Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Oct 1.
Published in final edited form as: J Am Geriatr Soc. 2017 Sep 11;65(10):2235–2243. doi: 10.1111/jgs.15022

Nursing Home Use Across the Spectrum of Cognitive Decline: Merging Mayo Clinic Study of Aging with CMS MDS Assessments

Jane A Emerson 1, Carin Y Smith 1, Kirsten Hall Long 2, Jeanine E Ransom 1, Rosebud O Roberts 1,3, Steven L Hass 4, Amy M Duhig 4, Ronald C Petersen 1,3, Cynthia L Leibson 1
PMCID: PMC5657551  NIHMSID: NIHMS887892  PMID: 28892128

Abstract

BACKGROUND/OBJECTIVES

Objective, complete estimates of nursing home (NH) use across the spectrum of cognitive decline are needed to help predict future care needs and inform economic models constructed to assess interventions to reduce care needs.

DESIGN

Retrospective longitudinal study

SETTING

Olmsted County, MN

PARTICIPANTS

Mayo Clinic Study of Aging participants assessed as Cognitively Normal (CN), Mild Cognitive Impairment (MCI), Previously-unrecognized-dementia, or Prevalent-dementia (Age=70–89 years; N=3,545)

MEASUREMENTS

Participants were followed in the Centers for Medicare and Medicaid Services Minimum Data Set (including Medicare/Medicaid/commercially insured/private pay) and Rochester Epidemiology Project provider-linked medical records for 1-year after assessment for days-of-observation, NH use (yes/no), NH days, NH days/days-of-observation, and mortality.

RESULTS

In the year after cognition was assessed, for persons categorized as CN, MCI, previously-unrecognized-dementia, and prevalent-dementia respectively, the percentages who died were 1.0%, 2.6%, 4.2%, 21%; the percentages with any NH use were 3.8%, 8.7%, 19%, 40%; for persons with any NH use, median NH days were 27, 38, 120, 305, and median percentages of NH days/days-of-observation were 7.8%, 12%, 33%, 100%. The year after assessment, among persons with prevalent-dementia and any NH use, >50% were a NH resident all days-of-observation. Pairwise comparisons revealed that each increase in cognitive-impairment-category exhibited significantly higher proportions with any NH use. One-year mortality was especially high for persons with prevalent-dementia and any NH use (30% versus 13% for those with no NH use); 58% of all deaths among persons with prevalent-dementia occurred while a NH resident.

CONCLUSIONS

Findings suggest reductions in NH use could result from quality alternatives to NH admission, both among persons with MCI and persons with dementia, together with suitable options for end-of-life care among persons with prevalent-dementia.

Keywords: dementia, mild cognitive impairment, nursing home, long-term-care

INTRODUCTION

The burden of dementia on individuals, family members, care-providers, and society is widely-recognized. As the baby-boom generation ages and life expectancy continues to rise, the numbers of affected individuals will increase, and demands for medical- and long-term care will escalate.[1] Predictions are disconcerting. There is little evidence that underlying pathology is modified by existing pharmaceuticals designed to treat dementia.[1,2] Thus, research is expanding beyond pharmacological treatment after symptoms become apparent to 1) enhancing quality of life, improving care, and reducing care needs for persons living with dementia,[3,4] and 2) preventing/postponing disease progression by detecting and treating disease at earlier stages.[1,2,5–10] While awaiting clinical-trial confirmation, disease-modifying effects at earlier disease stages have been reported in observational studies of exercise, social engagement, and diet.[11–13] However, as noted by Gustavsson et al.,[6] if any treatments are confirmed as having a positive effect on underlying pathology, such effects must also be translatable as having real-life relevance for patients, care providers, and society-at-large (eg., improved quality of life, greater independence, reduced care needs, and lower costs).

With respect to translating the efficacy of disease-modifying treatments specifically to consequences for long-term care, analyses will require economic modeling.[6,14–18] Presently, source data for such models have ascertained cognitive status using discharge diagnosis codes or screening instruments and/or estimated utilization/ costs using self-report or single-source administrative data that fail to include all payers (i.e., Medicare, Medicaid, commercial insurance, and private pay). The serious limitations of these approaches for estimating associations between healthcare utilization and cognitive impairment are demonstrated.[14,19–22] The need for objective, complete measures of utilization by persons carefully characterized for cognition across the spectrum of cognitive decline is increasingly recognized.[6,14,15,18]

We endeavored to partially address this need in a previous [23] and the present study. Both studies employed two resources, the Rochester Epidemiology Project (REP)[24] and the Mayo Clinic Study of Aging (MCSA).[25] REP is a population-based records-linkage infrastructure affording access to detailed clinical information and line-item billing data for essentially all medical-care encounters by Olmsted County, MN, residents.[24] MCSA is a population-based longitudinal investigation of prevalence, incidence, and natural history of cognitive decline.[25] MCSA used REP resources to construct age- sex-stratified-random samples from all Olmsted County residents age 70–89 years. All medical records of sampled individuals underwent neurologist’s review to identify persons with sufficient documentation within their medical records to qualify them as meeting Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for a clinical diagnosis of dementia.[26] Such individuals were categorized as having ‘prevalent-dementia’.[27] Persons not meeting criteria for prevalent-dementia based on record-review were invited to participate in prospective assessments. In-person assessments included extensive nurse/coordinator and physician neurologic evaluations, and neuropsychological testing to assess cognitive function in four key domains: memory, executive function, language and visuospatial skills.[25] Persons who met DSM-IV criteria for dementia upon prospective examination, but lacked sufficient medical-record documentation, were categorized as ‘previously-unrecognized-dementia’. Persons prospectively assessed as meeting Key Symposium Working Group on Mild Cognitive Impairment (MCI) criteria[28,29] were categorized as ‘MCI’, a transitional state between normal and dementia. Prospective determination of cognitively normal (CN) individuals was based on Mayo’s Older Americans Normative Studies.[30] Assignment of cognitive status was determined by consensus of study physicians, nurses and neuropsychologists.[25]

Our previous study[23] used REP provider-linked detailed medical records and acute-care billing data to estimate medical-care utilization and costs for MCSA participants categorized as CN, MCI, previously-unrecognized-dementia, and prevalent-dementia. Long-term care was not addressed. The present study aims to investigate nursing home (NH) use for persons in each MCSA category by merging REP and MCSA resources with Centers for Medicare and Medicaid Services (CMS) Minimum Data Set (MDS).[31] MDS was developed following Omnibus Budget Reconciliation Act requirements that all Medicare- and Medicaid-certified long-term-care facilities assess each resident (regardless of payer) at specified times throughout their stay.[32] Each individual is assigned a unique identifier maintained for all admissions across all facilities. Regularly-scheduled standardized assessments include admission and discharge dates, thus affording NH residency status on each day. MDS assessments are a gold-standard source for objective, population-based data on NH activity.

