Interventions to Improve Quality of Care: The Kaiser Permanente"Alzheimer's Association Dementia Care Project

, , , , ,
The American Journal of Managed Care, August 2004, Volume 10, Issue 8

Objective: To improve quality of dementia care in a KaiserPermanente service area through rigorous dissemination of practiceguidelines and social worker support for physicians andpatients.

Study Design: Pre-post design with practice behavior changeassessed by medical record review, and provider and caregiver satisfactionwith care assessed by surveys.

Methods: A diagnostic guideline and later a managementguideline were adopted for use by Kaiser Permanente physicians inmetropolitan Los Angeles. Physicians received training based onthe guidelines, and social workers provided ancillary support.Eighty-three community-dwelling dementia patients and their caregiverswere referred to the project by primary care physicians andthen were assessed and followed by social workers. Data wereabstracted from medical records to determine whether these interventionsled to improved quality of care as indicated by adherenceto key care processes derived from the adopted dementia guidelines.Chi-square and t tests were applied to compare guidelineadherence and satisfaction rates before and after the interventions.

Results: Compared with baseline, higher rates of provider andcaregiver satisfaction with Kaiser's system of dementia care werefound at the postintervention follow-up. There also were significantlyhigher rates of adherence to several practiceguideline–based quality measures: assessment of cognitive status;referrals to the Alzheimer's Association; and assessments of activitiesof daily living, decision-making capacity, depression, and wanderingrisk.

Conclusion: Quality of primary care for people with dementiacan be improved through guideline implementation with caremanagement support by social workers.

(Am J Manag Care. 2004;10:553-560)

Healthcare organizations are being challenged tocare for the growing number of older adults withchronic health conditions. Dementing diseases,such as Alzheimer's disease, vascular dementia, andother associated disorders, present particular challengesbecause there are strong social and behavioralcomponents to disease management. Physicians play acentral role in assessment, diagnosis, and treatment;but much of the management for noninstitutionalizedindividuals is done by families with the support of socialwork care managers and community resource organizations.Furthermore, primary care physicians are dailyfaced with a broad range of disorders and are challengedby dementia patients, who may compensate for cognitivelosses with retained social skills during the briefoffice visit.

Research suggests that Alzheimer's disease is neitherwell recognized nor systematically diagnosed.1-3According to Boise and colleagues,1 physicians diagnoseas few as 50% of dementia cases. Focus groups of primarycare physicians from 3 geographic areas found significantbarriers to the recognition and formal diagnosisof dementia. Physicians reported difficulty in recognizingpossible dementia. Many reported that they reliedon families to bring the disease to their attention. Therealso was a prevalent viewpoint that a formal diagnosticevaluation (including a comprehensive clinical historyand mental status examination, laboratory referrals torule out treatable conditions, and other procedures) isnot always necessary.

Yet delay in diagnosis also means delay in treatment.The last 10 years have seen approval of several medicationseffective in slowing progression of Alzheimer's disease,the most common type of dementia in people withmild or moderate cognitive impairment.4 Furthermore,people with undiagnosed dementia and their familiesare less likely to gain access to supportive services thatcan ameliorate caregiver burden and perhaps delayinstitutionalization.5,6

Clinical practice guidelines have been developed andpromoted as tools for improving dementia care.7-11However, practitioners often are not familiar with theseguidelines; and even when disseminated broadly, theguidelines are not implemented uniformly.3,12

Managed care organizations offer some unique possibilitiesfor quality improvement in dementia care. Theyprovide the communication lines for dissemination ofguidelines. Their accreditation is dependent on adoptionand implementation of guidelines-based qualityimprovement projects.13 Capitated payments can beused creatively to bring in the services of less expensiveprofessionals that may reduce costs or hold them steadywhile improving quality. A number of pilot projectshave been initiated to improve care for people withdementia in managed care settings 14; one such project,involving a large managed care plan in Los Angeles, isdescribed here.


Overview of Project and Description ofIntervention

In 1995, Kaiser Permanente's Metropolitan LosAngeles Service Area entered into a partnership with theAlzheimer's Association of Los Angeles to assess andimprove the quality of care provided to people withdementia. The Metropolitan Los Angeles Service Area ofKaiser Permanente serves a densely urban and ethnicallydiverse region. It includes 2 major medical centers andseveral satellite clinics. Based on extrapolated demographicestimates,15 up to 5000 people with dementiamay be served by Kaiser Permanente in this region.

