Evidence

Table 1. Detailed Description of Research Studies on Evaluation of Depression



# Article Objectives Study Design Main Findings Discussion

10 Goodwin PE, Smyer To determine the  Design: Retrospective.  Frequency of depression: Of the depressed residents, only 59%

MA frequency of  Database: 1990 National Medical  Mention of depression in medical were identified by nursing home

recognized and Expenditure Survey. charts = 11.7% (342 out of 2923). personnel. This indicates if a nursing

“Accuracy of unrecognized co-  Sources of Depression variables:  Frequency of comorbid depression: home resident is depressed, his/her

recognition and morbid depression  Aggregate variable of diagnosed  Constructed DSM-III-R variable = chance of being diagnosed is only a little

diagnosis of co- and their major depression (DMS-III-R). 10.9% (319 out of 2923). better than 50:50.

morbid depression in relationship in the  Behavioral symptom variables  Residents with a medical record

the nursing home” nursing home. from the Baseline Questionnaire of diagnosis of depression and met the Although both models have similar

the National Medical Expenditure constructed DSM-III-R criterion had success rates for depression recognition,

Aging and Mental To estimate how Survey (NMES). the highest number of depressive the constructed criteria model is more

Health well individual  Mention of the words “depressed” symptoms. parsimonious than the individual

behavior symptoms or “depression” in relation to a  Rate of agreement between the symptoms model (9 variables vs. 17

1999;3(4):340-50. are distinguished resident‟s behaviors in medical medical record diagnosis (theoretically variables determined from discriminant

between residents record. supplied by the physician) and the analysis).

with a recognized  Study sample: constructed DSM-III-R variable =

vs. unrecognized  N=2923. 2.8%. Limitations:

comorbid  Mean age = 81.7 ± 8.2 years  Accuracy of depression recognition:

depression.  Individual symptoms model:  The data are secondary; therefore, the

 98.1% of non-depressed constructed DSM-III-R variable is at best

**Comorbid residents correctly classified. a proxy for Major Depressive Disorder.

depression refers to  56.5% of those with constructed  The above variables were based on a

depression in the DSM-III-R criteria but no medical symptom checklist completed by a

presence of one or record depression. respondent untrained to diagnose

more physical or  4.23% of those with medical mental disorders.

psychiatric record depression but no  The length of time the symptom has

disorders. constructed DSM-III-R criteria been present and symptom intensity

were correctly classified. cannot be assessed.

 Total success rate = 85%.  There is concern about the uniformity of

 Constructed DSM-III-R criteria responses regarding the symptom

model: checklist because it was filled out by

 Similar rates as above model. different respondents.

 The total success rate = 83%.

Table 1. Detailed Description of Research Studies on Evaluation of Depression



# Article Objectives Study Design Main Findings Discussion

11 Burrows AB et al. To describe the  Design: Cross-sectional.  Cornell Scale rating scores: Nurses in a long-term care facility

clinical features of  Setting: A large academic, multi-  Mean scores rated by nurses: frequently observed depressive symptoms

“Depression in a long- untreated nursing level, long-term care facility.  Group I = 9.5. in residents who were not receiving

term care facility: home residents  Measurements:  Group II = 9.9. antidepressants.

Clinical features and whom nurses  DSM-III-R mood diagnoses.  Group III = 12.5.

discordance between identify as  Cornell Scale for Depression in  No significance among the three Nurses attributed an equivalent burden of

nursing assessment depressed. Dementia (two separate ratings by groups (F=1.20, df=2; p=0.31). depressive symptoms to residents who did

and patient interviews” nurses and residents).  Mean scores rated by residents: not meet DSM-III-R criteria for major

To compare nurse  Geriatric Depression Scale (GDS):  Group I = 4.1. depression as they did to residents who

Journal of the ratings of 30 items.  Group II = 6.9. did meet such criteria.

American Geriatrics depressed nursing  Study sample:  Group III = 15.9.

Society home residents with  N=37.  Significant differences (F=43.5, In residents whose interviews suggested

ratings from direct  Mean age = 88.4 (74 to 99) years. df=2; p10. especially for residents with non-major Limitations:

 Three diagnostic groups: depression (r=-0.20, p=0.40).

 Group I - Those meeting DSM-  Correlation between nurse-derived  The sample included only those

III-R criteria for major Cornell Scale scores and GDS scores untreated residents identified as

depressive disorder (Major was also low (r=0.17, p=0.32). depressed by routine nursing

Depression).  Correlation between resident-derived assessment.

 Group II - Those with other Cornell Scale scores and GDS scores  This study was restricted to residents

DSM-III-R mood disorders (non- was higher (r=0.48, p=0.003). without advanced dementia or

Major Depression).  For those thought to have major behavioral disturbances.

 Group III - Those without any depression, both nurse-derived  Potentially confounding influence of

apparent mood diagnosis (No (r=0.62, P=0.10) and resident-derived caregiver mood was not assessed.

Diagnosis). (r=0.71, P=0.05) Cornell scores

correlated well with GDS scores.

