Psychological therapies for depression in older adults residing in long-term care settings

In the single study reporting medium-term depression remission, participants allocated to the non-therapy comparator had an MDD prevalence increase from 25.0% at baseline to 73.1% at medium-term follow-up, a 48.1% increase.

“Amongst participants allocated to the non-therapy comparator, the prevalence of MDD at medium-term follow-up increased from 25.0% to 73.1% (48.1% increase).”

• cites Meeks 2015 sourceDisputed

  The Sample Characteristics section reports that 69.1% of all participants met criteria for a major depressive episode at baseline and that major depressive disorder was proportionally more common in the treatment-as-usual group than in the therapy group. That contradicts the claimed 25.0% baseline prevalence for the non-therapy comparator. The Diagnostic Recovery section reports follow-up improvement and remission outcomes, not a rise in comparator MDD prevalence to 73.1%. (LLM, conf=0.88)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

In the single study reporting long-term depression remission, participants allocated to the non-therapy comparator had an MDD prevalence increase from 25.0% at baseline to 65.2% at long-term follow-up, a 40.2% increase.

“Amongst participants allocated to the non-therapy comparator, the prevalence of MDD increased from 25.0% to 65.2% (40.2% increase).”

• cites Meeks 2015 sourceDisputed

  The Sample Characteristics section reports that 69.1% of the entire sample met criteria for a major depressive episode at baseline, with proportionally more major depressive disorder in the treatment-as-usual control group than in the BE-ACTIV group. That contradicts a 25.0% baseline MDD prevalence in the non-therapy comparator. The Diagnostic Recovery section reports follow-up recovery outcomes, not the claimed increase to 65.2%. Page numbers are unavailable in our copy. (LLM, conf=0.84)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

Only four studies reported independent assessments of treatment fidelity.

“Only four studies reported that independent assessments of treatment fidelity were conducted (Bailey 2017;Meeks 2008;Meeks 2015;Zerhusen 1991), although two of these studies had additional potential bias (Meeks 2008;Meeks 2015).”

• cites Bailey 2017 sourceNot checked — no full text of the source
• cites Meeks 2008 sourceNot checked — no full text of the source
• cites Meeks 2015 sourceDisputed

  The Discussion (page not supplied) explicitly identifies the lack of an independent evaluator of treatment adherence as a limitation. The Treatment Integrity passage says the first author rated adherence from session recordings. Thus, although the paper reports fidelity assessments, it contradicts the claim that its assessments were independent. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

One study reported that lower rates of major depressive disorder in the psychological therapy group compared with the control group continued through to long-term follow-up.

“although one study reported that the lower rates of major depressive disorder in the psychological therapy group compared with the control group continued through to long-term follow-up (Meeks 2015).”

• cites Meeks 2015 sourceDisputed

  The Diagnostic Recovery section reports an advantage for participants remaining in the study at three months, but says the groups were not significantly different at six months. The Discussion explicitly states that Be-actiV no longer had an advantage in diagnostic recovery at six months, contrary to the claimed persistence through long-term follow-up. Page numbers were not supplied for these passages. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

For Meeks 2015, MDD prevalence appeared unbalanced between groups (47.6% in the psychological therapy group vs 25% in the non-therapy comparator group).

“the prevalence of MDD appeared to be unbalanced between groups (47.6% in psychological therapy group vs 25% in non-therapy comparator group)”

• cites Meeks 2015 sourceDisputed

  The Sample Characteristics section reports a baseline imbalance in major depressive disorder, but in the opposite direction: proportionally fewer participants in the BE-ACTIV therapy group had the diagnosis than in the treatment-as-usual group. The Discussion reiterates that direction. Neither passage gives the claimed 47.6% versus 25% figures. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

Older people may respond to psychological therapy as well as, or better than, working-aged adults.

“In general, older adults are more likely to be treated with pharmacological than non-pharmacological interventions, despite indications that older people may respond to psychological therapy as well as, or better than, working-aged adults (Saunders 2021).”

• cites Saunders 2021 sourceSupported

Depression is common among older people living in long-term care facilities.

“Depression is common amongst older people residing in long-term care (LTC) facilities.”

• cites Seitz 2010 sourceNot checked — no full text of the source
• cites Fornaro 2020 sourceSupported

Residing in long-term care is significantly associated with increased depressive symptoms even after controlling for established risk factors.

“However, there is a significant association between residing in LTC and increased depressive symptoms, even when controlling for these risk factors (Anstey 2007;Pot 2005).”

• cites Anstey 2007 sourceNot checked — no full text of the source
• cites Pot 2005 sourceSupported

Several longitudinal studies in long-term care settings have reported persistence of depressive symptoms.

“The persistence of depressive symptoms has been reported in several longitudinal studies in LTC settings (Barca 2010;Davison 2021a;Smalbrugge 2006;Sutcliffe 2007).”

• cites Barca 2010 sourceNot checked — no full text of the source
• cites Davison 2021a sourceNot checked — no full text of the source
• cites Smalbrugge 2006 sourceSupported

The authors of one recently published systematic review of non-pharmacological interventions were unable to draw clear conclusions about the effectiveness of the interventions.

“The authors of one recently published systematic review of non-pharmacological interventions were unable to draw clear conclusions about the effectiveness of the interventions (Gramaglia 2021).”

• cites Gramaglia 2021 sourceSupported
  Experimental — accuracy not yet measured

Some previous reviews excluded participants with dementia despite the high prevalence of dementia in long-term care populations.

“Some reviews excluded participants with dementia (Gramaglia 2021;Simning 2017), despite the high prevalence of dementia in LTC populations (Seitz 2010).”

• cites Gramaglia 2021 sourcePartly supported

  The Methods and Discussion > Limitations (page not provided) show that Gramaglia et al. excluded studies involving residents with severe cognitive impairment (MMSE <20). The review allowed no or mild cognitive impairment, so it does not establish that it excluded all participants with dementia. The prevalence statement is attributed to Seitz, not to this source. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites Simning 2017 sourcePartly supported

  The Introduction says this review focused on nursing-facility residents without moderate or severe cognitive impairment. But the Methods (Inclusion and Exclusion Criteria; page unavailable) required only that at least 50% of a trial’s sample have mild impairment or better, so the review did not categorically exclude participants with dementia. The high-prevalence clause is attributed to a different citation. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites Seitz 2010 sourceNot checked — no full text of the source

Limited information is available on the efficacy of antidepressants among older people with cognitive impairment and dementia.

“Limited information is available on the efficacy of antidepressants amongst older people with cognitive impairment and dementia (Boyce 2012;Nelson 2011;Orgeta 2017), which is problematic given the high prevalence of these conditions in the LTC population.”

• cites Boyce 2012 sourceSupported
• cites Nelson 2011 sourceNot checked — no full text of the source
• cites Orgeta 2017 sourcePartly supported

  The Discussion and Implications for practice and research sections describe few trials and limited evidence on antidepressant efficacy for depression in people with Alzheimer’s disease. That supports the claim for this dementia subgroup, but the review does not establish that evidence is similarly limited across all older people with cognitive impairment or dementia. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

One review focused only on nursing homes and excluded studies of assisted-living facilities.

“One review focused only on nursing homes and excluded studies of lower-level dependency facilities (assisted-living facilities; Simning 2017).”

• cites Simning 2017 sourceSupported

Several reviews have reported promising results of psychological therapies for older people in the general community, particularly using CBT, problem-solving therapy, and reminiscence therapies.

“Several reviews have reported promising results of psychological therapies for older people in the general community, particularly using CBT, problem-solving therapy, and reminiscence therapies (Cuijpers 2014;Jonsson 2016;Wilson 2008).”

• cites Cuijpers 2014 sourceSupported
• cites Jonsson 2016 sourcePartly supported

  The Discussion reports support for problem-solving therapy and promising results for CBT and reminiscence/life review in older adults. But the review includes participants in long-term care as well as community settings, and cautions that generalizability across settings is unclear. It therefore does not establish the claim specifically for older people in the general community. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Wilson 2008 sourceNot checked — no full text of the source

A 2012 review of studies evaluating antidepressants in nursing home residents included two randomized controlled trials that did not demonstrate a significant benefit for antidepressant pharmacotherapy over placebo.

“One 2012 review of studies evaluating antidepressants in residents of nursing homes included two randomized controlled trials, which did not demonstrate a significant benefit for antidepressant pharmacotherapy over placebo (Boyce 2012).”

• cites Boyce 2012 sourceSupported

In seven studies, authors limited detection bias by blinding personnel collecting outcome measures and were at low risk.

“In seven studies, authors limited the risk of detection bias by blinding personnel collecting outcome measures and were at low risk of detection bias (Abraham 1992;Dozeman 2011;Hyer 2009;Karimi 2010;Meeks 2008;Meeks 2015;Tsai 2008).”

• cites Abraham 1992Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Dozeman 2011 sourceNot checked — no full text of the source
• cites Hyer 2009 sourceSupported

Two studies reported depression remission data.

“Two studies reported data on depression remission (Meeks 2008;Meeks 2015).”

• cites Meeks 2008 sourceNot checked — no full text of the source
• cites Meeks 2015 sourceSupported

Four studies were at low risk of selection bias for random sequence generation and allocation concealment.