The goal of the Dementia Care Project was to improvethe quality of care for people with dementia whileincreasing provider and consumer satisfaction. Thestudy used a longitudinal preintervention/postinterventiondesign, with evaluation data collected by medicalrecord review, caregiver interview, and provider survey.A consensus-based diagnostic practice guideline wasdeveloped by a multidisciplinary panel of KaiserPermanente practitioners and representatives of theAlzheimer's Association's Medical and ScientificAdvisory Board. The disciplines of neurology, psychiatry,geriatrics, family practice, psychology, and socialwork all were represented in the development group.This guideline sought to improve consistency in thediagnostic assessments done at Kaiser Permanente inthis region. The project later adopted a managementguideline published by Cummings et al.10 To implementthese guidelines as part of the Dementia Care Project,the following steps were taken:

  • Promotion by Kaiser Permanente's clinical leadership through letters to providers and presentations at meetings.
  • Broad dissemination of a laminated pocket version of the diagnostic guideline to all primary care physicians in this service area.
  • Annual in-service training for primary care providers that was notable for its inclusion of physician leaders and caregivers as speakers, as well as the participation of a theater troupe that enacted a physician visit and the administration of a mental status examination. The theater troupe was used to increase physician empathy for presenting families. It also was assumed that the novelty of this form of teaching would enhance learning.
  • Broad dissemination of a provider "tool kit" that contained the 2 guidelines and a variety of forms and tests of mental status and depression to support physicians as they followed guideline recommendations.

The Figure and Table 1 presentoutlines of the diagnostic and management guidelines used by the Dementia Care Project.

The Dementia Care Projectrelied heavily on social work caremanagers to support this guideline basedquality initiative. Two full-timeprofessionals were trained bythe Alzheimer's Association with a24-hour intensive course followedby 7 months of mentoring via caseconferences. Patients were referredto the social workers by their primarycare physicians. The socialwork care managers provided arange of services in support ofphysicians, including administrationof mental status exams, follow-upwith families on thepsychosocial aspects of care, homevisits, intervention with familiesregarding behavior managementissues, and facilitation of connectionto physicians for treatment ofcomorbid conditions. The caremanagers also provided guideline recommendedservices for familycaregivers, including assessmentsof patients, linkage to services withinKaiser Permanente and in thecommunity, caregiver supportgroups, family education, and ongoingcare management.


In order to participate in theDementia Care Project, the personwith dementia needed to have a primarycare physician in the designatedKaiser Permanente servicearea, have a guideline-based dementiadiagnosis, dwell in the community(not in an institution), and havean involved, nonpaid caregiver. Participants also neededto be referred for participation by their primary carephysicians. Data were collected on 83 persons withdementia and their caregivers. The dementia patientsranged in age from 63 to 93 years (mean age = 80 years).They were 52% male and ethnically diverse (55% AfricanAmerican, 21% Caucasian, 3% Latino, 1% Asian, and 19%other or not specified). This ethnic diversity reflects themembership of Kaiser Permanente in this geographicregion. Initial Folstein Mini-Mental Status Examinationscores16 recorded after the program intervention beganwere available for more than 90% of the patients andranged from 2 to 29 with a mean score of 17. Of the 83participating caregivers, 63% were spouses; 16%, daughters;11%, sons; and 10%, other. Their mean age was 66years. The ethnic breakdown of caregivers was 52%African American, 23% Caucasian, 4% Latino, 4% Asian,1% Native American, and 16% other or not specified.

Study Outcome Variables

The study was primarily focused on improving thequality of care for people with dementia in this targetedregion. The following indicators of improved qualitywere drawn from the practice guidelines:

  • Conducting a cognitive assessment with the Folstein Mini-Mental Status Examination.16
  • Obtaining laboratory tests (B12, Venereal Disease Research Laboratory/rapid plasma reagin [test forsyphilis], thyroid-stimulating hormone, complete blood count, and selected blood chemistries).
  • Referral to the Alzheimer's Association and to the Safe Return Program (for wandering risk) of theAlzheimer's Association.
  • Conducting assessments that included activities of daily living, decision-making capacity, depression,and wandering risk.