The present study’s primary goal was to provide longitudinal estimates of NH days/days-of-observation as a function of MCSA’s cognitive-impairment-categorization scheme. Individuals were followed from the date cognition was assessed until earliest of death, emigration, or 1-year to identify those with and without any NH use, and among those with any NH use, whether they were a NH resident on each day of follow-up.

To our knowledge, of the very few estimates of NH activity for which all individuals were characterized across the cognitive-impairment spectrum using recommended criteria, none considered NH days.[33–35] The longitudinal estimates of NH days/days-of-observation provided here extend beyond cohort analyses of time to NH admission and cross-sectional analyses of any NH use (yes/no). Estimates provide empirical data to help inform economic models constructed to investigate the cost-effectiveness of alternative approaches to both care delivery and disease-modifying treatments. Estimates can also inform predictions of and planning for long-term-care needs by individuals, care providers, and policy makers.

METHODS

This retrospective longitudinal study is set in Olmsted County, MN, and was approved by Mayo Clinic and Olmsted Medical Center Institutional Review Boards. Persons who declined authorization for use of medical records in research[36] were excluded.

Resources/Participants

REP and MCSA resources are briefly described above. Detailed descriptions are provided in Supplementary Text S1 and elsewhere.[24,25,37] For this study, MCSA sampling of all Olmsted County residents age 70–89 years was conducted on 10/01/2004 and/or 03/01/2008 (N=6,682). Supplementary Figure S1 provides a flow chart outlining study participant identification and categorization into cognitive categories. The final 3,545 participants included 461 with prevalent-dementia determined from record review plus 3,084 prospectively assessed as previously-unrecognized-dementia (N=118), MCI (N=528), or CN (N=2,438). Among prospectively-assessed persons, 2,421 were assessed in-person; 663 declined the in-person assessment but participated via telephone using the Modified Telephone Interview for Cognitive Status.[37–39]

Data Collection

Prospectively-assessed participants were assigned the cognitive status as of their initial assessment, with index date defined as assessment date. For participants with prevalent-dementia determined from record review, index date was defined as date of record review. Index dates ranged from 11/02/2004–08/02/2010.

Using MDS data from 10/01/1998–09/30/2010 (dates MDS 2.0 was obtained from CMS), we identified each unique NH resident with any assessment in any NH within or on the border of Olmsted County, MN (subsequently referred to as ‘local nursing homes’). The listing of local NH residents was then merged with the 3,545 MCSA study participants using information from REP and MDS assessments (Supplementary Figure S2, Text S2). Total NH days for each individual were accumulated 1-year after and 1-year before index separately; proportion of time in NH was estimated as total NH days/days-of-observation. All individuals had 365 days-of-observation before index. After index, days-of-observation was defined from index until earliest of death, emigration, or maximum of 1-year.

Data collected on or before index included certain characteristics associated with NH activity in the literature, including sex, age, race, marital status, education, and comorbidity.[40,41] Comorbidity was assessed using Johns Hopkins Adjusted Clinical Groups System® software to calculate a Resource Utilization Band (RUB) summary measure of comorbidity for each individual.[42]

Additional characteristics related to NH activity were obtained for a stratified-random subsample of participants. Due to small numbers, previously-unrecognized-dementia was combined with prevalent-dementia. Randomization was stratified by sex, age at index (70–79 and ≥80), and cognitive category to obtain approximately 50% CN, 25% MCI, and 25% dementia. Detailed provider-linked REP medical records of 400 individuals (CN=185, MCI=96, dementia=119) were reviewed for information on residential setting, gait-aid use, and dependency level by an experienced registered nurse abstractor [J.A.E.], blinded as to cognitive category. Abstraction began with the encounter closest to, but before index; absent sufficient information, review continued in a backward fashion. The choice of variables/values depended on whether information was frequently and reliably available within the medical record, eg., information on gait-aid use/type was routinely available; household income was unavailable. Values for ‘residential setting’ (lives alone versus with spouse/care-giver versus assisted living versus NH/other 24-hour-care facility), as well as variables ‘dependency level’ and ‘gait-aid use’ are indicative of availability of care, need for care, and mobility, each of which have been associated with NH activity in the literature.[40,41,43]

Statistical Analysis

Comparisons were conducted using chi-square tests for dichotomous variables, Mantel-Haenszel chi-square tests for ordinal variables, and Kruskal Wallis or Wilcoxon rank sum tests for continuous variables. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). Testing used a two-tailed alpha level of 0.05 to determine statistical significance.

RESULTS

Table 1 compares characteristics as of index for all 3,545 study participants. As cognitive-impairment category increased, age and comorbidity increased, education and percent married decreased. Pairwise analyses were conducted. Compared to CN individuals, persons with MCI were more likely male (p=0.03), older (p<0.001), and had less education (p<0.001) and greater comorbidity (p<0.001). Compared to persons with MCI, persons with previously-unrecognized-dementia were older (p=0.002) with less education (p=0.005) but similar comorbidity (p=0.73). Compared to persons with previously-unrecognized-dementia, persons with prevalent-dementia were similar for all available characteristics (p=0.20–0.84).

Table 1.

Characteristics as of Index Date by Cognitive Category for All 3,545 Study Participants

Cognitively Normal (2,438) Mild Cognitive Impairment (528) Previously Unrecognized Dementia (118) Prevalent Dementia (461) P Valuea
Male 0.10
 N (%) 1,137 (47%) 274 (52%) 52 (44%) 208 (45%)
Age, years <0.001
 Mean ± SD 79 ± 5.2 81 ± 5.0 83 ± 4.9 83 ± 4.5
 Median (IQR) 79 (74, 83) 82 (79, 85) 84 (80, 86) 84 (81, 87)
White race 0.053
 N (%) 2,401 (98%) 517 (98%) 113 (96%) 448 (97%)
Marriedb 0.006
 N (%) 1,505 (62%) 307 (58%) 57 (48%) NA
Education, years <0.001
 Mean ± SD 14 ± 3.0 13 ± 3.0 12 ± 3.3 12 ± 3.2
 Median (IQR) 13 (12, 16) 12 (12, 14) 12 (10, 14) 12 (11, 14)
RUBc <0.001
 N (%)
  No/ healthy/ low use 142 (5.8%) 20 (3.8%) 5 (4.2%) 29 (6.3%)
  Moderate use 892 (37%) 144 (27%) 36 (30%) 126 (27%)
  High use 781 (32%) 195 (37%) 38 (32%) 123 (27%)
  Very high use 623 (26%) 169 (32%) 39 (33%) 183 (40%)
Open in a new tab

SD = Standard deviation, IQR = Interquartile range, RUB = Resource Utilization Band

a

P values were calculated using Chi Square tests for categorical variables and Kruskal Wallis tests for continuous variables.

b

For CN individuals, persons with MCI and persons with previously-unrecognized-dementia, marital status was available from data collected prospectively for MCSA. For persons with prevalent-dementia based on record review, marital status was obtained electronically from REP medical records, and was missing for all but 32 individuals, 18 (56%) of whom were married.

