Advertisement

The Course of Depressive Symptoms Over 36 Months in 696 Newly Admitted Nursing Home Residents

  • Tom Borza
    Correspondence
    Address correspondence to Tom Borza, PhD, MD, Research Center for Age-related Functional Decline and Disease, Innlandet Hospital Trust, PO Box 68, 2313 Ottestad, Norway.
    Affiliations
    Research Center for Age-related Functional Decline and Disease, Innlandet Hospital Trust, Ottestad, Norway
    Search for articles by this author
  • Geir Selbæk
    Affiliations
    Norwegian National Center for Aging and Health, Vestfold Hospital Trust, Tønsberg, Norway

    Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway

    Department of Geriatric Medicine, Oslo University Hospital, Oslo, Norway
    Search for articles by this author
  • Bjørn Lichtwarck
    Affiliations
    Research Center for Age-related Functional Decline and Disease, Innlandet Hospital Trust, Ottestad, Norway
    Search for articles by this author
  • Jūratė Šaltytė Benth
    Affiliations
    Research Center for Age-related Functional Decline and Disease, Innlandet Hospital Trust, Ottestad, Norway

    Health Services Research Unit, Akershus University Hospital, Lørenskog, Norway

    Institute of Clinical Medicine, Campus Ahus, University of Oslo, Oslo, Norway
    Search for articles by this author
  • Sverre Bergh
    Affiliations
    Research Center for Age-related Functional Decline and Disease, Innlandet Hospital Trust, Ottestad, Norway

    Norwegian National Center for Aging and Health, Vestfold Hospital Trust, Tønsberg, Norway
    Search for articles by this author
Open AccessPublished:September 15, 2022DOI:https://doi.org/10.1016/j.jamda.2022.08.007

      Abstract

      Objectives

      To investigate the course of depressive symptoms in newly admitted nursing home (NH) residents and how resident characteristics were associated with the symptoms. To identify groups of residents following the same symptom trajectory.

      Design

      An observational, multicenter, longitudinal study over 36 months with 7 biannual assessments.

      Setting and Participants

      Representing 47 Norwegian NHs, 696 residents were included at admission to a NH.

      Methods

      Depressive symptoms were assessed with the Cornell Scale for Depression in Dementia (CSDD). We selected severity of dementia, functional impairment, physical health, pain, use of antidepressants, age, and sex as covariates. Time trend in CSDD score was assessed by a linear mixed model adjusting for covariates. Next, a growth mixture model was estimated to investigate whether there were groups of residents following distinct trajectories in CSDD scores. We estimated a nominal regression model to assess whether the covariates at admission were associated to group membership.

      Results

      There was a nonlinear trend in CSDD score. More severe dementia, a lower level of functioning, poorer physical health, more pain, use of antidepressants, and younger age at admission were associated with higher CSDD scores. Growth mixture model identified 4 groups: (1) persistent mild symptoms (32.6%), (2) persistent moderate symptoms (50.8%), (3) increasing symptoms (5.1%), and (4) severe but decreasing symptoms (11.6%). A lower level of functioning, poorer physical health, more pain, use of antidepressants, and younger age at admission were associated with higher odds for belonging to the severe but decreasing symptoms group compared with the persistent mild symptoms group.

      Conclusions and Implications

      Most NH residents were in trajectory groups with persistent mild or moderate depressive symptoms. Residents with more severe dementia, lower levels of functioning, poor physical health, severe pain, younger age at admittance, and who are using antidepressants should be monitored closely and systematically with respect to depression. Taking actions toward a more personalized treatment for depression in NHs is a priority and should be investigated in future studies.

      Keywords

      Depression and depressive symptoms are common among nursing home (NH) residents. According to a systemic review, the median prevalence of major depressive disorders and depressive symptoms were 10% and 29%, respectively
      • Seitz D.
      • Purandare N.
      • Conn D.
      Prevalence of psychiatric disorders among older adults in long-term care homes: A systematic review.
      Depressive symptoms have several potential etiologies in NH residents. The symptoms can, for example, be related to a mood disorder, may overlap with symptoms of dementia, may be linked to a medical disorder or medications taken for a disorder, or they may be related to leaving one's home and/or the loss of close relationships with family and friends and/or to loss of autonomy.
      • Snowdon J.
      Depression in nursing homes.
      ,
      • Custers A.F.
      • Cillessen A.H.
      • Westerhof G.J.
      • et al.
      Need fulfillment, need importance, and depressive symptoms of residents over the first eight months of living in a nursing home.
      Thus, depression in NH residents is multifaceted and difficult to diagnose.
      • Veltman E.
      • Kok A.
      • Lamers F.
      • et al.
      Stability and transition of depression subtypes in late life.
      ,
      • Matos Queirós A.
      • von Gunten A.
      • Martins M.
      • et al.
      The forgotten psychopathology of depressed long-term care facility residents: A call for evidence-based practice.
      Both depression and higher levels of depressive symptoms have negative consequences for NH residents. They are reportedly associated with more pain,
      • Erdal A.
      • Flo E.
      • Selbaek G.
      • et al.
      Associations between pain and depression in nursing home patients at different stages of dementia.
      ,
      • Jenny Wei Y.J.
      • Chen C.
      • Fillingim R.B.
      • et al.
      Uncontrolled pain and risk for depression and behavioral symptoms in residents with dementia.
      poorer quality of life,
      • Hurt C.
      • Bhattacharyya S.
      • Burns A.
      • et al.
      Patient and caregiver perspectives of quality of life in dementia. An investigation of the relationship to behavioural and psychological symptoms in dementia.
      progression of cognitive decline,
      • Rapp M.A.
      • Schnaider-Beeri M.
      • Wysocki M.
      • et al.
      Cognitive decline in patients with dementia as a function of depression.
      ,
      • Ulbricht C.M.
      • Rothschild A.J.
      • Hunnicutt J.N.
      • et al.
      Depression and cognitive impairment among newly admitted nursing home residents in the USA.
      greater disability,
      • Kaup B.A.
      • Loreck D.
      • Gruber-Baldini A.L.
      • et al.
      Depression and its relationship to function and medical status, by dementia status, in nursing home admissions.
      ,
      • Chau R.
      • Kissane D.W.
      • Davison T.E.
      Risk factors for depression in long-term care: A systematic review.
      and increased mortality.
      • Matos Queirós A.
      • von Gunten A.
      • Martins M.
      • et al.
      The forgotten psychopathology of depressed long-term care facility residents: A call for evidence-based practice.
      ,
      • Barca M.L.
      • Engedal K.
      • Laks J.
      • et al.
      A 12 months follow-up study of depression among nursing-home patients in Norway.
      ,
      • Watt J.A.
      • Goodarzi Z.
      • Veroniki A.A.
      • et al.
      Comparative efficacy of interventions for reducing symptoms of depression in people with dementia: Systematic review and network meta-analysis.
      Furthermore, such symptoms have an impact on the use of healthcare services.
      • Matos Queirós A.
      • von Gunten A.
      • Martins M.
      • et al.
      The forgotten psychopathology of depressed long-term care facility residents: A call for evidence-based practice.
      ,
      • Ulbricht C.M.
      • Rothschild A.J.
      • Hunnicutt J.N.
      • et al.
      Depression and cognitive impairment among newly admitted nursing home residents in the USA.
      ,
      • Smalbrugge M.
      • Jongenelis L.
      • Pot A.M.
      • et al.
      Incidence and outcome of depressive symptoms in nursing home patients in the Netherlands.
      ,
      • Rivera E.
      • Hirschman K.B.
      • Naylor M.D.
      Reported needs and depressive symptoms among older adults entering long-term services and supports.
      As depression and depressive symptoms are highly prevalent in NHs and have far-reaching consequences for residents and society, how depressive symptoms unfold over time is particularly understudied.
      • Matos Queirós A.
      • von Gunten A.
      • Martins M.
      • et al.
      The forgotten psychopathology of depressed long-term care facility residents: A call for evidence-based practice.
      Complex health conditions and high dropout rates, mainly because of death, complicate relevant studies.
      • Vossius C.
      • Selbæk G.
      • Šaltytė Benth J.
      • et al.
      Mortality in nursing home residents: A longitudinal study over three years.
      A recent systematic review of risk factors for depression in long-term care pointed to the general lack of knowledge of the interplay over time between depression and its risk factors.
      • Chau R.
      • Kissane D.W.
      • Davison T.E.
      Risk factors for depression in long-term care: A systematic review.
      As depression can be remediable, identifying groups of patients at particular risk is essential to tailor treatment programs.
      • Watt J.A.
      • Goodarzi Z.
      • Veroniki A.A.
      • et al.
      Comparative efficacy of interventions for reducing symptoms of depression in people with dementia: Systematic review and network meta-analysis.
      ,
      • Yuan Y.
      • Lapane K.L.
      • Baek J.
      • et al.
      Nursing home star ratings and new onset of depression in long-stay nursing home residents.
      ,
      • Leontjevas R.
      • Gerritsen D.L.
      • Smalbrugge M.
      • et al.
      A structural multidisciplinary approach to depression management in nursing-home residents: A multicentre, stepped-wedge cluster-randomised trial.
      There are few studies that include residents at NH admittance and assess their depressive trajectories over time, and such studies are warranted.
      • Matos Queirós A.
      • von Gunten A.
      • Martins M.
      • et al.
      The forgotten psychopathology of depressed long-term care facility residents: A call for evidence-based practice.
      ,
      • Yuan Y.
      • Lapane K.L.
      • Rothschild A.J.
      • et al.
      Changes in depressive symptoms and cognitive impairment in older long-stay nursing home residents in the USA: A latent transition analysis.
      Thus, in this 36-month longitudinal study of newly admitted NH residents with or without dementia we investigated the course of depressive symptoms and how resident characteristics were associated with the symptoms. We further aimed to identify groups of residents following the same symptom trajectory and to assess whether their characteristics at admittance were associated with belonging to such groups.

