A Dementia Care Management Intervention: Which Components Improve Quality?

Objectives:

To analyze whether types of providers and frequency of encounters are associated with higher quality of care within a coordinated dementia care management (CM) program for patients and caregivers.

Design:

Secondary analysis of intervention-arm data from a dementia CM cluster-randomized trial, where intervention participants interacted with healthcare organization care managers (HOCMs), community agency care managers (CACMs), and/ or healthcare organization primary care providers (HOPCPs) over 18 months.

Methods: Encounters of 238 patient/caregivers (dyads) with HOCMs, CACMs, and HOPCPs were abstracted from care management electronic records. The quality domains of assessment, treatment, education/support, and safety were measured from medical record abstractions and caregiver surveys. Mean percentages of met quality indicators associated with exposures to each provider type and frequency were analyzed using multivariable regression, adjusting for participant characteristics and baseline quality.

Results:

As anticipated, for all 4 domains, the mean percentage of met dementia quality indicators was 15.5 to 47.2 percentage points higher for dyads with HOCM-only exposure than for dyads with none (all P <.008); not anticipated were higher mean percentages with increasing combinations of provider-type exposure—up to 73.7 percentage points higher for safety (95% confidence interval 65.2%-82.1%) with exposure to all 3 provider types compared with no exposure. While greater frequency of HOCM-dyad encounters was associated with higher quality (P <.04), this was not so for other provider types.

Conclusions:

HOCMs’ interactions with dyads was essential for dementia care quality improvement. Additional coordinated interactions with primary care and community agency staff yielded even higher quality.

(Am J Manag Care. 2012;18(2):85-94)A successful dementia care management program providing high-quality dementia care to patients and family caregivers required care managers (nursing or social work background) trained to use dementia care management tools and protocols to yield substantial quality-of-care gains.

• More frequent encounters (in person, by phone, and by mail) with those care managers— embedded within the patient’s healthcare organization&mdash;were also associated with higher care quality.

•  Interactions with community agency staff and primary care physicians, with whom the algorithms and protocols were purposefully coordinated, were associated with even more meaningful increases in quality.

An estimated 5.4 million people living in the United States have Alzheimer disease (AD),¹ a number expected to increase 50% over the next 2 decades. The devastating effects of AD and other dementias impact an additional 15 million people who engage as informal family caregivers.¹ Because of accumulating functional disabilities and higher mortality, the need for healthcare delivery system supportive structures to assist patients and families is enormous. Several randomized controlled trials using coordinated care dementia management have provided strong evidence of meaningful health benefi ts for patients with dementia and for informal caregivers compared with usual care.^2-6

Effective chronic disease care management interventions typically include multiple components with diverse types of healthcare and service providers, which are tested in combination versus usual care.^2-4,6 These providers include clinicians, community agency staff, and specific care managers interacting with the patient-caregiver dyad, necessitating coordination among these entities. Some programs include nurse practitioners, for example, as care managers within specifi c clinics without built-in structural collaborations with outside agencies.^3,5 Whereas the focus of care management is driven by the clinical condition and is thus similar across programs, the means of delivery by provider type and the intensity of delivery are most likely to vary between programs. Given the complexity of comprehensive care management for dementia, broad translation of new knowledge about effi cacious interventions into diverse practice settings is more likely to occur when complexity is minimized, thereby maximizing sustainability and potentially reducing program costs.

We analyzed how exposure to different care management program provider types, singly or in combination, and exposure intensity were associated with positive effects on care quality with a dementia coordinated care management intervention.² This structured program emphasized collaborative care planning and enhanced communication across healthcare plan clinicians, social workers who served as care managers within healthcare organizations and community agencies, and the patient-caregiver dyad. We assessed the extent to which exposure to and interactions among these entities were associated with variation in achievement of higher dementia care quality.