c

RUB is a summary measure of co-morbidity, defined using aggregations of diagnostic groupings that have similar expected resource use, with values ranging from 0 (no encounters) to 5 (diagnosis codes associated with very high use)[42]; we calculated RUB values using diagnosis codes assigned in the year before index and excluding codes for dementia and MCI (ICD-9-CM codes 290.XX, 331.82, 331.83)

Table 2 provides information for the 400 individuals on whom characteristics before index were abstracted from medical records. Previously-unrecognized-dementia and prevalent-dementia were combined. Across-category differences in sex, age, and education were similar to those in Table 1. Pairwise analyses revealed differences between CN individuals and persons with MCI for residential status (p=0.002) and dependency (p<0.001); only 2.2% of CN individuals were in assisted living, NH, or other 24-hour-care facility versus 12% of persons with MCI; 19% of CN individuals received assistance from another individual versus 37% of persons with MCI. Both CN individuals and persons with MCI differed from persons with dementia for residency, gait-aid use, and dependency (all p<0.001). Among 119 persons with dementia, while nearly half were in a NH or other 24-hour-care facility, 13% lived alone; 55% used gait-aids; wheelchair/unable to walk was primary mode of mobility for 24%; 54% (n=64) needed 24-hour care/supervision; however, 42% (27/64) received that care outside a NH.

Table 2.

Characteristicsa by Cognitive Category for a Stratified Random Subset of 400 Individuals Drawn from All 3,545 Study Participants

Cognitively Normal (185) MCI (96) Dementia (119) P valueb
Male, N (%) 92 (50%) 45 (47%) 60 (50%) 0.86
Age, years <0.001
 Mean ± SD 79 ± 5.2 81 ± 5.0 82 ± 5.1
 Median (IQR) 79 (74, 83) 81 (77, 84) 82 (78, 86)
Education, years <0.001
 Mean ± SD 13.9 ± 2.9 12.6 ± 2.7 11.9 ± 3.3
 Median (IQR) 13 (12, 16) 12 (12, 14) 12 (10, 14)
Residential settinga, N (%) <0.001
 Alone 52 (28%) 34 (35%) 16 (13%)
 With spouse / care giver 124 (67%) 49 (51%) 36 (30%)
 Assisted living 2 (1.1%) 9 (9.4%) 11 (9.2%)
 NH or other 24-hour-care facility 2 (1.1%) 3 (3.1%) 56 (47%)
 Unknown 5 (2.7%) 1 (1.0%) 0 (0%)
Use of gait aida,c, N (%) <0.001
 No gait aid used 157 (85%) 68 (71%) 45 (38%)
 Cane 13 (7.0%) 10 (10%) 7 (5.9%)
 Walker 9 (4.9%) 11 (11%) 29 (24%)
 Wheelchair (including “unable to walk”) 3 (1.6%) 4 (4.2%) 29 (24%)
 Unknown 3 (1.6%) 3 (3.1%) 9 (7.6%)
Dependency levela,d, N (%) <0.001
 Grade 1 150 (81%) 60 (62%) 5 (4.2%)
 Grade 2 33 (18%) 32 (33%) 50 (42%)
 Grade 3 2 (1.1%) 4 (4.2%) 64 (54%)
Open in a new tab

SD = Standard deviation, IQR = Interquartile range, NH = Nursing Home

a

As determined based on manual review of Rochester Epidemiology Project provider-linked medical records. Information was obtained for encounter closest to but before index. Absent sufficient information from that encounter, abstraction continued in a backward fashion.

b

P values were calculated using chi-square tests for categorical variables and Kruskal Wallis tests for continuous variables

c

Gait aid was based on the primary mode of mobility.

d

Level of dependency was graded on a scale of one to three, based on need for assistance from another individual. Assistance with activities such as house cleaning, gardening, shopping, and transportation outside the home were not considered in our grading. Grade one was defined as able to perform usual activities without assistance. Grade two was assigned to individuals requiring assistance on a regular basis (including daily) with activities such as medication set-up, meal preparation, hygiene, or other activities of daily living, but who could be left alone in their residence. Grade three was defined as needing 24-hour care or supervision, either in-home or in another supervised setting.

Table 3 provides NH activity among all 3,545 individuals for years after and before index separately. Before index, all individuals had 365 days-of-observation. After index, increasing cognitive-impairment category was associated with fewer days-of-observation and higher mortality (21% of persons with prevalent-dementia died versus <5% in other categories). Within each cognitive category, the proportion with any NH use, and among persons with any NH use, median NH days/days-of-observation appeared generally higher the year after versus the year before index. After index, the proportion with any NH use increased >2-fold with each successive category; among persons with any NH use, median values for NH days/days-of-observation ranged from 7.8% for CN individuals to 100% for persons with prevalent-dementia, i.e., among the 40% of persons with prevalent-dementia and any NH use 1-year after index, >50% were a NH resident the entire time.

Table 3.

Nursing Home Activity for All 3,545 Study Participants in the Year After Index and the Year Before Index Separately.

Cognitively Normal (2,438) Mild Cognitive Impairment (528) Previously Unrecognized Dementia (118) Prevalent Dementia (461) P Valuea
Days-of-observationb
 After Index <0.001
  Mean ± SD 349 ± 59 342 ± 71 313 ± 97 324 ± 91
  Median (IQR) 366 (366, 366) 366 (366, 366) 366 (366, 366) 366 (366, 366)
Diedb, N (%)
 After Index 25 (1.0%) 14 (2.6%) 5 (4.2%) 95 (21%) <0.001
Any NH use, N (%)
 After Index 93 (3.8%) 46 (8.7%) 22 (19%) 183 (40%) <0.001
 Before Index 75 (3.1%) 24 (4.6%) 14 (12%) 162 (35%) <0.001
Total NH days (among those with any NH use)
 After Index <0.001
  Mean ± SD 62 ± 92 80 ± 109 161 ± 138 239 ± 139
  Median (IQR) 27 (15, 57) 38 (16, 77) 120 (31, 308) 305 (93, 366)
 Before Index <0.001
  Mean ± SD 49 ± 92 91 ± 132 154 ± 164 242 ± 142
  Median (IQR) 15 (9.0, 29) 26 (14, 79) 48 (26, 365) 357 (85, 365)
% NH day/days-of-observation (among those with any NH use)
 After Index <0.001
  Mean ± SD 19 ± 27 23 ± 29 44 ± 37 80 ± 34
  Median (IQR) 7.8 (4.1, 16) 12 (5.5, 24) 33 (9.1, 84) 100 (67, 100)
 Before Index <0.001
  Mean ± SD 13 ± 25 25 ± 36 42 ± 45 66 ± 39
  Median (IQR) 4.1 (2.5, 7.9) 7.3 (3.8, 22) 13 (7.1, 100) 98 (23, 100)
Open in a new tab

SD = Standard deviation, IQR = Interquartile range, NH = Nursing Home

a

P values were calculated using Chi Square tests for categorical variables and Kruskal Wallis tests for continuous variables.

b

During the period from index until earliest of death, emigration, or maximum of 1-year after index. By design, all subjects were alive as of index and in Olmsted County the full year before index.

