Archives of Gerontology and Geriatrics 59 (2014) 169–174

Geriatric conditions as predictors of increased number of hospitaladmissions and hospital bed days over one year: Findings of anationwide cohort of older adults from Taiwan

Hui-Hsuan Wang a, Ji-Tian Sheu a, Yea-Ing Lotus Shyu b,c, Hsing-Yi Chang d, Chia-Lin Li a,c,*a Department of Health Care Management, College of Management, Chang Gung University, 259 Wen-Hwa 1st Road, Kwei-Shan, Tao-Yuan 333, Taiwanb School of Nursing, College of Medicine, Chang Gung University, 259 Wen-Hwa 1st Road, Kwei-Shan, Tao-Yuan 333, Taiwanc Healthy Aging Research Center, Chang Gung University, 259 Wen-Hwa 1st Road, Kwei-Shan, Tao-Yuan 333, Taiwand Division of Preventive Medicine and Health Service Research, Institute of Population Health Sciences, National Health Research Institutes, #35, Keyan Road,

A3223, Zhunan Town, Maoli 350, Taiwan

A R T I C L E I N F O

Article history:

Received 26 September 2013

Received in revised form 21 January 2014

Accepted 3 February 2014

Available online 10 February 2014

Keywords:

Disability

Geriatric conditions

Hospital admission

Hospital days

Taiwan

A B S T R A C T

The main aim of the present study was to determine whether geriatric conditions independently predict

hospital utilizations after controlling for chronic diseases and disability among community dwelling

older adults. We analyzed data from a nationally representative sample of older adults aged 65 years and

above by linkage of 2005 Taiwan National Health Interview Survey data (including demographic

characteristics, chronic diseases, disability, and geriatric conditions such as depressive symptoms,

cognitive impairment, falls, and urinary incontinence), and 2006 National Health Insurance (NHI) claims

data (including hospital admissions and hospital bed days). A total of 1598 participants who consented

to data linkage, were successfully linked to NHI data, and had complete data for geriatric conditions were

eligible for analysis. The prevalence of depressive symptoms, cognitive impairment, falls, and urinary

incontinence were 20.6%, 26.1%, 21.3% and 23.9%, respectively. Overall, 18.2% (291/1598) of participants

had at least one hospital admission during 2006. After adjustment for demographics, prior

hospitalization, chronic diseases and functional disability, participants with geriatric conditions had

significantly more hospital admissions (incidence rate ratio = 1.34; 95% confidence interval = [1.02–

1.75]) and more hospital bed days (incidence rate ratio = 1.72; 95% confidence interval = [1.11–2.66])

than participants without geriatric conditions. Our results highlight the high prevalence (56.3%) of one or

more geriatric conditions and their independent association with excess hospital utilizations. Thus, it is

of critical importance to develop programs aimed at preventing or improving these conditions to reduce

hospital use in this population.

� 2014 Elsevier Ireland Ltd. All rights reserved.

Contents lists available at ScienceDirect

Archives of Gerontology and Geriatrics

jo ur n al ho mep ag e: www .e lsev ier . c om / lo cate /ar c hg er

1. Introduction

Hospitalizations are an important outcome in older adults. ATaiwanese report found that approximately 20% of people aged 65years and older had at least one hospitalization within one year (Li,Chang, Wang, & Bai, 2011). Previous studies investigatingpredictors of hospitalization in older adults have focused ondemographics, prior hospitalization, comorbidity, and functionalassessment (Dorr et al., 2006; Inouye et al., 2008). Geriatric

* Corresponding author at: Department of Health Care Management, College of

Management, Chang Gung University, 259 Wen-Hwa 1st Road, Kwei-Shan, Tao-

Yuan 333, Taiwan. Tel.: +886 3 2118800x5666; fax: +886 3 2118345.

E-mail address: clli@mail.cgu.edu.tw (C.-L. Li).

http://dx.doi.org/10.1016/j.archger.2014.02.002

0167-4943/� 2014 Elsevier Ireland Ltd. All rights reserved.

conditions (so-called geriatric syndromes) such as cognitiveimpairment, falls, and urinary incontinence are common in olderadults. These conditions have common features including sharedpredisposing risk factors, multifactorial causation, and theinvolvement of multiple organ systems. In addition, they have asubstantial effect on disability (Cigolle, Langa, Kabeto, Tian, &Blaum, 2007; Tinetti, Inouye, Gill, & Doucette, 1995). There isincreasing evidence to show that the co-occurrence of chronicdiseases and geriatric conditions in older adults is common andtheir combination has an impact on physical function, socialfunction, general health, and limitation of daily activities (Lee,Cigolle, & Blaum, 2009; Li et al., 2013; Rosso et al., 2011). A numberof studies have shown that older medical patients have a numberof geriatric conditions at hospital admission and that theseconditions are associated with a higher risk for adverse outcomes

H.-H. Wang et al. / Archives of Gerontology and Geriatrics 59 (2014) 169–174170

of hospitalization (Anpalahan & Gibson, 2008; Buurman et al.,2011; Lakhan et al., 2011; Laniece et al., 2008; Satish, Winograd,Chavez, & Bloch, 1996). Given that geriatric conditions, chronicdiseases, and disability are closely interrelated, a better under-standing of the independent predictive value of geriatric condi-tions on hospital utilization from population-based studies wouldassist health care providers in the development of strategies toidentify older persons at risk of adverse outcomes who may benefitthe most from early intervention.