      Methods

      Design

      This article is based on data from the Resource Use and Disease course in dementia–Nursing Home (REDIC-NH) study. The REDIC-NH is a Norwegian, observational, multicenter, longitudinal study of newly admitted NH residents.
      • Roen I.
      • Selbaek G.
      • Kirkevold O.
      • et al.
      Resourse Use and Disease Couse in dementia–Nursing Home (REDIC-NH), a longitudinal cohort study; design and patient characteristics at admission to Norwegian nursing homes.
      For this report, the residents were assessed biannually over the course of a study period of 36 months (ie, 7 assessments, T0–T6).

      Participants

      We included 696 newly admitted residents representing 47 NHs. All residents 65 years or older, or younger than 65 years with established dementia of any etiologic subtype, with an expected stay of 4 weeks or more were eligible to participate in this study. We excluded residents with life expectancy of less than 6 weeks.

      Data Collection

      Data at baseline (T0) were collected within a month of admission to the NH, and follow-up data were collected biannually (Table 1 details the attrition throughout the study period). Data were collected between March 2012 and November 2017. Healthcare workers in NHs, mainly registered nurses (74%), collected the data under the supervision of 10 research nurses. Prior to the study, the research nurses attended a 5-day training program to learn the study procedures and measures, and the data collectors attended a 2-day training program. Data were collected from resident records and through structured interviews with the residents, their next of kin, and caregivers in the NHs.
      • Roen I.
      • Selbaek G.
      • Kirkevold O.
      • et al.
      Resourse Use and Disease Couse in dementia–Nursing Home (REDIC-NH), a longitudinal cohort study; design and patient characteristics at admission to Norwegian nursing homes.
      ,
      • Callegari E.
      • Benth J.
      • Selbæk G.
      • et al.
      Does psychotropic drug prescription change in nursing home patients the first 6 months after admission?.
      Table 1Demographic and Clinical Characteristics of the Study Population and Attrition Throughout the Study Period of 36 months
      T0 (n = 696)T1 (n = 508)T2 (n = 427)T3 (n = 349)T4 (n = 294)T5 (n = 235)T6 (n = 194)
      Age in y—mean (SD)84.5 (7.5)84.7 (7.5)85.2 (7.7)85.4 (7.9)85.4 (7.9)85.8 (8.1)86.1 (8.0)
      Female sex (%)64.166.367.769.666.064.771.1
      CSDD—mean (SD)6.6 (5.3)5.8 (4.9)5.9 (4.8)6.5 (4.8)6.4 (4.8)6.6 (5.4)6.4 (5.0)
      CSDD >8 (%)
      Barca ML, Engedal K, Selbaek G. A reliability and validity study of the Cornell scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord. 2010;29:438–447.
      30.423.626.029.327.326.827.8
      CDR, sum of boxes score—mean (SD)10.3 (4.3)11.0 (4.3)11.7 (4.2)12.4 (4.3)12.9 (4.3)13.6 (3.9)13.6 (4.2)
      PSMS sum—mean (SD)15.3 (4.5)16.0 (4,8)16.7 (4.7)17.5 (4.8)18.4 (4.8)19.4 (4.7)19.8 (5.0)
      GMHR (fair/poor, %), as opposed to (excellent/good, %)52.753.960.464.968.372.273.1
      MOBID-2, mean (SD)2.1 (2.1)2.1 (2.0)2.2 (2.0)2.2 (2.0)2.3 (2.1)2.3 (1.9)2.4 (2.0)
      Use of antidepressants (yes, %)28.438.040.339.841.842.141.8
      Attrition:
      Deaths (n, cumulative)0115191261324371422
      Resident moving home or to another institution (n, cumulative)0324951556061
      Resident or nursing home withdrawn from the study (n, cumulative)061012121618
      Not analyzed at the specific follow-up due to protocol violation (n)035192311141
      CDR, Clinical Dementia Rating (scoring range 0-18); CSDD, The Cornell Scale for Depression in Dementia (scoring range 0-38); GMHR, The General Medical Health Rating Scale (excellent/good, fair/poor); MOBID-2, Mobilization-Observation–Behavior–Intensity–Dementia 2; PSMS, The Physical Self Maintenance Scale; SD, standard deviation; T0, Admittance to nursing home; T1, 6-month follow-up; T2, 12-month follow-up; T3, 18-month follow-up; T4, 24-month follow-up; T5, 30-month follow-up, T6, 36-month follow-up.
      Barca ML, Engedal K, Selbaek G. A reliability and validity study of the Cornell scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord. 2010;29:438–447.