“Four studies were at low risk of selection bias (random sequence generation and allocation concealment; Dozeman 2011; Reinhardt 2014; Travers 2017; Tsai 2008)”

• cites Dozeman 2011 sourceNot checked — no full text of the source
• cites Reinhardt 2014 sourcePartly supported

  The Measurement section says participants were randomized to problem-solving treatment or social contact. It does not describe how the random sequence was generated or whether allocation was concealed, so our copy does not establish the claimed low risk of selection bias for Reinhardt 2014. It cannot substantiate the assessment of the other three studies. (LLM, conf=0.83)

  Experimental — accuracy not yet measured
• cites Travers 2017 sourceNot checked — no full text of the source

Five studies restricted eligibility to participants meeting defined diagnostic criteria for a mood disorder, typically MDD, minor depression, and dysthymia, plus adjustment disorder with depression.

“Five studies restricted eligibility to participants who met defined diagnostic criteria for a mood disorder, which typically included MDD, 'minor depression', and dysthymia (Meeks 2008;Meeks 2015;Reinhardt 2014;Rosen 1997), in addition to adjustment disorder with depression (Hyer 2009).”

• cites Meeks 2008 sourceNot checked — no full text of the source
• cites Meeks 2015 sourcePartly supported

  The Inclusion and Exclusion Criteria section says participants had to meet DSM-IV criteria for a depressive disorder, and Sample Characteristics identifies major depressive disorder and minor depression among participants. Our copy does not identify dysthymia among this study’s eligible diagnoses. The sentence attributes adjustment disorder with depression to a different citation, not Meeks 2015; page numbers were not available in our copy. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites Reinhardt 2014 sourceSupported

In the single study reporting long-term depression remission, participants allocated to psychological therapy had an MDD prevalence increase from 47.6% at baseline to 55.5% at long-term follow-up, a 7.9% increase.

“Amongst participants allocated to psychological therapy, the prevalence of MDD increased from 47.6% at baseline to 55.5% at long-term follow-up (7.9% increase).”

• cites Meeks 2015 sourceCould not verify

  The Diagnostic Recovery section reports remission and improvement at follow-up, but not MDD prevalence in the treatment group at baseline and long-term follow-up. Searches for 47.6%, 55.5%, and follow-up MDD prevalence found no passage establishing the claimed 7.9-percentage-point increase. (LLM, conf=0.73)

  Experimental — accuracy not yet measured

In the single study reporting medium-term depression remission, participants allocated to psychological therapy had an MDD prevalence increase from 47.6% at baseline to 48.5% at medium-term follow-up, a 0.9% increase.

“Amongst participants allocated to psychological therapy, the prevalence of MDD increased from 47.6% at baseline to 48.5% at medium-term follow-up (0.9% increase).”

• cites Meeks 2015 sourceCould not verify

  The Diagnostic Recovery section reports improvement or remission at three months, but not MDD prevalence in the treatment group at baseline and medium-term follow-up. Searches for 47.6%, 48.5%, and the claimed prevalence increase found no supporting figures in our copy; the recovery results do not establish whether MDD prevalence increased. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

The net difference in change in MDD prevalence between psychological therapy and non-therapy comparator at long-term follow-up was 32.3%.

“The net difference in change in prevalence between conditions was 32.3%.”

• cites Meeks 2015 sourcePartly supported

  The Diagnostic Recovery section reports greater remission with Be-ACTIV than with treatment as usual at posttreatment (45.2% versus 15.0%). It does not report a 32.3-percentage-point net change in MDD prevalence at long-term follow-up; instead, it says the groups were not significantly different by six months. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

Interpretation of the treatment non-acceptability result requires caution because of inadequate reporting and potential unreliability, with very wide confidence intervals.

“Interpretation of the result for treatment non-acceptability requires particular caution due to inadequate reporting of this outcome and potential unreliability of the estimate, with very wide CIs for ORs (e.g.0.47 to 202.37 for Hamzehzadeh 2018, and 0.30 to 130.07 for Reinhardt 2014).”

• cites Hamzehzadeh 2018Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Reinhardt 2014 sourcePartly supported

  The Results (page not available) report difficulties with treatment acceptance and adherence, including loss to follow-up of 10/21 participants in the treatment group versus 1/16 in the comparison group. The paper also notes its small sample, but it does not report an odds ratio, the stated confidence interval of 0.30 to 130.07, or an assessment of that estimate's reliability. Those parts of the caution appear to be the citing authors' analysis rather than results reported by Reinhardt et al. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

The net difference in change in MDD prevalence between psychological therapy and non-therapy comparator at medium-term follow-up was 47.2%.

“The net difference in change in prevalence between conditions was 47.2%.”

• cites Meeks 2015 sourcePartly supported

  The Diagnostic Recovery section reports better improvement or remission with BE-ACTIV than with usual care among participants remaining at three months (66.7% versus 40.7%). It does not report a 47.2-percentage-point net change in MDD prevalence at medium-term follow-up; its posttreatment remission figures are 45.2% versus 15.0%. (LLM, conf=0.81)

  Experimental — accuracy not yet measured

Some previous reviews examined a broad range of non-pharmacological interventions and were not specific to the effectiveness of psychological therapies.

“some reviews examined a broad range of non-pharmacological interventions and were not specific to the effectiveness of psychological therapies”

• cites Gramaglia 2021 sourceSupported
• cites Hyer 2005 sourceNot checked — no full text of the source
• cites Snowden 2003 sourceNot checked — no full text of the source

Previous narrative and systematic reviews have had limited conclusions about the effectiveness of psychological therapies for LTC residents with depression because they used different methods.

“Conclusions regarding the effectiveness of psychological therapies for LTC residents with depression have been limited in previous narrative and systematic reviews due to different methods employed”

• cites Bharucha 2006 sourceNot checked — no full text of the source
• cites Cody 2013 sourceNot checked — no full text of the source
• cites Gramaglia 2021 sourcePartly supported

  The Discussion (Summary of Main Findings) says previous reviews used differing inclusion criteria and covered different types of studies. The Discussion (Limitations) says heterogeneity hinders clear conclusions about effectiveness, but applies that conclusion to this review of non-pharmacological interventions broadly—not specifically to conclusions in previous reviews about psychological therapies. (LLM, conf=0.79)

  Experimental — accuracy not yet measured

We lack clear evidence of the effectiveness of non-psychological interventions such as befriending and video calls in LTC settings.

“we lack clear evidence of the effectiveness of non-psychological interventions such as befriending (Doyle 2021) and video calls (Noone 2020) in LTC settings.”

• cites Doyle 2021 sourceSupported
• cites Noone 2020 sourceSupported

For Reinhardt 2014, the sample consisted of two potentially different groups of nursing home residents.

“study sample … was composed of 2 potentially different groups of nursing home residents”

• cites Reinhardt 2014 sourceSupported

For Meeks 2015, reported outcomes matched those in the trial registration posted in 2007.

“reported outcomes matched those in trial registration posted 2007”

• cites Meeks 2015 sourceCould not verify

  The Measures section identifies diagnostic recovery as the primary outcome, and the Discussion reports trial outcomes. Our copy does not give the outcomes listed in a 2007 trial registration or compare them with the reported outcomes (searched: registration, 2007, and registered outcomes). (LLM, conf=0.76)

  Experimental — accuracy not yet measured

Meeks 2015 was judged at high risk of other bias.

“Other bias High risk Judgement comment: the prevalence of MDD appeared to be unbalanced between groups (47.6% in psychological therapy group vs 25% in non-therapy comparator group).Author developed the intervention, so potential researcher allegiance.”

• cites Meeks 2015 sourcePartly supported

  The Sample Characteristics section reports a significant baseline imbalance in depressive diagnoses, with proportionally fewer participants with major depressive disorder in the BE-ACTIV group. The paper also describes the authors’ role in developing the intervention and acknowledges possible bias from a study author’s unblinded initial interviews (Discussion). It does not itself classify the trial as at high risk of other bias, and the cited 47.6% and 25% figures are not shown in the passages found. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

For Reinhardt 2014, the sample may not have been homogeneous.

“sample may not have been homogeneous”

• cites Reinhardt 2014 sourceSupported

The same depression remission results occurred at short-term follow-up (up to 3 months).

“The same results occurred at shortterm follow-up (up to 3 months).”

At short-term follow-up (up to 3 months), the mean quality of life or psychological well-being score was 0.51 standard deviations higher for the psychological therapy group than for non-therapy comparators (95% CI 0.19 to 0.82 higher).

“At short-term follow-up (up to 3 months), the mean score for the psychological therapy group was 0.51 standard deviations higher (95% CI 0.19 to 0.82 higher) than for non-thera-”

At end-of-intervention follow-up (4 to 13 weeks), the odds of treatment non-acceptability were 3.44 times higher for psychological therapies than for non-therapy comparators, based on 5 RCTs, with very low certainty.

“The odds of treatment non-acceptability for psychological therapies was 3.44 times higher (95% CI 1.19 to 9.93 higher) than for non-therapy comparators 313 (5 RCTs) ⊕⊝⊝⊝ Very low b,d”

• AssumptionThe authors state this as an assumption: Dropout rate following allocation is a valid proxy measure for treatment acceptability.

In the remaining 10 studies that reported gender, women represented between 58.6% and 90.2% of the sample.

“In the remaining 10 studies that reported on the gender of participants, women represented between 58.6% and 90.2% of the sample.”

Fifteen trials had two arms comparing a psychological therapy to treatment as usual (11 studies), wait list control (Rosen 1997), or attention/social contact control (Reinhardt 2014; Travers 2017; Tsai 2008).

“Fi een trials had two study arms, which compared a psychological therapy to treatment as usual (11 studies), a wait list control (Rosen 1997), or a condition that controlled for the effect of attention or social contact (Reinhardt 2014;Travers 2017;Tsai 2008).”

• AssumptionThe authors state this as an assumption: The control condition in the Sales 2015 study was treatment as usual.