METHODS

Context

The Alzheimer’s Disease Coordinated Care for San Diego Seniors (ACCESS) trial enrolled 408 patients with dementia paired with their informal caregivers (referred to as dyads). The intervention yielded higher receipt of recommended dementia care for 21 of 23 guideline-derived quality of care indicators at 18 months of follow-up, with the mean percentage per dyad of met indicators equal to 63.9% in the intervention arm versus 32.9% in the usual care arm. This cluster-randomized trial was a collaboration of 3 healthcare organizations and 3 community agencies in greater San Diego, California. A task force of clinical champions and community agency leaders met regularly to plan the care program including staffi ng structure and responsibilities, tools, and algorithms. Specially trained dementia care managers (primarily social workers) conducted structured assessments, identifi ed patient and caregiver problems, and further assessed problems to generate a care plan. Dyads were encouraged to participate in care planning. Dementia care managers coordinated and implemented chosen treatment actions and formally linked dyads to medical and community supports and services. The intervention also incorporated strategies for collaborative care planning and enhanced communication across healthcare plans and community agencies, with care coordination, ongoing followup, and decision support (Figure 1).^2,7 Primary care provider education on specifi c problem areas was provided to intervention clinics.

The ACCESS intervention’s Chronic Care Model—based components⁷ also included a leased Web-based care management software (CaseTrakker; IMA Technologies, Sacramento, California), which was then tailored to the study’s dementia care management protocols. Care managers embedded within the healthcare system as well as those working at study-engaged community agencies (San Diego Alzheimer’s Association, San Diego Caregiver Resource Center, Meals on Wheels San Diego) were trained in and used this software to track and manage case loads, administer assessments, organize care management activities into care plans, generate letters with appropriately merged assessment and care planning data to share with providers, and automate work flow to increase effi ciencies.

Subjects and Setting

Eligible subjects were Medicare recipients who received care from 1 of 3 San Diego area healthcare organizations. Most care recipients (80%) were enrolled in Medicare managed care. Eligibility characteristics and recruitment methods are detailed elsewhere.² The present analysis includes all 238 caregiver/care recipient dyads in the ACCESS intervention arm, including those who were randomized to care management but were unable to or elected not to receive it.

Sources of Data and Data Abstraction Methods

Exposures. All care management communications and encounters were recorded in CaseTrakker. Healthcare organization care managers used this Web-based system to record and guide initial assessments, reassessments, and all follow-up interactions. Community agency care managers (CACMs) accessed this system and recorded their interactions with dyads as well. All communications between care managers were also recorded in CaseTrakker, and every care manager relied upon this system as their electronic record of all individual and shared care management communications and encounters with dyads over the study intervention period.

Dementia Care Quality. Data sources we used for assessing dementia care quality were mailed caregiver surveys completed at baseline, 12 months, and 18 months, and abstracted medical records of enrolled patients with dementia.²

The CaseTrakker care management database was manually abstracted by 2 professional abstractors using an abstraction tool developed by study investigators.² All recorded notes from healthcare organization and community care managers were abstracted using on-screen or printed versions of these records. To assess interrater reliability, 10% of records were randomly selected and reabstracted. Agreement for the 3 primary predictor variables ranged from 94.2% to 100%, and kappa statistics for these variables ranged from 0.82 to 1.00.

Independent Variable Creation: Exposures to Types of Providers

The 3 primary independent variables were (1) healthcare organization care manager (HOCM), (2) CACM encounters with dyads, and (3) healthcare organization primary care provider (HOPCP) care management program—related activities. Data elements used to construct all 3 variables were abstracted from the care management software database. (The contents of all documentation in the care management software database were printed or reviewed on-screen during the abstraction process, and a notation that a phone call was made but no one was reached, for example, was not coded as an encounter.)

The HOCM activities coded as encounters for all interactions with dyads included scheduled initial assessments and 6-month reassessments (usually occurring as in-person meetings), as well as follow-up telephone calls, e-mails, and mailed written communications.

The CACM encounters included similar interactions with dyads. Community agency care manager encounters occurred only if there was a referral from the HOCM (which was not counted as a CACM-dyad encounter, since it might not have resulted in a CACM interaction with a dyad). Subsequent encounters between CACMs and dyads were based on CACM assessments and individualized need, and were coded as CACM encounters.