      Measures

      We used the Cornell Scale for Depression in Dementia (CSDD) to assess depressive symptoms.
      • Alexopoulos G.S.
      • Abrams R.C.
      • Young R.C.
      • et al.
      Cornell Scale for Depression in Dementia.
      The CSDD comprises 19 items that are scored as 0 (absent), 1 (mild or intermittent), 2 (severe); these can also be scored as “Item unable to evaluate.” Before calculating a sum score ranging from 0 to 38, where a higher score denotes more severe symptoms, any item marked as unable to evaluate or any missing values were replaced by a score of 0. This method of imputing the CSDD is in line with previous recommendation.
      • Leontjevas R.
      • Gerritsen D.L.
      • Vernooij-Dassen M.J.
      • et al.
      Comparative validation of proxy-based Montgomery-Asberg depression rating scale and Cornell scale for depression in dementia in nursing home residents with dementia.
      However, a sum score was not calculated if 20% or more of the items were missing.
      • Snowdon J.
      Depression in nursing homes.
      ,
      • Borza T.
      • Engedal K.
      • Bergh S.
      • et al.
      The course of depressive symptoms as measured by the Cornell scale for depression in dementia over 74 months in 1158 nursing home residents.
      The severity of dementia was assessed with the Clinical Dementia Rating (CDR) scale which covers 6 cognitive and functional domains.
      • Hughes C.P.
      • Berg L.
      • Danziger W.L.
      • et al.
      A new clinical scale for the staging of dementia.
      The scoring ranges from 0 (no dementia) to 3 (severe dementia). In the analysis, we used the CDR sum of boxes score, where the maximum score is 18 and a higher score denotes more severe dementia.
      • O'Bryant S.E.
      • Waring S.C.
      • Cullum C.M.
      • et al.
      Staging dementia using Clinical Dementia Rating Scale Sum of Boxes scores: A Texas Alzheimer's research consortium study.
      In addition, an expert panel including 2 of the authors (GS and SB), used all collected information to systematically diagnose dementia according to the International Classification of Diseases, Tenth Revision research criteria
      World Health Organization (WHO)
      The ICD-10 Classification of Mental and Behavioural Disorders.
      and mild cognitive impairment (MCI) according to the Winblad criteria.
      • Roen I.
      • Selbaek G.
      • Kirkevold O.
      • et al.
      Resourse Use and Disease Couse in dementia–Nursing Home (REDIC-NH), a longitudinal cohort study; design and patient characteristics at admission to Norwegian nursing homes.
      ,
      • 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.
      We used the Physical Self-Maintenance Scale (PSMS)
      • Lawton M.P.
      • Brody E.M.
      Assessment of older people: Self-maintaining and instrumental activities of daily living.
      to assess each resident ability to perform personal activities of daily living. The PSMS covers 6 domains (scored 1–5), where a higher score denotes a lower level of functioning and the maximum score is 30. The PSMS and CDR sum of boxes were imputed and totaled in the same way as the CSDD.
      Physical health was assessed with the General Medical Health Rating Scale (GMHR).
      • Lyketsos C.G.
      • Galik E.
      • Steele C.
      • et al.
      The General Medical Health Rating: A bedside global rating of medical comorbidity in patients with dementia.
      The GMHR rates the resident medical conditions and use of medication on a 4-point scale (excellent, good, fair, poor). In the analyses, the GMHR was dichotomized to “good” (excellent/good) and “poor” (fair/poor).
      The Mobilization-Observation-Behavior-Intensity-Dementia-2 (MOBID-2) Pain Scale was used to assess pain.
      • Husebo B.S.
      • Strand L.I.
      • Moe-Nilssen R.
      • et al.
      Mobilization-Observation-Behavior-Intensity-Dementia Pain Scale (MOBID): Development and validation of a nurse-administered pain assessment tool for use in dementia.
      In the analyses, we used the overall assessment of total pain on a 10-point visual analogue scale, where a higher score denotes more pain.
      Use of medication was registered with the Anatomical Therapeutic Chemical classification system. We dichotomized the use of antidepressants (N06A) (yes/no) during the study period for the analyses.
      The Supplementary Table 1 provides a more extensive description of the assessment scales.

      Ethical and Legal Considerations

      Nursing home personnel, including the physician, evaluated each resident's capacity to participate in the study. All residents with the capacity to participate gave written consent. For residents without the capacity to consent, their next of kin gave consent on their behalf. The Regional Ethics Committee for Medical Research in South-Eastern Norway approved the study (2011/1738a).

      Statistics

      We used the SPSS v 26 (IBM Corp), SAS v 9.4 (SAS Institute Inc), and STATA v 16 (StataCorp. LLC) to analyze data. Demographic and clinical characteristics were described as means and standard deviations (SDs) or frequencies and percentages.
      The intraclass correlation coefficient was estimated to assess whether there was a hierarchical structure present in the dataset as the residents were included from different wards within different NHs. Cluster-effect was identified on both levels. Therefore, the overall trend in CSDD scores through the 7 assessment points was assessed by an unadjusted linear mixed model (LMM) with random effects for residents nested within ward which was again nested within NH. The model included a third-order polynomial for time as fixed effect because of a nonlinear trend in CSDD scores. The model was further adjusted for preselected covariates (severity of dementia, functional impairment, physical health, pain, use of antidepressants, age, and sex), based on a recent systematic review of the most important risk factors for depression in NHs.
      • Chau R.
      • Kissane D.W.
      • Davison T.E.
      Risk factors for depression in long-term care: A systematic review.
      Whenever possible, covariates assessed longitudinally and simultaneously with CSDD were included, and interactions between each covariate and time were added as fixed effects into the model. The adjusted model was then reduced for exhaustive interactions by Bayes Information Criterion (BIC), where a smaller value means a better model. Among information criteria, BIC is considered as most appropriate in selecting a correct model.
      • Chakrabarti A.
      • Ghosh J.K.
      AIC, BIC and recent advances in model selection. Philosophy of statistics. Volume 7 in Handbook of the Philosophy of Science.
      Post-hoc analyses were carried out to derive mean changes in CSDD scores between time points.
      As a large share of residents died during the study period and mortality could have been informative with respect to CSDD scores, joint modeling assessing the longitudinal and survival processes simultaneously was performed. However, no evidence for informative mortality was identified; hence, only LMM results were reported.
      LMM assesses trend among all residents. To explore the data further, a growth mixture model was estimated.
      • Nagin D.S.
      Group-Based Modeling of Development.
      ,
      • Ram N.
      • Grimm K.J.
      Growth mixture modeling: A method for identifying differences in longitudinal change among unobserved groups.
      This approach attempts to identify unobserved groups of residents, each following a distinct trajectory in CSDD score, based on individual profiles and using a set of statistical criteria; BIC, nonoverlapping 95% confidence intervals of trajectories, reasonably large group size, and high average posterior probability of assignment. A nominal regression model was then estimated to assess whether the covariates at baseline were associated with group membership. Results with a P value of <.05 were considered statistically significant.

