Non-pharmacological interventions for adults with intellectual disabilities and depression: a systematic review

Background

Since the 1980s, there is awareness that psychiatric disorders can co-occur with intellectual disabilities (IDs) (Sovner & Hurley 1983; Marston et al. 1997; Cooper et al. 2007; Hurley 2008; Hermans et al. 2013). Nowadays, we know that depression is a common psychiatric disorder in adults with ID. The prevalence range of depression in the ID population varies from 2.2% to 7.6% (Deb et al. 2001; Smiley 2005; Cooper et al. 2007; Hermans et al. 2013). The prevalence is higher compared with that in the general population, despite the fact that depressive symptoms can be difficult to recognise in this population (Marston et al. 1997; Hurley 2008; Hermans et al. 2013). Depression is mainly characterised by sadness and loss of interest or pleasure (American Psychiatric Association 2013). Depression has a major impact on the quality of life (QoL) and leads to cognitive, social and physical problems (Coryell et al. 1993; Hays et al. 1995; Bijl & Ravelli 2000; Sprangers et al. 2000; Beekman et al. 2002; Alonso et al. 2004; Judd et al. 2008; Kober 2010; Horovitz et al. 2014; Rand & Malley 2017). Horovitz et al. (2014) found a significantly higher QoL in a group of adults with ID without Axis I diagnosis compared with adults with ID in the anxiety/mood disorder diagnosis group. Furthermore, adults with ID with higher levels of anxiety/depression are more likely to report a poor QoL (Rand & Malley 2017).

In the general population, antidepressants are frequently prescribed to treat depressive symptoms. Psychoactive medications, including antidepressants, are regularly prescribed in adults with ID, primarily to reduce challenging behaviour (Lott et al. 2004; Deb et al. 2009; Matson & Mahan 2010; Sheehan et al. 2015). There is some evidence that antidepressant medication can decrease depressive symptoms in adults with ID (Masi et al. 1997; Verhoeven et al. 2001; Janowsky et al. 2005). For example, in a group of 20 participants, Verhoeven et al. (2001) found Citalopram effective in decreasing depressive symptoms. Many adults with ID use more than one medication, and polypharmacy is common in adults with ID (Haider et al. 2014; Häβler et al. 2015; Bowring et al. 2017). Negative side effects (short term and long term) can appear when psychoactive medications are used in adults with ID (Deb et al. 2009; Mahan et al. 2010; Matson & Mahan 2010; Eady et al. 2015; Häβler et al. 2015). For example, physical complaints, neurological damage, movement side effects and physiological problems are mentioned (de Leon et al. 2009; Matson & Mahan 2010; Sheehan et al. 2017). Besides, adults with ID seem to be more amenable to develop side effects compared with the general population when psychoactive medications are used (Arnold 1993; Matson & Mahan 2010; Sheehan et al. 2017). Moreover, it can take a while for a psychoactive medication to work in the right daily dosage, and adults with ID may experience even more side effects when more than one psychotropic medication is used (Matson & Mahan 2010). Therefore, there is a need for evidence-based non-pharmacological treatments for depression in adults with ID.

In the general population, a wide range of systematic reviews on non-pharmacological interventions for depression have been published over the last couple of years (Merry et al. 2011; Cox et al. 2012; Catalan-Matamoros et al. 2016; Kvam et al. 2016; Lee et al. 2016; Stubbs et al. 2016). Unfortunately, the conclusions of these reviews (both positive and negative) cannot be generalised to the ID population because a large part of the non-pharmacological interventions for depression of the general population are not suitable for adults with ID. Next to cognitive limitations, adults with ID frequently have verbal limitations. Psychological interventions, for example, cognitive behavioural therapy (CBT), are too difficult for adults with a more severe ID and for those with verbal limitations. Furthermore, a large part of people with ID have physical limitations as well (Cooper et al. 2015). Consequently, exercise interventions can be too complicated to perform or physically impossible.

A few systematic reviews concerning non-pharmacological interventions for depression for adults with ID have been published. Some studies are investigating interventions for a part of the ID population. For example, Osugo & Cooper (2016) focused on interventions for adults with mild ID and mental ill health. They concluded that there was some evidence for group CBT (although larger trials are needed) but that in general the evidence-based interventions for people with mild ID and mental problems were limited. Koslowski et al. (2016) investigated in their systematic review and meta-analyses the effectiveness of interventions on mental health problems in adults with mild to moderate ID. They found no strong evidence for interventions aimed at improving mental health problems, including depression, and found a non-significant moderate effect size [d = 0.49, 95% confidence interval (CI)-0.05 to 1.03; P = 0.08] for depression interventions (psychotherapy only). The focus of other reviews in this research area is on specific treatments only. For instance, Vereenooghe & Langdon (2013) did a meta-analysis on psychological therapies for people with ID and mental health problems and found an overall moderate between-group effect size (g = 0.682, 95% CI 0.379 to 0.985). Furthermore, a subgroup meta-analysis indicated that individually psychological therapy (g = 0.778, 95% CI 0.110 to 1.445) was more effective than group-based psychological therapy (g = 0.558, 95% CI 0.212 to 0.903) and psychological interventions for depression had a moderate effect size (g = 0.742, 95% CI −0.116 to 1.599).