The HOPCP encounters were coded from documented communications between HOPCPs and HOCMs or CACMs, or as HOPCP interactions with dyads that were motivated by HOCM/CACM notifi cation. Healthcare organization primary care provider encounters typically occurred after an HOCM assessment, although HOPCPs could contact HOCMs for assistance and if recorded in the database, this is also included as an encounter.

Because of the highly skewed nature of encounter data, we coded these variables as 1/0 for any encounter by either specific care manager or primary care provider type (referred to as exposure) during the care management program duration for that dyad. Numbers of encounters are also coded as continuous integers (0-X) as an intensity measure, quantified as the average number of encounters per month of care management program duration for that dyad.

Dependent Variables: Dementia Care Quality

Table 1

Four domains of dementia care quality were developed and analyzed. Processes of dementia care quality, drawn from professional society guidelines,^8-10 were previously established by a steering committee of healthcare organization champions, community agency leadership, and ACCESS study team investigators. From these processes, 23 quality-of-care indicators were developed,^2,11 of which 22 are dichotomous and are grouped by content into 4 domains: assessment, treatment, education/support, and safety (). For each dyad, the mean percentage of met indicators over the follow-up period was calculated for each of the 4 domains. An analogous measure was derived for the 12-month period prior to enrollment (baseline).

Other Measures

All covariate measures, including care recipient and caregiver ethnicity (coded as ethnic minority [Latino, African American, Asian, First nation/Native Alaskan, Native Hawaiian/Pacific Islander] versus white), marital status, care recipient dementia severity (11-item Blessed-Roth Dementia Scale),¹² whether the caregiver was living with the care recipient and relationship to care recipient, whether the care recipient had a behavior problem in the past year, and caregiver’s education (coded as at least high school graduate vs less than high school graduate) were reported in the baseline caregiver survey. Caregivers also reported how large a problem it was to get caregiver assistance in the past 6 months (not a problem/did not need caregiving assistance, a small problem, or a big problem) in the baseline caregiver survey. Caregivers reported their own chronic health conditions according to a checklist and an open-ended question¹³ from which a weighted sum was created (possible range: 0-21).^14,15

Analyses

Data sources were linked by a common ID. Separate multivariable linear regression models were constructed for each of the 4 domains of dementia care quality (dependent variables). Independent variables in each model were provider type exposure, number of months that dyads were followed in the program, baseline quality in that domain, and patient/caregiver characteristics. Pearson correlations between each pair of the independent variables were tested to evaluate possible colinearity (defi ned as correlations greater than 0.55).

Primary predictor variables of exposure were sequentially added to account for all possible exposure combinations for each dyad, starting with HOCM alone, HOCM and HOPCP, HOCM and CACM, and all 3 exposure types occurring together. To account for the complex sample study design, all models were adjusted for healthcare organization site and clustered by clinic using Stata procedures for hierarchical data.¹⁶ As a sensitivity analysis, we reran all models including only the 154 dyads that were in the program the full 18 months.

The HOPCPs could attend 1 or more educational modules (attendance was recorded and coded). Because HOPCP encounters in the primary analysis did not count attendance, and some HOPCPs attended 1 or more sessions but had no recorded encounters (as defi ned for this analysis), a second sensitivity analysis tested the same models with inclusion (in addition to the original HOPCP variable) of an additional HOPCP exposure subgroup, dyads whose only exposure was HOPCP attendance at educational sessions (ie, they had no other encounters).

Intensity analyses were conducted with encounter variables modeled as continuous measures—the number of encounters (in-person, phone, e-mail, or mail) between care management providers and dyads or between care managers and primary care physicians (PCPs)&mdash;rather than dichotomous encounter (exposure) measures (any vs none). These analyses used the same covariates as were used in analyses of categorical (1/0) care management provider type exposure. Analyses were conducted using Stata 11 (StataCorp LP, College Station, Texas).