A cognitive-impairment or dementia exclusion criterion was unreported in only one study (Hussian 1979).

“Only one study did not report an exclusion criterion related to cognitive impairment or dementia (Hussian 1979).”

In 16 included studies, eligibility required a minimum level of depressive symptoms measured by a screening tool, most often the GDS-30, GDS-15, BDI, or SDS.

“In 16 studies, inclusion criteria specified a minimum level of depressive symptoms using a screening tool, most commonly, the GDS-30 (Yesavage 1983), GDS-15 (Sheikh 1986), BDI (Beck 1961), or the SDS (Zung 1965).”

• cites Yesavage 1983Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Sheikh 1986 sourceNot checked — no full text of the source
• cites Beck 1961 sourceNot checked — no full text of the source

Two studies restricted recruitment to people with mild-to-moderate cognitive impairment (MMSE 10 to 24) or dementia, and excluded older adults with normal cognition (Bailey 2017; Travers 2017).

“In contrast, two studies restricted recruitment to people with mildto-moderate cognitive impairment (MMSE 10 to 24) or dementia, and excluded older adults with normal cognition (Bailey 2017; Travers 2017).”

No included studies restricted eligibility to residents with MDD at baseline.

“No studies were restricted to residents with MDD at baseline.”

Studies commonly used standard cut-off scores for significant depressive symptoms, such as 10 or 11 on the GDS-30, 5, 6, or 8 on the GDS-15, and 50 on the SDS.

“Studies typically used standard cut-off scores indicative of a significant level of depressive symptoms, including a score of 10 or 11 on the GDS- 30 (Abraham 1992;Meeks 2008;Meeks 2015;Sales 2015;Tsai 2008), a score of 5, 6, or 8 on the GDS-15 (Daleo 1999;Hamzehzadeh 2018;Hyer 2009;Karimi 2010), and a score of 50 on the SDS (Dhooper 1993;Hansen 1976).”

• cites Abraham 1992Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Daleo 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Dhooper 1993Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It was usually not possible to classify baseline depressive symptom levels as mild, moderate, or severe because of outcome measure limitations, limited use of upper symptom limits, and limited reporting.

“It was usually not possible to classify baseline levels of depressive symptoms of participants in individual studies as mild, moderate, and severe, due to the nature of the outcome measures used (i.e. an absence of an accepted cut-off score to define severity), the limited use of upper limits of depression symptoms, and the limited information reported.”

The review did not identify any studies with a defined co-intervention control group.

“We did not identify any studies with a defined co-intervention control group, where there was a separate non-psychological treatment administered in both the control and intervention arms of the study.”

Twelve studies provided treatment non-acceptability data based on the review's definition of attrition.

“Twelve studies were judged as providing treatment non-acceptability data based on this definition (Abraham 1992;Daleo 1999;Dhooper 1993;Dozeman 2011;Hamzehzadeh 2018;Hansen 1976;Hyer 2009;Luo 2020;Meeks 2015;Reinhardt 2014;Travers 2017;Tsai 2008).”

• cites Abraham 1992Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Daleo 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Dhooper 1993Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Nearly all studies included in this review were conducted in high-income countries, particularly the USA.

“Nearly all studies included in this review were conducted in high-income countries, in particular, in the USA.”

Subgroup analysis showed no difference in effect on depressive symptoms by type of psychological therapy (Chi2 = 2.23, P = 0.33).

“Subgroup analysis showed no difference in effect by type of psychological therapy (Chi 2 = 2.23, P = 0.33; Analysis 3.1).”

Subgroup analysis showed no difference in effect by therapeutic contact based on number of sessions (Chi2 = 0.63, P = 0.43).

“number of sessions (Chi 2 = 0.63, P = 0.43; Analysis 7.2),”

Very low-certainty evidence showed higher odds of treatment nonacceptability for any psychological therapy compared with any non-therapy comparator at end-of-intervention.

“Very low-certainty evidence showed higher odds of treatment nonacceptability for any psychological therapies compared to any non-therapy comparator at end-of-intervention.”

Psychological therapy implemented by a therapist without involvement from LTC staff may have little to no effect on depressive symptoms at end-of-intervention follow-up compared with any non-therapy comparator, but the evidence is very uncertain (SMD -0.76, 95% CI -1.68 to 0.16; 5 RCTs, 176 participants).

“In contrast, psychological therapy implemented by a therapist without involvement from LTC staff may have little to no effect on depressive symptomatology at end-of-intervention follow-up compared to any non-therapy comparator, but the evidence was very uncertain (SMD -0.76, 95% CI -1.68 to 0.16; 5 RCTs, 176 participants).”

Depressive symptoms may be lower at end-of-intervention follow-up for any psychological therapy with greater than 9.5 total sessions compared with any non-therapy comparator, with a medium effect estimate, but the evidence is very uncertain (SMD -0.72, 95% CI -1.15 to -0.28; 7 RCTs, 271 participants).

“and lower for any psychological therapy greater than 9.5 total sessions, with a medium effect estimate (SMD -0.72, 95% CI -1.15 to -0.28; 7 RCTs, 271 participants) (Analysis 7.2); however, the evidence was very uncertain.”

Psychological therapy may reduce depressive symptoms at end-of-intervention follow-up compared with non-therapy comparator for studies with participants with low cognitive impairment, with a large effect estimate, but the evidence is very uncertain (SMD -1.30, 95% CI -2.26 to -0.34; 7 RCTs, 267 participants).

“Psychological therapy may reduce depressive symptomatology at end-of-intervention follow-up compared to non-therapy comparator for studies with participants with low cognitive impairment, with a large effect estimate (SMD -1.30, 95% CI -2.26 to -0.34; 7 RCTs, 267 participants),”

When Tsai 2008 was omitted because of study-design concerns, there was no evidence of an effect on depressive symptoms at medium-term follow-up (SMD -0.44, 95% CI -0.90 to 0.02).

“When Tsai 2008 was omitted due to concerns with study design, there was no evidence of an effect (SMD -0.44, 95% CI -0.90 to 0.02; Analysis 13.1).”

• cites Tsai 2008 sourceNot checked — no full text of the source

The effect of psychological therapies on quality of life and psychological well-being was not found at end-of-intervention or medium-term follow-up.

“and this effect was not found at end-of-intervention or medium-term follow-up.”

Studies typically recruited participants based on a score over a cut-off on a validated depression screening tool, indicating at least mild or subthreshold to moderate levels of depressive symptoms.

“Studies in this review typically recruited participants based on a score over a cut-off on a validated depression screening tool, indicating at least mild or subthreshold depression to moderate levels of depressive symptoms.”

Depressive symptoms may be lower at end-of-intervention follow-up for any psychological therapy of duration less than 8.5 weeks compared with any non-therapy comparator, with a medium effect estimate, but the evidence is very uncertain (SMD -0.74, 95% CI -1.34 to -0.14; 8 RCTs, 251 participants).

“Depressive symptomatology may be lower at end-of-intervention follow-up for any psychological therapy of duration less than 8.5 weeks (dichotomized at the median; see Subgroup analysis and investigation of heterogeneity) compared to any non-therapy comparator, with a medium effect estimate (SMD -0.74, 95%”

Subgroup analysis showed a difference by non-therapy comparator, with treatment as usual yielding greater effect estimates than non-specific attentional control (Chi2 = 7.87, P = 0.005).

“Subgroup analysis showed a difference by nontherapy comparator, whereby comparison to treatment as usual yielded greater effect estimates than comparison to non-specific attentional control (Chi 2 = 7.87, P = 0.005; Analysis 4.1).”

Subgroup analysis showed no difference in effect on depressive symptoms by baseline depression severity (Chi2 = 1.01, P = 0.31).

“Subgroup analysis showed no difference in effect by baseline depression severity (Chi 2 = 1.01,P = 0.31; Analysis 2.1).”

Due to data paucity, physical functioning, agitation, and adverse effects could not be examined at end-of-intervention.

“Due to data paucity, the following outcome measures were unable to be examined at end-of-intervention follow-up: physical functioning, agitation, and adverse effects.”

With Luo 2020 removed from analyses of quality of life and psychological well-being, psychological therapy appeared to have little to no effect on these outcomes at end-of-intervention and short-term follow-up.

“With this study removed, psychological therapy appeared to have little to no effect on quality of life and psychological well-being at end-ofintervention and short-term follow-up.”

• cites Luo 2020 sourceNot checked — no full text of the source

In Rosen 1997, improvement in depressive symptomatology among those allocated to psychological therapy was not greater than among those allocated to the non-therapy comparator.

“The study authors reported that amongst those randomly allocated to psychological therapy, the improvement in depressive symptomatology was not greater than for those allocated to the non-therapy comparator (Rosen 1997).”

• cites Rosen 1997 sourceNot checked — no full text of the source

Depressive symptoms may be lower at end-of-intervention follow-up for any psychological therapy of duration greater than 8.5 weeks compared with any non-therapy comparator, with a large effect estimate, but the evidence is very uncertain (SMD -1.25, 95% CI -1.99 to -0.51; 9 RCTs, 368 participants).

“and lower for any psychological therapy of duration greater than 8.5 weeks, with a large effect estimate (SMD -1.25, 95% CI -1.99 to -0.51; 9 RCTs, 368 participants) (Analysis 7.1); however, the evidence was very uncertain.”

Depressive symptoms may be lower at end-of-intervention follow-up for any psychological therapy with fewer than 9.5 total sessions compared with any non-therapy comparator, with a large effect estimate, but the evidence is very uncertain (SMD -1.08, 95% CI -1.86 to -0.30; 9 RCTs, 310 participants).