Table 1. Detailed Description of Research Studies on Evaluation of Depression



# Article Objectives Study Design Main Findings Discussion

12 Rovner BW To determine  Design: A longitudinal observational  Prevalence of depression: The findings from this study and other

prevalence of major study (1 year).  Depressive disorder = 12.6% studies indicate that there is a relatively

“Depression and depressive disorder  Setting: Baltimore nursing homes. (n=57). high rate of depression in the nursing

increased risk of and depressive  Assessment methods:  Depressive symptoms = 18.1% home population.

mortality in the symptoms and their  Modified Present State (n=82).

nursing home patient” relationship to Examination.  No depression = 69.4% (n=315). Residents with depressive disorder and

mortality in a  Rates of abnormalities in areas  Among those with depressive disorder depressive symptoms compared to those

American Journal of nursing home such as mood, thinking, and (n = 57): without depression:

Medicine population. perception.  Nurses identified 65% as depressed  Were less often demented.

 Mini-Mental State Examination. compared to the 69% identified by  Had higher levels of cognitive function.

1993;94(Suppl.  Diagnostic and Statistical Manual family members.  Tended to be less disabled (activities of

5A):19S-22S. of Mental Disorders (DSM-III-R).  Nursing home physicians identified daily living).

 Psychogeriatric Dependency only 14% of the group as having

Rating Scale. depressive disorder. Mortality was highest among those with

 Nursing staff and family  Only 21% of residents diagnosed with depressive disorder, followed by those

interviews about resident depressive disorder were receiving with no depression and those with

orientation, behavior, and antidepressant medications. depressive symptoms.

activities of daily living.  Mortality rates:

 Study sample:  31.1% (n=141) died within one year. Limitations:

 N=454.  Almost half (47.4%) of these

occurred in those with depressive  Relatively high attrition rate: 31.1% died

disorder compared with 24.4% within one year and 3.1% (14) were

among those with depressive discharged but were not available for

symptoms, and 29.8% among those follow up.

with no depression.  Small samples of residents receiving

 The likelihood of death increased by: and not receiving antidepressants, so

 About 52% for those with disability could not evaluate the effects of

in activities of daily living. treatment on mortality.

 88% for men.  The results may be dated since new

 59% for those with depressive therapies are now available.

disorder.

 44% those who were hospitalized.

Table 1. Detailed Description of Research Studies on Evaluation of Depression



# Article Objectives Study Design Main Findings Discussion

13 Teresi J et al. To estimate the  Design: Descriptive study.  Prevalence of probable or definite Compared to the recognition rate of

prevalence of  Setting: Downstate New York major depression = 14.4 ± 2.1% and depression by psychiatrists, nurse aides

“Prevalence of depression among nursing homes. of minor depression = 16.8 ± 2.3%. out-performed the other groups in

depression and nursing home  Assessment scales:  Prevalence of depressive disorder recognizing residents who were

depression residents and the  Psychiatric reported by: depressed. However, the extent of

recognition in nursing extent of  Cornell Scale for Depression in  Nurse aides = 32%. depression recognition by nurse aides was

homes” depression Dementia (CSDD).  Nurses = 29%. also the least in agreement with

recognition among  Feeling Tone Questionnaire  Social workers = 20%. psychiatrists.

Social Psychiatry & nursing and social (FTQ).  Psychiatrists = 44% (reference

Psychiatric work staff.  Hamilton Depression Rating group). Lack of overall agreement with psychiatric

Epidemiology Scale (Ham-D).  Significant differences (Cochran‟s ratings occurred among about one-third of

 Structured Clinical Interview for test=32.48, p 5 = depression.  There was no statistically significant

nursing and  Health of the Nation Outcome difference between GDS-15 scores for There were essentially no differences in

Journal of Clinical residential Scale for older people (HoNOS residents in nursing or residential the ability of qualified nursing staff

Nursing homes; 65+): indicators of depression homes. compared to other care staff in their ability

 Whether qualified (e.g., gloomy, minor changes in  Recognition of depression: to recognize depression (when qualified

2000;9(3):445-50. nursing staff were mood, mild but definite  First method: HoNOS ratings of 1 or nurses were presumed to have more

more likely to depression, moderate depression, higher (i.e., gloomy or worse) as psychological and psychiatric training).

recognize and severe depression). recognition of depression.

depression than  Study sample:  Overall staff: Identified 27.1% Limitations:

care staff with no  N=308 residents. (n=29) of the depressed, as

formal training;  Mean age = 82.9 ± 7.6 (65 to 101) indicated by the GDS-15.  Information about previous training and

and years.  Qualified nurses: Identified over qualifications of the participants was not

 The level of  N=332 care and nursing staff a third (36.4%) of GDS-15 cases. extracted.

training in (qualified nurses, nursing  Other care staff: Identified 20.6%  The differences between the two groups

psychological assistants, residential home of depressed residents. of nurses compared in the analysis may

care received by managers/supervisors and care  Second method: The item gloomy be reduced because it was possible that

staff within assistants). was not included (i.e., scores 2 or a small number of the senior care staff

residential and higher) as a depression rating. informants were qualified nurses who

nursing homes.  Overall staff: The recognition had chosen to work in residential

rate fell to 15% (n=16). homes.