These care processes were selected as study outcomesin part because they were guidelines targeted by theintervention. In addition, they have been identified andused with some frequency in studies of the quality of carefor dementia.17 It was predicted that each of these practiceswould increase as a result of project participation.

Kaiser Permanente physicians were surveyed abouttheir diagnostic practices, referral behaviors, perceptionsof services provided to patients with dementia, useof educational materials, and overall satisfaction withthe care of patients with dementia and their families.Satisfaction was assessed with a single question: "Howsatisfied are you with the treatment and support thatKaiser Permanente provides to patients with dementiaand their caregivers?" Similarly, participating caregiverswere surveyed regarding their satisfaction withdementia-related services at Kaiser Permanentethrough a series of questions including:

  • "Overall, how satisfied have you been with the dementia-related service you and the patient havereceived at Kaiser Permanente?"
  • "How satisfied have you been with the sensitivity and respect shown to you and the patient at KaiserPermanente?"
  • "Have Kaiser Permanente physicians and staff given you educational materials about dementia?"

Caregivers also were questioned regarding referrals they were given both within the organization and tocommunity services.


The study design and consent forms were reviewedand approved by the Kaiser Permanente institutionalreview board. All study participants gave informed consentfor the study. Three different sources of data wereused to assess the impact of the project on providerpractices and satisfaction with care:

  • Medical records, including primary care providerand social work files, were abstracted to assesswhether the quality indicators derived from theguidelines were being practiced. Preinterventionadherence and postintervention adherence tothese guidelines were measured and compared.
  • Surveys of primary care physicians, administeredbefore the program was implemented and again 3years later, were used to assess self-reported practicebehaviors and provider satisfaction withKaiser Permanente's system of care for people withdementia.
  • Interviews were conducted with caregivers atbaseline and between 3 and 6 months after theprogram was implemented to assess change in satisfactionwith dementia care.

Medical Record Abstractions.

Medical recordsincluding social work files were audited to examine thesystem of care before and after implementation of theDementia Care Project. Project coinvestigators designedand applied a structured chart abstraction form toreview the records of the 83 program participants. First,abstractors identified the first mention of dementia ormemory problems by medical staff after January 1998,the date on which the Dementia Care Project interventionwas initiated. After this index visit, abstractorsrecorded dates of documentation of the mental statusexam, assessments, and results of laboratory tests. Theabstractors also documented referrals to staff socialworkers and to outside resources such as theAlzheimer's Association.

Next, 42 of the 83 participants were randomlyselected for a study of preintervention versus postinterventioncare. Resource limitations dictated theneed for this more restricted sample. Records for thissubset of subjects were abstracted over 3 years priorto the onset of the intervention in January 1998 toevaluate differences in rates of adherence to the guidelines before the intervention compared with after theintervention.

Preintervention and Postintervention PhysicianSurveys.

With cover letters from the chiefs of internalmedicine and family practice, a total of 307 surveyswere mailed to primary care providers in theMetropolitan Los Angeles Service Area prior to projectimplementation; 112 surveys were returned, for a 36%response rate. Three years after the project was initiated,345 surveys were sent to primary care providers inthe same service area; 126 were returned, for a 37%response rate. (Because identifiers were not linked tothe surveys, it was not possible to determine the percentageof follow-up surveys received from the originalpreintervention sample.) As an incentive for participation,physicians who completed the follow-up surveywere entered into a lottery for a complementary dinnerfor 2 people.

Caregiver Interviews.

Caregivers of patients witha diagnosis of dementia who were enrolled in the projectparticipated in 2 structured telephone interviews.The first interview was conducted by a trained intervieweror the social work care manager at the time ofenrollment of the caregiver and patient in theDementia Care Project. A trained interviewer also conductedthe postintervention interview 3 to 6 monthsafter the caregiver and patient began receiving servicesfrom the social worker. All 83 participating caregiverscompleted the baseline and postinterventioninterviews.




Chi-square and tests were applied to compareguideline adherence rates prior to the intervention withthose after the intervention, based on the medicalrecord reviews and on the physician surveys. Chi-squareand tests were used to compare caregiver satisfactionwith care before and after the project intervention.


