptoms accurately, lack of caregiver education, caregiver burnout, and the challenge for health care providers to manage multiple conditions (Bynum et al., 2004; Zhao et al., 2008).

These situations provide both challenges and opportunities for improvement through systematic care management. Further research is needed to target specific interventions and models of care delivery to prevent hospitalization and morbidity by maximizing PWDs’ capabilities for living and addressing caregiver burden (Small, McDonnell, Brooks, & Papadopoulos, 2002). As the numbers and age of the nation’s elderly population increases, the prevalence of cognitive impairment will also increase. The need for a greater understanding for how to make interventions more effective in improving quality of life for PWDs and caregivers and decreasing disease-related costs will become even more important (Weiler, Lubben, & Chi, 1991).

A wide variety of dementia caregiver support programs and interventions have been found to generate positive impacts including increasing caregiver knowledge about dementia and care, reducing caregiver burden, decreasing caregivers’ depressive symptoms, and decreasing nursing home placement of PWD (Elliott, Burgio, & Decoster, 2010; Gitlin, Winter, & Dennis, 2010; Mittelman, Haley, Clay, & Roth, 2006; Sorensen, Pinquart, & Duberstein, 2012). Based in part on the application of the Stress/Health Model to caregiving (Schulz & Martire, 2004), the support and resources provided by such programs likely result in a safer environment, improved management of symptoms, and reduced stress on the caregivers—important outcomes that can decrease their panic response to call 911 or use ER or ambulance services (Mittelman et al., 2006).

The North Dakota Dementia Care Services Program (DCSP) is funded by the North Dakota Department of Human Services, Aging Services Division, and administered by the Alzheimer’s Association of MN-ND. Program information is publicized via various means including education classes, related social services agencies, health care providers, the Fargo Veterans’ Affairs Medical Center and community-based outreach clinics, senior centers, media, and word of mouth.

The program was modeled after a New York University (NYU) counseling and support intervention (Mittelman et al., 2006) effective in supporting dementia caregivers and families and delaying premature nursing home placement of PWDs. The NYU caregiver intervention recruited participants using specific dementia diagnostic criteria and consisted of an assessment followed by two individual and four family counseling sessions, voluntary attendance of a weekly support group, and availability of ad hoc telephone support. The DCSP is similar in that it provides care consultations that include a care plan with action steps, resources, and referrals. It is different in that it enrolls participants providing care to a person with either confirmed or suspected dementia and does not limit the number of counseling sessions. The DCSP’s goals are to enhance the caregiver’s knowledge of the disease process and develop strategies to address common obstacles faced by people caring for those with dementia. The DCSP strives to complete the care plan action steps within 3 visits, usually within 6 months of initial contact. This expedited process is intended to lessen time and travel constraints among caregivers, PWDs, and families residing in rural areas (average travel distance was 281 miles among DCSP caregivers who lived remote from their PWD). However, not all caregivers finish working with the DCSP within this visit/time framework. Some need to begin a new care plan right away, while others may wait a year or more before they again work with the DCSP. Thus, the DCSP does not use any formal discharge procedure as caregivers are encouraged to return at any time for additional assistance.

The Alzheimer’s Association of MN-ND employs the informational helpline and care consultants. The ND Department of Human Services has eight regional districts, served by DCSP’s five regional care consultants, who are distributed according to travel distance and population. Two project staff provide supervision and oversight to the regional care consultants who have backgrounds in the human services and receive orientation in the service delivery utilized in this project. The Alzheimer’s Association monitors their work from a clinical perspective and schedules periodic visits with each care consultant and supervisor to ensure quality and use of best practices. The DCSP implements improvements as needed—and measures client/program success through results-based feedback from an ongoing process and outcome evaluation conducted by an external, university-based, agency.

Caregiver support programs and interventions, including those related to dementia, may present unique challenges to analyzing health care cost outcomes. Some examples include short time spans working with caregivers, low numbers of program participants, low incidence of health care use among participants, absence of a study control group, and lack of patients’ medical claims and/or medical chart information. This study addresses these challenges in a rural context through the introduction and application of a model for measuring health care use and estimating cost changes associated with PWDs. Health services targeted for cost estimation include hospital, ambulance, ER, and 911 calls.