 Qualified nurses: 20.5% of cases  Staff training data were not collected at

were recognized. the time of the informant interview,

 Other care staff: 11.1% of cases making it impossible to relate

were recognized. depression recognition by an individual

 Of the 332 staff respondents, only 26 staff member to that person‟s training

(7.8%) had received any training in history.

psychological or psychiatric care while

employed.

Table 1. Detailed Description of Research Studies on Evaluation of Depression



# Article Objectives Study Design Main Findings Discussion

16 Horgas AL et al. To assess the  Design: A pilot study.  The MDS data identified: The prevalence of depressive and

prevalence of  Setting: 1 private, for-profit nursing  93% of the sample (n=126) as disruptive symptoms assessed by the

“Measuring behavioral behavior problems home in the Midwest. having cognition-related behavior MDS was comparatively lower than and

and mood disruptions as rated by the  Measurement scales: (MDS) problems. differs markedly from nursing assistants‟

in nursing home MDS and direct  Cognition Performance Scale  28.0% (n=39) as depressive. ratings (RMBPC) of the same behaviors.

residents using the caregivers (e.g., (CPS): 5 MDS items (0 to 7)  36.3% (n=49) as disruptive.

Minimum Data Set nursing assistants). (α=0.78).  The RMBPC indicated: (ratings by Nursing assistants‟ ratings were more

(MDS)”  Depressive symptoms: 10 MDS nursing assistants) congruent with physician diagnoses than

To examine the items (E1 and E2, 6 items  35.6% (n=48) of the sample had with MDS ratings and also more

Outcomes relationships excluded, 0 to 20) (α=0.83). cognition-related problems. consistent with empirical research.

Management for between the MDS  Disruptive behaviors: 5 MDS items  63.7% (n=86) as depressive.

Nursing Practice behavioral ratings, (Section E4; 0 to 15) (α=0.72).  71.9% (n=97) as disruptive. MDS ratings were only modestly

objective ratings  Measurement scales: (Revised  Diagnosed by physician: correlated with nursing assistants‟ ratings

2001;5(1):28-35. provided by nursing Memory and Behavior Problem  60% of residents (n=79) with of the same behavioral domains, but only

assistants, and Checklist [RMBPC], rated by nursing Alzheimer‟s disease or dementia. in non-demented (estimates not provided

physician- assistants)  21% (n=56) with depression. in the article) or non-depressed residents.

diagnosed  Frequency of 24 behaviors and  No significant group differences were

dementia and caregivers‟ reactions to them (5- found for the MDS subscales between Limitations:

depression. point scale, 0 to 96) (α=0.84). residents with and without a diagnosis

 Memory-related behavior: 7 of depression.  The study was conducted with residents

behaviors (0 to 28) (α=0.79).  The RMBPC depression score was from only one nursing home so limited

 Depressed behavior: 9 items (0 to also significantly higher for depressed generalizability.

36) (α=0.80). residents than those who were not (t =  The inter-rater reliabilities of the MDS,

 Disruptive behavior: 8 items (0 to - 3.2, df = 33.2, p=0.003). RMBPC, and the medical diagnoses

32) (α=0.67).  For those without depression (n=106): used in this study were not assessed,

 Study sample:  The MDS disruptive subscale which lead to overestimation or under-

 N=135. significantly correlated with RMBPC estimation of the prevalence rates

 Mean age = 84 ± 9.33 (53 to 99) memory subscale (r=0.30, p 5  Based on the GDS short form, there

quality indicator” quality indicator (probable depression). was no significant difference in the Both the GDS short form and the MDS

report.  Most recent Minimum Data Set GDS scores of residents who scored should be considered only as screening

Gerontologist assessment (Section E1 items A higher than 5 between the two sites. tools for depressive symptoms that

through P).  There was a significant difference warrant further follow-up evaluation (as in

2001;41(3):401-05.  Study sample: between the sites based on MDS the MDS-Resident Assessment

 N=91 (n=33 Site 1 and n=58 Site mood items only: Protocols).

2).  Site 1: 34% of 38.

 Mean age = 89.1 ± 6.9 years.  Site 2: 72% of 71. Limitations:

 Mean MMSE score =13.4 ± 7.7 (0  p 5:  Small sample size.

 Site 1: 25%.  This study was part of an ongoing

 Site 2: 78%. clinical trial not directly designed to

 2 (48) = 12.59, p 1.0 = those scoring 1 SD above  The inclusion and exclusion criteria for

the mean are ___ times more selecting participants were subjective.

likely to have the hazard of dying  Lack of measures of cognitive status.

than those scoring 1 SD below

the mean.