      Results

      Table 1 presents the demographic and clinical characteristics of the residents at different time points throughout the study period. Of the 696 residents included at T0, 446 (64.1%) were women and mean age was 84.5 (SD = 7.5) years, 583 (83.8%) had dementia, 96 had MCI (13.8%), 16 (2.3%) were cognitively healthy, and 1 (0.1%) could not be classified with respect to cognitive status.
      In the LMM, none of the interactions between covariates and time were kept in the model, implying that change in CSDD score was not associated with the covariates. There was a nonlinear trend in CSDD score during the study period in both unadjusted and adjusted models (Figure 1, Table 2, and Supplementary Table 2 details the post hoc analyses). More severe dementia, lower level of functioning, poorer physical health, more pain, use of antidepressants, and younger age at admission were associated with higher CSDD scores in both unadjusted and adjusted LMMs (Table 2).
      Figure thumbnail gr1
      Fig. 1Time trend of CSDD score (scoring range 0‒38) over 36 months according to linear mixed model.
      Table 2Results of Linear Mixed Model for CSDD Score Adjusted for Cluster Effect Within Ward Nested Within NH
      Unadjusted ModelsAdjusted Model
      RC (95% SE)P ValueRC (95% SE)P Value
      Intercept6.41 (0.26)<.0016.23 (1.64)<.001
      Time
      Third order polynomial.
      −0.14 (0.05).004−0.22 (0.05)<.001
      Time × time
      Third order polynomial.
      0.01 (0.004).0040.01 (0.003).001
      Time × time × time
      Third order polynomial.
      −0.0002 (0.00007).012−0.0002 (0.00007).011
      Covariate
       CDR-sum of boxes0.32 (0.03)<.0010.22 (0.03)<.001
       PSMS0.27 (0.02)<.0010.10 (0.03)<.001
       GMHR (good as ref.)1.71 (0.20)<.0011.04 (0.20)<.001
       MOBID-20.51 (0.05)<.0010.40 (0.05)<.001
       Use of antidepressants (“no” as ref.)1.25 (0.24)<.0011.11 (0.23)<.001
       Age−0.08 (0.02)<.001−0.06 (0.02).001
       Sex (female as ref.)0.03 (0.31).930−0.07 (0.28).799
      CDR, Clinical Dementia Rating; CSDD, The Cornell Scale for Depression in Dementia; GMHR, The General Medical Health Rating Scale; MOBID-2, The Mobilization-Observation–Behavior–Intensity–Dementia 2 Pain Scale; PSMS, The Physical Self Maintenance Scale; RC, Regression coefficient; ref., reference; SE, standard error.
      Bold indicates the P values which are significant.
      Third order polynomial.
      The growth mixture model for CSDD scores identified 4 different groups of residents each following distinct trajectories. Based on the severity and course of depressive symptoms over time, we labeled the groups as follows: persistent mild symptoms group (n = 225, 32.6%), persistent moderate symptoms group (n = 351, 50.8%), increasing symptoms group (n = 35, 5.1%), and severe but decreasing symptoms group (n = 80, 11.6%) (Figure 2 and Table 3). A low intercept (2.2) and persistent low CSDD scores throughout the study period characterized the group with persistent mild symptoms. The persistent moderate symptoms group had an intercept of 6.8 and the following CSDD scores were approximately 6‒7. The increasing symptoms group had an intercept of 5.8, then the CSDD scores increased successively before they flattened/decreased towards the end. The severe but decreasing group had a high intercept at 15.2, and the CSDD scores decreased successively throughout the study period but stayed above 10. The average within-group probability for all groups was 0.78 or higher.
      Figure thumbnail gr2
      Fig. 2Growth mixture model of CSDD scores (scoring range 0‒38) showing trajectory groups (n = 691; ie, patients with nonmissing values of CSDD at inclusion). CSDD, Cornell Scale for Depression in Dementia.
      Table 3Results of Growth Mixture Model (n = 691, ie, Patients With Nonmissing Values of CSDD at Inclusion) and Nominal Regression Model for Associations Between Clinical Characteristics and Group Belonging (n = 656, Group 1, Persistent Mild Symptoms Group as Reference)
      Growth Mixture Model
      Group 1 (n = 225, 32.6%)Group 2 (n = 351, 50.8%)Group 3 (n = 35, 5.1%)Group 4 (n = 80, 11.6%)
      RC (SE)P ValueRC (SE)P ValueRC (SE)P ValueRC (SE)P Value
      Intercept2.15 (0.28)<.0016.80 (0.33)<.0015.76 (0.91)<.00115.24 (0.50)<.001
      Linear0.005 (0.01).711−0.06 (0.04).1160.15 (0.20).457−0.14 (0.03)<.001
      Quadratic0.002 (0.001).0300.03 (0.01).030
      Cubic−0.001 (0.0003).004
      Average within-group probability0.810.780.850.86
      Nominal Regression Model
      Unadjusted ModelsAdjusted Model
      CovariateOR (95% CI)P ValueOR (95% CI)P Value
      CDR-sum of boxes
       Group 21.00 (1.00; 1.01).3821.00 (0.99; 1.01).829
       Group 30.98 (0.96; 1.01).1720.98 (0.96; 1.01).204
       Group 41.00 (0.99; 1.01).7551.00 (0.99; 1.01).640
      PSMS
       Group 21.06 (1.02; 1.10).0071.04 (0.99; 1.08).114
       Group 30.95 (0.87; 1.04).2990.97 (0.88; 1.07).567
       Group 41.13 (1.07; 1.20)<.0011.09 (1.02; 1.16).009
      GMHR
       Group 21.44 (1.02; 2,03).0371.20 (0.03; 1.75).320
       Group 30.63 (0.29; 1.37).2460.70 (0.31; 1.60).409
       Group 42.51 (1.44; 4.38).0011.84 (1.02; 3.35).045
      MOBID-2
       Group 21.21 (1.10; 1.32)<.0011.19 (1.08; 1.31)<.001
       Group 30.94 (0.76; 1.17).6050.99 (0.79; 1.24).941
       Group 41.31 (1.16; 1.48)<.0011.25 (1.10; 1.43).001
      Use of antidepressants (“yes”)
       Group 21.54 (1.03; 2.30).0371.36 (0.90; 2.07).150
       Group 31.68 (0.74; 3.80).2131.56 (0.68; 3.58).290
       Group 42.52 (1.43; 4.45).0022.18 (1.20; 3.98).011
      Age
       Group 20.95 (0.93; 0.98)<.0010.95 (0.92; 0.97)<.001
       Group 30.97 (0.92; 1.02).1830.97 (0.92; 1.02).247
       Group 40.94 (0.91; 0.97)<.0010.94 (0.91; 0.97).001
      Sex (male)
       Group 21.10 (0.77; 1.58).5881.02 (0.70; 1.50).907
       Group 30.87 (0.39; 1.94).7370.95 (0.42; 2.16).905
       Group 41.31 (0.76; 2.25).3321.20 (0.67; 2.14).547
      CDR, Clinical Dementia Rating; CI, confidence interval; Group 1, Persistent mild symptoms; Group 2, persistent moderate symptoms; Group 3, increasing symptoms; Group 4, severe but decreasing symptoms; CSDD, The Cornell Scale for Depression in Dementia; GMHR, The General Medical Health Rating Scale; MOBID-2, Mobilization-Observation–Behavior–Intensity–Dementia 2 Pain Scale; OR, odds ratio; PSMS, The Physical Self Maintenance; RC, regression coefficient; SE, standard error.
      Bold indicates the P values which are significant.
      According to adjusted nominal regression, lower level of functioning, poorer physical health, more pain, use of antidepressants, and younger age at admission were associated with higher odds for belonging to the group with severe but decreasing symptoms as compared with the group with persistent mild symptoms. In addition, more pain and younger age at admission were associated with higher odds for belonging to the group with persistent moderate symptoms compared with the group with persistent mild symptoms. Comparing the group with increasing symptoms with the group with persistent mild symptoms, there were no differences with respect to covariates (Table 3).