RESULTS

Table 2

Of the 238 intervention dyads, 68.5% of caregivers were female compared with 54.2% of care recipients (). Caregivers were most often spouses of care recipients (54.6%) but 39.5% were sons, daughters, or sons/daughters-in-law, and 70.2% lived with the care recipient. Mean duration of dementia symptoms was 4.1 years. Dementia severity included a range of all Blessed-Roth scores (0-17) and the mean score (5.7) refl ected moderate severity.

Table 3

A total of 74 dyads (31.1%) received exposure to all 3 care management provider types. Included in the sample of 238 dyads randomized to the intervention arm were 55 dyads (23.1%) that were offered but did not receive care management due to the caregiver declining it or an inability to be reached by telephone. Regarding the association between dementia care quality and exposure to type of provider (), there were signifi cantly higher mean percentages of met dementia quality indicators across all 4 domains for exposure to any care management provider type compared with no exposure (all P <.05). In the HOCM-only exposed groups (n = 51) compared with the no-exposure group (n = 55), the mean percentage with met indicators for the assessment domain was 47.2 percentage points higher (95% confi dence interval [CI] 42.9-51.5); for the treatment domain, 15.5 percentage points higher (95% CI, 3.9-29.1); for the education/support domain, 37.4 percentage points higher (95% CI, 22.7-52.1); and for the safety domain, 45.5 percentage points higher (95% CI, 33.3-57.8) (all P <.02; Table 3). For all quality domains, a gradient of better quality (ie, higher mean percentages of met indicators) was observed with exposure to increasing combinations of provider types, up to 73.7 percentage points higher (95% CI, 65.2-82.1) for safety with exposure to all 3 provider types compared with no care management exposure (Figure 2). Using a test of trend, the s uccessive addition of exposure types (HOCM, HOCM + HOPCP, HOCM + CACM, and HOCM + HOPCP + CACM) demonstrated a signifi cant increase in the mean percentage of indicators met within all 4 domains: for assessment, education/support, and safety the P values for trend all were <.001; for treatment, the P value for trend was .008.

Sensitivity analyses for the subset of dyads who received a full 18 months of care management were similar to those for the main analysis. Sensitivity analyses that included HOPCP educational session attendance as an exposure decreased the no-exposure proportion of the intervention arm sample from 55 (23.1%) to 31 (13.0%). Analyses with this revised categorization of provider type exposure demonstrated associations for the other exposure categories similar to those in the main analysis, but the HOPCP-only exposure group did not have higher quality than the revised “no exposure” group for any quality-of-care domain.

Modeled as encounters per month, there were an average of 0.78 HOCM encounters per month (per dyad) (median 0.67; range 0-4.5), 0.09 CACM encounters per month (per dyad) (median 0; range 0-1.08), and 0.12 HOPCP encounters per month (per dyad) (median 0.04; range 0-1.2) (Table 2). There was a statistically signifi cant association between higher levels of met indicators in all 4 quality domains and increasing frequency of HOCM encounters. For every additional full HOCM encounter per month (on average), the mean percentage of care indicators met in the assessment domain was 13.6 percentage points higher (95% CI, 4.6-22.6; P = .009); in the treatment domain it was 10.4 percentage points higher (95% CI, 0.7-20.0; P = .04); in the education/ support domain it was 14.2 percentage points higher (95% CI, 3.7-24.8; P = .01); and in safety domain it was 16.2 percentage points higher (95% CI, 7.2-25.2; P = .003). There was no association between quality and frequency of CACM or HOPCP encounters.

DISCUSSION

To our knowledge, this is the fi rst study to examine whether higher quality within a successful comprehensive dementia care management program is associated with exposure to different types of service providers or with the intensity of that exposure. These analyses were possible because of a Webbased software system used by intervention care managers that contained extensive documentation of intervention care management activities and communications. While HOCMs were the most critical provider type associated with improved quality, statistically and clinically signifi cant incremental gains in quality were seen with the addition of other provider types and thus, coordination between them. Program exposure in terms of greater numbers of encounters per month only conferred advantages with respect to HOCMs, but their overall encounter frequency was considerably higher than that for the other provider types, refl ecting a planned/intentional element of this program. That is, there might have been an association between more frequent encounters with a CACM and higher quality if a dedicated CACM had a more central role and more frequent encounters than the ACCESS staffi ng structure and responsibilities allowed. Moreover, the benefi t of multiple encounters between care managers and dyads for initiating and following up on problem-specifi c actions underscores the complexity of the dementia condition and the level of intervention required to make substantial quality gains.