“Depressive symptomatology may be lower at end-of-intervention follow-up for any psychological therapy fewer than 9.5 total sessions compared to any non-therapy comparator, with a large effect estimate (SMD -1.08, 95% CI -1.86 to -0.30; 9 RCTs, 310 participants),”

Due to data paucity, treatment non-acceptability, anxious symptomatology, physical functioning, agitation, and adverse effects could not be examined at short-term follow-up.

“Due to data paucity, the following outcome measures were unable to be examined at short-term follow-up: treatment non-acceptability, anxious symptomatology, physical functioning, agitation, adverse effects.”

Psychological therapy may reduce depressive symptomatology at end-of-intervention compared with any non-therapy comparator, with a large effect estimate but very uncertain evidence (SMD -1.04, 95% CI -1.49 to -0.58; 18 RCTs, 644 participants).

“Psychological therapy may reduce depressive symptomatology at end-of-intervention follow-up compared to any non-therapy comparator, with a large effect estimate, but the evidence was very uncertain (SMD -1.04, 95% CI -1.49 to -0.58; 18 RCTs, 644 participants; Analysis 1.1).”

Subgroup analysis found no evidence of a difference in effect estimates by participant level of cognitive function (Chi2 = 2.65, P = 0.10).

“However, subgroup analysis found no evidence of a difference in effect estimates by participant level of cognitive function (Chi 2 = 2.65, P = 0.10; Analysis 5.1).”

Subgroup analysis showed no difference in effect by therapeutic contact based on total therapy duration in weeks (Chi2 = 1.12, P = 0.29).

“Despite the above findings, subgroup analysis showed no difference by therapeutic contact based on total therapy duration in weeks (Chi 2 = 1.12,P = 0.29; Analysis 7.1),”

Depressive symptoms may be lower at end-of-intervention follow-up for any psychological therapy of total dose greater than 6.5 hours compared with any non-therapy comparator, with a large effect estimate, but the evidence is very uncertain (SMD -1.83, 95% CI -2.93 to -0.73; 6 RCTs, 134 participants).

“Depressive symptomatology may be lower at end-of-intervention follow-up for any psychological therapy of total dose greater than 6.5 hours compared to any non-therapy comparator, with a large effect estimate (SMD -1.83, 95% CI -2.93 to -0.73; 6 RCTs, 134 participants),”

When only RCTs that measured treatment fidelity were included, psychological therapy may reduce depressive symptoms at end-of-intervention follow-up compared with any non-therapy comparator, with a large effect estimate, but the evidence is very uncertain (SMD -1.06, 95% CI -1.91 to -0.21; 4 RCTs, 158 participants).

“When including only RCTs that measured treatment fidelity (Bailey 2017;Meeks 2008;Meeks 2015;Zerhusen 1991), psychological therapy may reduce depressive symptomatology at end-ofintervention follow-up for any psychological therapy compared to any non-therapy comparator, with a large effect estimate, but the evidence was very uncertain (SMD -1.06, 95% CI -1.91 to -0.21; 4 RCTs, 158 participants;”

Subgroup analysis showed no difference in effect by therapeutic contact based on total dose in hours (Chi2 = 2.72, P = 0.10).

“or total dose in hours (Chi 2 = 2.72, P = 0.10; Analysis 7.3).”

Subgroup analysis showed no evidence of a difference in effect estimates by LTC facility staff involvement (Chi2 = 0.22, P = 0.64).

“However, subgroup analysis showed no evidence of a difference in effect estimates by LTC facility staff involvement (Chi 2 = 0.22, P = 0.64; Analysis 6.1).”

Prospective trial registrations or protocol papers were located for only four studies in this review.

“We located prospective trial registrations or protocol papers for only four studies in this review.”

Only the subgroup of psychological therapies implemented by professional therapists with LTC staff involvement reduced symptoms of depression compared with the control.

“However, only the subgroup of psychological therapies implemented by professional therapists with LTC staff involvement reduced symptoms of depression compared with the control.”

Reporting of participant dropouts was poor in many included studies.

“the reporting of participant dropouts was poor in many studies”

Psychological therapies may have no effect on quality of life and psychological well-being at three to six months post-intervention compared with non-therapy controls, although the evidence is very uncertain.

“Psychological therapies may have no effect on quality of life and psychological well-being at three to six months' postintervention, compared to non-therapy controls, although the evidence is very uncertain.”

Sensitivity analyses indicated that initial estimates were robust given few changes in magnitude or statistical significance when studies with imputed missing data were omitted.

“Despite these concerns, sensitivity analyses indicated our initial estimates were robust given few changes in magnitude or statistical significance when studies with missing data imputed were omitted.”

Depressive symptomatology was consistently well reported using validated scale scores in the included studies.

“While depressive symptomatology was consistently well reported using scores on validated scales”

The effectiveness of psychological therapies for reducing depression for residents with high versus low cognitive impairment remains equivocal.

“The effectiveness of psychological therapies for reducing depression for residents with high versus low cognitive impairment remains equivocal.”

No evidence of non-reporting bias was found, whereby trials were selectively published.

“We found no evidence of non-reporting bias, whereby trials were selectively published (e.g.due to P values or magnitude of results).”

There were no subgroup differences in the effectiveness of psychological therapies implemented by professional therapists with and without the involvement of LTC staff.

“There were no subgroup differences found in the effectiveness of psychological therapies that were implemented by professional therapists with and without the involvement of LTC staff.”

When sensitivity analyses were conducted with only the six studies at low risk of bias for allocation concealment, the effect of psychological therapies was no longer observed.

“When sensitivity analyses were conducted with only these six studies, the effect of psychological therapies was no longer observed.”

This review observed a large effect estimate for psychological therapies in LTC residents on depressive symptomatology compared with no therapy.

“our review observed a large effect estimate for psychological therapies in LTC residents on depressive symptomatology compared with no therapy”

The superiority of psychological therapies for depression compared with non-therapy controls was not found at longer-term follow-up of more than six months post-intervention.

“The superiority of psychological therapies for depression compared to non-therapy controls was not found at longer term follow-up (more than six months' postintervention)”

There were no differences in the effectiveness of cognitive behavioural therapy, behavioural therapy, and reminiscence therapy in the subgroup analysis.

“There were no differences in the effectiveness of the three types of therapies in our subgroup analysis”

This review observed that psychological therapies had no effect on depressive symptoms compared with active comparators, namely non-specific attentional controls.

“We also similarly observed that psychological therapies had no effect on depressive symptoms when compared to active comparators (i.e.non-specific attentional controls) (Analysis 4.1).”

The effect estimate in this review was greater than the medium effect size estimated in the only known prior meta-analysis, which included 17 studies and 666 participants, including five RCTs also eligible for this review.

“Our estimate was greater than the medium effect size estimated in the only known prior meta-analysis (Cody 2013), which included 17 studies (666 participants), including five RCTs that were also eligible for our review.”

• cites Cody 2013 sourceNot checked — no full text of the source

For depressive symptomatology at end-of-intervention, the funnel plot did not indicate publication bias.

“For depressive symptomatology at end-of-intervention, the funnel plot did not indicate publication bias.”

When only studies that examined treatment fidelity were included, evidence of a difference between outcomes from psychological therapy and non-therapy comparators remained, with large estimates of effects.

“Similarly, when only including studies that examined treatment fidelity in the meta-analysis, evidence of a difference between outcomes from psychological therapy and non-therapy comparators remained, with large estimates of effects.”

Authors of included studies commonly did not identify the time point when individual participants stopped participating.

“Authors commonly failed to identify the time point in the trial when individual participants were no longer involved.”

Overall, psychological therapies did not appear to result in differential effects for individuals with higher or lower levels of depression at baseline.

“Overall, psychological therapies did not appear to result in differential effects for individuals with higher or lower levels of depression at baseline.”

Among 12 studies with participant attrition data, only five reported non-zero attrition.

“Of the 12 studies that provided data on participant attrition, only five reported non-zero attrition.”

The certainty of the evidence was very low for all outcomes.

“The certainty of the evidence was very low for all outcomes.”

Only two studies reported anxiety symptomatology as an outcome.

“Only two studies reported anxiety symptomatology as an outcome, which is concerning given indications in the literature of high levels of comorbid depression and anxiety (Creighton 2018).”

• cites Creighton 2018 sourceNot checked — no full text of the source

The duration and number of sessions of treatment did not appear to moderate the effects of psychological therapies.

“Also, the duration and number of sessions of treatment did not appear to moderate such effects.”

All participants in the included trials had at least mild symptoms of depression.

“all participants had at least mild symptoms of depression.”

This review lacked sufficient data to compare effect estimates for studies training LTC staff to implement psychological therapies with those implemented by professional therapists.

“we lacked sufficient data to enable us to compare the effect estimate for these studies compared to those where the intervention was implemented by a professional therapist”

The psychological therapies trialled in these studies were relatively brief, with most involving between two and 10 sessions.

“It was notable that the psychological therapies trialled in these studies were relatively brief, with most involving between two and 10 sessions.”

In this meta-analysis, only the subgroup of psychological therapies implemented by professional therapists with involvement of LTC staff reduced depressive symptoms at end-of-intervention compared with non-therapy comparators.

“In the current meta-analysis, only the subgroup of psychological therapies implemented by professional therapists (including graduate students) with involvement of LTC staff reduced depressive symptoms at end-of-intervention compared with nontherapy comparators (Analysis 6.1).”

This review expanded the finding to include that behavioural therapy may also be effective.

“however, we expanded this to include a finding that BT may also be effective (Analysis 3.1).”