Longitudinal or paired models are common in cost analyses and return on investment models. They offer distinct advantages in controlling for random error as measurements are repeated for each subject, have lower variance, and are more sensitive to detecting differences (Korn & Braubard, 1999). Cross-sectional or cross-time models do not require the same person to be measured in each time period but measure all subjects available across time. People can move freely in and out of the study. This allows more people in the study at each time frame, increasing the sample size but also increasing variance.

Both a cross-time and paired model were used to estimate cost savings based on utilization of hospital, ER, ambulance, and 911 calls. These savings were compared between the models and over time.

Method

Data Collection

Questions regarding health care use were asked with every caregiver contact, either through the information helpline or a care consultation. Caregivers answered questions regarding health care utilization on 98% of the contacts. Health care utilization was measured by asking the caregiver to recall the number of days during the past 3 months they and/or the PWDs were hospitalized, the number of times either used an ambulance or the ER, and the number of times they called 911 for assistance in the past 3 months. Caregivers were asked about both their and the PWDs’ health care utilization as this program is designed to help both. This study focuses on responses regarding PWDs.

Data collection spanned from January of 2010 through January of 2012. During this time, DCSP care-workers had irregular contact with caregivers, usually lasting for 6 to 9 months. Because caregivers were asked questions regarding the previous 3 months, 3-month periods were used to analyze the data. The time period based on answers given during the first 89 days of contact with DCSP was the baseline. The responses at initial contact (first time they are asked) referred to events before the initial contact with the DCSP (i.e., before intervention). There was a control for repeated answers. For example, if a caregiver reported at intake that in the past 3 months the PWD was hospitalized for 5 days, then at a care consultation 8 days later they reported 5 days of hospitalization in the past 3 months, it was recognized that this was the same event and only 5 days (not 10) were used for cost analysis.

In January 2011, 1 year after the DCSP began, there was enough data reported to begin tracking health care events. Two 3-month time periods (baseline, 1-89 days; Time 1, 90-179 days) were recorded by counting the number of PWDs reporting in each time period and the number of events they reported. An event was defined as days the PWD spent in hospital, times they used an ambulance or the ER, or times that 911 was called for them. In the subsequent months of April, July, and October of 2011 and January of 2012, number of PWDs and number of events during 3-month time periods (Time 2, 180-269 days; Time 3, 270-359 days; Time 4, 360-449 days) were recorded.

Self-reported data were used as this was a study using secondary nonidentifiable data provided by the DCSP. There was no capability of identifying medical records for the PWDs. No controls were in the model as the DCSP is a voluntary program, and we did not identify PWDs in the state that were not participating in the DCSP.

Cost Savings Estimation

Whenever possible we utilized cost estimates that directly pertained to the care of persons with AD or dementia and adjusted them according to U.S. medical cost inflation to the year 2010 to estimate associated health care costs. Cost figures, based on median cost for persons with an AD or dementia diagnosis, included US$1,977 per hospital day (HCUPnet, 2009); US$617 per urban ambulance transport and US$927 per rural ambulance transport in North Dakota (Wingrove, Nudell, Becknell, Patterson, & Staffan, 2008); US$568 per ER visit for persons 65 and older (Agency for Healthcare Research and Quality, 2011); and US$82 per 911 call in North Dakota (North Dakota Emergency Services Communication Coordinating Committee, 2010).

Within each time period an event rate (number of events in 3-month period per number of PWD) was estimated. To estimate cost savings, changes in event rates between time periods 1 and 4 and the baseline were multiplied by number of PWDs affecting the rate change and the cost per event. Ambulance rates used an initial step involving separate rural- and urban-specific cost estimates as their amounts differed dramatically.