Medical Record Abstractions


There was an increase in the reporting of Mini MentalStatus Examination scores from 16% prior to the projectto 93% after the project (chi-square, < .001) amongthe subset of 42 pre- and postintervention chartsreviewed. The first postintervention mental status examdocumented in records was most frequently administeredby a neurologist (27%), followed by a primary carephysician (23%) and then by a social worker (20%). Thepercentages of patients who obtained each of the followingguideline-supported laboratory tests during thepostintervention period were: thyroid-stimulating hormone,84.3%; Venereal Disease Research Laboratory/rapid plasma reagin, 67.5%; B12 , 69.9%; completeblood count, 94.0%; electrolytes, 84.3%; blood ureanitrogen, 81.9%; calcium, 75.9%; creatinine, 81.9%; andglucose, 89.2%.



Documentation of referrals to the Alzheimer'sAssociation increased from 3% prior to the project to76% afterwards (chi-square, < .001), and referrals tothe Alzheimer's Association's Safe Return programincreased from 0 to 29% (chi-square, < .001). Projectsocial workers made most of the referrals to theAlzheimer's Association (34%), followed in frequency byother staff social workers (19%).





The medical record abstraction identified increaseddocumentation of assessments of activities of daily living(13% preintervention, 93% postintervention; <.001), decision-making capacity (3% preintervention,19% postintervention; < .001), depression (11%preintervention, 57% postintervention; < .001), andwandering potential (8% preintervention, 74% postintervention; < .001). After the intervention, dementiaproject social workers were most likely to document anassessment of daily living (32%), followed by neurologists(26%). Project social workers also documented68% of all assessments for wandering risk. However,depression was noted most frequently by primary carephysicians (39%), followed by project social workers(20%) and neurologists (17%). Physicians more frequentlydocumented an assessment of decision-makingcapacity, (31% by neurologists, 19% by primary carephysicians).

Table 2 presents a summary of documented changesin the selected quality indicators from the preinterventionto the postintervention period.

Physician Surveys





A comparison of physician responses before and afterthe intervention showed a trend toward increased satisfactionwith the treatment and support KaiserPermanente provides to people with dementia ( test, = .067). Among the physicians in the postinterventionsample, 49% reported using a mental status testmore frequently than they had before the intervention.Prior to project implementation, 46% of physiciansreported that they never obtained laboratorytests as part of a diagnostic assessment of dementia;the percentage of physicians never obtaining thesetests fell to 14% after the intervention (chi-square, <.001). No significant change in referrals to theAlzheimer's Association were reported: prior to theintervention, 58% of physicians sometimes or alwaysreferred their patients to the Alzheimer's Association,whereas after the intervention, 63% reported makingthis referral at least some of the time (chi-square, = .440).





In an effort to assess the specific impact of providerworkshops on guideline adherence, responses of physicianswho reported attending the initial Dementia CareWorkshop used to launch this project in 1997 werecompared with those physicians who did not attend theworkshop. One hundred fifty-two physicians, primarilyin the family practice and internal medicine fields,attended the workshop; of these, 47 completed andreturned the postintervention physician survey.Workshop attendees who completed the survey weremore likely to report that they recalled the diagnosticguideline for dementia ( < .015), administered themental status exam ( < .005), provided educationalmaterial to families ( < .001), and referred to theAlzheimer's Association ( < .016) compared with the77 physicians who responded to the survey and who didnot attend the workshop.

Caregiver Interviews







A comparison of structured interviews administeredto caregivers before and after participation in theDementia Care Project indicated improvements in satisfactionand in several quality measures. At baseline,17% and 51% of caregivers reported being very satisfiedand satisfied with dementia care, respectively. In contrast,at follow-up, 40% and 39% of caregivers reportedbeing very satisfied and satisfied with dementia care,respectively ( < .05). Sensitivity and respect shownby Kaiser Permanente staff were rated more satisfactoryafter participation in the intervention, with 38%very satisfied at baseline comparedwith 70% very satisfied3 to 6 months after enrollment( < .001). Caregivers also indicatedthat they were more likelyto have been given educationalmaterial about dementia afterthe Dementia Care Project wasinitiated: 36% at baseline comparedwith 94% postintervention( < .001). Postintervention, ahigher proportion of caregiversreported receiving referrals tocommunity services such ashome health (2% preinterventionvs 13% postintervention; < .01);support groups (4% preinterventionvs 17% postintervention; < .01); and Meals onWheels (1% preintervention vs 12% postintervention; < .01).