P values for pairwise comparisons after index between persons with CN versus MCI, MCI versus previously-unrecognized-dementia, and previously-unrecognized-dementia versus prevalent-dementia respectively were 0.003, 0.36, and <0.001 for mortality; <0.001, 0.002, and <0,001 for percent with any NH use; and among persons with any NH use, 0.21, 0.008, and 0.02 for NH days, and 0.12, 0.008, and <0.001 for NH days/days-of-observation. Table 4 provides comparisons between persons with and without NH use within each cognitive category for both years combined. Sex was significantly associated with NH use for CN individuals only; older age was significantly associated with NH use for persons in each category, except prevalent-dementia. Being married was inversely associated with NH use for CN individuals and persons with MCI. Education was not significantly associated with NH use in any category. Except for persons with previously-unrecognized-dementia, the percentage with RUB comorbidity predictive of very high medical-care use was >50% for persons with NH use versus ≤30% for persons without NH use. Days-of-observation was similar between persons with and without NH use, except for persons with prevalent-dementia. Among persons with prevalent-dementia, mortality post-index differed greatly between persons with (30%) and without NH use (13%). Among persons with prevalent dementia, 58% of all deaths occurred while a NH resident.

Table 4.

Characteristics within Each Cognitive Category for All 3,545 Study Participants by Any Nursing Home Use (No/Yes)a during the Year Before and Year After Combined

Cognitively Normal (2,438) Mild Cognitive Impairment (528) Previously Unrecognized Dementia (118) Prevalent Dementia (461)
Nursing Home Use No(2,283) Yes (155) No (465) Yes(63) No(90) Yes(28) No(261) Yes(200)
Male P<0.001 P=0.13 P=0.88 P=0.20
 N (%) 1,090 (48%) 47 (30%) 247 (53%) 27 (43%) 40 (44%) 12 (43%) 111 (42%) 97 (48%)
Age, years P<0.001 P<0.001 P=0.01 P=0.16
 Median (IQR) 79(74,83) 82(79,85) 82(78,84) 84(81,87) 83 (80,85) 86 (82,87) 84 (80,87) 84 (81,87)
White race P=0.08 P=0.63 P=0.34 P=0.41
 N (%) 2,251 (99%) 150 (97%) 456 (98%) 61 (97%) 85 (94%) 28 (100%) 252 (96%) 196 (98%)
Married P<0.001 P=0.04 P=0.82 NAb
 N (%) 1,439 (63%) 66 (43%) 278 (60%) 29 (46%) 44 (49%) 13 (46%) - -
Education, years P=0.80 P=0.52 P=0.23 P=0.054
 Median (IQR) 13(12,16) 13(12,16) 12(12,14) 12(12,16) 12 (10,14) 12 (8,13) 12 (12,14) 12 (10,14)
RUBc P<0.001 P<0.001 P=0.40 P<0.001
 No/healthy/low use, N (%) 140 (6.1%) 2 (1.3%) 15(3.2%) 5(7.9%) 3 (3.3%) 2 (7.1%) 25 (9.6%) 4(2.0%)
 Moderate use, N (%) 862 (38%) 30 (19%) 138(30%) 6 (9.5%) 32 (36%) 4 (14%) 87(33%) 39 (20%)
 High use, N (%) 737 (32%) 44 (28%) 179(38%) 16(25%) 26 (29%) 12 (43%) 72(28%) 51 (26%)
 Very high use, N (%) 544 (24%) 79 (51%) 133 (29%) 36 (57%) 29 (32%) 10 (36%) 77 (30%) 106 (53%)
Days-of-observationd P=0.10 P=0.63 P=0.054 P<0.001
 Mean ± SD 715 ± 59 710 ± 62 706 ± 72 713 ± 64 669 ± 103 709 ± 68 700 ± 80 674 ± 103
 Median 731 731 731 731 731 731 731 731
Died year after index P<0.001 P=0.07 P=0.09 P<0.001
 N (%) 14 (0.6%) 11(7.1%) 10 (2.2%) 4 (6.4%) 2 (2.2%) 3 (11%) 35 (13%) 60 (30%)
Open in a new tab

SD = Standard deviation, IQR = Interquartile range, RUB = Resource Utilization Band

a

P values were calculated using chi-square tests for categorical variables and Wilcoxon rank sum tests for continuous variables. Mean and SD are provided for observation days as reference.

b

For persons with prevalent-dementia, marital status was obtained electronically from REP medical records and was missing for all but 32 individuals.

c

RUB is a summary measure of co-morbidity, defined using aggregations of diagnostic groupings that have similar expected resource use, with values ranging from 0 (no encounters) to 5 (diagnosis codes associated with very high use)[42]; we calculated RUB values using diagnosis codes assigned in the year before index and excluding codes for dementia and MCI (ICD-9-CM codes 290.XX, 331.82, 331.83)

d

Defined as days of Olmsted County residency, from 1-year before index until the earliest of death, emigration, or maximum of 1-year after index. By design, all subjects were alive as of index and in Olmsted County the full year before index.

DISCUSSION

This study provides complete, objective estimates of NH activity for persons well-characterized as to cognition. Estimates were afforded by merging data on MCSA participants who met criteria for CN, MCI, previously-unrecognized-dementia, or prevalent-dementia while resident of Olmsted County, MN, with CMS MDS NH assessments available for all local nursing home residents and for all payers (Medicare/Medicaid/commercial insurance/private pay).

Across-category comparisons 1-year after index revealed increasing cognitive-impairment category was associated with increasing proportions with any NH use, and among persons with any NH use, more NH days and higher NH days/days-of-observation. Results were unchanged following regression modeling adjusted for age, sex, education, and RUB co-morbidity [data not shown—available on request]. Pairwise comparisons between MCI and previously-unrecognized-dementia and between previously-unrecognized-dementia and prevalent-dementia revealed persons in the higher cognitive impairment category exhibited significantly higher values for each measure of NH activity. Additional pairwise comparisons for the subset of 400 individuals revealed the proportions in assisted living or NH/other 24-hour-care facility was less for CN individuals versus persons with MCI and for persons with MCI versus persons with dementia (previously-unrecognized and prevalent combined). While >50% of persons with dementia required 24-hour care/supervision, 13% of all persons with dementia lived alone; and among those requiring 24-hour care/supervision, 42% received that care outside a NH. These estimates have relevance for modeling the cost-effectiveness of community-care versus institutional-care. Existing models have focused almost exclusively on dementia,[4,6,18,44,45], with few exceptions,[4] dementia was determined using discharge-diagnosis codes. Future model construction can benefit from estimates provided here for persons well-characterized across the cognitive spectrum.