      Discussion

      This study showed a nonlinear trend in CSDD score over the study period of 36 months. There was a decrease in CSDD score the first 6 months, and from 12-month follow-up onward, the score increased successively but flattened and decreased somewhat toward the end of the study period. More severe dementia, lower level of functioning, poorer physical health, more pain, use of antidepressants, and younger age at admission were associated with more depressive symptoms. Furthermore, we were able to identify 4 groups following distinct trajectories in CSDD scores over 36 months: (1) persistent mild symptoms, (2) persistent moderate symptoms, (3) increasing symptoms, and (4) severe but decreasing symptoms.
      According to the LMM, the residents had more depressive symptoms upon entering the NH than they did 6 months later. Previous studies have described the first period in a NH as challenging for vulnerable older persons and have noted that NH residents can display more depressive symptoms during their first months in a NH.
      • Custers A.F.
      • Cillessen A.H.
      • Westerhof G.J.
      • et al.
      Need fulfillment, need importance, and depressive symptoms of residents over the first eight months of living in a nursing home.
      ,
      • Smalbrugge M.
      • Jongenelis L.
      • Pot A.M.
      • et al.
      Incidence and outcome of depressive symptoms in nursing home patients in the Netherlands.
      ,
      • Yuan Y.
      • Lapane K.L.
      • Rothschild A.J.
      • et al.
      Changes in depressive symptoms and cognitive impairment in older long-stay nursing home residents in the USA: A latent transition analysis.
      In a 20-year study, Villenueve et al discussed the phenomenon of “relocation stress syndrome” in persons who entered NHs.
      • Villeneuve R.
      • Meillon C.
      • Dartigues J.F.
      • et al.
      Trajectory of quality of life before and after entering a nursing home: A longitudinal study.
      In addition, depression can be a reason for NH placement.
      • Hajek A.
      • Brettschneider C.
      • Lange C.
      • et al.
      Longitudinal predictors of institutionalization in old age.
      ,
      • Giebel C.
      • Sutcliffe C.
      • Verbeek H.
      • et al.
      Depressive symptomatology and associated factors in dementia in Europe: Home care versus long-term care.
      Resident adaptions to the NH setting could be a factor to the decrease in depressive symptoms we observed 6 months after inclusion in our study, as also seen in other studies.
      • Smalbrugge M.
      • Jongenelis L.
      • Pot A.M.
      • et al.
      Incidence and outcome of depressive symptoms in nursing home patients in the Netherlands.
      ,
      • Villeneuve R.
      • Meillon C.
      • Dartigues J.F.
      • et al.
      Trajectory of quality of life before and after entering a nursing home: A longitudinal study.
      In the present study, we observed an increase in depressive symptoms after 12-month follow-up. There are few longitudinal studies for comparison. However, in a previous longitudinal study, we included residents with a mean stay of 941 days in a NH and had different assessment intervals, but we also saw increasing CSDD symptoms after 12-month follow-up.
      • Borza T.
      • Engedal K.
      • Bergh S.
      • et al.
      The course of depressive symptoms as measured by the Cornell scale for depression in dementia over 74 months in 1158 nursing home residents.
      This was related to an increase in nonmood symptoms of the CSDD (lack of joy, irritability, agitation, retardation, loss of interest, appetite loss, weight loss, lack of energy, diurnal variation, difficulty falling asleep, multiple awakening, and early morning awakening).
      • Borza T.
      • Engedal K.
      • Bergh S.
      • et al.
      The course of depressive symptoms as measured by the Cornell scale for depression in dementia over 74 months in 1158 nursing home residents.
      ,
      • Barca M.L.
      • Selbaek G.
      • Laks J.
      • et al.
      The pattern of depressive symptoms and factor analysis of the Cornell Scale among patients in Norwegian nursing homes.
      Consequently, in the present study, the increase in depressive symptoms after 12 months may be related to increasing nonmood CSDD symptoms. Moreover, Kurlowicz et al demonstrated how CSDD could label symptoms as depressive that were not necessarily due to a depressive disorder among frail NH residents with multimorbidity and functional impairment.
      • Kurlowicz L.H.
      • Evans L.K.
      • Strumpf N.E.
      • et al.
      A psychometric evaluation of the Cornell Scale for Depression in Dementia in a frail, nursing home population.
      Indeed, our study population became more cognitively and functionally impaired and more residents experienced poorer physical health during the study period. Greater awareness of depressive symptoms over time in NH studies may also be a factor related to the increase in observer-rated depressive symptoms.
      • McCusker J.
      • Cole M.G.
      • Voyer P.
      • et al.
      Observer-rated depression in long-term care: Frequency and risk factors.
      The flattening and decrease of CSDD score toward the end of the study period may, at least partly, be related to a survivor effect. Authors have previously described that survivors in NHs may display fewer depressive symptoms.
      • Borza T.
      • Engedal K.
      • Bergh S.
      • et al.
      The course of depressive symptoms as measured by the Cornell scale for depression in dementia over 74 months in 1158 nursing home residents.
      ,
      • Boorsma M.
      • Joling K.
      • Dussel M.
      • et al.
      The incidence of depression and its risk factors in Dutch nursing homes and residential care homes.
      Although we have identified significant differences in CSDD scores during the study period, the clinical impact of these differences lying in the range of 1–2 points is debatable. The minimum clinically important difference (MCID) for CSDD has not been established.
      • Watt J.A.
      • Veroniki A.A.
      • Tricco A.C.
      • et al.
      Using a distribution-based approach and systematic review methods to derive minimum clinically important differences.
      In our analyses, we included covariates mostly assessed longitudinally, and the findings were strikingly clear and robust. Our findings support conclusions from a systematic review of risk factors for depression in NH residents that cognitive and functional impairments are the most consistently documented risk factors. However, there are fewer studies and previous findings are more mixed with respect to physical health, pain, use of antidepressants, and age at admission as risk factors.
      • Jenny Wei Y.J.
      • Chen C.
      • Fillingim R.B.
      • et al.
      Uncontrolled pain and risk for depression and behavioral symptoms in residents with dementia.
      ,
      • Chau R.
      • Kissane D.W.
      • Davison T.E.
      Risk factors for depression in long-term care: A systematic review.
      Further, studies have pointed to bidirectional relationships between depression and covariates in our study that can complicate the picture.
      • Erdal A.
      • Flo E.
      • Selbaek G.
      • et al.
      Associations between pain and depression in nursing home patients at different stages of dementia.
      ,
      • Chau R.
      • Kissane D.W.
      • Davison T.E.
      Risk factors for depression in long-term care: A systematic review.
      ,
      • Parmelee P.A.
      • Katz I.R.
      • Lawton M.P.
      Incidence of depression in long-term care settings.
      ,
      • Bergh S.
      • Selbaek G.
      • Engedal K.
      Discontinuation of antidepressants in people with dementia and neuropsychiatric symptoms (DESEP study): Double blind, randomised, parallel group, placebo controlled trial.
      The CSDD mean scores in our study were around 6 or 7 at all assessments but with relatively large SDs, indicating high between-patient variation. The growth mixture model identified 4 groups with distinct CSDD trajectories, with intercepts varying from 2.2 to 15.2. The overall increase in CSDD score from 12- to 30-month follow-up in the linear mixed model likely relates to the group with increasing symptoms and a small increase in CSDD score among the largest group with persistent moderate symptoms. Likewise, the decrease in CSDD score the first 6 months likely relates to the group severe but decreasing symptoms with highest intercept. There could be that a regression to the mean effect to some extent influenced the results. Another study of depressive symptom trajectories in NHs applied a method different from ours (hierarchical clustering) that complicates comparison. The study used the Geriatric Depression Scale to monitor depressive symptoms over 6 months, excluded residents with moderate to severe cognitive impairment, and found 3 homogeneous groups with distinct trajectories separated according to symptom severity.
      • McCusker J.
      • Cole M.G.
      • Voyer P.
      • et al.
      Six-month trajectories of self-reported depressive symptoms in long-term care.
      In our study, approximately 5% followed the trajectory with increasing symptoms that crossed other trajectories. It would be of interest to characterize this group further to better identify residents at increasing risk for depression, but the nominal regression model was not able to demonstrate any differences compared with the group with persistent mild symptoms. Lack of statistical power could be a reason as only 35 (5.1%) residents were in the group with increasing symptoms. About 12% were in the group with severe symptoms that decreased during the study period. Residents with lower levels of functioning, poorer physical health, more pain, who used antidepressants, and at younger age at admittance to a NH had higher odds for belonging to this group compared with the group with persistent mild symptoms. This supports the notion that the presence of these factors at admittance is associated with more depressive symptoms over time. The fact that use of antidepressants was associated with more severe symptoms in several analyses in our study points to an opportunity for more-personalized antidepressant treatment and other possible treatments for depression in NH residents.
      • Watt J.A.
      • Goodarzi Z.
      • Veroniki A.A.
      • et al.
      Comparative efficacy of interventions for reducing symptoms of depression in people with dementia: Systematic review and network meta-analysis.
      ,
      • Bergh S.
      • Selbaek G.
      • Engedal K.
      Discontinuation of antidepressants in people with dementia and neuropsychiatric symptoms (DESEP study): Double blind, randomised, parallel group, placebo controlled trial.
      ,
      • Meeks S.
      • Van Haitsma K.
      • Schoenbachler B.
      • et al.
      BE-ACTIV for depression in nursing homes: Primary outcomes of a randomized clinical trial.
      • Dudas R.
      • Malouf R.
      • McCleery J.
      • et al.
      Antidepressants for treating depression in dementia.
      • Davison T.E.
      • McCabe M.P.
      • Busija L.
      • et al.
      A cluster randomised trial of the program to enhance adjustment to residential living (PEARL): A novel psychological intervention to reduce depression in newly admitted aged care residents.
      This study has limitations. Depressive symptom, particularly in patients with dementia can fluctuate in severity and duration,
      • Alexopoulos G.S.
      • Abrams R.C.
      • Young R.C.
      • et al.
      Cornell Scale for Depression in Dementia.
      ,
      • Bergh S.
      • Selbaek G.
      • Engedal K.
      Discontinuation of antidepressants in people with dementia and neuropsychiatric symptoms (DESEP study): Double blind, randomised, parallel group, placebo controlled trial.
      and the study might have benefitted from more frequent assessments. Many different assessors might yield variations in the collected data that could influence the results. However, the assessors and the study nurses participated in standardized training to reduce variability, and the statistical models adjusted for potential cluster effect. Our study included well-documented risk factors as covariates, but did not deal with distinct psychological factors, such as perceived support, or NH-specific characteristics.
      • Chau R.
      • Kissane D.W.
      • Davison T.E.
      Risk indicators for depression in aged care: The contribution of a meaningful life, mastery and environmental fit.
      It is a weakness of the study that only 38 of the 47 NHs were able to report data on eligible residents and in these 38 NH less than 50% of eligible residents were included.
      • Roen I.
      • Selbaek G.
      • Kirkevold O.
      • et al.
      Resourse Use and Disease Couse in dementia–Nursing Home (REDIC-NH), a longitudinal cohort study; design and patient characteristics at admission to Norwegian nursing homes.
      Furthermore, previous studies reporting results from the REDIC-NH study discussed the representativeness of the study because of the exclusion of respite residents with shorter expected stays and more women opposed to men, and a somewhat younger mean age among residents who declined to participate compared with those who participated.
      • Vossius C.
      • Selbæk G.
      • Šaltytė Benth J.
      • et al.
      Mortality in nursing home residents: A longitudinal study over three years.
      ,
      • Roen I.
      • Selbaek G.
      • Kirkevold O.
      • et al.
      Resourse Use and Disease Couse in dementia–Nursing Home (REDIC-NH), a longitudinal cohort study; design and patient characteristics at admission to Norwegian nursing homes.
      ,
      • Callegari E.
      • Benth J.
      • Selbæk G.
      • et al.
      Does psychotropic drug prescription change in nursing home patients the first 6 months after admission?.
      The strengths of this study are the longitudinal design involving residents at admittance to a NH, a reasonably large sample size, and a long follow-up period. The residents were assessed at regular intervals with well-established scales. Participants who drop out and missing data are inherent challenges to studies of this type. We used advanced statistical models including all available data in the analyses, also data from dropouts. Moreover, the joint modeling did not find evidence for informative mortality with respect to CSDD scores. The associations between CSDD scores and the selected covariates align with previous knowledge, which we see as a strength for our study. We included residents with all degrees of cognitive impairment, and only residents with life expectancy of less than 6 weeks were excluded. Thus, the study is of high clinical relevance for NHs.