Statistical Analysis

Two methods of cost estimation were used, paired where nonzero rates from a pair of times were used and cross-time where all observations were used to estimate rates. The paired analysis used paired t-tests between average rates at baseline and Time 1, baseline and Time 2, baseline and Time 3, and baseline and Time 4 for caregivers who reported an event in at least one of the time periods tested. Observations with reports of no events at both times were excluded from the paired t-tests to avoid any nonnormal underlying distributions.

The cross-time analysis used all responses (including reports of no events) from the different time frames to estimate rate of events by dividing the number of events by the number of PWDs reporting during that time. Overall rates from each of the four time periods were compared to the baseline rate using a normal approximation to the binomial. The two methods were compared by the number of PWDs they included, the standard errors of the rate difference estimates, the power of the tests, and the cost estimates. Cost estimates were analyzed within each type of utilization and across time. The cross-time data were split into two groups, PWDs with comorbidities and those without, and compared as to the percentage of PWDs with comorbidities over time and the utilization rates between the two groups over time and between health care types.

Results

The number of PWDs reporting health care utilization events increased from 162 in January 2011 to 348 in January 2012. The 348 PWDs in January of 2012 had an average age of 77.8 years (range 30 to 98 years; 5.8% of the PWDs were young onset). 54.3% of the PWDs were female. 49.0% of the PWDs lived in rural areas and 84.5% were living at home. 87.7% of the PWDs had a diagnosis of AD or dementia (10.2% had a suspected diagnosis) and 14.4% had a comorbid condition. 20.2% were at stage 1, 67.8% at stage 2, and 12.0% at stage 3 of the disease. 25.2% were still driving.

There were 440 caregivers for the 348 PWDs in January of 2012 (53 PWDs had multiple caregivers working with the DCSP). The average age of the caregivers was 62.7 years (range 26 to 94). Three-quarters (76%) were female, 48% were children of the PWD (39% were daughters), 20% were wives, and 11% were husbands of the PWD.

Caregivers of PWDs spent a relatively short time with the DCSP. For the 348 in this study, half did not have visits after 59 days or about 2 months. The first quartile was 15 days (about 2 weeks) and the third quartile was 154 days (about 5 months). Of the caregivers 88% did not work with the DCSP past 9 months. Caregivers of rural PWDs were less likely to work a long time (more than 5 months) than those of urban PWDs. No other association was found between demographics or severity with time working with DCSP.

Table 1 shows the number of PWDs whose caregivers responded to questions regarding their health care utilization from January 2011 to January 2012. The number of PWDs in the baseline increased an average of 15.5 people each month. As the time period for collecting data was further from the baseline, the average increase in number of PWDs whose caregivers responded to questions regarding health care utilization of the PWDs dropped dramatically, from 7.4 people per month during T1 to 1.7 people per month during T4.

Table 1.

Number of PWDs and PWDs With Paired Data Reporting Hospital Events From January 2011 to January 2012

	Baseline 1-3 months	Time 1 4-6 months	Time 2 7-9 months	Time 3 10-12 months	Time 4 13-15 months
Number of PWDs
Jan 2011	162	52
Apr 2011	210	79	40
Jul 2011	263	103	61	29
Oct 2011	314	127	74	38	28
Jan 2012	348	141	79	43	33
Increase/month	15.5	7.4	4.3	2.3	1.7
Number of PWDs with pairs
Jan 2011		5
Apr 2011		10	2
Jul 2011		14	6	2
Oct 2011		19	8	3	3
Jan 2012		20	8	4	5
Increase/month		1.25	0.67	0.33	0.67

Baseline includes the 3-month period before working with the DCSP.

Table 1 also shows the number of PWDs with paired data or PWDs who had an event either during the baseline time period or a further time period. These numbers are much smaller (ranging from 1 to 20) as they do not include PWDs whose caregivers repeatedly reported no events. They also increased much more slowly over time. While the number of PWDs increased 7.4 per month during Time 1, the number of PWDs with paired data only increased 1.25 per month. As of January 2012 only 20 PWDs reported an event at either the baseline or the first time period Time 1. Cost savings for Times 3 and 4 were unreliable in the paired model (seven were negative, four had no standard error) due to insufficient sample size. The total cost savings at Time 1 was US$180,102 and US$37,971 at Time 2.