We observed higher mortality for persons with versus those without any NH use in each cognitive category; differences reached significance for CN individuals and persons with prevalent-dementia. The risk of death among NH residents is known to be especially high for persons with dementia,[46] We observed that 30% of persons with prevalent-dementia and any NH use died the year after index; 58% of all deaths among persons with prevalent-dementia occurred while a NH resident. Among the 40% of individuals with prevalent-dementia and any NH use the year after index, >50% were a NH resident the full time. The high proportion with any NH use, extended stays, and high mortality reinforce the importance of existing studies and need for additional research exploring various end-of-life care options for NH residents with dementia.[3,47,48]

Our observation of higher NH activity by persons with prevalent- versus previously-unrecognized-dementia reflects others’ findings that NH use 1–2 years before persons received a clinical diagnosis of dementia was higher than NH use in that period among persons not subsequently diagnosed.[49–51] While our and others’ findings might suggest that earlier detection/treatment of dementia could result in reduced NH use, findings may reflect ascertainment bias. Others’ studies used clinically-diagnosed dementia; we used medical-record-documentation of DSM-IV criteria to define ‘prevalent-dementia’. Absence of clinical diagnosis/documentation could occur if individuals avoided frequent medical contact or were reluctant to mention dementia-related symptoms. Importantly, such bias does not explain higher NH activity we observed in pairwise comparisons for persons prospectively assessed as CN versus MCI versus previously-unrecognized-dementia; no record review was involved. Nevertheless, findings fail to directly substantiate suggestions that earlier detection and disease-modifying interventions that delay progression from CN to MCI or MCI to previously-unrecognized-dementia would result in reduced NH activity;[16,17,52] the independent contribution of cognition to NH activity could not be assessed absent potentially confounding characteristics (eg., household income, long-term-care insurance).

Few investigations of NH activity have identified cognitive status using preferred criteria for either MCI[28,29] or Cognitive Impairment No Dementia (CIND).[53] Tukko et al.[33] assessed Canadian Study of Health and Aging participants for cognition at baseline; 5-years later, persons with CIND at baseline were more likely to have been institutionalized than persons without cognitive impairment at baseline. Gnjdic et al.[34] assessed Australian men at baseline as normal or MCI; between baseline and 3.4 years, the two groups exhibited similar times to NH admission. However, time to admission 3.4 years to end-of-follow-up (post-baseline average=5 years) was shorter for persons assessed at baseline as MCI versus normal. As Gnjidic et al. recognized,[34] differences between persons assessed at baseline as normal versus MCI may be overestimated if MCI at baseline progressed to dementia over extended observation.[54] We attempted to offset concerns regarding progression of MCI after index by limiting post-index NH activity to 1-year and separately analyzing NH activity 1-year before index, i.e., during which time MCI may have progressed from CN.

While Tuokko et al.’s[33] and Gnjdic et al.’s[34] findings are broadly consistent with our results from pairwise comparisons between CN and MCI, our results differ from a recent report by Ton et al.[35] using Aging, Demographics, and Memory Study (ADAMS) data. ADAMS is a nationally-representative sample, with all participants prospectively assessed as normal or MCI using methodology similar to that used in MCSA; participants are also prospectively assessed as mild, moderate, or severe dementia.[55] Ton et al.[35] focused on amnestic MCI (aMCI) and mild, moderate, and severe Alzheimer’s dementia (AD). NH activity was limited to self-report of any NH use 2-years before assessment. When analyzed across all categories, the proportion of persons with any NH use increased significantly; the proportion with any NH use was significantly lower for persons with aMCI versus persons with severe AD. However, in contrast to our findings of higher NH use for persons with MCI versus CN individuals and for persons with previously-unrecognized-dementia versus MCI. Ton et al’s. point estimates were lower for aMCI versus normal individuals; confidence intervals overlapped when comparing aMCI with normal individuals and when comparing aMCI with mild and moderate AD individuals; the authors recognized the limitations of small sample sizes and the cross-sectional design, noting the need for future studies to examine utilization after assessing cognition.[35]

Strengths

Study strengths include accurate assignment of CN, MCI, and previously-unrecognized-dementia;[25,37] and objective, essentially complete data on NH use for all study participants. Merging MCSA, REP, and MDS resources allowed consideration of loss-to-follow-up, days-of-observation, and the proportion of time spent in a NH. All NH days were captured, including multiple admissions and short stays. Resulting estimates address limitations recognized in the literature,[14,15,19–22] and afford information on NH activity beyond cross-sectional estimates of the proportion of NH residents with cognitive impairment, which can be biased toward long-term residents, as well as findings limited to any NH use (yes/no) or time to NH admission.

Limitations

Estimates are for a single population (86% white in 2010). Compared with Minnesota and all other upper-mid-west states, Olmsted County residents exhibit higher income and education, but very similar chronic disease prevalence rates and age-, sex-, and racial-distributions.[56] Low racial diversity and higher income/education levels could compromise the generalizability of findings to different racial and socioeconomic groups.

Medical and NH care is delivered by few providers, potentially compromising the generalizability of findings to different health-care environments. However, published data reveal 680 Olmsted County residents were in a nursing facility/skilled-nursing facility in 2010, which was 4.1% of the County population aged 65+.[57] The rate is similar to 3.7% of MN residents age 65+ in a NH in 2011[58] and 3.9% for US white Medicare beneficiaries living in a long-term-care facility in 2011.[59]

Prevalent-dementia was based on neurologist’s application of DSM-IV criteria following detailed medical-record review;[27] reliable information on dementia severity were unavailable. Prospective assessments were limited to eligible persons who agreed to participate; some assessments were conducted via telephone (Supplemental Figure S1). Previous investigations reveal that participants who refused participation were older, more likely male, with higher comorbidity (eg, diabetes); participation was not associated with history of stroke, hypertension, coronary heart disease, marital status, or prior clinical diagnosis of MCI or dementia.[25] Furthermore, persons who declined prospective participation but did not refuse use of medical records in research[36] were followed in REP records for median 3.9 years after declining prospective participation to estimate dementia incidence using neurologist’s application of DSM-IV criteria; rates were similar to rates for participants followed prospectively in MCSA serial examinations.[60]

Characteristics associated with NH use did not include 1) several variables identified in the literature as contributing to NH use or 2) standardized criteria for residential setting, gait-aid use, and dependency level. Findings do not include NH costs, household income/expenditures, or informal (unpaid) care use.

Conclusions/Implications

Compared to CN individuals, persons with MCI exhibited higher mortality and a greater proportion with any NH use. For every measure of NH activity under consideration, each pairwise comparison between persons categorized as MCI versus previously-unrecognized-dementia and between previously-unrecognized-dementia versus prevalent-dementia revealed higher activity for the higher cognitive-impairment category. Findings suggest reductions in NH use could result from identifying 1) quality alternatives to NH admission for both persons with MCI and persons with dementia, as well as 2) preferable options for reducing the proportion of time spent as a NH resident by persons with dementia, especially for those with terminal illness. Further investigation is needed to explore reasons underlying the higher NH activity that was observed with increasing cognitive-impairment category. Findings provided here support arguments for earlier detection to help individuals, care providers, and policy makers predict and prepare for future care needs. Estimates can also help inform future efforts to model potential consequences both of disease-modifying treatments for reducing NH utilization/costs and alternative approaches to care-delivery.

