Non‐pharmacological interventions for caregivers of stroke survivors

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

We aim to provide the most reliable summary of the effect of interventions targeted towards informal caregivers of stroke survivors or targeted towards informal caregivers and the care recipient (the stroke survivor). The specific questions are as follows.

 

  1. What are the effects of interventions targeted towards informal caregivers of stroke survivors?
  2. Is the evidence of benefit greater in any pre‐defined subgroup?

 

Background

An informal carer (or unpaid carer) has been defined as ‘a person of any age who provides unpaid help and support to a relative, friend or neighbour who cannot manage to live independently without the carers help due to frailty, illness, disability or addiction’ (SGHD 2005).

Informal caregivers play an important and sizeable part in the total care provided to stroke survivors. Informal caregivers often provide significant amounts of assistance with personal and instrumental activities of daily living: they monitor signs and symptoms and general health; store, control and appropriately administer medications; organise and co‐ordinate care among health and social care providers; act as an advocate for the care recipient; and provide emotional and psychosocial support. Therefore, the caregiving scenario can be complex, demanding and challenging.

Providing informal unpaid care, help, or support to stroke survivors who live in the community can be burdensome (McCullagh 2005) and stressful (Smith 2003), and can have an adverse effect on the caregiver’s psychological wellbeing (Berg 2005Kotila 1998) and physical health (Carnwath 1987).

A range of healthcare interventions targeted towards stroke survivors and their family or other informal caregivers has been tested in randomised controlled trials (RCTs).

We propose that a systematic review and meta‐analysis be performed to determine more clearly the effects of interventions directed towards the informal caregiver, the caregiving working conditions (e.g. typical hours of care provided; flexibility; the nature of caregiving tasks performed; the physical environment; the physical, emotional and mental demands; autonomy in decision making; training opportunities; availability of professional support; etc) or interventions that target combined caregiver and caregiving working conditions on a range of outcomes.

Description of the condition

An increasing number of studies has found an association between stress in informal caregivers and immune dysregulation (Gouin 2008), an increased risk of mortality (Schulz 1999), elevated blood pressure (King 1992), impaired wound healing (Kiecolt‐Glaser 1995), increased risk of coronary heart disease (Lee 2003), and poorer cognitive function (Lee 2004) among women who provide care to their disabled or ill spouses.

The hypothesis is that when the demands placed on the informal caregiver are at variance with the needs, expectations and capacity of the caregiver, this stress can predispose the caregiver to ill health.

Description of the intervention

We are interested in any interventions targeted towards the caregiver or the caregiving working conditions, or interventions that target the combination of caregiver and caregiving working conditions.

How the intervention might work

These interventions might work to reduce the caregiving demands through:

 

  • changing the knowledge, beliefs, attitudes or behaviours of the caregiver; or
  • temporarily reducing or removing the caregiver’s responsibility for the stroke survivor; or
  • addressing ongoing psychological and social problems.

 

Why it is important to do this review

Aspects of the health of caregivers are addressed in several Cochrane Reviews; however, the caregivers are not the primary focus of any review. Given that caregivers provide a substantial amount of the overall care delivered to stroke survivors, and are likely to be at risk of adverse health outcomes, it would be useful for healthcare professionals, informal caregivers and those responsible for the disbursement of health and social care resources to have easy access to relevant information.

Objectives

We aim to provide the most reliable summary of the effect of interventions targeted towards informal caregivers of stroke survivors or targeted towards informal caregivers and the care recipient (the stroke survivor). The specific questions are as follows.

 

  1. What are the effects of interventions targeted towards informal caregivers of stroke survivors?
  2. Is the evidence of benefit greater in any pre‐defined subgroup?

 

Methods

Criteria for considering studies for this review

Types of studies

We will seek all truly randomised controlled trials of non‐pharmacological interventions targeted towards informal caregivers of stroke survivors with the aim of either: changing knowledge, beliefs, attitude or behaviours of the informal caregiver, or temporarily reducing or removing the caregiver’s responsibility for the stroke survivor. We will not exclude studies where the stroke survivor is included in the intervention unless the intervention is primarily targeted at the stroke survivor.