      Conclusions and Implications

      Findings from our study contribute to more knowledge about depressive symptoms in NHs that may help to prevent depression and individualize treatment. Most (83%) NH residents were in trajectory groups with persistent mild or moderate depressive symptoms. Residents with more severe dementia, lower levels of functioning, poor physical health, severe pain, younger age at admittance, and who use antidepressants should be monitored closely and systematically with respect to depression. Our finding that the use of antidepressants was noticeably associated with more severe depressive symptoms calls for closely monitoring effects and for taking action toward more-personalized antidepressant treatment and other treatments for depression in the NH setting. Further research should aim to investigate the importance of personalized treatment programs for depression in NH residents.

      Acknowledgments

      We thank the patients and their next of kin for their participation in the study. We also thank the participating workers in the nursing homes and the research nurses who collected the data.

      Supplementary Data

      Supplementary Table 1Description of Assessment Scales Used to Collect Information for the Measures in the Study
      Clinical FeaturesAssessment Scale/MethodsDescription
      Depressive symptomsCornell Scale for Depression in Dementia (CSDD)The CSDD evaluates the following 19 items: anxiety, sadness, lack of joy
      Nonmood factors of the CSDD.†
      , irritability
      Nonmood factors of the CSDD.†
      , agitation
      Nonmood factors of the CSDD.†
      , retardation
      Nonmood factors of the CSDD.†
      , multiple physical complaints, loss of interest
      Nonmood factors of the CSDD.†
      , appetite loss
      Nonmood factors of the CSDD.†
      , weight loss
      Nonmood factors of the CSDD.†
      , lack of energy
      Nonmood factors of the CSDD.†
      , diurnal variation
      Nonmood factors of the CSDD.†
      , difficulty falling asleep
      Nonmood factors of the CSDD.†
      , multiple awakening
      Nonmood factors of the CSDD.†
      , early morning awakening
      Nonmood factors of the CSDD.†
      , suicidal ideation, poor self-esteem, pessimism, and delusion.
      Barca, ML, Selbaek, G, Laks, J, et al. The pattern of depressive symptoms and factor analysis of the Cornell Scale among patients in Norwegian nursing homes. Int J Geriatr Psychiatry. 2008;23:1058‒1065.