Depression can occur in all levels of ID. Hence, an overview of evidence-based non-pharmacological interventions for depression for the whole ID population is needed, as the severe and profound ID population got no or little attention in previous reviews. Therefore, the aim of this review is to evaluate non-pharmacological treatments for adults with ID (all levels) and depression. Our first research question is: Which non-pharmacological interventions have been studied in adults with ID and depression? Our second research question is: What were the results of these non-pharmacological interventions?

Method

We have used the Preferred Reporting Items for Systematic Reviews and Meta-analyses checklist to perform this study (Moher et al. 2009). The study is registered in the PROSPERO database (PROSPERO 2016: CRD42016051524).

Study selection

Inclusion criteria

Inclusion criteria have been clearly defined before the start of the study. All papers published in English before 3 October 2016, mentioning non-pharmacological interventions for treating depression (of any type) or depressive symptoms in adults (aged ≥16 years) with ID (IQ ≤ 70), have been selected. Outcome measures on depressive symptoms must be mentioned in the paper to be included. Because the aim of this study is to find all non-pharmacological interventions for adults with ID and depression, no exclusion on type of study design has been made. Therefore, all different kinds of study designs, from case study to randomised controlled trial (RCT), were included. The choice to include all study designs contributes to the presentation of the current state of evidence for the different types of non-pharmacological interventions, and possible gaps of knowledge can be exposed.

When a combined population of children and adults was described in a paper, the results of the adult population must be separately presented to be included. The same applies to level of ID: when the study population also contained adults with an IQ of more than 70, the population with IQ ≤ 70 must be separately presented to be included. In some papers, combined interventions (pharmacological and non-pharmacological) are described. Papers are only included when the results of both of these interventions are studied separately. (Systematic) reviews of non-pharmacological interventions, as well as narrative papers without results and congress abstracts, were not included in the current review.

Process of the study selection

After identifying papers through the electronic search, duplicates have been removed. Then, title and abstract of the papers have been scanned by two reviewers (PH and HH) independently, according to the predefined inclusion criteria mentioned previously. After the selection of all relevant records on title and abstract, the databases of the two reviewers were merged to see if there was any disagreement about included papers. Any disagreement was solved by discussion in a consensus meeting. Reference lists of recent systematic reviews were studied by both reviewers (PH and HH) for relevant studies. Hereafter, full texts of all remaining papers have been assessed for eligibility by two reviewers (PH and HH) independently. Hand searches in reference lists of included papers were done to search for relevant papers to include. The potential relevant papers of both hand searches were discussed by both reviewers (PH and HH) before including the papers in the final database. After this step, the final database was created. See Fig. 1 for the flow chart of the selection of studies. Data extraction of the included studies has been performed by one reviewer (PH) and checked by the second reviewer (HH). See Tables 1a and 1b for details of the study characteristics.