Most empirical research has tended to focus on the associationbetween individual geriatric conditions and hospitalization. Forexample, Chodosh et al. (2004) linked hospital utilization data todata from the MacArthur Research Network on Successful AgingCommunity Study and concluded that high-functioning olderpersons who experienced decline in overall cognitive functionwere more likely to be hospitalized over 3 years. Satish et al. (1996)have reported that certain geriatric conditions such as falls areassociated with total hospital days, and depression is associatedwith poor one year survival among acutely hospitalized malepatients aged 65 years and over, even after controlling forlimitation of daily activities. Anpalahan and Gibson (2008) havefurther demonstrated that geriatric conditions such as a history ofrecurrent falls is more strongly associated with increased hospitalstays, and recurrent falls and incontinence are more stronglyassociated with admission to residential care than admissiondiagnoses or medical comorbidities among older patients aged 75years and over in Australia. Recently, Wang, Shamliyan, Talley,Ramakrishnan, and Kane (2013) performed a systematic review ofthe association of eight geriatric syndromes with hospitalization ornursing home admission and concluded that multiple morbidity,frailty, disabilities, and cognitive impairment were associated withhospitalization and nursing home admission. Given that mostolder adults suffering from geriatric conditions live in thecommunity and possibly go unrecognized (Cigolle et al., 2007;Lee et al., 2009), these abovementioned research findings highlightthe importance of early identification of geriatric conditionsamong community dwelling older adults, so that early interven-tions to reduce the risk of further adverse events can beimplemented.

Inouye, Studenski, Tinetti, and Kuchel (2007) have proposed aunifying conceptual model for geriatric syndromes to demonstratethat shared risk factors may lead to these syndromes and to theoverarching geriatric syndrome of frailty. These geriatric syn-dromes may also feedback and lead to the development of moregeriatric syndromes (Inouye et al., 2007). Given the growinginterest in the study of geriatric conditions, however, consensus onthe definitions, as well as a definite list, of geriatric conditions isstill lacking and results in variation in the conditions includedbetween different studies in the literature. In this study, weanalyzed data from a nationally representative sample of olderadults aged 65 years and above by linkage of 2005 Taiwan NationalHealth Interview Survey (NHIS) data and 2006 National HealthInsurance (NHI) claims data. We included all geriatric conditions(depressive symptoms, cognitive impairment, falls, and urinaryincontinence) for which NHIS survey data were available. There isconsensus that depressive symptoms (Buurman et al., 2012; Hajjaret al., 2009; Li et al., 2013; Rosso et al., 2011), cognitive impairment(Buurman et al., 2012; Cigolle et al., 2007; Kojima et al., 2012), fallsand urinary incontinence (Cigolle et al., 2007; Inouye et al., 2007;Lee et al., 2009; Tinetti et al., 1995) are geriatric conditions. Thepresent study was intended to explore whether geriatric condi-tions remained an independent predictor of hospital utilizationsafter controlling for chronic diseases and disability. Our studydiffers from most previous studies in that we classified participantsinto two groups based on the presence or absence of geriatricconditions. We hypothesized that older adults with one or more

geriatric conditions would have a significantly increased numberof admissions and total hospital stay over one year, independent ofchronic diseases and functional disability. We also aimed todescribe the distribution of geriatric conditions among older adultsand their associated factors.

2. Methods

2.1. Study population

This was a prospective study involving participants in the NHISin Taiwan, 2005. The study sample was drawn from the NationalRegistry Database through a complex multistage design, which hasbeen reported in detail previously (NHIS Working Group, 2006).Ethical approval was obtained from the Institutional Review Boardof the National Health Research Institutes. All participantsprovided signed informed consent. Participants were asked toprovide consent for data linkage to the NHI claims database. Theoriginal sample comprised 27,726 participants (response rate80.6%), including 2727 individuals aged 65 years and above. Out ofthese, a total of 1760 provided consent for data linkage. Of these,we excluded 16 individuals who were unable to be linked toinsurance data and 302 with incomplete data regarding geriatricconditions, leaving 1598 eligible participants. We compared thecharacteristics of study participants who were included (N = 1598)and excluded (N = 1129) to assess the degree of respondent bias.