 85 years with multiple

residents”  Measurements: antidepressant therapy: diagnoses, particularly cancer, were less

To identify resident  Receipt of any antidepressant.  Received SSRIs = 46%. likely to receive pharmacologic treatment

Journal of the characteristics that  Receipt of a selective serotonin  Received tricyclic antidepressants for depression.

American Geriatrics predict receipt of reuptake inhibitor (SSRI) among (TCAs) = 36%.

Society any treatment and those treated.  Receiving less than manufacturer- Residents with severe impairment in

choice of  Study sample: recommended minimum effective cognitive functioning and dementia were

2002;50(1):69-76. antidepressant  N=428,055. dose for depression: less likely to receive an SSRI. In contrast,

used.  SSRIs = 1%. residents with moderate and severe

 TCAs = 56%. impairment in physical functioning were

 All others combined = 68%. more likely to receive SSRI therapy.

 Predictors of antidepressant therapy:

 Age 85 and over: Limitations:

 Less likely to receive

antidepressant therapy.  The analysis was limited by the lack of

 OR=0.93 (95% CI=0.88-0.98). information on non-pharmacological

 Most severely cognitively impaired: methods of treating depression.

 Least likely to receive  It was highly probable that most of the

antidepressants. sample was not evaluated by

 OR=0.69 (95% CI=0.64-0.75). psychiatrists because data were

 More severe functional impairment: collected from the MDS.

 Less likely to receive an SSRI.  There was no actual documented

 OR=0.81 (95% CI=0.72-0.91). indication for the medication (this

 Decreased likelihood of receiving information was obtained from the

antidepressant therapy MDS).

 Cancer: OR=0.90 (95% CI=0.84-

0.96).

 Six or more clinical conditions

OR=0.89 (95% 0.83-0.95).

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

17 Badger TA et al. To evaluate the  Design: Repeated-measure (pre-  Total scores on the posttests were Although there were no statistically

effects of and posttest and follow-up at six significantly higher than pretest significant differences found for the

“Knowledge of depression months). scores (administrations of the 20 item number of patients assessed as

depression and education on  Setting: 12 programs in 3 DAQ) [F(2, 721)=125.09, p=0.001]. depressed and the number referred or

application to practice: depression southwestern states.  Results from 241 chart audits showed treated at the 6-month follow-up survey,

a program evaluation” knowledge and its  Participants: Registered nurses from that there were significant differences greater numbers of patients were reported

clinical application. general medical-surgical units, between the pre-program and post- as depressed and were started on

Issues in Mental nursing homes, or community health program records [F(2,238)=18.69, medication than in the initial survey across

Health Nursing settings. p=0.001]. all 12 training groups.

 Depression awareness quiz (DAQ):  Depression knowledge and clinical

1996;17(2):93-109.  20 multiple-choice items (0 to 20). application: The increased knowledge and clinical

 Kuder-Richardson=0.92.  Initially, one-third (108) reported application were maintained 6 months

 Chart audit instrument: less than 10% of their patients as after the program‟s conclusion.

 Intra-rater=0.90. depressed. Assessment skills remained stable at 6

 Inter-rater reliability=0.92.  19% (70) reported between 11% months.

 Score=0 to 25. and 25% of patients as depressed.

 Intervention:  17% (61) reported between 26% Limitations:

 Three-day workshop on and 50% of their patients as

depression: depressed.  There might have been sample bias

 Biological theories.  10% (36) reported over 76% of because participants who attended the

 Psychological theories. their patients were depressed. program already recognized the high

 Cultural aspects.  No estimates for the frequency of incidence of depression in their patient

 Developmental considerations. assessments and referrals by study population, yet many reported a lack of

 Depression assessment and participants six months after the skill or feeling unsure of their

screening. intervention were reported by the assessment skills.

 Suicide assessment. authors.  Many participants reported failing to

 Treatment.  52% of the participants (192) did not document assessments, so the results

 Study sample: refer their depressed patients. from chart audit must be used with

 N=363.  Only 10% routinely referred their caution.

 Age range = 23 to 74 years. depressed patients.

 Majority had either an associate or  65% of the participants recommended

baccalaureate degree in nursing. counseling to less than 10% of their

 76 participants reported degrees depressed patients. A combination

above the baccalaureate level in approach was rarely recommended.

nursing or in other fields.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

28 Svarstad BL, Mount JK To identify risk  Design: Retrospective.  About half (55%) of the residents Demand for benzodiazepine was

factors associated  Setting: 18 randomly selected occasionally or usually demonstrated significantly higher among alert residents

“Effects of residents‟ with the nursing homes in Wisconsin behavioral problems, 34% had pain, and those with diagnosed depression and

depression, sleep, and prescription and between 1986 and 1989. and 31% had sleep problems. pain.

demand for medication use of  Measurements:  Benzodiazepine prescription rates

on benzodiazepine use benzodiazepines in  A current prescription for a were significantly higher for those: Depression was found to be a risk factor

in nursing homes” nursing homes. benzodiazepine during the index  Under 85 years of age. for chronic benzodiazepine use. A total of

month.  Who did not have dementia. 44% of residents with diagnosed

Psychiatric Services To assess the  Any as needed orders for  Who had depression, pain, or sleep depression and prescriptions for a

prevalence and benzodiazepines during the same problems. benzodiazepine did not have prescriptions