The cross-time approach where all reports from caregivers are included is shown in Table 2. Rates from each time period were compared to the baseline. One insignificant rate increase was found (−0.081, hospital at Time 4) and two rate decreases were insignificant (0.052 and 0.039, ambulance at Times 3 and 4). The pooled standard errors increased gradually, approximately doubling from Time 1 to Time 4, as the sample size decreased. The cost savings gradually decreased over time. Total cost savings were US$143,118 at Time 1, US$72,611 at Time 2, US$36,304 at Time 3, and −US$969 at Time 4.

Table 2.

Cross-Time Estimation of Event Rates and Cost Savings, January 2012

	Baseline	Time 1	Time 2	Time 3	Time 4
Hospital
N PWDs	348	141	79	43	33
Rate of events	0.647	0.184	0.278	0.279	0.727
Rate difference		0.462 *	0.368 *	0.367 *	−0.081
SE		0.050	0.062	0.079	0.087
Power		0.999	0.999	0.999
Cost savings		US$128,829	US$57,486	US$31,240	−US$5,266
Ambulance
N PWDs	341	140	78	42	33
Rate of events	0.100	0.043	0.013	0.048	0.061
Rate difference		0.057 *	0.087 *	0.052	0.039
SE		0.028	0.035	0.048	0.054
Power		0.685	0.925	0.235	0.114
Cost savings		US$5,335	US$5,108	US$1,567	US$1,025
ER
N PWDs	348	141	79	43	33
Rate of events	0.218	0.121	0.013	0.093	0.061
Rate difference		0.098 *	0.206 *	0.125 *	0.158 *
Standard error		0.039	0.048	0.065	0.074
Power		0.822	0.999	0.640	0.778
Cost savings		US$7,834	US$9,232	US$3,062	US$2,958
911 Calls
N PWDs	348	141	79	43	33
Rate of events	0.147	0.050	0.025	0.023	0.030
Rate difference		0.097 *	0.121 *	0.123 *	0.116 *
Standard error		0.032	0.041	0.055	0.063
Power		0.951	0.982	0.871	0.655
Cost savings		US$1,120	US$785	US$435	US$315
Total		US$143,118	US$72,611	US$36,304	−US$969

Note: Baseline includes the 3-month period before working with the DCSP.

Rate difference is the difference between the baseline rate and the rate for the time period. Rate differences were tested for significance and are shown in bold. Cost savings are estimated by taking the rate difference times the cost of one event (e.g., 1 day in the hospital, one use of the ambulance) times the number of PWDs in the time period.

p < .05.

Total estimated cost savings, using the cross-time model, were found for the five times shown in Table 1, January 2011 to January 2012. Figure 1 shows the estimated cost savings as they increase over time. Savings from Time 1 increased as the number of PWDs increased. Time 2 savings were first estimated in April 2011 but were not substantial until July 2011. Savings from Time 3 estimations were fairly consistent from July 2011 onward. Time 4 savings were so small that they were negligible to the graph and not shown. Total cost savings rose from US$30,839 in January 2011 to US$251,064 in January 2012.

Figure 1.

Estimated total cost savings, January 2011 to January 2012

The caregivers did report if the PWD had comorbidities. The data for the cross-time model were split into PWDs with comorbidities, and those without (Table 3). Table 3 shows there is no consistent increase or decrease in the amount of PWDs with comorbidities over time. PWDs who worked with the DCSP for longer periods of time were not more likely to have comorbidities. Second, Table 3 shows the health care utilization rate for both groups. The difference in rates was not consistent either over time or between types of health care.

Table 3.