Table 1a. Characteristics of included studies

Author (year)	Study design	Sample size	Participants (age, gender and level of ID)	Intervention	Measured depression outcome	Results on depressive symptoms
McCabe et al. (2006)	RCT	n = 34: 19 EG, 15 CG. (The CG group later also took part of the EG). Total EG: 34	Mean age: 34.1 (EG), 39.8 (CG)	Group CBT, 5 weeks, 2 h sessions. Control group: waiting list.	BDI-II, ATQ-R	Intervention significantly decreased depressive symptoms and negative automatic thoughts.
			22 M (16 EG + 6 CG)/27 F (18 EG + 9 CG)			Follow-up (3 months, n = 18): impact of the intervention on depressive symptoms sustained over time; no further improvement found.
			Mild–moderate ID
McGillivray et al. (2008)	RCT	n = 47: 20 TG, 27 CG. Control group received treatment as well after follow-up (TG 2).	Mean age: 38.4 (TG), 31.2 (CG)	Staff administered group CBT programme, 12 weeks, 2 h sessions.	BDI-II, ATQ-R	Significant decrease in depressive symptoms and negative automatic thoughts in the CBT group.
						Follow-up (3 months): positive effects maintained.
			32 M (TG 13, CG 19)/15 F (TG 7, CG 8)
			Mild ID
Ghafoori et al. (2010)	Pilot study, one group (pre, post and follow-up)	n = 8	Mean age: 20.0	Cognitive behavioural group therapy programme, 9 weeks, 1.5 h sessions.	SCL-90-R	Significant improvements in six primary dimensions of the SCL-90-R, including ‘depression’.
			2 M/6 F
						Follow-up (4 months): no significant treatment effects maintained at follow-up.
			Mild–moderate ID
Hassiotis et al. (2013)	RCT	n = 32: 16 TG, 16 CG.	Mean age: 33.7 (M-iCBT), 38.3 (TAU)	M-iCBT, 16 weeks, 1 h sessions. Control: treatment as usual.	BDI-Y	No significant treatment effect.
						Follow-up (6 months): no significant effects on depression.
			12 M/20 F
			Mild–moderate ID
McGillivray & Kershaw (2013)	Controlled trial (pre, post and follow-up)	n = 82: 32 G1, 24 G2, 26 G3.	Mean age overall = 37*	Staff administered group CBT programme with (1) a staff-initiated referral to a GP, (2) staff administered group CBT programme only, (3) referral to GP only.	BDI-II, ATQ-R	CB-only group and CB with GP referral group: greatest reduction in depression symptoms directly after the programme. Significant reduction in frequency of negative automatic only in the CB-only group.
			47 M/35 F*
			Mild ID
						Follow-up (8 months): CB strategies (particularly CB with referral to GP) appeared effective in reducing depressive symptoms and negative automatic thoughts.
McGillivray & Kershaw (2015)	Controlled trial (pre, post and follow-up)	n = 70: 23 G1, 23 G2, 24 G3.	Mean age overall: 36.0*	Cognitive and behavioural strategies (group 1), cognitive focused strategies (group 2), behavioural focused strategies (group 3).	BDI-II, ATQ-R.	The mean depression scores decreased in all three intervention groups after the intervention. No significant difference between groups.
			42 M/28 F*
			Mild ID			Follow-up (6 months): group 1, all individuals indicated improvement; group 2, 67% maintained improvement; group 3, 47% maintained improvement.
Lindsay et al. (2015)	Controlled trial (pre, post and follow-up)	n = 24: 12 TG, 12 CG	Mean age: 28.9 (TG), 33.1 (CG)	Experimental group: individual CBT.	BSI, GDS	No significant effect on BSI depression score. Statistically significant reductions in self-reported depression (GDS) and carer-reported depression (GDS).
				Control group: waitlist (TAU).
			12 M (6 TG + 6 CG)/12 F (6 TG + 6 CG)
						Follow-up (3 to 6 months follow-up; treatment group only): significant decrease on the GDS maintained.
			Mild ID
Jahoda et al. (2015)	Feasibility study (one group: pre, post and follow-up)	n = 21	Mean age = 42.2	Behavioural activation, 10–12 sessions.	GDS-LD#	Significant reduction in self-report depressive symptoms. Positive change for informant reports on depressive symptoms.
			12 M/9 F		IDDS#
			Mild–moderate–severe ID
						Follow-up (3 months): reduction depressive symptoms maintained.
Heller et al. (2004)	RCT	n = 53: 32 TG, 21 CG	Mean age: 39.41 (TG), 40.22 (CG)	TG: 12 weeks, 3 days per week health promotion programme, 2 h a day (1 h exercise + 1 h health education). CG: no training.	CDI	Participants in the intervention group were less depressed than those in the control group (marginally significant).
			24 M/29 F			No follow-up.
			Mild–moderate ID
Carraro & Gobbi (2014)	RCT	n = 27: 14 EG, 13 CG	Mean age overall: 40.1*	Short-term group-based exercise programme, 12 weeks, two times a week, 1 h sessions.	Zung Self-rating Depression Scale	Significant reduction of depressive symptoms in the exercise group compared with the control group.
			16 M/11 F*	Control group: painting activities.		No follow-up.
			Mild–moderate ID

• * Not specified per group.
• # Measurements especially developed for people with ID.
• ATQ-R, Automatic Thoughts Questionnaire–Revised; BDI, Beck Depression Inventory; BDI-Y, Beck Depression Inventory Youth; BSI, Brief Symptom Inventory; CBT, cognitive behaviour therapy; CDI, Children's Depression Inventory; F, female; GDS-LD, Glasgow Depression Scale for people with learning disabilities; ID, intellectual disability; IDDS, Intellectual Disabilities Depression Scale; M, male; M-iCBT, manualised individual cognitive behavioural therapy; RCT, randomised controlled trial; SCL-90, Symptom Checklist-90–Revised; TAU, treatment as usual.