      The 19 items are scored 0 (absent), 1 (mild or intermittent), 2 (severe), yielding a maximum score of 38, and higher score denotes more severe symptoms. A cut-off point of 8/9 can be indicative for depression.
      Barca, ML, Engedal, K, Selbaek, G. A reliability and validity study of the Cornell Scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord. 2010;29:438‒447.
      CognitionClinical Dementia Rating (CDR)The CDR evaluates the following 6 cognitive and functional domains: memory, orientation, judgement and problem solving, community affairs, home and hobbies, and personal care.

      The 6 items are rated from 0 (no dementia) to 3 (severe dementia). The CDR sum of boxes score ranges from 0‒18, and a higher score denotes more severe cognitive impairment.
      Personal activities of daily livingPhysical Self-Maintenance Scale (PSMS)The PSMS evaluates toileting, feeding, dressing, grooming, physical ambulation, and bathing.

      The six items are rated from 1 to 5 (total scoring range 6‒30), and a higher score denotes lower level of functioning.
      Physical healthGeneral Medical Health Rating Scale (GMHR)The GMHR takes into account the number of general medical conditions, the severity of the conditions, and the use of medications.

      The GMHR allows ratings of excellent/good/fair/poor.
      PainMobilization-Observation-Behavior-Intensity-Dementia Pain Scale (MOBID-2)The MOBID-2 is proxy rated and evaluates pain intensity based on 5 guided movements and observations of five body organs the previous days.

      The 10 items are rated on a 10-point visual analogue scale and the scales yields a final overall assessment of pain, range 0‒10 where a higher score denotes more pain.
      Nonmood factors of the CSDD.
      Barca, ML, Selbaek, G, Laks, J, et al. The pattern of depressive symptoms and factor analysis of the Cornell Scale among patients in Norwegian nursing homes. Int J Geriatr Psychiatry. 2008;23:1058‒1065.
      Barca, ML, Selbaek, G, Laks, J, et al. The pattern of depressive symptoms and factor analysis of the Cornell Scale among patients in Norwegian nursing homes. Int J Geriatr Psychiatry. 2008;23:1058‒1065.
      Barca, ML, Engedal, K, Selbaek, G. A reliability and validity study of the Cornell Scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord. 2010;29:438‒447.
      Supplementary Table 2Results of Post-Hoc Analyses Assessing Change in CSDD Score After LMM
      Time PointsT1T2T3T4T5T6
      T0−0.5 (−0.9; −0.1), 0.018−0.4 (−1.0; 0.1), 0.125−0.1 (−0.8; 0.6), 0.7450.3 (−0.5; 1.1), 0.5050.5 (−0.4; 1.3), 0.2640.3 (−0.7; 1.2), 0.561
      T10.04 (−0.2; 0.3), 0.7430.4 (−0.1; 0.9), 0.1310.8 (0.1; 1.4), 0.0291.0 (0.2; 1.7), 0.0100.8 (0.01; 1.5), 0.046
      T20.3 (0.1; 0.6), 0.0170.7 (0.2; 1.2), 0.0040.9 (0.4; 1.5), 0.0020.7 (0.05; 1.4), 0.035
      T30.4 (0.2; 0.6), 0.0010.6 (0.2; 1.0), 0.0010.4 (−0.3; 1.0), 0.241
      T40.2 (0.01; 0.40), 0.0400.0 (−0.7; 0.7), 0.997
      T5−0.2 (−0.7; 0.3), 0.426
      T0, admittance to nursing home; T1, 6-month follow-up; T2, 12-month follow-up; T3, 18-month follow-up; T4, 24-month follow-up; T5, 30-month follow-up; T6, 36-month follow-up.
      Numbers are mean changes, 95% confidence intervals, and P values. Bold indicates the P values which are significant.