Effect of Comorbidity on Time With DCSP and Rate of Health Care Utilization

	Comorbidity, N	No comorbidity, N	% With comorbidity	Comorbidity utilization rate	No comorbidity utilization rate	% Change in rate
Hospital
Baseline	48	285	14.41%	1.92	0.47	310.71%
T1	20	116	14.71%	0.35	0.16	113.68%
T2	9	67	11.84%	0	0.18	−100.0%
T3	6	37	13.95%	0.17	0.30	−43.94%
T4	4	29	12.12%	1.25	0.66	90.79%
Ambulance
Baseline	48	278	14.72%	0.08	0.09	−10.90%
T1	20	115	14.81%	0.00	0.04	−100.0%
T2	9	66	12.00%	0.00	0.02	−100.0%
T3	6	36	14.29%	0.17	0.03	500.00%
T4	4	29	12.12%	0.00	0.07	−100.0%
ER
Baseline	48	285	14.41%	0.27	0.22	22.52%
T1	20	116	14.71%	0.20	0.10	93.33%
T2	9	67	11.84%	0.00	0.01	−100.0%
T3	6	37	13.95%	0.33	0.05	516.67%
T4	4	29	12.12%	0.25	0.03	625.00%
911 Calls
Baseline	48	285	14.41%	0.08	0.16	−49.47%
T1	20	116	14.71%	0.05	0.04	16.00%
T2	9	67	11.84%	0.11	0.01	644.44%
T3	6	37	13.95%	0.00	0.03	−100.0%
T4	4	29	12.12%	0.00	0.03	−100.0%

Baseline refers to the 3-month period before working with the DCSP.

Discussion

The DCSP is a program designed to provide assistance quickly for caregivers of PWD. By providing resources, referrals, and care consultations, they hope to help ease the burden of care, empower the caregiver, and reduce health care costs. It is designed to provide assistance to caregivers for a few months, though caregivers are always able to contact the DCSP after any time and work with them again. This makes it difficult to have adequate numbers of PWDs to follow after 9 or even 6 months, as illustrated by the drop in numbers for Times 3 and 4 (Table 1). This can reduce the number of PWDs available for cost estimation, especially after 9 months. Having follow-up data (obtaining utilization information 1 year or more after PWDs have left the program) may provide limited additional data, assuming the caregivers can be contacted, they are still willing to provide information, and the PWD has not been placed.

Much of the health care utilization data consists of reports of no events by PWDs. No events were reported from 89.7% of the time (baseline) to 97.8% of the time (Time 2). Not utilizing health care is just as important to note as when it is utilized. As dementia is a progressive disease, PWDs who are not using health care over large time expanses should be noted. The second cross-time cost estimation method shown incorporates reports of no events in the model, as well as keeps the sample size high enough so not to decrease the power of the test.

Estimated cost savings were similar for Times 1 and 2 in both models though higher in the paired model. In Table 3, ER cost savings at Time 2 (7-9 months US$9,232) were higher than those at Time 1 (3-6 months US$7,834). All other cost savings decreased over time, partially due to smaller numbers of PWDs at later time intervals. Reductions in hospital events comprised 80% to 90% of the cost savings. However, reducing emergency and 911 service use can have further repercussions by reducing the stress of the caregiver and improving their health.

Limitations

The models shown are based on self-reported data, not medical records that may provide a more accurate report of timing and cost but are difficult to obtain. However, this self-reported data were found to be internally consistent. In 12 months of data, there were only one caregiver who had discrepancies in reporting health care utilization events. The numbers of events reported were consistent within time periods and event types.

The cost models described did not control for other factors such as severity or insurance status. Though there was an association between rural/urban status and time with the DCSP (rural PWDs had shorter time), there was no association between PWD demographics or severity and having an event, or having a change (increase or decrease) in events. It is possible that unmeasured factors, such as accessibility (insurance or travel distance) are influencing the costs, especially the decline in savings over time. There may also be unmeasured severity markers that keep a PWD working with the DCSP for longer periods of time and thus causing a decrease in savings. Also, two of the health care cost estimates (i.e., hospital and ER) used in this model were derived from national data due to their unavailability at the state level, though the hospital costs used were for people with a diagnosis of AD or dementia.