Types of participants

We will include trials that recruit informal caregivers of stroke patients. A broad definition of an informal caregiver is ‘a person of any age who provides one or more hours of unpaid help and support per week to a stroke survivor’. However, for the purpose of this review, we will accept the investigators’ definition. We will exclude trials of mixed aetiology if the percentage of stroke patients was less than 80%.

Types of interventions

We are interested in reviewing trials of non‐pharmacological interventions, compared with no care or routine care, that have the following features:

 

  • delivered to an informal caregiver of a stroke survivor;
  • delivered to an informal caregiver and a stroke survivor as a dyad, that is, both informal caregiver and stroke survivor are randomised;
  • where there is an intention to have an impact on caregivers’ knowledge, beliefs, attitude or behaviours.

 

We are also interested in reviewing trials of non‐pharmacological interventions where there is an intention to reduce or remove the responsibility for caregiving, for example, through the provision of external support services (such as home help, day care, respite care, support groups, etc), or the means by which to employ external support.

We will consider any non‐pharmacological intervention regardless of who provided the intervention (e.g. OT, PT, nurse, etc) or the type of intervention (e.g. educational, counselling, etc) or amount of intervention delivered.

We would expect the interventions to fall within the following categories.

 

  1. Education and training
  2. Counselling (formal)
  3. General advice and support (including problem solving)
  4. Provision of external support services
  5. Combination
  6. Other

 

Types of outcome measures

Primary outcomes

 

  1. Informal caregiver stress and strain (e.g. Caregiver Strain Index, Caregiver Burden Scale, Caregiver Reaction Assessment, Bakas Caregiving Outcome Scale, Relatives Stress Scale, Burden Interview, Sense of Competence Questionnaire, Questionnaire on Resources and Stress) at the end of scheduled follow up.
  2. Informal caregiver wellbeing at the end of scheduled follow up (e.g. Caregiver Well‐Being Scale)

 

Secondary outcomes

 

  1. Global measures of stress or distress. Lying above or below the median cut‐off point (Willmott 2004) on global measures of stress or psychological distress (e.g. General Health Questionnaire (GHQ), Global measures of Perceived Stress Scale (Cohen 1988), etc). If cut‐off values are not available, then we will use the available mean scores and standard deviations.
  2. Measures of anxiety. Lying above or below the cut‐off point (e.g. Hospital Anxiety and Depression Scale (HADS) cut‐off point greater than 11, that is a ‘severely’ disordered state of anxiety). If cut‐off values are not available then we will use the available mean scores and standard deviations.
  3. Measures of depression (e.g. HADS cut‐off point greater than 11, that is a ‘severely’ disordered state of depression). If cut‐off values are not available then we will use the available mean scores and standard deviations.
  4. Informal caregiver health‐related quality of life at the end of scheduled follow up (e.g. Nottingham Health Profile (NHP)).
  5. Informal caregiver satisfaction.
  6. Informal caregiver mortality.

 

Search methods for identification of studies

See the ‘Specialized register’ section in the Cochrane Stroke Group module.

Electronic searches

We will search the Cochrane Stroke Group Trials Register. In addition, we will search the following databases.

 

  • The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, latest issue)
  • MEDLINE (1950 to present) (Appendix 1)
  • EMBASE (1974 to present)
  • AMED (Allied and Complementary Medicine) (1985 to present)
  • AARP (Ageline) (1987 to present)
  • CINAHL (1982 to present)
  • Dissertation abstracts (1861 to present)
  • HMIC Health Management and Information Consortium (1983 to present)
  • PsycINFO (1967 to present)
  • Social Work Abstracts (1977 to present)
  • Wilson Social Sciences Abstracts (1984 to present)
  • Science Citation Index Expanded (SCI‐Expanded)(ISI Web of Science 1900 to present)
  • Social Sciences Citation Index (SSCI)(ISI Web of Science 1956 to present)
  • Arts & Humanities Citation Index (A&HCI)(ISI Web of Science 1975 to present)
  • Conference Proceedings Citation Index ‐ Science (CPCI‐S)(ISI Web of Science 1990 to present)
  • Conference Proceedings Citation Index ‐ Social Sciences & Humanities (CPCI‐SSH)(ISI Web of Science 1990 to present)

 

We will also search the following databases of clinical trials:

 

  • Internet  Stroke  Center: The Stroke Trials Registry (http://www.strokecenter.org/trials/)
  • Clinical trials.gov (http://clinicaltrials.gov/)

 

Searching other resources

We will search reference lists and, using ISI Web of Science Cited Reference Search, we will undertake forward citation checks of all included trials and relevant reviews identified. We will contact experts in the field and authors of included trials for advice as to other potentially relevant studies.