Supplementary Material

Supp FigS1. Supplementary Figure S1.

Flow chart outlining the process used to identify the subset consisting of 3,545 Mayo Clinic Study of Aging participants in this investigation, including enumeration and characterization by cognitive impairment category.

Supp FigS2. Supplementary Figure S2.

Flow chart outlining the process used to merge the 3,545 Mayo Clinic Study of Aging participants included in this investigation with Centers for Medicare and Medicaid Services Minimum Data Set information on all nursing home activity within local nursing homes among Olmsted County, MN, residents from 10/01/1998 through 09/30/2010.

Supp TextS1. Supplementary Text S1.

Detailed descriptions of two data sources, Rochester Epidemiology Project and Mayo Clinic Study of Aging.

Supp TextS2. Supplementary Text S2.

Text describing data resources and variables used to merge the 3,545 Mayo Clinic Study of Aging participants included in this investigation with Centers for Medicare and Medicaid Services Minimum Data Set nursing home assessments. Text also includes details regarding how nursing home stays and days were determined for each individual.

Acknowledgments

We thank Ms. Deborah Strain for expert assistance with manuscript preparation and submission. We also wish to especially thank the MCSA study participants for their generous contributions of time and information. As corresponding author, I affirm that I have listed everyone who contributed significantly to the work and, with the exception of study participants, have obtained written consent from all contributors who are not authors and are named in the Acknowledgment section.

Funding:

This work was supported by AbbVie, Department of Health Economics and Outcomes Research (HEOR) and the Mayo Clinic Study of Aging [National Institutes of Health grant number U01 AG006786]. Some study data were obtained from the Rochester Epidemiology Project, which is supported by the National Institute on Aging [National Institutes of Health grant number R01 AG034676].

Disclosures:

All authors, except SLH, AMD, and KHL, were funded in part by AbbVie, Department of HEOR. SLH and AMD are former employees of AbbVie, Department of HEOR. KHL had a subcontract with Mayo Clinic on the AbbVie, Department of HEOR-funded study. Ronald C. Petersen also notes the following disclosures: Roche, Inc., Merck, Inc., Genentech, Inc., Biogen, Inc., Eli Lilly and Company, consultant. All authors have provided full disclosure of financial, personal, and potential conflicts of interest. With the exception of items noted here and in the manuscript file, none of the authors have additional conflicts of interest.

Authors Contributions:

The following criteria were met by the listed co-authors:

  1. All authors (Cynthia Leibson, Jane Emerson, Steven Hass, Amy Duhig, Ronald Petersen, Carin Smith, Jeanine Ransom, Kirsten Hall Long, Rosebud Roberts) made substantial contributions—specific contributions are as follows:

    1. Conception and design—all authors

    2. Acquisition of data—Emerson, Smith, Roberts, Petersen

    3. Analysis and interpretation of data—Smith, Ransom, Long

  2. All authors were involved with drafting the article or revising it critically for important intellectual content.

  3. All authors gave final approval of the version to be published.

Sponsors’ Roles:

In support of the manuscript, the National Institutes of Health had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. AbbVie was involved with the initial study design; AbbVie had no role in the collection, analysis, or interpretation of data or in the preparation, review, or approval of the manuscript.

Footnotes

Parts of this manuscript were presented at the National Academy of Neuropsychology 33rd Annual Conference October 16-19, 2013, San Diego, CA.