2.2. Demographic characteristics, prior hospitalization, chronic

diseases, disability, and outcome measure

We assessed factors considered to be associated with medicalservice use in older adults including demographics, priorhospitalization (have been hospitalized in 2005, data from 2005NHI claims database), chronic diseases, and disability. Trainedinterviewers used standard questionnaires to collect baseline datafrom participants on age, sex, years of education, marital status,and chronic diseases including diabetes, heart disease, hyperten-sion, dyslipidemia, stroke, chronic obstructive pulmonary disease,and cancer. For each disease, participants were asked whether thediagnosis had been confirmed by a medical professional. Partici-pants reported their ability to perform six activities of daily living(ADL) (eating, bathing, dressing, using the toilet, getting in or out ofbed, and walking across a small room). Participants were askedwhether they could perform these activities with no difficulty,some difficulty, much difficulty, or were unable to perform them.ADL disability was dichotomized as being able to perform theseactivities with no difficulty or some difficulty vs. much difficulty orbeing unable to perform one or more ADLs. The outcome measurein this study was hospitalization for any cause, obtained throughdata linkage to the 2006 NHI claims database.

2.3. Geriatric conditions assessment

Geriatric conditions were self-reported, with conditions chosenfor the present study determined by the questions included in the2005 NHIS (depressive symptoms, cognitive impairment, falls, andurinary incontinence). The 10 item version of the Center forEpidemiologic Studies Depression Scale (CES-D) was used to assessdepressive symptoms (Andresen, Malmgren, Carter, & Patrick,1994; Radloff, 1977). Participants with scores from 10 to 30 weredefined as having depressive symptoms (Andresen et al., 1994).The Mini-Mental State Examination (MMSE) was used to assesscognitive function after obtaining permission from the Psycholog-ical Assessment Resources (PAR), Inc. This scale provides a totalscore ranging from 0 to 30 points, with higher scores representingbetter cognitive function (Folstein, Folstein, & McHugh, 1975).

Table 1Characteristics of study participants.

Overall Excluded participants Includeda participants P-valueb

N 2727 1129 1598

Age (%) <0.001

65–74 59.3 53.1 63.8

75+ 40.7 46.9 36.2

Sex (% female) 50.5 57.8 45.4 <0.001

Education (%) <0.001

0 years 41.4 54.2 32.3

1–6 years 38.1 30.3 43.7

7+ years 20.5 15.5 24.0

Marital status (%married

or living with partner)

62.9 55.2 68.3 <0.001

Hospitalization in 2005 (% yes) – – 16.9 –

Diabetes (% yes) 17.4 18.7 16.4 0.116

Heart disease (% yes) 18.2 18.6 17.8 0.599

Hypertension (% yes) 42.4 41.8 42.9 0.573

Dyslipidemia (% yes) 20.7 18.2 22.5 0.006

Stroke (% yes) 7.7 10.9 5.5 <0.001

Chronic obstructive

pulmonary disease (% yes)

7.2 7.8 6.9 0.362

Cancer (% yes) 2.4 2.9 2.1 0.150

Number of diseasesc (%) 0.780

0 36.5 36.9 36.2

1 30.1 29.4 30.7

2+ 33.4 33.7 33.1

ADL disability (% yes) 9.1 16.6 3.8 <0.001

Depressive symptoms (% yes) 20.6 25.3 18.1 <0.001

Cognitive impairment (% yes) 26.1 35.1 21.4 <0.001

Falls (% yes) 21.3 24.0 19.3 0.004

Urinary incontinence (% yes) 23.9 25.5 22.8 0.097

Number of geriatric conditionsd (%) <0.001

0 43.7 37.6 46.8

1 33.1 34.7 32.3

2+ 23.2 27.7 20.9

a Included participants were limited to those who had provided consent for data linkage, were successfully linked to NHI data, and had complete data for geriatric

conditions.b Categorical variables were compared using Pearson’s Chi-square test and shown as percentages.c Number of diseases including diabetes, heart disease, hypertension, dyslipidemia, stroke, chronic obstructive pulmonary disease, and cancer.d Number of geriatric conditions including depressive symptoms, cognitive impairment, falls, and urinary incontinence.

H.-H. Wang et al. / Archives of Gerontology and Geriatrics 59 (2014) 169–174 171

Previous studies have indicated that level of education significant-ly affects the points obtained in the Chinese version of the MMSE(Katzman et al., 1988). Thus, the cut-point for cognitiveimpairment in this study was set according to educational level(Katzman et al., 1988). The cut-points were: 17 for participantswho were illiterate and had had no schooling, 20 for those with 1–6years of education, and 24 for those with 7 or more years ofeducation (Katzman et al., 1988). Fallers were defined as havinghad at least one fall during the previous year. Urinary incontinencewas defined as reporting any urine leakage during the previousyear.