Finally, ambulatory care costs were unavailable for DCSP participants. PWDs can benefit from use of primary and preventive health care services to effectively manage their health and reduce the need for complex, intensive, and expensive emergency and inpatient care (Grumbach & Bodenheimer, 2002). Future studies should incorporate other costs such as those for long-term care, prescriptions, loss of work, and indirect (e.g., related to stress). These costs can be compared with the cost of the program for a return on investment model.

Conclusion

The short duration of the DCSP and a low occurrence of health care events make it difficult to obtain a large enough sample to do a longitudinal or paired analysis. Excluding PWDs who report no events disregards PWDs showing benefits from program participation by not utilizing these health care services. Measuring no use of health care is as important as measuring decreased use in health care.

The cross-time model that includes all reports of events is a practical and efficient way to measure cost savings with a program such as the DCSP where many of the caregivers report no events and relatively few of the caregivers participate in the program for 1 year or longer. It provides a cost savings estimate when it is not feasible to wait for multiple years to have enough caregivers reporting events in multiple time periods. The estimated cost savings from such a model are similar to paired models and with less variance. It also allows for a more accurate data interpretation as it encompasses all caregivers including those who report no events. Follow-up of PWDs after one or more years of program participation could improve these estimates.

The DCSP has shown consistently high estimated health care cost savings for PWDs in a very rural and sparsely populated state, totaling US$251,064 after 2 years. This partially offsets the total cost of the DCSP, US$1.2 million are funded for a 2-year period. Other potential cost savings from this program include long-term care cost avoidance, measured as shown in Klug, Volkov, Muus, and Halaas (2012). Estimates of potential long-term care cost avoidance from the program initiation (January 2011) to May 2012 are US$29.3 million (95% CI [US$21.4, 60.9 million]).

Cost estimation models similar to the one shown here can provide program evaluators and administrators the opportunity to conveniently access and analyze health cost information on a periodic basis to promote program understanding decision-making, and performance improvement. Finally, these health care cost analysis models can be incorporated into a return on investment model when combined with the analysis of program costs, long-term care cost savings, and benefits from reducing caregiver stress.

Footnotes

Acknowledgements

Jan Mueller, Heidi Haley-Franklin, Krista Headland, and Gretchen Dobervich provided review of the article (Alzhimer’s Association of Minnesota/North Dakota, Dementia Care Services Program); Erica Lien (Alzhimer’s Association of Minnesota/North Dakota, Dementia Care Services Program) provided data entry; and Brad Gibbens (University of North Dakota School of Medicine and Health Sciences, Center for Rural Health) provided assistance with design.

Authors’ Note

Requirement for international review board was waived as data are nonidentifiable and there is no interaction with the subjects.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project is supported by funding granted through the North Dakota Department of Human Services, Aging Services Division, grant number 190-08687.

References

Agency for Healthcare Research and Quality. (2011, October 26). Emergency room services—Mean and median expenses per person with expense and distribution of expenses by source of payment: United States, 2008. Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from Medical Expenditure Panel Survey Household Component Database.

Alzheimer’s Association. (2012). 2012 Alzheimer’s disease facts and figures. Alzheimer’s & Dementia, 8, 131-168.

Butler

S. S.

Turner

Kaye

L. W.

Ruffin

Downey

(2005). Depression and caregiver burden among rural elder caregivers. Journal of Gerontological Social Work, 46, 47-63.

Bynum

J. P. W.

Rabins

P. V.

Weller

Niefeld

Anderson

G. F.

A. W.

(2004). The relationship between a dementia diagnosis, chronic illness, Medicare expenditures, and hospital use. Journal of the American Geriatric Society, 52, 187-194.

Elliott

A. F.

Burgio

L. D.

DeCoster

(2010). Enhancing caregiver health: Findings from the resources for enhancing Alzheimer’s caregiver health II intervention. Journal of the American Geriatric Society, 58, 30-37.

Frytak

J. R.

Henk

H. J.

Zhao

Bowman

Flynn

J. A.

Nelson

(2008). Health service utilization among Alzheimer’s disease patients: Evidence from Managed care. Alzheimer’s & Dementia, 4, 361-367.