References

  • 1.National Plan to Address Alzheimer’s Disease: 2016 Update. U.S. Department of Health & Human Services (online); [Accessed April 24, 2017]. Available at: https://aspe.hhs.gov/report/national-plan-address-alzheimers-disease-2016-update. [Google Scholar]
  • 2.Selko D. Preventing Alzheimer’s Disease. Science. 2012;337:1488–1492. doi: 10.1126/science.1228541. [DOI] [PubMed] [Google Scholar]
  • 3.Callahan CM, Sachs GA, LaMantia MA. Redesigning systems of care for older adults with Alzheimer’s disease. Health Aff (Millwood) 2014;33:626–632. doi: 10.1377/hlthaff.2013.1260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Long KH, Moriarty JP, Mittelman MS, et al. Estimating the potential cost savings from the New York University Caregiver Intervention in Minnesota. Health Aff (Millwood) 2014;33:596–604. doi: 10.1377/hlthaff.2013.1257. [DOI] [PubMed] [Google Scholar]
  • 5.Petersen RC. Challenges of epidemiological studies of mild cognitive impairment. Alzheimer Dis Assoc Disord. 2004a;18:1–2. doi: 10.1097/00002093-200401000-00001. [DOI] [PubMed] [Google Scholar]
  • 6.Gustavsson A, Green C, Jones RW, et al. Current issues and future research priorities for health economic modelling across the full continuum of Alzheimer’s disease. Alzheimers Dement. 2017;13:312–321. doi: 10.1016/j.jalz.2016.12.005. [DOI] [PubMed] [Google Scholar]
  • 7.Weiner MW, Veitch DP, Aisen PS, et al. Recent publications from the Alzheimer’s Disease Neuroimaging Initiative: Reviewing progress toward improved AD clinical trials. Alzheimers Dement. 2017;13:e1–e85. doi: 10.1016/j.jalz.2016.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ahmed RM, Paterson RW, Warren JD, et al. Biomarkers in dementia: clinical utility and new directions. J Neurol Neurosurg Psychiatry. 2014;85:1426–1434. doi: 10.1136/jnnp-2014-307662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Karakaya T, Fußer F, Schröder J, et al. Pharmacological treatment of mild cognitive impairment as a prodromal syndrome of Alzheimer’s Disease. Curr Neuropharmacol. 2013;11:102–108. doi: 10.2174/157015913804999487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Baumgart M, Snyder HM, Carrillo MC, et al. Summary of the evidence on modifiable risk factors for cognitive decline and dementia: A population-based perspective. Alzheimers Dement. 2015;11:718–726. doi: 10.1016/j.jalz.2015.05.016. [DOI] [PubMed] [Google Scholar]
  • 11.Belleville S, Gilbert B, Fontaine F, et al. Improvement of episodic memory in persons with mild cognitive impairment and healthy older adults: evidence from a cognitive intervention program. Dement Geriatr Cogn Disord. 2006;22:486–499. doi: 10.1159/000096316. [DOI] [PubMed] [Google Scholar]
  • 12.Boyle PA, Buchman AS, Wilson RS, et al. Effect of purpose in life on the relationship between Alzheimer’s disease pathological changes in cognitive function in advanced age. Arch Gen Psychiatry. 2012;69:499–504. doi: 10.1001/archgenpsychiatry.2011.1487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Morris MC, Tangney CC, Wang Y, et al. MIND diet slows cognitive decline with aging. Alzheimers Dement. 2015;11:1015–1022. doi: 10.1016/j.jalz.2015.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Lin PJ, Neumann PJ. The economics of mild cognitive impairment. Alzheimers Dement. 2013;9:58–62. doi: 10.1016/j.jalz.2012.05.2117. [DOI] [PubMed] [Google Scholar]
  • 15.Wimo A, Winblad B. Pharmacoeconomics of mild cognitive impairment. Acta Neurol Scand Suppl. 2003;179:94–99. doi: 10.1034/j.1600-0404.107.s179.13.x. [DOI] [PubMed] [Google Scholar]
  • 16.Barnett JH, Lewis L, Blackwell AD, et al. Early intervention in Alzheimer’s disease: a health economic study of the effects of diagnostic timing. BMC Neurol. 2014;14:101–109. doi: 10.1186/1471-2377-14-101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Zissimopoulos J, Crimmins E, St Clair P. The value of delaying Alzheimer’s disease onset. Forum Health Econ Policy. 2014;18:25–39. doi: 10.1515/fhep-2014-0013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Lueke S, Hoffmann W, Flebetaa S. Transitions between care settings in dementia: are they relevant in economic terms? Value Health. 2014;17:679–685. doi: 10.1016/j.jval.2014.06.008. [DOI] [PubMed] [Google Scholar]
  • 19.Taylor DH, Jr, Ostbye T, Langa KM, et al. The accuracy of Medicare claims as an epidemiological tool: the case of dementia revisited. J Alzheimers Dis. 2009;17:807–815. doi: 10.3233/JAD-2009-1099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Murman DL, Von Eye A, Sherwood PR, et al. Evaluated need, costs of care, and payer perspective in degenerative dementia patients cared for in the United States. Alzheimer Dis Assoc Disord. 2007;21:39–48. doi: 10.1097/WAD.0b013e31802f2426. [DOI] [PubMed] [Google Scholar]
  • 21.van Dalen MT, Suijker JJ, MacNeil-Vroomen J, et al. Self-report of healthcare utilization among community-dwelling older persons: a prospective cohort study. [Accessed April 24, 2017];PLoS One. 2014 doi: 10.1371/journal.pone.0093372. (online) Available at: http://doi.org/10.1371/journal.pone.0093372. [DOI] [PMC free article] [PubMed]
  • 22.Hunger M, Schwarzkopf L, Heier M, et al. Official statistics and claims data records indicate non-response and recall bias within survey-based estimates of health care utilization in the older population. BMC Health Serv Res. 2013;13:1–11. doi: 10.1186/1472-6963-13-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Leibson CL, Long KH, Ransom JE, et al. Direct medical costs and source of cost differences across the spectrum of cognitive decline: a population-based study. Alzheimers Dement. 2015;11:917–932. doi: 10.1016/j.jalz.2015.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.St Sauver JL, Grossardt BR, Yawn BP, et al. Data resource profile: the Rochester Epidemiology Project (REP) medical records-linkage system. Int J Epidemiol. 2012a;41:1614–1624. doi: 10.1093/ije/dys195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Roberts RO, Geda YE, Knopman DS, et al. The Mayo Clinic Study of Aging: design and sampling, participation, baseline measures and sample characteristics. Neuroepidemiology. 2008;30:58–69. doi: 10.1159/000115751. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) 4. Washington, DC: American Psychiatric Association; 1994. [Google Scholar]
  • 27.Knopman DS, Petersen RC, Rocca WA, et al. Passive case-finding for Alzheimer’s disease and dementia in two U.S. communities. Alzheimers Dement. 2011;7:53–60. doi: 10.1016/j.jalz.2010.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Winblad B, Palmer K, Kivipelto M, et al. Mild cognitive impairment – beyond controversies, towards a consensus: report of the International Working Group on Mild Cognitive Impairment. J Int Med. 2004;256:240–246. doi: 10.1111/j.1365-2796.2004.01380.x. [DOI] [PubMed] [Google Scholar]
  • 29.Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med. 2004b;256:183–194. doi: 10.1111/j.1365-2796.2004.01388.x. [DOI] [PubMed] [Google Scholar]
  • 30.Ivnik RJ, Malec JF, Smith GE, et al. Mayo’s Older Americans Normative Studies: WAIS-R, WMS-R and AVLT norms for ages 56 through 97. Clin Neuropsychol. 1992;6:1–104. [Google Scholar]
  • 31.Minimum Data Set (MDS) – Version 2.0 for Nursing Home Resident Assessment and Care Screening. Centers for Medicare & Medicaid Services (online); [Accessed April 24,2017]. Available at http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/NursingHomeQualityInits/downloads/MDS20MDSAllForms.pdf. [Google Scholar]
  • 32.Esquire Turnham H. [Accessed April 24, 2017];Federal Nursing Home Reform Act from the Omnibus Budget Reconciliation Act of 1987 or simply OBRA ‘87 SUMMARY(online) Available at: http://www.ncmust.com/doclib/OBRA87summary.pdf.
  • 33.Tuokko H, Frerichs R, Graham J, et al. Five-year follow-up of cognitive impairment with no dementia. Arch Neurol. 2003;60:577–582. doi: 10.1001/archneur.60.4.577. [DOI] [PubMed] [Google Scholar]
  • 34.Gnjidic D, Stanaway FF, Cumming R, et al. Mild cognitive impairment predicts institutionalization among older men: a population-based cohort study. PloS ONE. 2012;7:e46061. doi: 10.1371/journal.pone.0046061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ton TGN, DeLeire T, May SG, et al. The financial burden and health care utilization patterns associated with amnestic mild cognitive impairment. Alzheimers Dement. 2017;13:217–224. doi: 10.1016/j.jalz.2016.08.009. [DOI] [PubMed] [Google Scholar]
  • 36.Melton LJ., 3rd The threat to medical-records research. N Engl J Med. 1997;337:1466–1470. doi: 10.1056/NEJM199711133372012. [DOI] [PubMed] [Google Scholar]
  • 37.Roberts RO, Knopman DS, Geda YE, et al. Coronary heart disease is associated with non-amnestic mild cognitive impairment. Neurobiol Aging. 2010;31:1894–1902. doi: 10.1016/j.neurobiolaging.2008.10.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Knopman DS, Roberts RO, Geda YE, et al. Validation of the telephone interview for cognitive status-modified in subjects with normal cognition, mild cognitive impairment, or dementia. Neuroepidemiology. 2010;34:34–42. doi: 10.1159/000255464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Cook SE, Marsiske M, McCoy KJ. The use of the Modified Telephone Interview for Cognitive Status (TICS-M) in the detection of amnestic mild cognitive impairment. J Geriatr Psychiatry Neurol. 2009;22:103–109. doi: 10.1177/0891988708328214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Andel R, Hyer K, Slack A. Risk factors for nursing home placement in older adults with and without dementia. J Aging Health. 2007;19:213–228. doi: 10.1177/0898264307299359. [DOI] [PubMed] [Google Scholar]
  • 41.Luppa M, Luck T, Weyerer S, et al. Prediction of institutionalization in the elderly. A systematic review Age Ageing. 2010;39:31–38. doi: 10.1093/ageing/afp202. [DOI] [PubMed] [Google Scholar]
  • 42.The Johns Hopkins ACG® System: Excerpt from Technical Reference Guide, Version 9.0. Johns Hopkins Bloomberg School of Public Health (online); Dec, 2009. [Accessed April 24, 2017]. Available at: http://www.healthpartners.com/ucm/groups/public/@hp/@public/documents/documents/dev_057914.pdf. [Google Scholar]
  • 43.von Bonsdorff M, Rantanen T, Laukkanen P, et al. Mobility limitations and cognitive deficits as predictors of institutionalization among community-dwelling older people. Gerontology. 2006;52:359–365. doi: 10.1159/000094985. [DOI] [PubMed] [Google Scholar]
  • 44.Schwarzkopf L, Menn P, Leidl R, et al. Are community-living and institutionalized dementia patients cared for differently? Evidence on service utilization and costs of care from German insurance claims data. BMC Health Serv Res. 2013;13:2–10. doi: 10.1186/1472-6963-13-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Kane RL, Lum TY, Kane RA, et al. Does home- and community-based care affect nursing home use? J Aging Soc Policy. 2013;25:146–160. doi: 10.1080/08959420.2013.766069. [DOI] [PubMed] [Google Scholar]
  • 46.Mitchell SL, Teno JM, Miller SC, et al. A national study of the location of death for older persons with dementia. J Am Geriatr Soc. 2005;53:299–305. doi: 10.1111/j.1532-5415.2005.53118.x. [DOI] [PubMed] [Google Scholar]
  • 47.Hanson LC, Zimmerman S, Song MK, et al. Effect of the goals of care intervention for advanced dementia: a randomized clinical trial. JAMA Intern Med. 2017;177:24–31. doi: 10.1001/jamainternmed.2016.7031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Gozalo P, Teno JM, Mitchell SL, et al. End-of-life transitions among nursing home residents with cognitive issues. New Eng J Med. 2011;365:1212–1221. doi: 10.1056/NEJMsa1100347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Lin P-J, Zhong Y, Fillit HM, et al. Medicare expenditures of individuals with Alzheimer’s disease and related dementias or mild cognitive impairment before and after diagnosis. J Am Geriatr Soc. 2016;64:1549–1557. doi: 10.1111/jgs.14227. [DOI] [PubMed] [Google Scholar]
  • 50.Suehs BT, Davis CD, Alvir J, et al. The clinical and economic burden of newly diagnosed Alzheimer’s disease in a medicare advantage population. Am J Alzheimers Dis Other Demen. 2013;28:384–392. doi: 10.1177/1533317513488911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Zhu CW, Cosentino S, Ornstein K, et al. Medicare utilization and expenditures around incident dementia in a multiethnic cohort. J Gerontol A Biol Sci Med Sci. 2015;70:1448–1453. doi: 10.1093/gerona/glv124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Gaugler JE, Hovater M, Roth DL, et al. Analysis of cognitive, functional, health service use, and cost trajectories prior to and following memory loss. J Gerontol B Psychol Sci Soc Sci. 2013;68:562–567. doi: 10.1093/geronb/gbs078. [DOI] [PubMed] [Google Scholar]
  • 53.Graham JE, Rockwood K, Beattie BL. Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet. 1997;349:1793–1796. doi: 10.1016/S0140-6736(97)01007-6. [DOI] [PubMed] [Google Scholar]
  • 54.Roberts RO, Knopman DS, Mielke MM, et al. Higher risk of progression to dementia in mild cognitive impairment cases who revert to normal. Neurology. 2014;82:317–325. doi: 10.1212/WNL.0000000000000055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Plassman BL, Langa KM, Fisher GG, et al. Prevalence of dementia in the United States: The Aging, Demographics, and Memory Study. Neuroepidemiology. 2007;29:125–132. doi: 10.1159/000109998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.St Sauver JL, Grossardt BR, Leibson CL, et al. Generalizability of epidemiological findings and public health decisions: an illustration from the Rochester Epidemiology Project. Mayo Clin Proc. 2012b;87:151–160. doi: 10.1016/j.mayocp.2011.11.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.City-Data. [Accessed April 24, 2017];Olmsted County, Minnesota (MN) (online) Available at: http://www.city-data.com/county/Olmsted_County-MN.html.
  • 58.Nursing Home Data Compendium 2013 Edition. Centers for Medicare and Medicaid Services (online); [Accessed April 24, 2017]. Available at: https://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/CertificationandComplianc/downloads/nursinghomedatacompendium_508.pdf. [Google Scholar]
  • 59.Health, United States, 2014. With Special Feature on Adults Aged 55–64. Wahington, D.C: U.S. Department of Health and Human Services; 2015. [Accessed April 24, 2017]. (online). Available at: http://www.cdc.gov/nchs/data/hus/hus14.pdf. [Google Scholar]
  • 60.Knopman DS, Roberts RO, Pankratz VS, et al. Incidence of dementia among participants and nonparticipants in a longitudinal study of cognitive aging. Am J Epidemiol. 2014;180:414–423. doi: 10.1093/aje/kwu103. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp FigS1. Supplementary Figure S1.

