A Coping Strategy for Carers of Family Members With Dementia
The Supplementary File provides the full protocol of this pragmatic multicentre randomised controlled trial. Our intervention, based on the US Coping with Caregiving programme, was individual and manualised and required active participation. Our primary objective was to determine the clinical effectiveness (measured by the hospital anxiety and depression scale) and cost effectiveness (reported in an accompanying paper) of eight sessions of a manual based coping strategy, delivered over 8-14 weeks by supervised psychology graduates to carers of family members with dementia, compared with usual service provision, over eight months.
The secondary outcomes were depression and anxiety caseness on the hospital anxiety and depression scale; quality of life of both the carer and the recipient of the care; and abusive behaviour by the carer. We plan on analysing time to entry to 24 hour care of the family member with dementia at longer term follow-up (at two and seven years).
We recruited carers to the trial from 4 November 2009 to 8 June 2011. The first four month follow-up took place on 4 March 2010, with the final eight month follow-up on 7 February 2012.
We recruited through disparate settings: two mental health trusts’ memory services (Camden and Islington Foundation Trust, urban setting; North Essex Partnership Foundation Trust, suburban and rural); the North East London Foundation Trust Admiral nurse, suburban (specialist nurses for carers of family members with dementia); and the Dementia Research Centre-National Hospital for Neurology and Neurosurgery, a tertiary service with a high rate of referrals for young people with early onset dementia.
We included carers of family members referred in the previous year who provided emotional or practical support at least weekly and identified themselves as the primary carer of a family member with dementia not living in 24 hour care. We excluded carers who were unable to give informed consent to the trial, were currently taking part in a randomised controlled trial in their capacity as a carer, or who lived more than 1.5 hours travelling time from the researchers' base. We administered the mini-mental state examination to carers aged 60 or over only at baseline. If they scored less than 24 the research assistant discussed the participant with GL or CC to see whether this was related to cognition, mood, or education. If carers were judged to have a dementia they were not included in the study and we informed the referring clinician.
Prospective participants were initially approached by a clinician and given or sent an information sheet. Those interested in participating were referred to the research team. The referral gave the name, sex, and relationship to the family member of the prospective participant as well as the patient's sex. The researchers telephoned the carer 24 hours or more after they received the information sheet. The researchers answered any questions and then arranged to meet those who agreed to take part to obtain their informed consent and complete baseline assessment before randomisation.
To conceal allocation we used an online computer generated randomisation system to allocate participants to the intervention or to treatment as usual. This system was set up and maintained by an independent clinical trials unit and accessed by the START trial manager. Randomisation was stratified by trust using random permuted blocks. To allow for potential clustering effects in the intervention arm we used an allocation ratio of 2:1 (intervention: treatment as usual). A member of the therapy team then phoned the participants and informed them of their allocation, either to treatment as usual when they would be contacted for a four month follow-up or to the intervention when an appointment was made for the therapy to start. Allocation within the individual teams was according to workload.
Carers were interviewed at baseline and at four and eight months after randomisation, usually in their own home, unless they preferred to come to the research team base in University College London. We have continued to follow up carers, asking them to remain in the study for two years even if the recipient of their care had been placed in a care home or died. Results of this longer term follow-up will be reported separately. Information collected at baseline consisted of sociodemographic details about the carer and recipient of the care; and clinical and resource use items (as detailed in the accompanying paper). At both the four and eight month follow-up we repeated the collection of clinical and resource use information.
Sociodemographic details obtained at baseline included age, sex, ethnicity, relationship to the recipient of care (for example, spouse, child), level of education, last occupation, and living situation.
Measures regarding the carer's health and wellbeing collected at all three study time points were:
At all time points, we also asked carers for information about the recipient of their care:
We blinded outcome assessors to randomisation status, but it was not possible to blind the study participants. The researchers worked in two teams, each assessing outcomes for approximately half the participants and providing therapy to those allocated to treatment in the half of participants they were not assessing. Assessors asked participants at the beginning of each interview not to disclose their allocation group.
With the first author's permission, we developed an individual therapy programme (START, STrAtegies for RelaTives) based on the Coping with Caregiving programme from the United States. We adapted it for UK use for individual carers of family members with dementia over eight sessions ( Box ). The therapy took place where the carers preferred, usually in their homes, without the family member with dementia in the room. The therapy was carried out with an interpreter if the carer did not speak English fluently.
We employed and trained psychology graduates with no clinical training to deliver the intervention. The training programme had a strong practical focus on how to deliver the therapy, potential clinical dilemmas, working with interpreters, empathic listening skills, effective use of supervision, and when to ask for help. We trained the therapists to adhere to the manual and required them to demonstrate, by role play, competence in delivering each session of the intervention. Our clinical psychologist (PR) met with each team of therapists for 1.5 hours of group clinical supervision every fortnight. She also had one hour of dedicated time per week for individual consultation as needed by the therapists. The therapists recorded one therapy session per participant, selected at random, and a researcher not involved in the therapy used a standard checklist to rate the session for fidelity to the manual. Overall fidelity scores ranged from 1 to 5, with 5 being high. If fidelity scores were not high the supervising clinical psychologist discussed this in supervision.