Given that the great majority of dementia care management programs focus on psychosocial support, behavioral management, and educational needs, the need for medical provider engagement is likely to be less than the need for counseling and support; the frequency of PCP encounters supports this. However, even limited engagement of PCPs produced substantial and statistically signifi cant gains in quality in our study. Care management programs with an even greater degree of direct

medical provider engagement have also achieved improvements in dementia care quality, but the feasibility of expansion of PCP involvement may limit such programs’ spread.^3,5

Our finding that exposure to multiple provider types (HOPCPs and CACMs in addition to HOCMs) signifi cantly increased quality of care attests to the value-added importance of care integration and coordination for patients with dementia and their caregivers. The degree of integration of disease management programs with PCPs has been argued to be a distinguishing factor in the effectiveness of such programs.¹⁷ Our results also support those of the Medicare Coordinated Care

Demonstration evaluation, which concluded that a strong working relationship and communication between care managers and physicians was one of the keys to effective programs’ success.^18,19

The greatest contribution of additional provider types beyond HOCMs occurred for safety-related care. Attention to safety may require a greater collaborative effort or trigger engagement of additional community agencies and PCPs for successful completion. In particular, with respect to monitoring for risk of abuse and the need to report driving status to the county health department (a legal requirement in California, where this study was conducted), community agencies or medical providers are more likely to get involved. Care managers are likely to alert medical providers when such concerns are raised or some follow-up action is required.

Interactions and coordination between program components appeared to be greatly supported by Web-based care management software. With increased concern over reducing unnecessary healthcare utilization through better advance care planning,²⁰ transitional care,^21,22 and medical error reduction, ²³ the need for shared information technology and interdisciplinary care is rapidly expanding.²⁴ In fact, investment in health information technology has been promoted at the US presidential level since 2004^25,26 and most recently by the President’s Council of Advisors on Science and Technology, which has called for the federal government to facilitate widespread development and adoption of a “universal exchange language” to allow information transfer across facilities.²⁷

Regarding study limitations, measures of program activities were dependent upon care managers’ documentation of such activity. Healthcare organization care managers may have more reliably documented their activities than did those in community agencies, and might have under-recorded interactions with medical providers. However, documentation served to guide all activities, facilitate follow-up, and provide needed information for communication with others, ensuring that most or nearly all activities were likely to be recorded. This analysis was not powered to detect smaller but potentially meaningful between-group differences or effect sizes. Finally, this study was conducted in 1 US geographical area and might not generalize to other regions.

Dementia care management is becoming an increasingly recognized essential program for achieving high-quality dementia care. The recently released California State Plan for Alzheimer’s Disease and Related Dementias endorses the need for dementia care management programs,²⁸ and Healthy People 2020 now includes a focus on dementia, including Alzheimer disease, and cites the importance of medical and community services coordination to improve quality and outcomes of care.²⁹ Successful dissemination of such programs is increasingly dependent upon knowledge of how to retain their effectiveness while minimizing complexity and cost.Acknowledgments

The authors thank Brian Mittman, PhD, Theodore G. Ganiats, MD, Robert W. DeMonte Jr, MD, Richard Della Penna, MD, Lisa E. Heikoff, MD, Lorie Van Tilburg, Roger Bailey, PhD, Margo Fox Picou, Tom Gillette, PhD, and Roberto Valesquez for their collaboration and support. The authors also acknowledge the collaboration and contributions of clinical and support staff and providers at all participating Alzheimer’s Disease Coordinated Care for San Diego Seniors (ACCESS) study sites: Kaiser Permanente San Diego, Scripps Clinic, University of California, San Diego HealthCare, Alzheimer’s Association San Diego Chapter, Meals on Wheels Greater San Diego, and Southern Caregiver Resource Center. Finally, the authors thank our research staff, who helped with data abstraction and collection: Abdulrahman Hajar, BS, Marianne Doyle, MSW, Jessika Herrera, Liz Aguirre-Giron, Cassandra Hodo, Cecilia Huang, Jennifer Larkin, and Anna Dickey.