This review found no overall difference in effect size between therapies implemented by professional therapists with and without staff input.

“However, we found no overall difference in effect size between therapies implemented by professional therapists with and without staff input.”

The effect of involving LTC facility staff in implementing the interventions is unclear.

“The effect of involving LTC facility staff in implementing the interventions is unclear.”

This review of 19 studies provides evidence that psychological therapies may be effective for depression in LTC.

“This review of 19 studies provides evidence that psychological therapies may be effective for depression in LTC.”

Only two studies were identified for this review with a follow-up period of six months or more.

“Only two studies were identified for this review with a follow-up period of six months or more.”

Most studies included in the current review were prone to bias.

“Most of the studies included in the current review were prone to bias.”

Researcher involvement in the evaluation of interventions they developed created potential conflicts of interest for several included trials.

“Researcher involvement in the evaluation of interventions they developed created potential conflicts of interest for several of the included trials.”

The absence of sufficient evidence on longer-term impacts of psychological therapies is a clear gap in research.

“the absence of sufficient evidence on longer-term impacts of psychological therapies is a clear gap in research.”

Other studies claimed to be RCTs but failed to provide sufficient details to determine adequacy of the randomization process.

“In contrast, others claimed to be RCTs but failed to provide sufficient details to determine adequacy of the randomization process.”

The assessment of treatment fidelity was often omitted in the included trials.

“Finally, the assessment of treatment fidelity was o en omitted”

Psychological therapies may not be superior to providing residents with non-specific attention, such as participation in a discussion or education group or a social visit from a volunteer or student.

“However, psychological therapies may not be superior to providing residents with non-specific attention, such as participation in a discussion or education group or a social visit from a volunteer or student.”

In many trials, a large proportion of the sample had been prescribed an antidepressant medication and continued its use, but authors did not typically attempt to control statistically for the effect of this medication use.

“In many trials, while a large proportion of the sample had been prescribed an antidepressant medication, and continued their use, authors did not typically attempt to control statistically for the effect of this medication use.”

Subgroup analyses suggested that psychological therapies may be superior to treatment as usual.

“Subgroup analyses suggested that psychological therapies may be superior to treatment as usual, which typically includes assistance with activities of daily living and an opportunity to participate in a programme of activities organized by the LTC facility, and may include prescription of antidepressant medications.”

• AssumptionRests on an assumption: Treatment as usual in the included trials typically includes assistance with activities of daily living and an opportunity to participate in facility-organized activities, and may include prescription of antidepressant medications.

This review collated evidence from 19 randomized controlled trials conducted between 1976 and 2020.

“This review collated evidence from 19 randomized controlled trials (RCTs) conducted between 1976 and 2020.”

Some studies were labelled as quasi-experimental but reported randomized allocation of participants to treatment conditions.

“Some studies were labelled as quasi-experimental but reported randomized allocation of participants to treatment conditions.”

The studies included in this review paid little attention to broader treatment outcomes such as quality of life and psychological well-being.

“The studies included in this review paid little attention to broader treatment outcomes such as quality of life and psychological wellbeing.”

Lack of support for research evaluating psychological treatments for depression in LTC settings was evident.

“lack of support for research evaluating psychological treatments for depression in LTC settings was evident.”

The difference between the two number-of-sessions subgroups was not statistically significant (Chi2 = 0.63, df = 1, P = 0.43, I2 = 0%).

“Test for subgroup differences: Chi² = 0.63, df = 1 (P = 0.43), I² = 0%”

With the Luo 2020 study omitted, the pooled short-term effect on depressive symptoms was -1.07 (95% CI -1.58 to -0.56) across 15 studies with 462 participants.

“12.1.2Short-term 15 462 Std.Mean Difference (IV, Random, 95% CI) -1.07 [-1.58, -0.56]”

Among participants with high cognitive impairment, psychological therapy does not significantly reduce depressive symptoms at end of intervention (standardised mean difference -0.41, 95% CI -0.89 to 0.07; 5 studies, 176 participants).

“5.1.1High cognitive impairment 5 176 Std.Mean Difference (IV, Random, 95% CI) -0.41 [-0.89, 0.07]”

In the analysis labelled 5.1, the pooled effect across subgroups was statistically significant (Z = 2.90, P = 0.004).

“Total (95% CI) Heterogeneity: Tau² = 0.69; Chi² = 67.49,df = 11 (P < 0.00001); I² = 84% Test for overall effect: Z = 2.90 (P = 0.004)”

Psychological therapy does not significantly reduce anxious symptoms compared with non-therapy comparators at the end of intervention (standardised mean difference -0.68, 95% CI -2.50 to 1.14; 2 studies, 115 participants).

“1.4 Anxious symptomatology 2 115 Std.Mean Difference (IV, Random, 95% CI) -0.68 [-2.50, 1.14]”

The difference between the two total-treatment-dose subgroups was not statistically significant (Chi2 = 2.72, df = 1, P = 0.10, I2 = 63.3%).

“Test for subgroup differences: Chi² = 2.72, df = 1 (P = 0.10), I² = 63.3%”

The difference between the two long-term care staff involvement subgroups was not statistically significant (Chi2 = 0.22, df = 1, P = 0.64, I2 = 0%).

“Test for subgroup differences: Chi² = 0.22, df = 1 (P = 0.64), I² = 0%”

In the sensitivity analysis restricted to studies with low attrition, the pooled effect on depressive symptoms was -1.29 (95% CI -1.83 to -0.75) across 14 studies with 499 participants.

“11.1 Depressive symptomatology 14 499 Std.Mean Difference (IV, Random, 95% CI) -1.29 [-1.83, -0.75]”

Behavioural therapy reduces depressive symptoms at end of intervention compared with non-therapy comparators (standardised mean difference -0.69, 95% CI -1.21 to -0.18; 8 studies, 303 participants).

“3.1.2BT 8 303 Std.Mean Difference (IV, Random, 95% CI) -0.69 [-1.21, -0.18]”

In the low cognitive impairment subgroup of analysis 5.1, the pooled effect was statistically significant (Z = 2.66, P = 0.008).

“5.1.2Low cognitive impairment Daleo 1999 Dhooper 1993 Dozeman 2011 Hyer 2009 Reinhardt 2014 Sales 2015 Tsai 2008 Subtotal (95% CI) Heterogeneity: Tau² = 1.36;Chi² = 58.59,df = 6 (P < 0.00001); I² = 90% Test for overall effect: Z = 2.66 (P = 0.008)”

The effect of psychological therapy on depressive symptoms does not differ significantly by type of therapy (Chi2 = 2.23, df = 2, P = 0.33, I2 = 10.3%).

“Test for subgroup differences: Chi² = 2.23, df = 2 (P = 0.33), I² = 10.3%”

Reminiscence therapy reduces depressive symptoms at end of intervention compared with non-therapy comparators (standardised mean difference -1.65, 95% CI -2.83 to -0.47; 6 studies, 203 participants).

“3.1.3RT 6 203 Std.Mean Difference (IV, Random, 95% CI) -1.65 [-2.83, -0.47]”

The effect of psychological therapy on depressive symptoms differs significantly by type of non-therapy comparator (Chi2 = 7.87, df = 1, P = 0.005, I2 = 87.3%).

“Test for subgroup differences: Chi² = 7.87, df = 1 (P = 0.005), I² = 87.3%”

At medium-term follow-up, the effect of psychological therapy on quality of life and psychological well-being is not statistically significant (standardised mean difference 0.26, 95% CI -0.95 to 1.48; 2 studies, 68 participants).

“1.3.3Medium-term 2 68 Std.Mean Difference (IV, Random, 95% CI) 0.26 [-0.95, 1.48]”

At long-term follow-up, the effect of psychological therapy on depressive symptoms is not statistically significant (standardised mean difference -0.16, 95% CI -0.58 to 0.27; 2 studies, 92 participants).

“1.1.4Long-term 2 92 Std.Mean Difference (IV, Random, 95% CI) -0.16 [-0.58, 0.27]”

Among participants with low cognitive impairment, psychological therapy reduces depressive symptoms at end of intervention (standardised mean difference -1.30, 95% CI -2.26 to -0.34; 7 studies, 267 participants).

“5.1.2Low cognitive impairment 7 267 Std.Mean Difference (IV, Random, 95% CI) -1.30 [-2.26, -0.34]”

Across 14 studies with 470 participants, psychological therapies reduced depressive symptoms relative to non-therapy comparators in long-term care residents, with a standardised mean difference of -0.93 (95% CI -1.43 to -0.43).

“6.1 Depressive symptomatology (endof-intervention) 14 470 Std.Mean Difference (IV, Random, 95% CI) -0.93 [-1.43, -0.43]”

In the sensitivity analysis restricted to studies with treatment fidelity, the pooled effect on depressive symptoms at end of intervention was -1.06 (95% CI -1.91 to -0.21) across 4 studies with 158 participants.

“9.1 Depressive symptomatology (endof-intervention) 4 158 Std.Mean Difference (IV, Random, 95% CI) -1.06 [-1.91, -0.21]”

Across 16 studies with 581 participants, the pooled effect of psychological therapies on depressive symptoms by number of sessions was -0.87 (95% CI -1.31 to -0.42).

“7.2 Depressive symptomatology, therapeutic contact (number of sessions) 16 581 Std.Mean Difference (IV, Random, 95% CI) -0.87 [-1.31, -0.42]”

Among participants with baseline depression severity below 54.8 points on a standardised 100-point scale, psychological therapy reduces depressive symptoms at end of intervention (standardised mean difference -0.81, 95% CI -1.38 to -0.25; 9 studies, 371 participants).