In the treatment as usual group, services were based around the family member with dementia. Standard treatment concerns medical, psychological, and social issues. Thus the treatment consisted of assessment, diagnosis, and information; drug treatment; cognitive stimulation therapy; practical support; treatment of neuropsychiatric and cognitive symptoms; and carer support. In each setting, treatments aimed to be in line with the clinical guidelines for good dementia care of the National Institute for Health and Care Excellence.
This study was originally powered for a primary outcome of anxiety score on the hospital anxiety and depression scale based on data from a cross sectional pilot study of carers of family members with dementia. Mean anxiety scores for this group were 7.2 (SD 4) points. We considered a decrease of 2 points in mean score and 0.5 change in standard deviation to be clinically significant (expert consensus). To detect such a difference with 90% power at a 5% significance level, we required 75 participants in each group. To account for therapist clustering, we used a design effect of 1.87 for the intervention group, assuming an average of 30 carers per therapist and an intracluster correlation of 0.03. Based on these calculations and inflating for 20% attrition, we planned to recruit 90 participants in the treatment as usual group (no clustering) and 168 participants in the intervention group (clustering).
After recruitment, the research team (with approval from the funding body while the database was still locked) agreed that the primary outcome should be changed to the total score on the hospital anxiety and depression scale as this has been shown to have better sensitivity and positive predictive value than either of the individual anxiety and depression scores in identifying depression. We calculated that the sample size then available (87 treatment as usual, 173 intervention group) would be sufficient to detect a mean difference in total score on the hospital anxiety and depression scale of at least 2.4 points (with 80% power, 5% significance), which was considered to be clinically important. This calculation assumed a standard deviation for the total score of 7.4 (as seen in pilot data), allowed for analysis of covariance (with assumed correlation 0.5), and repeated follow-up measurements at four and eight months (assumed correlation 0.7). We factored in drop-out rates at 10% (based on that observed), and we applied a revised design effect of 1.4 for the intervention arm (using an intracluster correlation of 0.03 and the observed average cluster size of 15 carers for each therapist).
In the primary analysis we used regression methods to estimate group differences in total score on the hospital anxiety and depression scale over the eight month follow-up. We used random effects models to account for the therapist clustering in the intervention arm and repeated measurements at four and eight months. We adjusted for baseline total score and centre (on which randomisation was stratified) and also on factors believed to affect affective symptoms (carer's age, sex, carer burden, and neuropsychiatric symptoms of the recipient of the care). We carried out all analyses by intention to treat but excluded carers with data missing at both the four and the eight month follow-up.
We used sensitivity analyses to reanalyse the primary outcome and to assess robustness of our conclusions. Analyses considered adjustment for imbalances in baseline characteristics between the randomised groups and the differential effects of treatment over time (treatment by time interaction). Using logistic regression we also investigated the extent to which missing outcomes varied by baseline characteristics; we then repeated the main analyses adjusting for those factors associated with missingness.
We applied similar approaches for analysis of the secondary outcomes. For binary outcomes we used random effects logistic regression. We compared entry of the family member with dementia to 24 hour care between groups using a simple comparison of proportions (not allowing for clustering) because of small numbers.
All statistical analyses followed a predefined analysis plan and were carried out using STATA version 11.
Methods
The Supplementary File provides the full protocol of this pragmatic multicentre randomised controlled trial. Our intervention, based on the US Coping with Caregiving programme, was individual and manualised and required active participation. Our primary objective was to determine the clinical effectiveness (measured by the hospital anxiety and depression scale) and cost effectiveness (reported in an accompanying paper) of eight sessions of a manual based coping strategy, delivered over 8-14 weeks by supervised psychology graduates to carers of family members with dementia, compared with usual service provision, over eight months.
The secondary outcomes were depression and anxiety caseness on the hospital anxiety and depression scale; quality of life of both the carer and the recipient of the care; and abusive behaviour by the carer. We plan on analysing time to entry to 24 hour care of the family member with dementia at longer term follow-up (at two and seven years).
Recruitment and Follow-up
We recruited carers to the trial from 4 November 2009 to 8 June 2011. The first four month follow-up took place on 4 March 2010, with the final eight month follow-up on 7 February 2012.
Setting
We recruited through disparate settings: two mental health trusts’ memory services (Camden and Islington Foundation Trust, urban setting; North Essex Partnership Foundation Trust, suburban and rural); the North East London Foundation Trust Admiral nurse, suburban (specialist nurses for carers of family members with dementia); and the Dementia Research Centre-National Hospital for Neurology and Neurosurgery, a tertiary service with a high rate of referrals for young people with early onset dementia.