2002;53(9):1159-65. determinants of period.  Factors that predicted for an antidepressant.

demand for  Number of as needed doses used benzodiazepine prescription:

benzodiazepines during the 14-day observation  Female gender (OR = 1.30, 95%CI These results suggest that inadequate

among nursing period. = 1.01 – 1.67). treatment of primary problems that can

home residents.  Chronic use = ≥10 continuous  Depression (OR = 1.68, 95%CI = induce insomnia support other findings

days for hypnotics and ≥4 months 1.25 – 2.27). that alert residents with diagnosed

**Only sections on for anxiolytics.  Pain (OR = 1.20, 95%CI = 1.04 – depression often receive benzodiazepines

depression will be  Study sample: 1.40). as the sole treatment for depression.

discussed here.  N=2060.  Sleep problems (OR = 2.23, 95%CI

 Mean age = 82.3 ± 10.8 years. = 1.93 – 2.57). Limitations:

 Adjusted odds of chronic

benzodiazepine use were twice as  Data were collected between 1986 and

high among residents with: 1989.

 Depression (OR = 2.04, 95%CI =  Had to rely on recorded diagnoses and

1.30 – 3.20) and nurses‟ assessments (secondary data).

 Multiple prescriptions (OR = 2.08, Further studies are needed to validate

95%CI = 3.32). these measures.

 Adjusted odds for any demand for  Inability to survey physicians or examine

benzodiazepine were also twice as previous treatment efforts.

high among those with:  Although study sample is similar to other

 Depression (OR = 2.42, 95%CI = nursing homes, results may not be

1.41 – 4.16). generalizable to other states.

 Multiple prescriptions (OR = 2.36,  Relationships among depression, pain,

95%CI = 1.36 – 4.11). and sleep, demand, and benzodiazepine use

 Pain problem (OR = 2.51, 95%CI = are complex and subject to alternative

1.93 – 3.25). explanations.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

37 McCurren C et al. To evaluate the  Design: Experimental with  Treatment group showed greater The comprehensive assessments

effectiveness of an randomization (randomized block by improvement in depressive performed by the geropsychiatric nurse

“Depression among intervention initial GDS score). symptomatology than control group: indicate that much of the detected

nursing home elders: strategy that uses a  Setting: Three nursing home  Measured by changes in mean depression was related to losses, which

Testing an intervention master‟s prepared facilities. GDS scores: Time 2 support the findings from this study that

strategy” geropsychiatric  Measurements:  Control = - 0.176. individualized strategies (e.g., provision of

nurse and trained  Geriatric Depression Scale (GDS):  Treatment = - 5.027. emotional and social support, identification

Applied Nursing older adult 30 yes/no items.  Compared: t = 4.94, p

 Aversive Feeding Behavior Scale were there any significant differences 0.45).

(AFBS): 25 presence/absence between the treatment and placebo

items (0 to 25). groups in the proportion showing at Limitations:

 Facial behaviors. least a 50% improvement.

 Intervention:  Limited generalizability because of the

 Week 1-2: 25mg sertraline per more controlled environment with clinical

day. trials.

 Week 3-4: 50mg per day.  The results might have been influenced

 Week 5-8: 100mg per day. by the expectation of nursing home staff.

 Study sample:  The small treatment/placebo difference

 N=31 residents with dementia. might have been due to the inclusion of

 n=17 sertraline group. a relatively large number of residents

 n=14 placebo group. with minor depression.

 The study design was potentially

weighted against finding a significant

treatment effect.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

40 Oslin DW et al. To compare nursing  Design: Comparison (open label,  Treatment response rate for The results indicate that sertraline is less

home residents cross over) study. cognitively intact intent-to-treat effective than nortriptyline for the treatment

“Heuristic comparison treated with open  Setting: Eight community nursing residents: of depression in elderly nursing home

of sertraline with label sertraline in homes.  Regular-dose nortriptyline group = residents who were depressed. More

nortriptyline for the doses up to 100mg  Measurements: 51.5% (17 of 33). specifically regular-dose nortriptyline

treatment of per day with those  Mini-Mental State Exam.  Low-dose group = 18.8% (3 of 16). seemed the most appropriate for

depression in frail treated in a double-  Cumulative Illness Rating Scale  Sertraline group = 18.2% (4 of 22). depressed residents with intact cognition

elderly patients” blind randomized (CIRS).   = 8.646; p=0.013.

2

while low-dose nortriptyline was the most

study of low vs.  Physical Self-Maintenance Scale  Treatment response rate for appropriate for those who are cognitively

American Journal of regular doses of (PSMS): functional status. cognitively impaired intent-to-treat impaired.