“2.1.2< 54.8 points on a standardised 100-point scale 9 371 Std.Mean Difference (IV, Random, 95% CI) -0.81 [-1.38, -0.25]”

With the Luo 2020 study omitted, the pooled effect on depressive symptoms at end of intervention was -1.07 (95% CI -1.57 to -0.58) across 17 studies with 594 participants.

“12.1.1End-of-intervention 17 594 Std.Mean Difference (IV, Random, 95% CI) -1.07 [-1.57, -0.58]”

The effect of psychological therapy on depressive symptoms does not differ significantly between higher and lower baseline depression severity subgroups (Chi2 = 1.01, df = 1, P = 0.31, I2 = 1.4%).

“Test for subgroup differences: Chi² = 1.01, df = 1 (P = 0.31), I² = 1.4%”

Across 12 studies with 326 participants, the pooled effect of psychological therapies on depressive symptoms by total treatment dose was -1.20 (95% CI -1.87 to -0.54).

“7.3 Depressive symptomatology, therapeutic contact (total treatment dose) 12 326 Std.Mean Difference (IV, Random, 95% CI) -1.20 [-1.87, -0.54]”

For trials with fewer than 9.5 total sessions, the pooled effect on depressive symptoms was -1.08 (95% CI -1.86 to -0.30) across 9 studies with 310 participants.

“7.2.1 < 9.5 total sessions 9 310 Std.Mean Difference (IV, Random, 95% CI) -1.08 [-1.86, -0.30]”

When a professional therapist delivered therapy with long-term care staff involvement, the pooled effect on depressive symptoms was -1.02 (95% CI -1.62 to -0.43) across 9 studies with 303 participants.

“6.1.2Professional therapist with LTC staff involvement 9 303 Std.Mean Difference (IV, Random, 95% CI) -1.02 [-1.62, -0.43]”

The difference between the two therapeutic-contact-in-weeks subgroups was not statistically significant (Chi2 = 1.12, df = 1, P = 0.29, I2 = 10.8%).

“Test for subgroup differences: Chi² = 1.12, df = 1 (P = 0.29), I² = 10.8%”

Psychological therapy does not significantly reduce depressive symptoms at end of intervention compared with non-specific attentional control (standardised mean difference -0.30, 95% CI -0.83 to 0.23; 6 studies, 208 participants).

“4.1.1Non-specific attentional control 6 208 Std.Mean Difference (IV, Random, 95% CI) -0.30 [-0.83, 0.23]”

In the sensitivity analysis restricted to studies with missing data, the pooled effect on depressive symptoms at end of intervention was -1.11 (95% CI -1.71 to -0.51) across 12 studies with 469 participants.

“8.1 Depressive symptomatology (endof-intervention) 12 469 Std.Mean Difference (IV, Random, 95% CI) -1.11 [-1.71, -0.51]”

The benefit of psychological therapy over non-therapy comparators for depressive symptoms persists in the short term (standardised mean difference -1.03, 95% CI -1.49 to -0.56; 16 studies, 512 participants).

“1.1.2Short-term 16 512 Std.Mean Difference (IV, Random, 95% CI) -1.03 [-1.49, -0.56]”

Cognitive behavioural therapy reduces depressive symptoms at end of intervention compared with non-therapy comparators (standardised mean difference -1.02, 95% CI -1.99 to -0.04; 5 studies, 125 participants).

“3.1.1CBT 5 125 Std.Mean Difference (IV, Random, 95% CI) -1.02 [-1.99, -0.04]”

In older adults living in long-term care, psychological therapy of any type, compared with any non-therapy comparator, lowers depressive symptoms at the end of intervention (standardised mean difference -1.04, 95% CI -1.49 to -0.58; 18 studies, 644 participants).

“1.1.1End-of-intervention 18 644 Std.Mean Difference (IV, Random, 95% CI) -1.04 [-1.49, -0.58]”

For trials with more than 6.5 total hours of treatment, the pooled effect on depressive symptoms was -1.83 (95% CI -2.93 to -0.73) across 6 studies with 134 participants.

“7.3.2> 6.5 total hours 6 134 Std.Mean Difference (IV, Random, 95% CI) -1.83 [-2.93, -0.73]”

For trials with less than 8.5 weeks of therapeutic contact, the pooled effect on depressive symptoms was -0.74 (95% CI -1.34 to -0.14) across 8 studies with 251 participants.

“7.1.1< 8.5 weeks 8 251 Std.Mean Difference (IV, Random, 95% CI) -0.74 [-1.34, -0.14]”

For trials with more than 9.5 total sessions, the pooled effect on depressive symptoms was -0.72 (95% CI -1.15 to -0.28) across 7 studies with 271 participants.

“7.2.2 > 9.5 total sessions 7 271 Std.Mean Difference (IV, Random, 95% CI) -0.72 [-1.15, -0.28]”

Psychological therapy is associated with higher treatment non-acceptability than non-therapy comparators at the end of intervention (odds ratio 3.44, 95% CI 1.19 to 9.93; 12 studies, 585 participants).

“1.2 Treatment non-acceptability 12 585 Odds Ratio (M-H, Random, 95% CI) 3.44 [1.19, 9.93]”

For trials with more than 8.5 weeks of therapeutic contact, the pooled effect on depressive symptoms was -1.25 (95% CI -1.99 to -0.51) across 9 studies with 368 participants.

“7.1.2> 8.5 weeks 9 368 Std.Mean Difference (IV, Random, 95% CI) -1.25 [-1.99, -0.51]”

At the end of intervention, psychological therapy does not significantly improve quality of life and psychological well-being compared with non-therapy comparators (standardised mean difference 0.40, 95% CI -0.02 to 0.82; 6 studies, 195 participants).

“1.3.1 End-of-intervention 6 195 Std.Mean Difference (IV, Random, 95% CI) 0.40 [-0.02, 0.82]”

Across 17 studies with 619 participants, the pooled effect of psychological therapies on depressive symptoms by therapeutic contact in weeks was -0.98 (95% CI -1.45 to -0.52).

“7.1 Depressive symptomatology, therapeutic contact (weeks) 17 619 Std.Mean Difference (IV, Random, 95% CI) -0.98 [-1.45, -0.52]”

Psychological therapy reduces depressive symptoms at end of intervention compared with treatment as usual (standardised mean difference -1.47, 95% CI -2.10 to -0.84; 12 studies, 430 participants).

“4.1.2Treatment as usual 12 430 Std.Mean Difference (IV, Random, 95% CI) -1.47 [-2.10, -0.84]”

Among participants with baseline depression severity above 54.8 points on a standardised 100-point scale, psychological therapy reduces depressive symptoms at end of intervention (standardised mean difference -1.31, 95% CI -2.08 to -0.53; 9 studies, 273 participants).

“2.1.1 > 54.8 points on a standardised 100-point scale 9 273 Std.Mean Difference (IV, Random, 95% CI) -1.31 [-2.08, -0.53]”

With the Tsai 2008 study omitted, the pooled effect on depressive symptoms at end of intervention was -1.13 (95% CI -1.59 to -0.67) across 17 studies with 581 participants.

“13.1.1End-of-intervention 17 581 Std.Mean Difference (IV, Random, 95% CI) -1.13 [-1.59, -0.67]”

With the Tsai 2008 study omitted, the pooled short-term effect on depressive symptoms was -1.12 (95% CI -1.57 to -0.68) across 15 studies with 449 participants.

“13.1.2Short-term 15 449 Std.Mean Difference (IV, Random, 95% CI) -1.12 [-1.57, -0.68]”

In older adults with high baseline depression severity residing in long-term care, psychological therapy reduced depressive symptomatology at end of intervention by 1.31 standard deviations (95% CI 2.08 to 0.53 lower) compared with non-therapy comparators, based on 9 RCTs and 273 participants.

“The mean score for the psychological therapy group was 1.31 standard deviations lower (95% CI 2.08 to 0.53 lower) than for non-therapy comparators 273 (9 RCTs)”

Behavioural therapy reduced depressive symptomatology at end of intervention by 0.69 standard deviations (95% CI 1.21 to 0.18 lower) compared with non-therapy comparators, based on 8 RCTs and 303 participants.

“The mean score for the behavioural therapy group was 0.69 standard deviations lower (95% CI 1.21 to 0.18 303 (8 RCTs)”

In a sensitivity analysis excluding Daleo 1999, the standardized mean difference for depressive symptomatology at end-of-intervention was -0.86 (95% CI -1.26 to -0.46) across 17 studies with 619 participants.

“14.1.1End-of-intervention 17 619 Std.Mean Difference (IV, Random, 95% CI) -0.86 [-1.26, -0.46]”

Across 18 studies and 644 participants, the pooled effect of psychological therapy versus non-therapy comparators on depressive symptomatology was a standardised mean difference of -1.04 (95% CI -1.49 to -0.58).

“15.1 Depressive symptomatology 18 644 Std.Mean Difference (IV, Random, 95% CI) -1.04 [-1.49, -0.58]”

• Also stated in: § Cochrane Database of Systematic Reviews · plain

  “Total (95% CI) Heterogeneity: Tau² = 0.76; Chi² = 108.72,df = 17 (P < 0.00001); I² = 84% Test for overall effect: Z = 4.44 (P < 0.00001)”

In the sensitivity analysis omitting Daleo 1999, the pooled effect of psychological therapy versus non-therapy comparators on depressive symptomatology at end of intervention was a standardised mean difference of -0.86 (95% CI -1.26 to -0.46).