Table 1b. Characteristics of included case studies

Author (year)	Study design	Sample size	Participants (age, gender and level of ID)	Intervention	Measured depression outcome	Results on depressive symptoms
Lindsay et al. (1993)	Case studies	n = 2	Mean age: 24	Cognitive therapy (duration/frequency of therapy is unclear).	Zung Depression Scale	Both subjects improved on the Zung Depression Scale. Suicidal thoughts decreased in one participant.
			1 M/1 F
			Mild ID
						Follow-up (6 weeks): reductions on depression scores maintained.
Matson (1982)	Case studies	n = 4	Mean age: 33.5	Behavioural therapy, 10–35 sessions.	Self-rating Depression Scale BDI	Significant decrease of depressive symptoms on both scales.
			3 F/1 M			Follow-up (4–6 months): treatment effect maintained.
			Mild–moderate ID
Stuart et al. (2014)	Case study	n = 1	Age: 40.0	Seven sessions of therapy over 3 months (simplified behavioural activation and daily audio-bases progressive muscular relaxation).	GDS-LD#	Decrease in depressive symptoms on the GDS-LD but still above the cut-off point.
			1 F
			Mild ID
						No follow-up.
Anderson & Kazantzis (2008)	Case studies	n = 3	Age range: 19–52 (no mean published)	Social problem-solving skills training, 15 individual sessions.	Adapted Zung Depression Scale#	Pretreatment—follow-up: depression showed a 40% change in one case and 31% change in another case.
			2 M/1 F			Follow-up (4 weeks): improvement maintained.
			Mild ID
Altabet et al. (2002)	Case studies	n = 3	Mean age: 57	LT sessions, 30 min (between 08:00 and 10:00 h), 10.000 lux, 5 days per week for 12 weeks.	DASH# (depression sub-scale)	Positive effects on mood and sleep patterns.
						Follow-up (3 weeks): treatment gains maintained; (8 weeks): increased depressive symptoms.
			1 M/2 F		ABC# (lethargy and irritability sub-scale)
			Profound ID
					Mood chart

• # Measurements especially developed for people with ID.
• ABC, Aberrant Behavior Checklist; BDI, Beck Depression Inventory; DASH, Diagnostic Assessment for the Severely Handicapped; F, female; GDS-LD, Glasgow Depression Scale for people with learning disabilities; ID, intellectual disability; LT, light therapy; M, male.

Results

The electronic search identified 6023 papers. After removing duplications, 4267 papers have been included in the initial database. These 4267 papers have been screened on title and abstract by two reviewers independently (PH and HH). The reference lists of recent systematic reviews (Flynn 2012; Sturmey 2012; Chen 2013; Hwang & Kearney 2013; Matson 2013; Vereenooghe & Langdon 2013; Jennings & Hewitt 2015; Koslowski et al. 2016; Maber-Aleksandrowicz et al. 2016; Osugo & Cooper 2016; Unwin et al. 2016) were studied for relevant papers by these two reviewers as well. One relevant new paper was found. Both reviewers read 113 full-text articles and screened these papers on the inclusion criteria. The main exclusion reason was the absence of study results (e.g. narrative articles or study results on depressive symptoms were not published) (Fig. 1). A total of 15 papers have been included after full-text screening. Hand searches in reference lists of these included papers (also done by reviewers PH and HH) did not reveal other relevant papers to include in the final database. Therefore, the final database contained 15 papers.

Discussion

The current systematic review contains 15 studies evaluating the effect of a total of five different non-pharmacological interventions to decrease depressive symptoms in adults with ID. These five different types of non-pharmacological interventions are similar to those found by Holvast et al. (2017) in the elderly population with depression in primary care. Based on our study, we can conclude that CBT is an effective non-pharmacological intervention to decrease depressive symptoms in adults with mild or moderate ID (McCabe et al. 2006; McGillivray et al. 2008; McGillivray & Kershaw 2013; Lindsay et al. 2015; McGillivray & Kershaw 2015). However, these results must be interpreted with caution because of the methodological problems of some studies as seen in the quality assessment. In the general population, CBT is a widely used effective treatment for depression (Butler et al. 2006). CBT can be used in the mild to moderate ID population to decrease depressive symptoms as well, although more RCTs are needed to establish its usefulness in clinical practice. The main part of the included studies in this paper includes interventions for people with mild or moderate ID. In only two studies, people with severe or profound ID have been included, even though it is known that they can suffer from depression as well (Cooper et al. 2007; Hermans et al. 2013). In general, conducting intervention studies in the ID population is challenging. For instance, ethical dilemmas, specific living conditions of people with ID, dependence on professional staff, a difficult informed consent procedure, the burden of the measurements and challenging behaviour are issues researchers are confronted with when conducting intervention studies in this population (Oliver et al. 2002; Hamers et al. 2017). This might be an explanation why intervention studies with adults with severe ID are even more scarce. In the 1980s, Matson already published about behavioural therapy for adults with ID and depression. Unfortunately, none of the studies on behavioural therapy included in this review used control groups (Matson 1982; Stuart et al. 2014; Jahoda et al. 2015). Therefore, we cannot conclude with certainty that behavioural therapy is responsible for the decrease in depressive symptoms. However, as positive results are published in these papers, it seems promising.