2.4. Statistical analysis

We used the Pearson’s Chi-Square test to compare thecharacteristics of study participants who were included andexcluded. We also used the Pearson’s Chi-Square test to examinefactors associated with one or more geriatric conditions amongolder adults. In our study, the distributions of number of hospitaladmissions and hospital bed days were positively skewed and bothcould be considered as rare events. Consequently, the traditionalleast square technique is inefficient and produces inconsistent andbiased estimates. In addition as there was only one parameter, it istoo simple to express as a Poisson distribution. Thus, we chosenegative binomial regression models to account for overdispersionof count data for the number of hospital admissions and hospitalbed days in our study (Cameron & Trivedi, 1998). The adjustedincidence rate ratios (IRR) and 95% confidence intervals (95% CI) fornumber of admissions and hospital bed days were estimated usingmultiple negative binomial regression models. All analyses were

conducted using SAS statistical software, version 9.1 (SAS Institute,Cary, NC) and Stata statistical software, version 10.1 (Stata Corp.).

3. Results

Table 1 compares the baseline characteristics of studyparticipants who were included and excluded. Participants weresignificantly more likely to be younger, male, married or livingwith a partner, have higher education levels, and less likely to haveADL disability or geriatric conditions.

Table 2 shows that those who had one or more geriatricconditions were significantly older, more likely to be women, morelikely to have lower education levels, less likely to be married orliving with a partner, and more likely to be hospitalized in 2005,and have a history of diabetes, heart disease, hypertension,dyslipidemia, stroke, and ADL disability.

Table 3 shows the distribution of admissions and hospital beddays from 2006 NHI claims data. Overall, 18.2% (291/1598) ofparticipants had at least one hospital admission during 2006.Participants with geriatric conditions had a significantly higherrate of hospital admission (19.9% vs. 16.3%), and were significantlymore likely to stay in hospital for 28 or more days (3.7% vs. 1.3%)than those without geriatric conditions.

Table 4 shows the IRRs and 95% CIs for hospital admissions andhospital bed days. After adjustment for age, sex, education, maritalstatus, prior hospitalization, number of diseases, and ADLdisability, participants with geriatric conditions had significantlymore hospital admissions (IRR = 1.34; 95% CI = [1.02–1.75]) andmore hospital bed days (IRR = 1.72; 95% CI = [1.11–2.66]) thanparticipants without geriatric conditions during 2006.

Table 2Characteristics of participants by the presence of geriatric conditions.

Number of geriatric

conditionsa

P-valueb

0 1+

N 748 850

Age (%) 0.002

65–74 67.8 60.2

75+ 32.2 39.8

Sex (% female) 32.6 59.7 <0.001

Education (%) <0.001

0 years 23.7 39.9

1–6 years 45.6 42.1

7+ years 30.8 18.0

Marital status (%married

or living with partner)

76.9 60.7 <0.001

Hospitalization in 2005 (% yes) 14.5 19.0 0.017

Diabetes (% yes) 13.5 19.0 0.003

Heart disease (% yes) 14.4 20.9 0.001

Hypertension (% yes) 39.5 45.9 0.010

Dyslipidemia (% yes) 19.5 25.0 0.009

Stroke (% yes) 2.6 8.0 <0.001

Chronic obstructive

pulmonary disease (% yes)

5.8 7.8 0.107

Cancer (% yes) 1.5 2.6 0.115

Number of diseasesc (%) <0.001

0 41.7 31.3

1 30.1 31.2

2+ 28.2 37.5

ADL disability (% yes) 0.3 6.9 <0.001

Depressive symptoms (% yes) – 34.0 –

Cognitive impairment (% yes) – 40.2 –

Falls (% yes) – 36.4 –

Urinary incontinence (% yes) – 42.8 –

a Number of geriatric conditions including depressive symptoms, cognitive

impairment, falls, and urinary incontinence.b Categorical variables were compared using Pearson’s Chi-square test and

shown as percentages.c Number of diseases including diabetes, heart disease, hypertension, dyslipi-

demia, stroke, chronic obstructive pulmonary disease, and cancer.

Table 3Number of hospital admissions and total hospital bed days during 2006.

Overall Number of

geriatric conditionsa

P-valueb

0 1+

N 1598 748 850

Those ever

hospitalized in 2006 (N)

291 122 169

Number of admissions

in 2006 (N; %)

0.005

0 1307 (81.8) 626 (83.7) 681 (80.1)

1 196 (12.3) 87 (11.6) 109 (12.8)

2 55 (3.4) 27 (3.6) 28 (3.3)

3+ 40 (2.5) 8 (1.1) 32 (3.8)

Total hospital bed days

in 2006 (N; %)

0.010

0 1307 (81.8) 626 (83.7) 681 (80.1)

1–27 days 250 (15.6) 112 (15.0) 138 (16.2)

28+ days 41 (2.6) 10 (1.3) 31 (3.7)

a Number of geriatric conditions including depressive symptoms, cognitive

impairment, falls, and urinary incontinence.b Categorical variables were compared using Pearson’s Chi-square test and

shown as numbers and percentages.