“Subtotal (95% CI) Heterogeneity: Tau² = 0.52; Chi² = 78.91,df = 16 (P < 0.00001); I² = 80% Test for overall effect: Z = 4.18 (P < 0.0001)”

• Also stated in: § Cochrane Database of Systematic Reviews · plain

  “-0.86 [-1.26 , -0.46]”

Cognitive behavioural therapy reduced depressive symptomatology at end of intervention by 1.02 standard deviations (95% CI 1.99 to 0.04 lower) compared with non-therapy comparators, based on 5 RCTs and 125 participants.

“The mean score for the cognitive behavioural therapy group was 1.02 standard deviations lower (95% CI 1.99 to 0.04 lower) than for non-therapy comparators.125 (5 RCTs)”

In the same sensitivity analysis excluding Daleo 1999, at short-term follow-up the standardized mean difference was -0.84 (95% CI -1.22 to -0.45) across 15 studies with 487 participants.

“14.1.2Short-term 15 487 Std.Mean Difference (IV, Random, 95% CI) -0.84 [-1.22,-0.45]”

Psychological therapy reduced depressive symptomatology at end of intervention by 1.47 standard deviations (95% CI 2.10 to 0.84 lower) compared with treatment as usual, based on 12 RCTs and 430 participants.

“The mean score for the psychological therapy group was 1.47 standard deviations lower (95% CI 2.10 to 0.84 lower) than for the treatment as usual group 430 (12 RCTs)”

Many older people prefer non-pharmacological treatment approaches for depression.

“older people typically report a preference for psychological treatments for depression over antidepressant medications (Gum 2006;Luck-Sikorski 2017).”

• cites Gum 2006 sourceNot checked — no full text of the source
• cites Luck-Sikorski 2017 sourceNot checked — no full text of the source

Most residents in long-term care who receive treatment for depression are prescribed antidepressant medications.

“Currently, most residents treated for depression are prescribed antidepressant medications, despite the potential availability of psychological therapies that are suitable for older people and a preference amongst many older people for non-pharmacological treatment approaches.”

• cites George 2007 sourceNot checked — no full text of the source

A meta-analysis found significantly higher rates of depression among older people in long-term care facilities than among those in the community.

“One meta-analysis comparing rates of depression amongst older people living in LTC facilities versus those living in the community reported significantly higher rates in people in LTC facilities (Xiu-Ying 2012).”

• cites Xiu-Ying 2012 sourceNot checked — no full text of the source

Established risk factors for late-life depression, including medical comorbidity, disability and functional decline, and cognitive impairment, are ubiquitous in long-term care settings.

“Established risk factors for late-life depression, including medical comorbidity, disability and functional decline, and cognitive impairment are ubiquitous in LTC settings (Davison 2012).”

• cites Davison 2012 sourceNot checked — no full text of the source

In a review of over one million newly admitted long-term care residents in the USA, less than 1% received any amount or type of psychological therapy.

“To illustrate, in one review of over one million newly admitted LTC residents in the USA, less than 1% received any amount or type of psychological therapy (Ulbricht 2017).”

• cites Ulbricht 2017 sourceNot checked — no full text of the source

Generalizing findings from community studies to the long-term care population is problematic.

“Generalizing findings from community studies to the LTC population is therefore problematic and a review of research specifically targeting LTC residents is required.”

When a trial used more than one depression symptom instrument, clinician-rated scales were prioritized over informant-rated and self-rated scales because of concerns about the reliability of self-reported symptoms in the long-term care population.

“If a trial used more than one instrument to assess depression symptoms, then we prioritized these in the following order: clinician-rated scale, informant-rated scale, and self-rated scale (given concerns regarding the reliability of self-reported symptoms of depression in the LTC population; Davison 2009).”

• cites Davison 2009 sourceNot checked — no full text of the source

Surveys indicate that many long-term care residents continue to have high symptom counts despite taking antidepressants.

“Surveys have indicated that many LTC residents continue to have high symptom counts, despite taking antidepressants (Davison 2007;O'Connor 2010).”

• cites Davison 2007 sourceNot checked — no full text of the source
• cites O'Connor 2010 sourceNot checked — no full text of the source

There is a lack of clear evidence on the effectiveness of psychological therapies for depression among older people living in long-term care facilities.

“Despite this preference, there continues to be a lack of clear evidence of the effectiveness of psychological therapies for depression amongst older people living in LTC facilities.”

Previous reviews, both narrative and systematic, have indicated that psychological therapies may have an impact on depressive symptomatology in long-term care residents.

“A number of previous reviews, both narrative and systematic, have indicated that psychological therapies may have an impact on depressive symptomatology in LTC residents (Bharucha 2006;Cody 2013;Hyer 2005;Powers 2008;Simning 2017;Snowden 2003).”

• cites Bharucha 2006 sourceNot checked — no full text of the source
• cites Cody 2013 sourceNot checked — no full text of the source
• cites Hyer 2005 sourceNot checked — no full text of the source

Older people typically report a preference for psychological treatments for depression over antidepressant medications.

“As is the case in broader adult samples (Hanson 2016), older people typically report a preference for psychological treatments for depression over antidepressant medications (Gum 2006;Luck-Sikorski 2017).”

• cites Gum 2006 sourceNot checked — no full text of the source
• cites Luck-Sikorski 2017 sourceNot checked — no full text of the source

Cody and Drysdale reported a medium effect size of psychotherapy on symptoms of depression in long-term care from a meta-analysis of studies.

“However, Cody and Drysdale reported a medium effect size of psychotherapy on symptoms of depression in LTC from a meta-analysis of studies (Cody 2013).”

• cites Cody 2013 sourceNot checked — no full text of the source

All 19 included studies had at least one risk-of-bias domain rated unclear because information was missing from the report or protocol/registration.

“All 19 included studies had at least one or more risk of bias domains rated as 'unclear' because information was missing from the study report or trial protocol/registration, or both (see Figure 2; Figure 3).”

Based on eligibility criteria and sample descriptions, the review concluded that in most studies the majority of participants had mild-to-moderate depressive symptoms.

“Based on information from eligibility criteria and descriptions of study samples, we concluded that, in most studies, the majority of participants had mild-to-moderate levels of depressive symptoms.”

Baseline depression severity subgroup analyses used only two categories, high versus low baseline symptoms, based on a standardized 100-point scale dichotomized at ±54.8 points.

“Thus, subgroup analyses for baseline depression severity were conducted with two categories only: high versus low baseline symptoms of depression, based on scores on a standardized 100-point scale (dichotomized at ± 54.8 points).”

Most other studies failed to provide sufficient information about random sequence generation or allocation concealment to allow a rating.

“Most other studies failed to provide sufficient information about either the random sequence generation process or allocation concealment to allow a rating to be made.”

Risk of reporting bias could not be ascertained in the remaining 16 studies due to absence of a published protocol or trial registration detailing outcomes.

“Risk of reporting bias could not be ascertained in the remaining 16 studies, due to the absence of an available published protocol paper or trial registration detailing outcomes.”

Thirteen studies were at high risk of other bias.

“Thirteen studies were at a high risk of other bias.”

The degree to which the studies represent the proportion of residents with different levels of depression symptoms in LTC settings is unclear.

“While there was variation in symptom severity across studies, the degree to which the studies represent the proportion of residents with different levels of symptoms of depression in LTC settings is unclear.”

Sensitivity analyses omitting Daleo 1999, Luo 2020, and Tsai 2008 had minimal impact on the short-term depressive symptomatology conclusion.

“Sensitivity analyses (omitting Daleo 1999, Luo 2020, and Tsai 2008, due to concerns about study methods) had minimal impact on this overall conclusion (Analysis 12.1; Analysis 13.1; Analysis 14.1).”

• cites Daleo 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Luo 2020 sourceNot checked — no full text of the source
• cites Tsai 2008 sourceNot checked — no full text of the source

One study was at high risk of reporting bias because the primary outcome from a published protocol was not reported in the efficacy paper.

“One study was at high risk of reporting bias, with the primary outcome from a prospectively published protocol paper not reported in the efficacy paper (Dozeman 2011).”

• cites Dozeman 2011 sourceNot checked — no full text of the source

Omitting Luo 2020 from the end-of-intervention quality-of-life analysis had minimal impact on the conclusion (SMD 0.25, 95% CI -0.17 to 0.68).

“Omitting Luo 2020 due to concerns with the tool used for participant eligibility had minimal impact on this conclusion (SMD 0.25, 95% CI -0.17 to 0.68; Analysis 12.2).”

• cites Luo 2020 sourceNot checked — no full text of the source

Sensitivity analyses omitting Daleo 1999, Luo 2020, and Tsai 2008 had minimal impact on the end-of-intervention depressive symptomatology conclusion.

“Sensitivity analyses (omitting Daleo 1999, Luo 2020, and Tsai 2008, due to concerns about study methods) had minimal impact on this overall conclusion (Analysis 12.1; Analysis 13.1; Analysis 14.1).”

• cites Daleo 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Luo 2020 sourceNot checked — no full text of the source
• cites Tsai 2008 sourceNot checked — no full text of the source

When Luo 2020 was omitted from the short-term quality-of-life analysis, the effect estimate became small and crossed the line of no effect (SMD 0.35, 95% CI -0.02 to 0.72).

“When Luo 2020 was omitted due to concerns with the tool used for participant eligibility, the effect estimate became small and crossed the line of no effect (SMD 0.35, 95% CI -0.02 to 0.72; Analysis 12.2).”

• cites Luo 2020 sourceNot checked — no full text of the source

Information about LTC facility characteristics was not typically provided in sufficient detail to determine how well the studies reflected LTC facilities more broadly.