      References

        • Seitz D.
        • Purandare N.
        • Conn D.
        Prevalence of psychiatric disorders among older adults in long-term care homes: A systematic review.
        Int Psychogeriatr. 2010; 22: 1025-1039
        • Snowdon J.
        Depression in nursing homes.
        Int Psychogeriatr. 2010; 22: 1143-1148
        • Custers A.F.
        • Cillessen A.H.
        • Westerhof G.J.
        • et al.
        Need fulfillment, need importance, and depressive symptoms of residents over the first eight months of living in a nursing home.
        Int Psychogeriatr. 2014; 26: 1161-1170
        • Veltman E.
        • Kok A.
        • Lamers F.
        • et al.
        Stability and transition of depression subtypes in late life.
        J Affect Disord. 2020; 265: 445-452
        • Matos Queirós A.
        • von Gunten A.
        • Martins M.
        • et al.
        The forgotten psychopathology of depressed long-term care facility residents: A call for evidence-based practice.
        Dement Geriatr Cogn Dis Extra. 2021; 11: 38-44
        • Erdal A.
        • Flo E.
        • Selbaek G.
        • et al.
        Associations between pain and depression in nursing home patients at different stages of dementia.
        J Affect Disord. 2017; 218: 8-14
        • Jenny Wei Y.J.
        • Chen C.
        • Fillingim R.B.
        • et al.
        Uncontrolled pain and risk for depression and behavioral symptoms in residents with dementia.
        J Am Med Dir Assoc. 2021; 22: 2079-2086.e2075
        • Hurt C.
        • Bhattacharyya S.
        • Burns A.
        • et al.
        Patient and caregiver perspectives of quality of life in dementia. An investigation of the relationship to behavioural and psychological symptoms in dementia.
        Dement Geriatr Cogn Disord. 2008; 26: 138-146
        • Rapp M.A.
        • Schnaider-Beeri M.
        • Wysocki M.
        • et al.
        Cognitive decline in patients with dementia as a function of depression.
        Am J Geriatr Psychiatry. 2011; 19: 357-363
        • Ulbricht C.M.
        • Rothschild A.J.
        • Hunnicutt J.N.
        • et al.
        Depression and cognitive impairment among newly admitted nursing home residents in the USA.
        Int J Geriatr Psychiatry. 2017; 32: 1172-1181
        • Kaup B.A.
        • Loreck D.
        • Gruber-Baldini A.L.
        • et al.
        Depression and its relationship to function and medical status, by dementia status, in nursing home admissions.
        Am J Geriatr Psychiatry. 2007; 15: 438-442
        • Chau R.
        • Kissane D.W.
        • Davison T.E.
        Risk factors for depression in long-term care: A systematic review.
        Clin Gerontol. 2019; 42: 224-237
        • Barca M.L.
        • Engedal K.
        • Laks J.
        • et al.
        A 12 months follow-up study of depression among nursing-home patients in Norway.
        J Affect Disord. 2010; 120: 141-148
        • Watt J.A.
        • Goodarzi Z.
        • Veroniki A.A.
        • et al.
        Comparative efficacy of interventions for reducing symptoms of depression in people with dementia: Systematic review and network meta-analysis.
        BMJ. 2021; 372: n532
        • Smalbrugge M.
        • Jongenelis L.
        • Pot A.M.
        • et al.
        Incidence and outcome of depressive symptoms in nursing home patients in the Netherlands.
        Am J Geriatr Psychiatry. 2006; 14: 1069-1076
        • Rivera E.
        • Hirschman K.B.
        • Naylor M.D.
        Reported needs and depressive symptoms among older adults entering long-term services and supports.
        Innov Aging. 2020; 4: igaa021
        • Vossius C.
        • Selbæk G.
        • Šaltytė Benth J.
        • et al.
        Mortality in nursing home residents: A longitudinal study over three years.
        PLoS One. 2018; 13: e0203480
        • Yuan Y.
        • Lapane K.L.
        • Baek J.
        • et al.
        Nursing home star ratings and new onset of depression in long-stay nursing home residents.
        J Am Med Dir Assoc. 2019; 20: 1335-1339.e1310
        • Leontjevas R.
        • Gerritsen D.L.
        • Smalbrugge M.
        • et al.
        A structural multidisciplinary approach to depression management in nursing-home residents: A multicentre, stepped-wedge cluster-randomised trial.
        Lancet. 2013; 381: 2255-2264
        • Yuan Y.
        • Lapane K.L.
        • Rothschild A.J.
        • et al.
        Changes in depressive symptoms and cognitive impairment in older long-stay nursing home residents in the USA: A latent transition analysis.
        Aging Ment Health. 2021; 25: 1903-1912
        • Roen I.
        • Selbaek G.
        • Kirkevold O.
        • et al.
        Resourse Use and Disease Couse in dementia–Nursing Home (REDIC-NH), a longitudinal cohort study; design and patient characteristics at admission to Norwegian nursing homes.
        BMC Health Serv Res. 2017; 17: 365
        • Callegari E.
        • Benth J.
        • Selbæk G.
        • et al.
        Does psychotropic drug prescription change in nursing home patients the first 6 months after admission?.
        J Am Med Dir Assoc. 2021; 22: 101-108.e101
        • Alexopoulos G.S.
        • Abrams R.C.
        • Young R.C.
        • et al.
        Cornell Scale for Depression in Dementia.
        Biol Psychiatry. 1988; 23: 271-284
        • Leontjevas R.
        • Gerritsen D.L.
        • Vernooij-Dassen M.J.
        • et al.
        Comparative validation of proxy-based Montgomery-Asberg depression rating scale and Cornell scale for depression in dementia in nursing home residents with dementia.
        Am J Geriatr Psychiatry. 2012; 20: 985-993
        • Borza T.
        • Engedal K.
        • Bergh S.
        • et al.
        The course of depressive symptoms as measured by the Cornell scale for depression in dementia over 74 months in 1158 nursing home residents.
        J Affect Disord. 2015; 175: 209-216
        • Hughes C.P.
        • Berg L.
        • Danziger W.L.
        • et al.
        A new clinical scale for the staging of dementia.
        Br J Psychiatry. 1982; 140: 566-572
        • O'Bryant S.E.
        • Waring S.C.
        • Cullum C.M.
        • et al.
        Staging dementia using Clinical Dementia Rating Scale Sum of Boxes scores: A Texas Alzheimer's research consortium study.
        Arch Neurol. 2008; 65: 1091-1095
        • World Health Organization (WHO)
        The ICD-10 Classification of Mental and Behavioural Disorders.
        Diagnostic Criteria for Research, 1993
        • 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 Intern Med. 2004; 256: 240-246
        • Lawton M.P.
        • Brody E.M.
        Assessment of older people: Self-maintaining and instrumental activities of daily living.
        Gerontologist. 1969; 9: 179-186
        • Lyketsos C.G.
        • Galik E.
        • Steele C.
        • et al.
        The General Medical Health Rating: A bedside global rating of medical comorbidity in patients with dementia.
        J Am Geriatr Soc. 1999; 47: 487-491
        • Husebo B.S.
        • Strand L.I.
        • Moe-Nilssen R.
        • et al.
        Mobilization-Observation-Behavior-Intensity-Dementia Pain Scale (MOBID): Development and validation of a nurse-administered pain assessment tool for use in dementia.
        J Pain Symptom Manage. 2007; 34: 67-80
        • Chakrabarti A.
        • Ghosh J.K.
        AIC, BIC and recent advances in model selection. Philosophy of statistics. Volume 7 in Handbook of the Philosophy of Science.
        Elsevier, 2011
        • Nagin D.S.
        Group-Based Modeling of Development.
        Harvard University Press, 2005
        • Ram N.
        • Grimm K.J.
        Growth mixture modeling: A method for identifying differences in longitudinal change among unobserved groups.
        Int J Behav Dev. 2009; 33: 565-576
        • Villeneuve R.
        • Meillon C.
        • Dartigues J.F.
        • et al.
        Trajectory of quality of life before and after entering a nursing home: A longitudinal study.
        J Geriatr Psychiatry Neurol. 2022; 35: 102-109
        • Hajek A.
        • Brettschneider C.
        • Lange C.
        • et al.
        Longitudinal predictors of institutionalization in old age.
        PLoS One. 2015; 10: e0144203
        • Giebel C.
        • Sutcliffe C.
        • Verbeek H.
        • et al.
        Depressive symptomatology and associated factors in dementia in Europe: Home care versus long-term care.
        Int Psychogeriatr. 2016; 28: 621-630
        • Barca M.L.
        • Selbaek G.
        • Laks J.
        • et al.
        The pattern of depressive symptoms and factor analysis of the Cornell Scale among patients in Norwegian nursing homes.
        Int J Geriatr Psychiatry. 2008; 23: 1058-1065
        • Kurlowicz L.H.
        • Evans L.K.
        • Strumpf N.E.
        • et al.
        A psychometric evaluation of the Cornell Scale for Depression in Dementia in a frail, nursing home population.
        Am J Geriatr Psychiatry. 2002; 10: 600-608
        • McCusker J.
        • Cole M.G.
        • Voyer P.
        • et al.
        Observer-rated depression in long-term care: Frequency and risk factors.
        Arch Gerontol Geriatr. 2014; 58: 332-338
        • Boorsma M.
        • Joling K.
        • Dussel M.
        • et al.
        The incidence of depression and its risk factors in Dutch nursing homes and residential care homes.
        Am J Geriatr Psychiatry. 2012; 20: 932-942
        • Watt J.A.
        • Veroniki A.A.
        • Tricco A.C.
        • et al.
        Using a distribution-based approach and systematic review methods to derive minimum clinically important differences.
        BMC Med Res Methodol. 2021; 21: 41
        • Parmelee P.A.
        • Katz I.R.
        • Lawton M.P.
        Incidence of depression in long-term care settings.
        J Gerontol. 1992; 47: M189-M196
        • Bergh S.
        • Selbaek G.
        • Engedal K.
        Discontinuation of antidepressants in people with dementia and neuropsychiatric symptoms (DESEP study): Double blind, randomised, parallel group, placebo controlled trial.
        BMJ. 2012; 344: e1566
        • McCusker J.
        • Cole M.G.
        • Voyer P.
        • et al.
        Six-month trajectories of self-reported depressive symptoms in long-term care.
        Int Psychogeriatr. 2016; 28: 71-81
        • Meeks S.
        • Van Haitsma K.
        • Schoenbachler B.
        • et al.
        BE-ACTIV for depression in nursing homes: Primary outcomes of a randomized clinical trial.
        J Gerontol B Psychol Sci Soc Sci. 2015; 70: 13-23
        • Dudas R.
        • Malouf R.
        • McCleery J.
        • et al.
        Antidepressants for treating depression in dementia.
        Cochrane Database Syst Rev. 2018; 8: Cd003944
        • Davison T.E.
        • McCabe M.P.
        • Busija L.
        • et al.
        A cluster randomised trial of the program to enhance adjustment to residential living (PEARL): A novel psychological intervention to reduce depression in newly admitted aged care residents.
        BMC Geriatr. 2020; 20: 98
        • Chau R.
        • Kissane D.W.
        • Davison T.E.
        Risk indicators for depression in aged care: The contribution of a meaningful life, mastery and environmental fit.
        Australas J Ageing. 2020; 39: e368-e374