In the two studies (RCTs) investigating exercise as a non-pharmacological treatment to decrease depressive symptoms, intervention groups as well as control groups have been used (Heller et al. 2004; Carraro & Gobbi 2014). Both studies reported positive results in decreasing depressive symptoms. In the study of Heller et al. (2004), participants got health education and exercise, and in the study of Carraro & Gobbi (2014), participants got exercise only, which makes it hard to compare these two exercise studies. Despite this fact, exercise interventions seem promising interventions to decrease depression in adults with ID without psychical limitations and should be further studied.

The study of Anderson & Kazantzis (2008) was the only study in this review focusing on social problem-solving skills programme. Unfortunately, this was a multiple single-case study with only three participants, which makes it hard to draw any conclusions about the effect of this intervention on decreasing depressive symptoms (Anderson & Kazantzis 2008). BLT in adults with profound ID is studied by Altabet et al. 2002. Positive effects were seen on mood, but no conclusions can be drawn because of the small sample size and no control group (Altabet et al. 2002), so more research is needed. The recent published pilot study with promising results of Hermans et al. (2017) is the first step towards more insight into the effect of BLT as a treatment for depression in adults with ID.

A large part of the 15 included studies of this review are case reports or studies with a small sample size. Some of the reviewed studies also have methodological problems, for example, no control group or no follow-up. Despite the fact that a large number of adults with ID suffer from depressive symptoms, limited well-conducted studies are carried out to evaluate the effect of non-pharmacological interventions to decrease depressive symptoms.

The strength of the current systematic review is that the whole study selection (from title/abstract to full text) and the quality assessment are done by two reviewers independently. Another strength of this study is that there was no restriction on publication year, so all relevant studies published before the start of this study are screened. Besides, the study protocol of this systematic review was registered at the start of the study, which makes the current systematic review transparent. The methodological quality of the current study was assessed with the AMSTAR checklist (Shea et al. 2007) by an independent researcher and received an AMSTAR score of 8 (out of 11).

A limitation of the current systematic review is the small number of papers that could be included because of the inclusion criteria. Many studies were excluded because of lack of data on study methods and outcome measures. For example, in several papers, the IQ level of participants was not mentioned. Further, depressive outcome measures were not reported in quite a few papers, for example, in the case series of Tsiouris (2007). Because of the small number of papers included in our review, which are spread over five different kinds of interventions, a meta-analysis on the effect of the non-pharmacological treatments was unfortunately not possible. We did not use the ‘grey literature’ in this systematic review. So papers could have been missed. Third, this review is limited by only including papers published in English. Fourth, the included studies are very different in study design, which makes them hard to compare with each other.

The used tool to evaluate the quality of the studies (the Cochrane Risk of Bias Tool) is actually designed for (randomised) controlled trials. As for a more pragmatic approach, we used it to evaluate all the study designs of the included papers. Eventually, the use of this specific tool emphasised the poor quality of most of the studies. It is well known that there is a possibility of publication bias of papers with positive outcomes (Dickersin et al. 1987; Easterbrook et al. 1991; Turner et al. 2008; Luijendijk & Koolman 2012) and therefore, papers with negative outcomes can be missed, which may have influenced the results of the current review.

In conclusion, currently CBT is the most well-studied non-pharmacological intervention for depression in people with ID, which seems to be effective as well in the mild and moderate ID population. Other promising interventions are exercise and possibly behavioural therapy and BLT. Although it is known that performing an RCT in adults with ID (and depression) can be challenging, we emphasise that further research, preferably RCTs, is needed to grow the evidence-base for non-pharmacological interventions for people with ID and depression. In this way, the non-pharmacological treatment options in this population can be expanded, which is especially important for those with severe or profound ID who can often only rely on pharmacological treatments.

Conflict of Interest

There are no conflicts of interest.