“However, this information was not typically provided in sufficient detail to determine the degree to which the studies reflected LTC facilities more broadly.”

The evidence in this review may not apply to older people with more than mild-to-moderate cognitive impairment or dementia.

“Thus, the evidence in this review may not apply to older people with more than mild-to-moderate cognitive impairment or dementia.”

A funnel plot revealed no indications of publication bias for depressive symptomatology.

“A funnel plot revealed no indications of publication bias for depressive symptomatology (Figure 4).”

Confidence in all estimates was downgraded two levels due to risk of bias.

“Subsequently, our confidence in all estimates were downgraded two levels due to risk of bias.”

Older adults living in LTC typically have more complex and comorbid health and mental health profiles than those living in the community.

“Older adults living in LTC typically have more complex and comorbid health and mental health profiles than those living in the community (Amare 2020)”

• cites Amare 2020 sourceNot checked — no full text of the source

The Almeida 2022 trial found that behavioural activation delivered by trained LTC facility staff had no effect on symptoms of depression.

“and found the intervention had no effect on symptoms of depression (Almeida 2022).”

• cites Almeida 2022Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

One previous review focused on interventions designed to improve well-being rather than depression specifically.

“one review focused on interventions designed to improve well-being, rather than depression specifically”

• cites Bharucha 2006 sourceNot checked — no full text of the source

Difficulty in accessing missing data resulted in several studies with unclear risk of bias ratings for some domains.

“The difficulty in accessing missing data resulted in several studies with unclear risk of bias ratings for some domains.”

Owing to considerable statistical heterogeneity, the majority of estimates were downgraded due to inconsistency.

“Owing to considerable amounts of statistical heterogeneity, the majority of estimates were downgraded due to inconsistency.”

Some estimates included fewer than five trials and should be interpreted with caution as per prior recommendations.

“Some of our estimates included fewer than five trials and thus should be interpreted with caution as per prior recommendations (Saueressig 2021).”

• cites Saueressig 2021 sourceNot checked — no full text of the source

The meta-analysis of treatment non-acceptability was based on only 22 events.

“The results of the analysis were based on a total of only 22 events”

Operationalizing treatment non-acceptability as participant dropout was problematic in this review because many authors did not provide enough detail to classify dropouts as intervention-related or unrelated.

“The operationalization of treatment nonacceptability as participant dropout was problematic in this review”

Older adults in LTC may not adhere as easily to psychological treatments developed for community-dwelling older adults.

“and hence many not adhere as easily to psychological treatments developed for community-dwelling older adults.”

The evidence about the effects of psychological therapies on all outcomes is very uncertain.

“However, it is important to note that the evidence about the effects of psychological therapies on all outcomes is very uncertain.”

The risk of selective outcome reporting bias was unclear for most studies due to absence of available prospectively published information.

“However, the risk of this bias due to selective outcome reporting was unclear for most studies due to the absence of available prospectively published information.”

Prior reviews failed to restrict eligibility criteria to studies of participants with depressive symptoms or disorders, or both.

“the failure to restrict eligibility criteria to studies of participants with depressive symptoms or disorders, or both.”

The difference in attrition rates between psychological therapy and treatment as usual may reflect more opportunities for therapy participants to communicate a desire to withdraw rather than poor acceptance of the intervention.

“Thus, the difference in attrition rates may simply reflect more opportunities to communicate a desire to withdraw participation, rather than indicating poor acceptance of the intervention itself.”

• AssumptionRests on an assumption: Participants receiving psychological therapies typically have more contacts with the research or treatment team than treatment-as-usual participants, who are contacted only for data collection.

Drawing conclusions from prior reviews about the effectiveness of psychological therapies for depression is difficult.

“Therefore, drawing conclusions from prior reviews regarding the effectiveness of psychological therapies for depression is difficult.”

These results suggest that the effects of psychological therapy may not be as clear in residents with high cognitive impairment and dementia.

“These results suggest that the effects of psychological therapy may not be as clear in residents with high cognitive impairment and dementia”

There are currently no clinical guidelines that stipulate which psychological interventions are effective for depression in LTC populations.

“there are no clinical guidelines that stipulate which psychological interventions are effective for depression in such populations.”

The certainty of evidence for all comparisons in this review was very low.

“However, it is important to note the very low certainty of evidence for all comparisons conducted in our review.”

The findings of the effectiveness of psychological therapies may not apply equally to LTC residents with more complex presentations.

“Thus, the findings of the effectiveness of psychological therapies may not apply equally to LTC residents with more complex presentations.”

The evidence about the effect of treatment characteristics is very uncertain.

“While it is important to note that the evidence is very uncertain about the effect of these treatment characteristics”

In the Daleo 1999 trial, no participants were excluded from analyses.

“no participants were excluded from analyses.”

In the Abraham 1992 trial, 76 participants were randomised but only 42 completed the intervention and the follow-up data collection.

“76 participants randomized but only 42 completed the intervention and the follow-up data collection.”

For Daleo 1999, baseline mean scores for cognition, orientation, and depression did not differ between groups.

“Pretreatment: no differences in mean scores on cognition, orientation, or depression at baseline.”

• cites Daleo 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

In the Daleo 1999 trial, it is unlikely that participants and personnel were blinded.

“It is unlikely that participants and personnel were blinded.”

In the Abraham 1992 trial, the small number of clusters (7 clusters allocated to 3 conditions) may have introduced bias.

“A few clusters (7 clusters allocated to 3 conditions) may have introduced bias.”

Meeks 2008 was judged at low risk of selection bias for allocation concealment.

“Allocation concealment (selection bias) Low risk Judgement comment: author stated a biostatistician generated a list of numbers that determined which group each participant went in (personal correspondence).Only the biostatistician and lead author (also pretreatment assessor and therapist) knew which participant was assigned to which group.”

• cites Meeks 2008 sourceNot checked — no full text of the source

The effect of psychological therapies on secondary outcomes for older people without depression at baseline may differ from the current results.

“The effect of psychological therapies evaluated in these trials on these outcomes for older people without depression at baseline (e.g.people with anxiety but not depression) may differ from the current results.”

In the Abraham 1992 trial, there were no group differences at pretreatment on depression, hopelessness, dissatisfaction with life, cognitive functioning, or demographic data.

“Pretreatment: no group differences on depression, hopelessness, dissatisfaction with life, cognitive functioning, or demographic data”

For Meeks 2008, only the biostatistician and lead author knew which participant was assigned to which group.

“Only the biostatistician and lead author (also pretreatment assessor and therapist) knew which participant was assigned to which group.”

• cites Meeks 2008 sourceNot checked — no full text of the source

This finding questions whether at least part of the effects associated with psychological therapies could be attributed to companionship, attention, and social engagement offered to residents in treatment rather than treatment-specific mechanisms.

“this finding is provocative as it questions whether at least part of the effects associated with psychological therapies could be attributed to the companionship, attention, and social engagement that is offered to residents in treatment, rather than the purported mechanisms of change specific to the treatment model.”

In the Daleo 1999 trial, no formal independent ratings of treatment fidelity were conducted using the session transcripts.

“there was no formal, independent ratings of treatment fidelity conducted using these transcripts.”

There were major limitations in the reporting of the methods used in the Hamzehzadeh 2018 trial.

“Major limitations in the reporting of the methods used in the trial.”

• cites Hamzehzadeh 2018Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Research on psychological therapies is bedevilled with small sample sizes, particularly when strict inclusion criteria are used.

“In general, research on psychological therapies is bedevilled with small sample sizes, particularly when strict inclusion criteria are used.”

For Dhooper 1993, baseline mean scores for cognition, orientation, and depression did not differ between groups.

“Pretreatment: no differences in mean scores on cognition, orientation, or depression at baseline.”

• cites Dhooper 1993Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

For Meeks 2008, participants and personnel were likely not blinded due to the nature of the intervention.

“participants and personnel were likely not blinded due to nature of intervention”

• cites Meeks 2008 sourceNot checked — no full text of the source

Studies of psychological therapies are likely to be at risk of bias due to difficulty identifying a true placebo condition and difficulty blinding participants to the condition.

“studies of psychological therapies are likely to be at risk of bias due to the difficulty of identifying a true placebo condition and the difficulty of blinding participants to the condition.”

The capacity to draw firm conclusions on the efficacy of psychological therapies in LTC settings is hindered by several methodological limitations.

“our capacity to draw firm conclusions on the efficacy of psychological therapies in LTC settings is hindered by several methodological limitations”

In the Daleo 1999 trial, outcome measures were administered by personnel who were unlikely to be blinded to condition.

“outcome measures were administered within the group therapy sessions (or at approximately the same time for those in the control condition) by personnel who were unlikely blinded to condition.”

The pooled estimates indicated substantial to considerable statistical heterogeneity.

“The estimate indicated substantial to considerable statistical heterogeneity.Downgraded one level for inconsistency.”

Meeks 2015 had high risk of attrition bias because roughly half of the PST group finished treatment and follow-up interviews, compared with the social contact group.

“High risk Quote:”

All included RCTs had a high risk of bias in at least one risk-of-bias category.

“All RCTs had a high risk of bias for at least one category.Downgraded two levels for risk of bias.”

The certainty of the evidence for the effect of psychological therapy on depressive symptomatology in older adults in long-term care was very low.

“⊕⊝⊝⊝ Very low a,b,c”

• Also stated in: § Cochrane Database of Systematic Reviews · plain

  “⊕⊝⊝⊝ Very low a,b”

Meeks 2015 had high risk of performance bias because participants and personnel were not blinded.

“High risk Quote (personal correspondence from author):”