Data collection and analysis

Selection of studies

One review author (LL) will screen all the titles, abstracts and keywords of publications identified by the searches to assess their eligibility. At this stage, we will exclude studies that do not meet the inclusion criteria. Another review author (PL or LS) will independently review the decision to include or exclude. We will obtain a paper copy of the full publication for every study that is potentially relevant. Two review authors will assess all possibly relevant studies according to pre‐specified selection criteria. We will resolve any disagreements by consensus.

Data extraction and management

We will use data reported in the published sources and unpublished data in this review. Two review authors will independently extract the data using a standard data recording form. We will evaluate concealment of randomisation, blinding in outcome evaluation, and intention‐to‐treat analysis and will grade these as present or unclear. We will resolve any disagreements by discussion.

Assessment of risk of bias in included studies

For each included trial we will extract information about the method of randomisation and allocation concealment, blinding of outcome assessment and whether all the randomised patients were accounted for in the analysis. We will base sensitivity analyses on these variables.

Measures of treatment effect

We will present dichotomous outcomes as odds ratios (OR) with 95% confidence intervals (CI). We will present continuous outcomes, if possible, as mean differences (MD) with 95% CI.

Unit of analysis issues

We will focus on trials that have randomised individual caregivers or caregiver and stroke survivor dyads. Where we have clustered designs in which participants are randomised at group level we will use the intra‐cluster correlation coefficient (ICC) to estimate the effective sample size.

Dealing with missing data

Our primary aim is to obtain standardised data through collaboration with the original trialists. If data are missing from a published report we will contact the primary investigators in an attempt to get this information. Incomplete data are relatively common in trials of rehabilitation. It is difficult to impute missing values for continuous outcomes. However, if we have a data set with missing dichotomous outcome data, we will perform a ‘what if’ sensitivity analysis to explore what would have happened if all the missing data went for or against the main hypothesis.

Assessment of heterogeneity

If there is substantial methodological, statistical or clinical heterogeneity, we will not combine the study results into a meta‐analysis. We will identify heterogeneity by visual examination of the forest plots, by using the Chi2 test and a significance level of alpha = 0.1, because of the low power of this test. We will examine the quantification of heterogeneity with the I2 statistic, ranging from 0% to 100% including its 95% CI (Higgins 2002).

Assessment of reporting biases

If possible, we will assess publication bias using a Rank Correlation test and a funnel plot.

Data synthesis

We will summarise the data statistically if they are available, sufficiently similar, and of sufficient quality. For dichotomous outcomes we will use Peto odds ratios. For continuous data we will use MD. We will use a fixed‐effect model (primary), and we will also use a random‐effects model (DerSimonian 1986) to assess the sensitivity of the results to the choice of model.

Subgroup analysis and investigation of heterogeneity

The effect of interventions may vary across trials in the meta‐analysis because each intervention provided is likely to be different in terms of whether the caregiver is the target of the intervention or both the informal caregiver and stroke survivor; who is providing the intervention; how it is provided (individualised or group‐based); where it is provided; when it is provided; the characteristics of the intervention being delivered; and the purpose of the intervention.

We will only perform subgroup analyses if there is a statistically significant treatment effect in one of the primary outcomes. We will conduct subgroup analysis of treatment effect based on:

 

  • the type of intervention (i.e. education and training, counselling (formal), general advice and support (including problem solving), provision of external support services, combination, other);
  • the target of the intervention (i.e. caregiver, caregiver and stroke survivor dyad).

 

If insufficient numbers of trials appear in any of the categories we may need to collapse categories.

Sensitivity analysis

We will carry out sensitivity analyses to explore the influence of study design factors (method of randomisation, allocation concealment, blinding of final outcome assessment, and the presence of an intention‐to‐treat analysis) to determine how robust the analyses are. We will selectively exclude studies based on the following.

 

  1. The adequacy of method of randomisation or allocation concealment.
  2. The adequacy of blinding of final outcome assessment.
  3. The presence or absence of intention‐to‐treat analysis.