Geriatric Psychiatry nortriptyline (an  Hamilton Rating Scale for residents (same order as above):

exploratory study to Depression (Ham-D): 21 items. 7.1% (1 of 14), 66.7% (4 of 6), and Limitations:

2000;8(2):141-49. determine the effect  Those with decreases in Ham-D 2

0% (0 of 6) ( = 11.434; p=0.0033).

size of treatment for scores of at least 33% were  Treatment response rate for those  Potential bias associated with allowing

depression). considered „responders.‟ cognitively intact who completed the some residents to receive sertraline

 Measured endpoints: study: under open-label conditions and those

 Week 10 assessments = residents  Regular-dose = 68.4% (13 of 19). not eligible for nortriptyline were entered

who completed the study protocol.  Low-dose = 18.2% (2 of 11). into the sertraline group

 Assessments conducted at the  Sertraline group = 33.3% (4 of 12).  A 33% reduction in Ham-D scores was

time of termination = worsened   = 8.059; p=0.018.

2 used rather than the typically 50%

depression.  Treatment response for cognitively reduction used in other studies. This

 Final assessment = ratings were impaired completers, (same order as may lead to underestimating treatment

not confounded by adverse drug above): 16.7% (1 of 6), 80% (4 of 5), response.

events (i.e., when ADEs led to 2

and 0% (0 of 3) ( = 6.886; p=0.032).

early termination).

 Study sample:

 N=92.

 The average dose of nortriptyline

at the endpoint:

 Regular-dose group = 49.0 ±

23.5mg.

 Low-dose group = 9.25 ±

3.80mg.

 Sertraline = 83.1 ± 25.2mg.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

41 Rosen J et al. To assess the  Design: Open clinical trial.  There were no reported adverse In this preliminary study, sertraline

safety and  Setting: Riverview Center for Jewish events that led to medication appeared to be safe in elderly nursing

“Sertraline in the tolerability of an Seniors in Pittsburg, Pennsylvania. discontinuation in any of the patients. home residents.

treatment of minor antidepressant trial  Measurements: Sertraline doses were increased in

depression in nursing in this frail  Hamilton Rating Scale. three residents at week 4 (none of Limitations are associated with those for a

home residents: A pilot population.  Global Assessment Scale (GAS). these experienced side effects). pilot study:

study”  Udvalg for Kliniske Undersogelsen  The mean Hamilton score improved

(UKU) Side Effect Rating Scale: significantly (mean score decrease =  Small sample size.

International Journal of 48 items assessing potential 5.9 ± 4.4, p 10 or a decline in the Hamilton

determined whether increase to score by 50%).

100mg/day for the remaining 2

weeks.

 Study sample:

 N=12.

 Mean age = 83.2 ± 7.8 (72 to 92)

years.

 Nursing home residents with

DSM-IV minor depressive

disorder.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

42 Trappler B, Cohen CI To examine the  Design: Open-label and prospective  Overall, there was a significant Overall, the participants showed significant

effects of three drug trial. decline (36%) in Ham-D scores, and response to all three SSRIs, as indicated

“Use of SSRIs in very selective  Setting: A non-profit, 271-bed, long- 42% showed a ≥ 50 decrease in by the overall decline in Ham-D scores.

old depressed nursing serotonergic term care facility in Brooklyn. Ham-D scores. However, there were no significant

home residents” reuptake inhibitors  Measurements:  Persons with non-CNS-associated differences in the response rates to or

(SSRIs) –  Hamilton Depression Scale (Ham- major depressive disorder (MDD) rates of serious adverse effects of the

American Journal of fluoxetine, D): 21 items (administered initially showed significantly more SSRIs among the residents who took the

Geriatric Society sertraline, or and at 12 weeks). improvement than those with CNS- respective anti-depressive medications.

paroxetine – on  A successful drug response is at associated MDD (e.g., vascular

1998;6(1):83-89. depressed nursing least a 50% improvement on the dementia and Alzheimer‟s). There were significantly fewer residents

home patients. Ham-D.  93% vs. 20% improvement. with CNS-associated depression

  = 20.27, p0.05). of depression.

Journal of Clinical combination, on  Geriatric Depression Scale: 30  Hypothesis 2: Verbalization will have

Geropsychology depression, self- yes/no items. no effect on levels of depression, self- Limitation:

esteem, and life  Reid‟s Desired Locus of Control esteem, life satisfaction, and locus of

1999;5(4):291-300. satisfaction in older Scale. control. The study sample was a small

nursing home  Interventions: (groups of 8)  Verbalization had a significant convenience sample.

residents.  Group 1: only the presence of the effect on life satisfaction (no t value

researcher. reported, p0.05).

intermittently only.  Hypothesis 3: Expressive touch and

touched by the  Group 3: the researcher-resident verbalization combined will have no

researcher (usually interaction = verbal only. effect on levels of depression, self-

on the hand, arm,  Group 4: the researcher-resident esteem, life satisfaction, and locus of

shoulder, or back) interaction = a combination of control.

during a 15-minute expressive touch and talking.  The combination therapy had a

period.  Study sample: significant effect on depression

 N=24. (p0.05).