Table 4Negative binomial regression of number of hospital admissions and total hospital bed

Number of hospital admissions

IRR 95% CI

Age

65–74 Reference

75+ 1.49 [1.14–1.95]

Sex

Female Reference

Male 1.40 [1.05–1.87]

Education

0 years Reference

1–6 years 0.88 [0.64–1.21]

7+ years 0.90 [0.62–1.30]

Marital status

Other Reference

Married or living with partner 0.98 [0.74–1.31]

Hospitalization in 2005

No Reference

Yes 2.85 [2.13–3.82]

Number of diseasesa

0 Reference

1 0.89 [0.65–1.23]

2+ 1.33 [0.98–1.80]

ADL disability

No Reference

Yes 2.10 [1.23–3.60]

Number of geriatric conditionsb

0 Reference

1+ 1.34 [1.02–1.75]

a Number of diseases including diabetes, heart disease, hypertension, dyslipidemia,

b Number of geriatric conditions including depressive symptoms, cognitive impairm

H.-H. Wang et al. / Archives of Gerontology and Geriatrics 59 (2014) 169–174172

4. Discussion

Our findings confirm our hypothesis that older adults with oneor more geriatric conditions have significantly more hospitaladmissions and more hospital bed days in the subsequent year,independent of chronic diseases and disability. Moreover, amongolder adults aged 65 and above in Taiwan, as many as 56.3% hadone or more geriatric conditions. The high prevalence of geriatricconditions and their independent association with excess hospitalutilizations underscores the critical importance of preventing orimproving these conditions to reduce hospital utilization in thispopulation.

While the association between geriatric conditions at hospitaladmission and further functional decline (Buurman et al., 2011,

days in 2006.

Total hospital bed days

P-value IRR 95% CI P-value

Reference

0.004 2.15 [1.36–3.40] 0.001

Reference

0.024 1.83 [1.13–2.96] 0.014

Reference

0.433 0.74 [0.43–1.29] 0.293

0.581 0.56 [0.30–1.04] 0.066

Reference

0.914 1.10 [0.68–1.77] 0.700

Reference

<0.001 4.10 [2.34–7.17] <0.001

Reference

0.492 0.89 [0.53–1.49] 0.664

0.067 1.79 [1.07–3.00] 0.027

Reference

0.007 2.01 [0.67–6.02] 0.215

Reference

0.037 1.72 [1.11–2.66] 0.015

stroke, chronic obstructive pulmonary disease, and cancer.

ent, falls, and urinary incontinence.

H.-H. Wang et al. / Archives of Gerontology and Geriatrics 59 (2014) 169–174 173

2012; Lakhan et al., 2011), hospital readmission (Laniece et al.,2008; Winograd et al., 1991), nursing home utilization (Anpalahan& Gibson, 2008; Satish et al., 1996; Winograd et al., 1991), andmortality (Buurman et al., 2011; Satish et al., 1996; Winograd et al.,1991) has been well studied, few studies have looked specifically atother predictive values of geriatric conditions, such as number ofhospital admissions and hospital bed days in community dwellingolder adults. Our findings add to the understanding of factorsleading to hospital utilizations by demonstrating that communitydwelling older adults who have one or more geriatric conditions(depressive symptoms, cognitive impairment, falls, and urinaryincontinence), have a 34% increase in expected number ofadmissions and a 72% increase in expected hospital bed daysover one year compared to those without geriatric conditions, evenafter controlling for demographic characteristics, prior hospitaliza-tion, chronic diseases, and functional disability (Table 4). Theseresults are similar to those of previous studies of older medicalinpatients (Anpalahan & Gibson, 2008; Kojima et al., 2012; Satishet al., 1996). Kojima et al. (2012) investigated factors associated withprolonged hospital stay in a geriatric ward in Japan and concludedthat the number of geriatric syndrome components (including falls,cognitive impairment, urinary incontinence, constipation, insomnia,etc.) was significantly associated with hospital stay, independent ofmultiple diseases and ADL dependence.

In this study, we found that older adults with geriatricconditions had a significantly higher prevalence of co-occurrenceof two or more chronic diseases (37.5%) than older adults withoutgeriatric conditions (28.2%) (Table 2). Rosso et al. (2011)demonstrated an additive synergistic interaction between thepresence of one or more geriatric syndromes (including urinaryincontinence, falls, and depression) and the presence of one ormore cardiometabolic diseases (including coronary artery disease,coronary heart failure, and diabetes) on functional measures.Buurman et al. (2012) conducted a prospective cohort studyamong patients aged 65 years and older in the Netherlands andfound that geriatric conditions (including fall risk, cognitiveimpairment, depressive symptoms, incontinence, ADL im-pairment, etc.) were highly prevalent in acutely hospitalized olderpatients and associated with functional decline three and twelvemonths after admission. Lakhan et al. (2011) performed aprospective cohort study of geriatric syndromes (including falls,pressure ulcer, delirium, ADL limitation, instrumental activity ofdaily living impairment, cognitive impairment, bladder inconti-nence, and bowel incontinence) among older medical patientsaged 70 years and older in three hospitals in Australia and foundthat almost half (49.4%) experienced one or two syndromes duringthe premorbid period. Many patients did not return to theirpremorbid level of function at discharge and had acquired newgeriatric conditions during hospitalization. These findings demon-strate that geriatric conditions can be associated with furtherfunctional decline among older adults after hospital admission andthis combination of comorbid diseases and geriatric conditions hasa substantial impact on recovery of function and risk of developingnew conditions, which may result in longer hospital stays and anincreased risk of readmission. These observations provide aplausible biological pathway behind our findings that older adultswith geriatric conditions had an increased number of admissionsand hospital bed days.