Participants
We included carers of family members referred in the previous year who provided emotional or practical support at least weekly and identified themselves as the primary carer of a family member with dementia not living in 24 hour care. We excluded carers who were unable to give informed consent to the trial, were currently taking part in a randomised controlled trial in their capacity as a carer, or who lived more than 1.5 hours travelling time from the researchers' base. We administered the mini-mental state examination to carers aged 60 or over only at baseline. If they scored less than 24 the research assistant discussed the participant with GL or CC to see whether this was related to cognition, mood, or education. If carers were judged to have a dementia they were not included in the study and we informed the referring clinician.
Procedure
Prospective participants were initially approached by a clinician and given or sent an information sheet. Those interested in participating were referred to the research team. The referral gave the name, sex, and relationship to the family member of the prospective participant as well as the patient's sex. The researchers telephoned the carer 24 hours or more after they received the information sheet. The researchers answered any questions and then arranged to meet those who agreed to take part to obtain their informed consent and complete baseline assessment before randomisation.
Allocation to Trial Groups
To conceal allocation we used an online computer generated randomisation system to allocate participants to the intervention or to treatment as usual. This system was set up and maintained by an independent clinical trials unit and accessed by the START trial manager. Randomisation was stratified by trust using random permuted blocks. To allow for potential clustering effects in the intervention arm we used an allocation ratio of 2:1 (intervention: treatment as usual). A member of the therapy team then phoned the participants and informed them of their allocation, either to treatment as usual when they would be contacted for a four month follow-up or to the intervention when an appointment was made for the therapy to start. Allocation within the individual teams was according to workload.
Assessments
Carers were interviewed at baseline and at four and eight months after randomisation, usually in their own home, unless they preferred to come to the research team base in University College London. We have continued to follow up carers, asking them to remain in the study for two years even if the recipient of their care had been placed in a care home or died. Results of this longer term follow-up will be reported separately. Information collected at baseline consisted of sociodemographic details about the carer and recipient of the care; and clinical and resource use items (as detailed in the accompanying paper). At both the four and eight month follow-up we repeated the collection of clinical and resource use information.
Sociodemographic details obtained at baseline included age, sex, ethnicity, relationship to the recipient of care (for example, spouse, child), level of education, last occupation, and living situation.
Measures regarding the carer's health and wellbeing collected at all three study time points were:
The hospital anxiety and depression scale, a self completed scale, which has been validated for all age groups and settings, in people who are physically well or unwell, and in Asian and African ethnic groups. The scale determines caseness of depression and anxiety with scores ranging from 0 to 21 and as a total score ranging from 0 to 42 (higher scores indicating more symptoms). We chose the total score as our primary outcome because it has a better sensitivity and positive predictive value than either of the individual scales in identifying depression when compared with the international classification of diseases criteria. The anxiety and depression score was also dichotomised as "case" and "non-case," with a cut-off point of 8/9.
The Zarit burden interview, a 22 item self report questionnaire, is the most consistently used measure of burden in carers; scores range from 0 to 88, with higher scores indicating more burden.
The modified conflict tactics scale is a self completed measure of potentially abusive behaviour by carers towards the recipient of their care. Ten behaviours are scored as to whether, during the previous three months, these have occurred never (0), almost never (1), sometimes (2), most of the time (3), or all of the time (4), and these items can be added to make a score. These behaviours range from shouting to threatening to shaking or slapping. A score of 2 or more on any one of the items is classified as an abusive behaviour. If participants scored this on any item, we discussed the score with a supervising clinician and if it was judged that the recipient of care was at risk, permission was asked to inform the clinical team so that the carer and recipient of care could have appropriate help.
Health status questionnaire, mental health domain, measures health related quality of life throughout the age ranges and is sensitive to change. It is summarised as a continuous score, ranging from 0 to 100, with higher scores indicating better outcome.
The brief COPE, a self completed measure of coping strategies by the carer, validated in carers of family members with dementia, with subscales that measure problem focused, emotion focused, and dysfunctional coping.
At all time points, we also asked carers for information about the recipient of their care:
The neuropsychiatric inventory is a validated instrument with 12 symptom domains that are scored for their severity and frequency and summarised as a single continuous score (higher scores indicating worse symptoms). We included this tool as neuropsychiatric symptoms have been shown to be associated with psychological morbidity of carers.
The clinical dementia rating, which we used as an informant instrument, grades the level of impairment of someone with dementia (categories: healthy, very mild, mild, moderate, severe).
Quality of life-Alzheimer's disease was rated by the carers, to assess the family member's overall quality of life. The total score ranges from 13 to 52, with higher scores indicating better outcome.