 Hypothesis 4: Expressive therapy and

verbal communication either alone or

in combination will have no effect on

residents‟ level of comfort as

determined by insight and intuition.

 Not supported by empirical

evidence but by insight and

intuition.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

46 Crowley-Robinson P et To assess whether  Design: Longitudinal study with  Nursing home with a visiting dog: Depression scores decreased significantly

al. tension, experimental design (randomization  Overall, there were no changes in for residents in nursing homes with a

depression, anger, to 2 treatment groups and 1 control the proportion of low, medium, and resident dog and no dog at any time.

“A long-term study of vigor, fatigue or group). high depression scores and

elderly people in confusion scores of  Setting: Three nursing homes. depression scores over the six There was a similar trend evident in the

nursing homes with elderly nursing  Returned Services League War assessment periods. nursing home with a visiting dog. This

visiting and resident home residents as Veterans Home = resident dog.  Nursing home with a resident dog: suggests that a visiting person can have a

dogs” measured by the  Moreton Bay Nursing Care Unit = There were significant changes in the similar positive effect as a resident dog on

Profile of Mood visiting dog. proportion of low, medium, and high these mood states.

Applied Animal States (POMS) are  Wheeler Garden Settlement = no depression scores over six

Behaviour Science influenced by a dog but visited by researcher. assessment periods: The results suggest that there are many

visiting dog, a  Measurements:  High depression scores: a benefits from having a resident dog in a

1996;47(1-2):137-48. resident dog, or no  POMS questionnaire: significant reduction in frequency. nursing home (i.e., reducing anger, vigor,

dog.  65 words related to feelings  Low depression scores: a confusion, and tension, in addition to

experienced within last week. significant increase in frequency. depression).

**This summary  Each word rated as 0=not at all  The trend continued after the

focuses on to 4=extremely. removal of the dog. Limitations:

depression only.  Administered Period 1 (pretest)  Nursing home with no dog at any time

= 4 months before introduction (control):  No discussion of possible challenges

of the dog.  There were significant differences with having a dog in residence.

 Subsequent administrations in the proportion of low, medium  No discussion of limitations of the study.

every 3 months (Periods 2 to 5). and high depression scores and  The three sites might not have been

 Administered at Period 6 significant differences in the comparable (i.e., proportion of males

(posttest) = 3 months after depression scores over the six and females and other facility

removing dog. assessment periods. characteristics).

 Interventions:  Study conducted in Australia so may not

 A resident dog vs. a visiting dog be applicable to the US.

with researchers weekly.

 No dog at any time but visits from

the researchers at each of the six

assessment periods.

 Study sample:

 N=95.

 Mean age = 82.3 ± 8.2 (56 to 105)

years.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

47 Fitzsimmons S To determine if  Design: Experimental design with  31% of all participants were without a There was a significant decrease in

participation in a randomization, control and treatment chart diagnosis of depression. depression as measured by posttest GDS

“Easy rider wheelchair therapy biking groups, and pre- and post-testing.  51% of participants were not in the treatment group compared to the

biking: A nursing- program had an  Setting: New York State Home for receiving any treatment for control group.

recreation therapy effect on the degree Veterans (long-term facility in depression.

clinical trial for the of depression in upstate New York).  Changes in mean GDS scores: The results of this study also showed a

treatment of older adults.  Conceptual framework:  Control group: increased from 7.95 positive effect of an interdisciplinary

depression”  Roy Adaptation Model. to 8.65 at posttest. psychosocial intervention on depression

 Objectives of the program:  Treatment group: decreased from levels.

Journal of  To increase positive coping 7.68 to 4.21 at posttest.

Gerontological Nursing mechanisms through social  There was a significant difference From the observations made during the

interactions. between the changes in mean GDS study, the biking program reportedly gave

2001; 27(5):14-23.  To provide enjoyable scores for the two groups ( in all participants an opportunity to have a

experiences to minimize, score = - 3.47; t = - 5.095; greater sense of their surroundings and

reduce, or eliminate depressive p<0.001). was “the talk of the facility.”

moods.  A subgroup analysis showed that the

 Depression scale: intervention was more successful for Limitations:

 Geriatric Depression Scale (GDS) those with dementia than those

– short form with a score ≥ 5 without.  The gender mix of the sample is not

indicative of depression.  The subgroup with the next highest similar to that in most nursing homes

 Intervention: drop in depression scores was the (i.e., more male residents – 67% in this

 Therapeutic bike program for 1 group of individuals without a sample).

hour a day for 5 days a week for 2 depression diagnosis and not  Many variables influence the levels of

weeks. currently on any anti-depressive depression in the elderly population, but

 Part 1: a discussion group medication. this study examined only one (i.e.,

about bike riding. ambulation status).

 Part 2: a 15-minute ride around  This sample includes older adults taking

the facility. various types of medications, which

 Study sample: were not controlled.