Fried, Ferrucci, Darer, Williamson, and Anderson (2004)provided a review of the concepts of disability, frailty, andcomorbidity and concluded that comorbidity was associated withhigh health care utilization and expenditures. However, it isnotable that those who had two or more chronic diseasespresented with only a borderline significant increased risk ofgreater hospital admissions in our study (Table 4). In this study,chronic diseases included diabetes, heart disease, hypertension,

dyslipidemia, stroke, chronic obstructive pulmonary disease, andcancer. We cannot exclude the possibility that some otherunderlying diseases that may have been the cause of an admissionwere not included in this study, thereby leading to the lack of anassociation between comorbidity and an increased number ofadmissions. It is also possible that lack of data on diseases severitymay have confounded our results and resulted in the failure toreach statistical significance.

In the present study, the geriatric conditions assessed werelimited to those that had been surveyed in the NHIS, including themain neuro-psychological domain conditions (depressive symp-toms and cognitive impairment), and somatic domain conditions(falls and urinary incontinence). Despite this limitation, we believethat these four conditions are important, not only because theyhave been associated with increased risk of further adverse events,but also because they are closely interrelated and can be preventedor delayed (Chodosh, Martinez, Aneshensel, Wight, & Karlamangla,2010; Ganz, Bao, Shekelle, & Rubenstein, 2007; Goveas et al., 2012;Holroyd-Leduc & Straus, 2004; Liu-Ambrose, Ashe, Graf, Beattie, &Khan, 2008; Tinetti et al., 1995). Our observations indicate that thepresence of at least one of these conditions could be a marker forincreased risk of excess hospital utilization among communitydwelling older adults, independent of chronic diseases and ADLdisability. Further investigation is needed to explore whether suchindividuals could benefit from interventions to prevent excesshealth care utilization.

Our study has other limitations. The geriatric conditions weexamined are based on self-report, and therefore recall bias isunavoidable. In addition, although the initial sample of 2727individuals is a nationally representative sample of adults aged 65years and over, our analytic sample could be biased due to includedparticipants being limited to those who consented to data linkage,were successfully linked to NHI data, and had complete data forgeriatric conditions. Comparison of characteristics of respondentswho were included and excluded (Table 1) suggest that our samplecould be biased toward participants who are younger, male, havehigher education levels, are married or living with a partner, andhave better health status including being less likely to have ADLdisability, and less likely to have geriatric conditions. Therefore, wemay have underestimated rates of admissions and the associationbetween geriatric conditions and admissions and hospital beddays. As we were able to demonstrate a statistically significantassociation despite this potential underestimation, it suggests thatthe associations between the presence of one or more geriatricconditions in older adults and an increased number of admissionsand hospital bed days are substantial. Lastly, as the NHI data wereobtained for the year 2006 and the NHIS was conducted in 2005,there would be a time lag between the geriatric conditionsassessed and commencement of enumeration of hospital admis-sions.

Despite these limitations, we consider our results importantbecause of their practical implications. We found that depressivesymptoms, cognitive impairment, falls, and urinary incontinencewere as highly prevalent as other chronic diseases in older adults(Table 1). It is notable that participants with at least one geriatriccondition were significantly more likely to have 3 or more hospitaladmissions (3.8% vs. 1.1%) and were significantly more likely tostay in hospital for 28 or more days (3.7% vs. 1.3%) over one year(Table 3). Because of a lack of clinical data in our study, we were notable to determine whether geriatric conditions were an early signof poor disease control among community dwelling older persons,which has subsequently led to hospital admission. However, ourobservation that 56.3% of respondents have one or more geriatricconditions highlights that a substantial number of people aged 65years and older living in the community are at higher risk of excesshospital utilizations and therefore require more care. We also

H.-H. Wang et al. / Archives of Gerontology and Geriatrics 59 (2014) 169–174174

found that the likelihood of having one or more geriatric conditionsin older adults was higher in those of older age, women, those withlower education levels, those who were unmarried and livingalone, those who had been hospitalization in 2005, those withchronic diseases, and those with ADL disability (Table 2). We hopeour findings will stimulate further effort to develop strategies forthe systematic assessment of geriatric conditions among commu-nity dwelling older persons which could identify older persons athigher risk of adverse outcomes. Clinicians should assess geriatricconditions at regular clinic attendances. This would help identifyearly those who would benefit most from interventions aimed atpreventing and improving geriatric conditions before hospitaladmission has occurred, thus leading to reduced utilization ofhospital services.