Author Affiliations: From Geriatric Research Education and Clinical Center (JC), VA Greater Los Angeles Healthcare System, Sepulveda, CA; Department of Medicine, Division of Geriatrics (JC) and Department of Neurology (KIC, SDV, BGV), David Geffen School of Medicine at UCLA, Los Angeles, CA; RAND Health (MJP), Santa Monica, CA; Parkinson’s Disease Research, Education, and Clinical Center (KIC, SDV, BGV), VA Greater Los Angeles Healthcare System, Los Angeles, CA; University of Michigan Medical School (MK), Ann Arbor, MI; Health Services Research and Development Service, Sepulveda Center of Excellence (MLL), VA Greater Los Angeles Healthcare System, Sepulveda, CA.

Funding Source: This study was supported by the California Department of Public Health (contract 07-65800). The ACCESS project was funded by the State of California Department of Aging (IG-0001-22); the California HealthCare Foundation (contract 99-3020); the State of California Department of Health Services Alzheimer’s Disease Education Initiative; and the State of California Department of Public Health (contract 06-55314) for the University of California, Los Angeles Alzheimer’s Disease Research Center.

Author Disclosures: The authors (JC, MLP, KIC, SDV, MK, MLL, BGV) report no relationship or fi nancial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (JC, MLP, KIC, MLL, BGV); acquisition of data (JC, MLP, KIC, BGV); analysis and interpretation of data (JC, MLP, KIC, SDV, MLL, BGV); drafting of the manuscript (JC, KIC, SDV, MK, BGV); critical revision of the manuscript for important intellectual content (JC, MLP, KIC, SDV, MK, MLL, BGV); statistical analysis (JC, SDV, MLL); obtaining funding (JC, MLP, BGV); administrative, technical, or logistic support (KIC, SDV, MK, BGV); and supervision (JC, KIC, BGV).

Address correspondence to: Joshua Chodosh, MD, MSHS, 11301 Wilshire Blvd (11G), Los Angeles, CA 90073. E-mail: jchodosh@mednet.ucla.edu.1. Alzheimer’s Association. 2011 Alzheimer’s Disease Facts and Figures. www.alz.org/downloads/Facts_Figures_2011.pdf. Accessed December 12, 2011.

2. Vickrey BG, Mittman BS, Connor K, et al. The effect of a disease management intervention on quality and outcomes of dementia care: a randomized controlled trial. Ann Intern Med. 2006;145(10):713-726.

3. Callahan CM, Boustani MA, Unverzagt FW, et al. Effectiveness of collaborative care for older adults with Alzheimer disease in primary care: a randomized controlled trial. JAMA. 2006;295(18):2148-2157.

4. Belle SH, Burgio L, Burns R, et al; Resources for Enhancing Alzheimer’s Caregiver Health (REACH) II Investigators. Enhancing the quality of life of dementia caregivers from different ethnic or racial groups: a randomized, controlled trial. Ann Intern Med. 2006;145(10):727-738.

5. Ganz DA, Koretz BK, Bail JK, et al. Nurse practitioner comanagement for patients in an academic geriatric practice. Am J Manag Care. 2010;16(12):e343-e355.

6. Judge KS, Bass DM, Snow AL, et al. Partners in dementia care: a care coordination intervention for individuals with dementia and their family caregivers. Gerontologist. 2011;51(2):261-272.

7. Wagner EH, Austin BH, Von Karff M. Organizing care of patients with chronic illness. Milbank Q.1996;74(4):511-544.

8. American Medical Association. Diagnosis, Management and Treatment of Dementia: A Practical Guide for Primary Care Physicians. Chicago, IL: American Medical Association; 1999.