The Dementia Care Project was a collaborative effortof the Alzheimer's Association and Kaiser Permanentein Los Angeles to improve the quality of care given topeople with dementia and their caregivers. Practiceguidelines were the framework used to establish qualitygoals for the system of care. They were reinforced byeducational programs and materials for primary careproviders, and by social workers, who provided caremanagement, support, and other services to patientsand their families.

Over the course of this project's implementationfrom 1997 through 2000, the medical record review andsurvey data both revealed greater use of mental statusexams to screen patients for dementia. The finding thatphysicians who attended the initial educational workshopreported that they were more likely to administerthis exam further supported this outcome. How muchthe reported increase in use of this test was due to theworkshops versus differences in characteristics ofphysicians who do and do not attend such workshops(in terms of their receptivity to change and their interestin dementia care) cannot be determined from thisstudy. Significantly, however, despite an increase inphysician administration of the mental status exam,20% of project participants were given their initial mentalstatus exam by the dementia project social workerrather than a physician.

Examination of quality indicators for postdiagnosticmanagement indicated that the specially trained socialworkers were clearly an asset to quality dementia care.The chart audits revealed that these professionals wereimportant to guideline implementation. Most frequentlythey were the professionals who evaluated patientsfor ability to perform activities of daily living and riskfor wandering. Less frequently, they provided assessmentsof patient depression and decision-makingcapacity. As was similarly noted by the HMOWorkgroup on Care Management of the AmericanAssociation of Health Plans Foundation,18 the caremanagers were the professionals most likely to referfamilies to the Alzheimer's Association and its SafeReturn Program as well as to other community organizationsfor supplementary support services.

The number of study subjects was low (n = 83) relativeto the estimated number of people with dementia inthe targeted service area (n = 5000). Limited researchfunding and a relatively short period of data collection(less than 2 years) were contributing factors. In addition,referrals to the study were initially low and grewover time as the primary care providers became morefamiliar and comfortable with the intervention.Changing physician referral patterns takes time.Following the management guideline promoted throughthis project, physicians were not expected to makecommunity referrals themselves, but rather they wereencouraged to refer to a social work care manager. Thephysicians needed to have confidence in the socialworkers. Over time, this confidence developed. Usingpreliminary outcome data from this project as evidenceof value, Kaiser Permanente has since established abroad care management system for memory-impairedand frail elders in this service area. Since 2001, therehave been more than 3000 inquiries or requests for thecare management program to provide services; to meetthis demand, the number of care managers was expandedto the current 6. The Senior Care ManagementProgram currently averages 140 inquiries a month, andthe care managers carry an active caseload of 210patients, of whom 75% to 80% have dementia.

Kaiser Permanente is a staff model-type managedcare organization. Some components of this interventionmight be more feasible to implement in this typeof healthcare system than in solo or small practices, orin loosely organized practice models. Another limitationto be taken into consideration when interpretingevaluation findings is that our pre-post study design didnot include a comparison group (ie, a group that didnot receive the intervention). Thus, although thisstudy's findings suggest that the quality of dementiacare improved for intervention participants, other concurrentevents also may have been factors in study outcomes.For example, cholinesterase inhibitors becamemore common during the intervention period. Bothpharmaceutical company interventions and the availabilityof a treatment may have increased providerattention to these disorders. In addition, dementia is aprogressive condition, and changes in the severity ofdementia over time certainly occurred in the studysample. Without a control group, it is impossible toassign improvements in care solely to the intervention,as the worsening course of the disease itself could haveprompted clinicians to conduct additional assessmentsand management changes over time. Future studies ofrefinements to this dementia care intervention appliedin other healthcare settings should include a controlgroup.

Based on the experience of developing and implementingthe Dementia Care Project, we make the followingrecommendations to other providers ofdementia services:

  1. Chronic diseases such as Alzheimer's disease andthe other dementias demand a multidisciplinary,biopsychosocial approach to care. The addition ofsocial workers specially trained in dementia careand partnership with community-based organizationslike the Alzheimer's Association can helpensure that the multifaceted needs of thesepatients and their family supporters are met.Primary care providers alone cannot provide therange of care needed by people with dementia.
  2. Evidence- and consensus-based practice guidelinesare useful tools for targeting and improvingquality of care in a managed care setting, especiallywhen they are reinforced through opinionleader endorsement, provider-training programs,and provider practice support tools (eg, tool kits,computer-generated prompts), and by bringing inallied medical staff as part of a disease managementprogram.