 Size: N=39.  Both residents with and without

 Mean age = 80.49 (67 to 99) dementia were included in the study,

years. and the GDS may not have been the

 All those with a diagnosis or most valid tool for the residents with

symptoms of depression. dementia.

 Long-term effects of the biking program

could not be measured.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

48 Sumaya IC et al. To investigate the  Design: Placebo-controlled,  The mean pretest GDS score was Although none of the participants were

effects of bright crossover design. highly correlated with months of clinically diagnosed with depression, all

“Bright light treatment light treatment on  Setting: Long-term care facility. institutionalization (r=0.81, p<0.01), pretest scores were indicative of moderate

decreases depression depression among  Measurements: showing greater depression with to severe depression.

in institutionalized institutionalized  Geriatric Depression Scale (GDS): longer institutionalization.

older adults: A older adults. measured at pre-intervention and  The results showed that depression Participants in the present study who were

placebo-controlled post-intervention for all but control scores remained unchanged during institutionalized longer tended to score

crossover study” groups. the control and placebo conditions, higher on the GDS at baseline, and the

 Interventions: but depression scores decreased most depressed residents tended to

Journal of  Participants were placed under significantly for the treatment group experience the greatest improvement from

Gerontology: varying light conditions. Each of (pretest GDS score=14.9 ± 0.97 vs. the treatment.

Psychological the light conditions consisted of posttest GDS score=11.3 ± 0.74,

Sciences light treatment for 30 minutes per p<0.01). Limitations:

day for 5 days a week.  GDS pretest scores were also

2001;56A(6):M356-60.  1 week of treatment at therapeutic significantly correlated with difference  The inclusion and exclusion criteria were

dose; 1 week of placebo; and 1 scores between pre- and posttest relatively stringent, so generalization of

week of no treatment (control). treatment GDS scores (r=0.62, the findings is limited to moderately to

 Study sample: p<0.05), where more improvement severely depressed (whether diagnosed

 N=11 wheelchair-bound older was related to higher depression or not) institutionalized older adults with

adults. before treatment. good cognitive and retinal functioning.

 Mean age = 83.8 ± 9.56 (67 to 94)  The study sample was small and only

years. one facility was sampled.

 Seasonal variations of mood and

behavior were not assessed.

Table 2. Detailed Description of Research Studies on Management of Depression



# Article Objectives Study Design Main Findings Discussion

50 Rosen J et al. To determine:  Design: Randomized comparison  None of the residents in the control  The intervention appeared to be

trial. group showed significant associated with improvements in sample

“Control-relevant  The effect of the  Setting: A 290-bed, long-term care improvement compared to 36% (4 of residents‟ depression; however, the

intervention in the psychosocial facility in Pittsburgh, Pennsylvania. 11) of those in the treatment group positive effects were not sustained in the

treatment of minor and intervention on  Measurements: (P=0.045; Fisher‟s exact test, one follow-up.

major depression in a depressed  Structured Clinical Diagnostic tailed).  The intervention did not appear to

long-term care facility.” residents. Interview (SCDI).  There was a significant decrease in significantly reduce the need for

 The durability of  DSM-III-R depression criteria: Ham-D scores for the treatment group psychotropic use in the treatment group.

American Journal of any positive major depression. (t= -2.23, P<0.50) compared to the  Future studies are needed to further

Geriatric Psychiatry effect after  DSM-IV: minor depression. control group (t= -1.25, P=0.24). investigate the relationship between

withdrawal of the  Hamilton Rating Scale for  However, the improvement in Ham-D cohesiveness and depression in elderly

1996;5(3):247-257. intervention Depression (Ham-D): 17 items. for the treatment group was not nursing home residents.

 The association  Geriatric Depression Scale (GDS): significantly different from the control

of clinical, 30 items. group when evaluated with a Limitations:

demographic, or  Sheltered Care Environmental repeated-measures ANOVA

diagnostic Scale (SCES). [F(1.20)=0.28, P=0.60].  50% of the residents who met the

features with  Intervention:  During the intervention: cognitive requirements refused to

potential positive  A 2-month intervention program of  Significant improvements in Ham-D participate.

response. planned activities that matched scores (t= -4.25, P<0.001)  The sample residents were concurrently

residents‟ leisure interests and  Significant improvements in GDS using psychotropic medications.

preferred socialization method scores (t= -2.46, P<0.02)  A review of the medical history of these

(treatment group).  Two months after the intervention residents showed that many

 A “wait list” control condition for 2 (follow-up): experienced severe depression several

months before intervention  Ham-D scores worsened (t= 2.84, months or years before the start of the

(control group). P<0.01). study and most likely partially responded

 Study sample:  GDS scores worsened (t= 1.73, to medication management.

 N=31. P=0.10).  Small sample size.

 Mean age (range) = 78.7±10.2 (50  “Cohesion” (community support)

to 96) years. measured by the SCES was the only

 Those who had both major and independent variable found to be

minor depression. significantly different between the

responders and non-responders.

 The responders showed a

significant improvement in the

cohesion score during the

intervention (t=2.22, P<0.046).