Gitlin

L. N.

Winter

Dennis

M. P.

(2010). A biobehavioral home-based intervention and the well-being of patients with dementia and their caregivers: The COPE randomized trial. Journal of the American Medical Association, 304, 983-991.

Grumbach

Bodenheimer

(2002). A primary care home for Americans: Putting the house in order. Journal of the American Medical Association, 288, 889-893.

Gutterman

E. M.

Markowitz

J. S.

Lewis

Fillit

(1999). Cost of Alzheimer’s disease and related dementia in managed Medicare. Journal of the American Geriatric Society, 47, 1065-1071.

10.

HCUPnet. (2011, November 15). 2009 National statistics—principal diagnosis only—outcomes by patient and hospital characteristics for ICD-9-CM principal diagnosis codes 331.0, 290.0-290.9. Rockville, MD: Agency for Healthcare Research and Quality. Retrieved from http://hcupnet.ahrq.gov/HCUPnet.jsp.

11.

Hill

J. W.

Futterman

Duttagupta

Mastey

Lloyd

J. R.

Fillit

(2002). Alzheimer’s disease and related dementias increase costs of comorbidities in managed Medicare. Neurology, 58, 62-70.

12.

Klug

M. G.

Volkov

Muus

Halaas

G. W.

(2012). Deciding when to put grandma in the nursing home: Measuring inclinations to place persons with dementia. American Journal of Alzheimers Disease and Other Dementias, 27, 223-227.

13.

Korn

E. L.

Graubard

B. I.

(1999). Analysis of health surveys. New York, NY: Wiley.

14.

Mittelman

M. S.

Haley

W. E.

Clay

O. J.

Roth

D. L.

(2006). Improving caregiver well-being delays nursing home placement of patients with Alzheimer’s disease. Neurology, 67, 1592-1599.

15.

National Alliance for Caregiving, Schulz

Cook

(2011, November). Caregiving costs: Declining health in the Alzheimer’s caregiver as dementia increases in the care recipient. Bethesda, MD: National Alliance for Caregiving. Retrieved from http://www.caregiving.org/pdf/research/Alzheimers_Caregiving_Costs_Study_FINAL.pdf.

16.

North Dakota Emergency Services Communication Coordinating Committee. (2010). Emergency services communication in North Dakota: A status report, 2010. Bismarck, ND: North Dakota Emergency Services Communication System.

17.

Schulz

Martire

L. M.

(2004). Family caregiving of persons with dementia: Prevalence, health effects, and support strategies. American Journal of Geriatric Psychiatry, 12, 240-249.

18.

Shriver

(2010). The Shriver report: A woman’s nation takes on Alzheimer’s. Chicago, IL: Alzheimer’s Association.

19.

Small

G. W.

McDonnell

D. D.

Brooks

R. L.

Papadopoulos

(2002). The impact of symptom severity on the cost of Alzheimer’s disease. Journal of the American Geriatric Society, 50, 321-237.

20.

Sorensen

Pinquart

Duberstein

(2002). How effective are interventions with caregivers? An updated meta-analysis. Gerontologist, 42, 356-372.

21.

U.S. Department of Health and Human Services. (2012, July 2). Dementias, including Alzheimer’s disease. Retrieved from http://healthypeople.gov/2020/topicsobjectives2020/objectiveslist.aspx?topicId=7

22.

Weiler

P. G.

Lubben

J. E.

Chi

(1991). Cognitive impairment and hospital use. American Journal of Public Health, 81, 1153-1157.

23.

Wingrove

Nudell

Becknell

Patterson

Staffan

(2008). The rebasing of North Dakota medical service payment rates (provider type: ambulance). Tinton Falls, NJ: SafeTech Solutions.

24.

Zhao

Kuo

T. C.

Weir

Kramer

M. S.

Ash

A. S.

(2008). Healthcare costs and utilization for Medicare beneficiaries with Alzheimer’s. BMC Health Services Research, 8, 108.