From the Alzheimer's Association of Los Angeles, Calif (DLC, MP); the Department ofNeurology and the Alzheimer's Disease Center, University of California, Los Angeles (BGV);Kaiser Permanente, Southern California Region, Pasadena, Calif (LS); the Alzheimer'sAssociation, Chicago, Ill (EH); and Kaiser Permanente, Metropolitan Los Angeles Region,Calif (RY).

Funding for this study was provided by the Garfield Foundation and the Alzheimer'sAssociation, Chicago, Ill.

Address correspondence to: Debra L. Cherry, PhD, Alzheimer's Association, 5900Wilshire Boulevard, Suite 1100, Los Angeles, CA 90036. E-mail:


1. Boise L, Camicioli R, Morgan DL, Rose JH, Congleton L. Diagnosing dementia:perspectives of primary care physicians. 1999;39:457-464.

Ann Intern Med.

2. Callahan CM, Hendrie HC, Tierney WM. Documentation and evaluation ofcognitive impairment in elderly primary care patients. 1995;122:422-429.

Am J Alzheimer Dis.

3. Fortinsky RH, Wasson JH. How do physicians diagnose dementia? Evidencefrom clinical vignette responses. March/April 1997:51-61.


4. Doody RS, Stevens JC, Beck C, et al. Practice parameter: management ofdementia (an evidence-based review). . 2001;56:1154-1166.


5. Mittelman M, Ferris S, Shulman E, et al. Family intervention to delay nursinghome placement of patients with Alzheimer's disease. 1996;276(21):1725-1731.

J Gerontology

6. Zarit SH, Stephans M, Townsend A, Greene R. Stress reduction for family caregivers:effects of adult day care use. 1998;53(5):S267-S277.

7. Clinical Practice Guideline Number 19: Recognition and Initial Assessment ofAlzheimer's Disease and Related Dementias. Rockville, Md: US Dept of Healthand Human Services, Agency for Health Care Policy and Research; November1996. AHCPR publication 97-0702.

Am J Psychiatry

8. American Psychiatric Association. Practice guideline for the treatment ofpatients with Alzheimer's disease and other dementias of late life. .May 1997;154(suppl):1-39.

Diagnosis, Management and Treatment ofDementia: A Practice Guide for Primary Care Physicians.

9. Guttman R, Seleski M, eds. Chicago, Ill: AmericanMedical Association; 1999.

Am Fam Physician

10. Cummings JL, Frank J, Cherry DL, et al. Guidelines for managing Alzheimer'sdisease, part 1: assessment: . 2002;65(11):2263-2272.

Dis Manag HealthOutcomes

11. Maslow K, Selstad J, Denman SJ. Guidelines and care management issues forpeople with Alzheimer's disease and other dementias. . 2002;10(11):693-706.

Alzheimer Dis Assoc Disord.

12. Rosen CS, Chow HC, Greenbaum MA, et al. How well are clinicians followingdementia practice guidelines? 2002;16:15-23.

13. O'Kane M. A letter from the NCQA [National Committee for QualityAssurance] president. Available at: Accessed June 8, 2004.


14. Cherry DL. Dementia services in managed care: issues and trends.. 1999;23:24-29.


15. Evans DA, Funkenstein HH, Albert MS, et al. Prevalence of Alzheimer'sdisease in a community population of older persons. . 1989;262(18):2551-2556.

J Psychiatr Res

16. Folstein MF, Folstein SE. Mini Mental State: a practical method for grading thecognitive state of patients for the clinician. . 1985;12:189-198.

Ann Intern Med

17. Chow TW, MacLean CH. Quality indicators for dementia in vulnerable community-dwelling and hospitalized elders. . 2001;135:668-676.

Establishing Relations withCommunity Resource Organizations: An Imperative for Managed CareOrganizations Serving Medicare Beneficiaries

18. HMO Workgroup on Care Management. . Washington, DC: AmericanAssociation of Health Plans Foundation; 1999.