5. Conclusions

In summary, our finding of an independent association betweengeriatric conditions and increased hospital utilizations highlightsthe importance of early identification of geriatric conditions.Further investigations are required to explore how systematicassessment and management of geriatric conditions can result intargeted early interventions aimed at reducing risk of adverseoutcomes including excess hospital utilizations.

Conflict of interest statement

None.

Acknowledgements

This study is based (in part) on data from the National HealthInterview Survey Original Database provided by the Bureau ofHealth Promotion, the Department of Health and the NationalHealth Research Institutes. The interpretation and conclusionscontained herein do not represent those of the Bureau of HealthPromotion, the Department of Health or the National HealthResearch Institutes. This study was supported (in part) by grants(NSC100-2410-H-182-004-MY2 and NSC102-2410-H-182-008)from the National Science Council of the Republic of China,Taiwan and a grant (EMRPD1C0301) from the Healthy AgingResearch Center, Chang Gung University, Taiwan.

References

Andresen, E. M., Malmgren, J. A., Carter, W. B., & Patrick, D. L. (1994). Screening fordepression in well older adults: Evaluation of a short form of the CES-D (Center forEpidemiologic Studies Depression Scale). American Journal of Preventive Medicine,10(2), 77–84.

Anpalahan, M., & Gibson, S. J. (2008). Geriatric syndromes as predictors of adverseoutcomes of hospitalization. Internal Medicine Journal, 38(1), 16–23.

Buurman, B. M., Hoogerduijn, J. G., de Haan, R. J., Abu-Hanna, A., Lagaay, A. M., Verhaar,H. J., et al. (2011). Geriatric conditions in acutely hospitalized older patients:Prevalence and one-year survival and functional decline. PLoS ONE, 6(11), e26951.

Buurman, B. M., Hoogerduijn, J. G., van Gemert, E. A., de Haan, R. J., Schuurmans, M. J., &de Rooij, S. E. (2012). Clinical characteristics and outcomes of hospitalized olderpatients with distinct risk profiles for functional decline: A prospective cohortstudy. PLoS ONE, 7(1), e29621.

Cameron, A. C., & Trivedi, P. K. (1998). Regression Analysis of Count Data. New York:Cambridge University Press.

Chodosh, J., Martinez, D. M., Aneshensel, C. S., Wight, R. G., & Karlamangla, A. S. (2010).Depressive symptoms, chronic diseases, and physical disabilities as predictors ofcognitive functioning trajectories in older Americans. Journal of the AmericanGeriatrics Society, 58(12), 2350–2357.

Chodosh, J., Seeman, T. E., Keeler, E., Sewall, A., Hirsch, S. H., Guralnik, J. M., et al. (2004).Cognitive decline in high-functioning older persons is associated with an increasedrisk of hospitalization. Journal of the American Geriatrics Society, 52(9), 1456–1462.

Cigolle, C. T., Langa, K. M., Kabeto, M. U., Tian, Z., & Blaum, C. S. (2007). Geriatricconditions and disability: The Health and Retirement Study. Annals of InternalMedicine, 147(3), 156–164.

Dorr, D. A., Jones, S. S., Burns, L., Donnelly, S. M., Brunker, C. P., Wilcox, A., et al. (2006).Use of health-related, quality-of-life metrics to predict mortality and hospitaliza-tions in community-dwelling seniors. Journal of the American Geriatrics Society,54(4), 667–673.

Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state. A practicalmethod for grading the cognitive state of patients for the clinician. Journal ofPsychiatric Research, 12(3), 189–198.

Fried, L. P., Ferrucci, L., Darer, J., Williamson, J. D., & Anderson, G. (2004). Untangling theconcepts of disability, frailty, and comorbidity: Implications for improved target-ing and care. Journals of Gerontology – Series A Biological Sciences and MedicalSciences, 59(3), 255–263.

Ganz, D. A., Bao, Y., Shekelle, P. G., & Rubenstein, L. Z. (2007). Will my patient fall?Journal of the American Medical Association, 297(1), 77–86.

Goveas, J. S., Hogan, P. E., Kotchen, J. M., Smoller, J. W., Denburg, N. L., Manson, J. E., et al.(2012). Depressive symptoms, antidepressant use, and future cognitive health inpostmenopausal women: The Women’s Health Initiative Memory Study. Interna-tional Psychogeriatrics, 24(8), 1252–1264.

Hajjar, I., Yang, F., Sorond, F., Jones, R. N., Milberg, W., Cupples, L. A., et al. (2009). Anovel aging phenotype of slow gait, impaired executive function, and depressivesymptoms: Relationship to blood pressure and other cardiovascular risks. Journalsof Gerontology – Series A Biological Sciences and Medical Sciences, 64(9), 994–1001.