9. Cummings JL, Frank JC, Cherry D, et al. Guidelines for managing Alzheimer’s disease: part I: assessment. Am Fam Physician. 2002; 65(11):2263-72.

10. Cummings JL, Frank JC, Cherry D, et al. Guidelines for managing Alzheimer’s disease: part II: treatment. Am Fam Physician. 2002;65 (12):2525-2234.

11. Chodosh J, Berry E, Lee M, et al. A dementia quality of care intervention: impact on physician knowledge, attitudes, and perceptions of care quality [published correction appears in J Am Geriatr Soc. 2006;54(9):1479]. J Am Geriatr Soc. 2006;54(2):311-317.

12. Blessed G, Tomlinson BE, Roth M. Blessed-Roth Dementia Scale (DS). Psychopharmacol Bull. 1988;24(4):705-708.

13. Hays RD, Wells KB, Sherbourne CD, Rogers W, Spritzer K. Functioning and well-being outcomes of patients with depression compared with chronic general medical illnesses. Arch Gen Psychiatry. 1995;52 (1):11-19.

14. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.

15. Vickrey B, Edmonds Z, Shatin D, et al. General neurologist and subspecialist care for multiple sclerosis: patients’ perceptions. Neurology. 1999;53(6):1190-1197.

16. Bell RM, McCaffrey DF. Bias reduction in standard errors for linear regression with multi-stage samples. Survey Methodology. 2002;28: 169-179.

17. Geyman JP. Disease management: panacea, another false hope, or something in between? Ann Family Med. 2007;5(3):257-260.

18. Peikes D, Chen A, Schore J, Brown R. Effects of care coordination on hospitalization, quality of care, and health care expenditures among Medicare benefi ciaries: 15 randomized trials. JAMA. 2009;301(6): 603-618.

19. Brown R. The Promise of Care Coordination Models That Decrease Hospitalizations and Improve Outcomes for Medicare Benefi ciaries With Chronic Illnesses. Report by the National Coalition of Care Coordination. Princeton, NJ: Mathematica Policy Research; March 2009.

20. Engelhardt JB, Rizzo VM, Della Penna RD, et al. Effectiveness of care coordination and health counseling in advancing illness. Am J Manag Care. 2009;15(11):817-825.

21. Coleman EA, Parry C, Chalmers S, Min SJ. The care transitions intervention: results of a randomized controlled trial. Arch Intern Med. 2006;166(17):1822-1828.

22. Naylor MD, Brooten D, Campbell R, et al. Comprehensive discharge planning and home follow-up of hospitalized elders: a randomized clinical trial. JAMA. 1999;281(7):613-620.

23. Shekelle PG, Pronovost PJ, Wachter RM, et al. Advancing the science of patient safety. Ann Intern Med. 2011;154(10):693-696.

24. Bergeron DF, Hyrkas K. Developing a clinical informatics system which supports a model of partnership among caregivers, patients, and families. Stud Health Technol Inform. 2009;146:754-755.

25. Scott JS. Why stop at HIPAA? getting back on the healthcare information highway. Healthc Financ Manage. 2004;58(9):34, 36.

26. Amatayakul M. The path to EHR. Healthc Financ Manage. 2004;58(8): 98-99.

27. President’s Council of Advisors on Science and Technology. President’s Council of Advisors on Science and Technology releases report on networking and information technology. http://www.whitehouse .gov/administration/eop/ostp/pressroom. Published December 16, 2010. Accessed June 16, 2011.

28. California Health and Human Services Agency. California’s State Plan for Alzheimer’s Disease: An Action Plan for 2011-2021. http://www .chhs.ca.gov/Documents/State%20Plan%20-%20FINAL.pdf. Accessed June 16, 2011.

29. US Department of Health and Human Services. Dementias, including Alzheimer’s disease. Objectives. Healthy People 2020. http://www .healthypeople.gov/2020/topicsobjectives2020/overview.aspx?topicid=7. Accessed June 16, 2011.