Blinding
We blinded outcome assessors to randomisation status, but it was not possible to blind the study participants. The researchers worked in two teams, each assessing outcomes for approximately half the participants and providing therapy to those allocated to treatment in the half of participants they were not assessing. Assessors asked participants at the beginning of each interview not to disclose their allocation group.
Therapy Intervention
With the first author's permission, we developed an individual therapy programme (START, STrAtegies for RelaTives) based on the Coping with Caregiving programme from the United States. We adapted it for UK use for individual carers of family members with dementia over eight sessions ( Box ). The therapy took place where the carers preferred, usually in their homes, without the family member with dementia in the room. The therapy was carried out with an interpreter if the carer did not speak English fluently.
Training and Delivery
We employed and trained psychology graduates with no clinical training to deliver the intervention. The training programme had a strong practical focus on how to deliver the therapy, potential clinical dilemmas, working with interpreters, empathic listening skills, effective use of supervision, and when to ask for help. We trained the therapists to adhere to the manual and required them to demonstrate, by role play, competence in delivering each session of the intervention. Our clinical psychologist (PR) met with each team of therapists for 1.5 hours of group clinical supervision every fortnight. She also had one hour of dedicated time per week for individual consultation as needed by the therapists. The therapists recorded one therapy session per participant, selected at random, and a researcher not involved in the therapy used a standard checklist to rate the session for fidelity to the manual. Overall fidelity scores ranged from 1 to 5, with 5 being high. If fidelity scores were not high the supervising clinical psychologist discussed this in supervision.
Treatment as Usual
In the treatment as usual group, services were based around the family member with dementia. Standard treatment concerns medical, psychological, and social issues. Thus the treatment consisted of assessment, diagnosis, and information; drug treatment; cognitive stimulation therapy; practical support; treatment of neuropsychiatric and cognitive symptoms; and carer support. In each setting, treatments aimed to be in line with the clinical guidelines for good dementia care of the National Institute for Health and Care Excellence.
Power Calculation
This study was originally powered for a primary outcome of anxiety score on the hospital anxiety and depression scale based on data from a cross sectional pilot study of carers of family members with dementia. Mean anxiety scores for this group were 7.2 (SD 4) points. We considered a decrease of 2 points in mean score and 0.5 change in standard deviation to be clinically significant (expert consensus). To detect such a difference with 90% power at a 5% significance level, we required 75 participants in each group. To account for therapist clustering, we used a design effect of 1.87 for the intervention group, assuming an average of 30 carers per therapist and an intracluster correlation of 0.03. Based on these calculations and inflating for 20% attrition, we planned to recruit 90 participants in the treatment as usual group (no clustering) and 168 participants in the intervention group (clustering).
After recruitment, the research team (with approval from the funding body while the database was still locked) agreed that the primary outcome should be changed to the total score on the hospital anxiety and depression scale as this has been shown to have better sensitivity and positive predictive value than either of the individual anxiety and depression scores in identifying depression. We calculated that the sample size then available (87 treatment as usual, 173 intervention group) would be sufficient to detect a mean difference in total score on the hospital anxiety and depression scale of at least 2.4 points (with 80% power, 5% significance), which was considered to be clinically important. This calculation assumed a standard deviation for the total score of 7.4 (as seen in pilot data), allowed for analysis of covariance (with assumed correlation 0.5), and repeated follow-up measurements at four and eight months (assumed correlation 0.7). We factored in drop-out rates at 10% (based on that observed), and we applied a revised design effect of 1.4 for the intervention arm (using an intracluster correlation of 0.03 and the observed average cluster size of 15 carers for each therapist).
Statistical Analysis
In the primary analysis we used regression methods to estimate group differences in total score on the hospital anxiety and depression scale over the eight month follow-up. We used random effects models to account for the therapist clustering in the intervention arm and repeated measurements at four and eight months. We adjusted for baseline total score and centre (on which randomisation was stratified) and also on factors believed to affect affective symptoms (carer's age, sex, carer burden, and neuropsychiatric symptoms of the recipient of the care). We carried out all analyses by intention to treat but excluded carers with data missing at both the four and the eight month follow-up.
We used sensitivity analyses to reanalyse the primary outcome and to assess robustness of our conclusions. Analyses considered adjustment for imbalances in baseline characteristics between the randomised groups and the differential effects of treatment over time (treatment by time interaction). Using logistic regression we also investigated the extent to which missing outcomes varied by baseline characteristics; we then repeated the main analyses adjusting for those factors associated with missingness.
We applied similar approaches for analysis of the secondary outcomes. For binary outcomes we used random effects logistic regression. We compared entry of the family member with dementia to 24 hour care between groups using a simple comparison of proportions (not allowing for clustering) because of small numbers.
All statistical analyses followed a predefined analysis plan and were carried out using STATA version 11.
SHARE