Holroyd-Leduc, J. M., & Straus, S. E. (2004). Management of urinary incontinence inwomen: Clinical applications. Journal of the American Medical Association, 291(8),996–999.

Inouye, S. K., Studenski, S., Tinetti, M. E., & Kuchel, G. A. (2007). Geriatric syndromes:Clinical, research, and policy implications of a core geriatric concept. Journal of theAmerican Geriatrics Society, 55(5), 780–791.

Inouye, S. K., Zhang, Y., Jones, R. N., Shi, P., Cupples, L. A., Calderon, H. N., et al. (2008).Risk factors for hospitalization among community-dwelling primary care olderpatients: Development and validation of a predictive model. Medical Care, 46(7),726–731.

Katzman, R., Zhang, M. Y., Ouang-Ya-Qu, Wang, Z. Y., Liu, W. T., Yu, E., et al. (1988). AChinese version of the Mini-Mental State Examination: Impact of illiteracy in aShanghai dementia survey. Journal of Clinical Epidemiology, 41(10), 971–978.

Kojima, T., Akishita, M., Kameyama, Y., Yamaguchi, K., Yamamoto, H., Eto, M., et al.(2012). Factors associated with prolonged hospital stay in a geriatric ward of auniversity hospital in Japan. Journal of the American Geriatrics Society, 60(6), 1190–1191.

Lakhan, P., Jones, M., Wilson, A., Courtney, M., Hirdes, J., & Gray, L. C. (2011). Aprospective cohort study of geriatric syndromes among older medical patientsadmitted to acute care hospitals. Journal of the American Geriatrics Society, 59(11),2001–2008.

Laniece, I., Couturier, P., Drame, M., Gavazzi, G., Lehman, S., Jolly, D., et al. (2008).Incidence and main factors associated with early unplanned hospital readmissionamong French medical inpatients aged 75 and over admitted through emergencyunits. Age and Ageing, 37(4), 416–422.

Lee, P. G., Cigolle, C., & Blaum, C. (2009). The co-occurrence of chronic diseases andgeriatric syndromes: The health and retirement study. Journal of the AmericanGeriatrics Society, 57(3), 511–516.

Li, C. L., Chang, H. Y., Wang, H. H., & Bai, Y. B. (2011). Diabetes, functional ability, andself-rated health independently predict hospital admission within one year amongolder adults: A population based cohort study. Archives of Gerontology and Ger-iatrics, 52(2), 147–152.

Li, C. L., Chiu, Y. C., Chang, H. Y., Hsu, K. H., Bai, Y. B., & Wang, H. H. (2013). Association ofgeriatric conditions and cardiovascular diseases with disability in older adults withdiabetes: Findings from a nationally representative survey. Geriatrics and Geron-tology International, 13(3), 563–570.

Liu-Ambrose, T. Y., Ashe, M. C., Graf, P., Beattie, B. L., & Khan, K. M. (2008). Increased riskof falling in older community-dwelling women with mild cognitive impairment.Physical Therapy, 88(12), 1482–1491.

NHIS Working Group. (2006). 2005 Taiwan National Health Interview and MedicationSurvey, Sampling Design (In Chinese). Taiwan National Health Interview SurveyResearch Brief, Taipei (pp. 2–).

Radloff, L. (1977). The CES-D Scale: A self-report depression scale for research in thegeneral population. Applied Psychological Measurement, 1(3), 385–401.

Rosso, A. L., Eaton, C. B., Wallace, R., Gold, R., Curb, J. D., Stefanick, M. L., et al. (2011).Combined impact of geriatric syndromes and cardiometabolic diseases on mea-sures of functional impairment. Journals of Gerontology – Series A Biological Sciencesand Medical Sciences, 66A(3), 349–354.

Satish, S., Winograd, C. H., Chavez, C., & Bloch, D. A. (1996). Geriatric targeting criteria aspredictors of survival and health care utilization. Journal of the American GeriatricsSociety, 44(8), 914–921.

Tinetti, M. E., Inouye, S. K., Gill, T. M., & Doucette, J. T. (1995). Shared risk factors for falls,incontinence, and functional dependence: Unifying the approach to geriatricsyndromes. Journal of the American Medical Association, 273(17), 1348–1353.

Wang, S. Y., Shamliyan, T. A., Talley, K. M., Ramakrishnan, R., & Kane, R. L. (2013). Notjust specific diseases: Systematic review of the association of geriatric syndromeswith hospitalization or nursing home admission. Archives of Gerontology andGeriatrics, 57(1), 16–26.

Winograd, C. H., Gerety, M. B., Chung, M., Goldstein, M. K., Dominguez, F., & Vallone, R.(1991). Screening for frailty: Criteria and predictors of outcomes. Journal of theAmerican Geriatrics Society, 39(8), 778–784.