Flow chart outlining the process used to identify the subset consisting of 3,545 Mayo Clinic Study of Aging participants in this investigation, including enumeration and characterization by cognitive impairment category.

NIHMS887892-supplement-Supp_FigS1.pdf (14.9KB, pdf)
Supp FigS2. Supplementary Figure S2.

Flow chart outlining the process used to merge the 3,545 Mayo Clinic Study of Aging participants included in this investigation with Centers for Medicare and Medicaid Services Minimum Data Set information on all nursing home activity within local nursing homes among Olmsted County, MN, residents from 10/01/1998 through 09/30/2010.

NIHMS887892-supplement-Supp_FigS2.pdf (188.1KB, pdf)
Supp TextS1. Supplementary Text S1.

Detailed descriptions of two data sources, Rochester Epidemiology Project and Mayo Clinic Study of Aging.

NIHMS887892-supplement-Supp_TextS1.docx (17.8KB, docx)
Supp TextS2. Supplementary Text S2.

Text describing data resources and variables used to merge the 3,545 Mayo Clinic Study of Aging participants included in this investigation with Centers for Medicare and Medicaid Services Minimum Data Set nursing home assessments. Text also includes details regarding how nursing home stays and days were determined for each individual.

NIHMS887892-supplement-Supp_TextS2.docx (13.9KB, docx)

RESOURCES