Featured examples of the CSRI
Goal-oriented rehabilitation in early-stage dementia (GREAT)
This study is currently underway and involves people with early-stage Alzheimer’s disease, vascular or mixed dementia, and their carers. The CSRI is administered through interview, with recruitment from 2013-2015.
Generic mental health CSRI
This CSRI was developed to assist with comparing across interventions for similar mental health problems.
Arthritis study CSRI
This particular CSRI was self-completed at baseline and administered via post at other time points. Participants were predominantly older people. Participants were recruited from 2000-2003.
Homestart CSRI
This CSRI collected service receipt data for all family members, for families with a child under the age of five experiencing stress. Recruitment took place from 2001-2002.
Economic cost of severe antisocial behaviour in children CSRI
With respect to this study, the CSRI was adapted in order to be completed by the main carer of the child included in the study. Study data collection took place from 1995-1999.
Chronic pain CSRI
This CSRI collected service receipt data for adolescents with chronic pain, and was self-completed via post. Participants were recruited in 2004.
Goal-oriented rehabilitation in early-stage dementia (GREAT)
This is an ongoing randomised controlled trial of an estimated 480 dyads (people with early-stage dementia and carer pairs). The aim is to identify if goal-oriented cognitive rehabilitation (ten sessions over 3 months followed by 4 maintenance sessions over 6 months) is effective and cost-effective for people with early-stage dementia and their carers. The protocol for the trial can be found here.
In this case, the CSRI, which may be found here, is administered via interview.
One of the features of this CSRI is the fact that it uses ‘signposting’, prompting researchers to skip particular sections if prior answers reveal redundancy. One common application of this relates to how questions about unpaid carer inputs are asked. Co-resident carers are directed to a question providing percentage bands for an average day, whereas carers not living with their care recipient are asked more simply for the hours of care per week they provide, since this is simpler for them to estimate.
This CSRI also makes use of cards, for example when requesting information on receipt of direct payments. This prompt requests that the researcher shows the respondent the definition of direct payments, in order to obtain an accurate figure.
The CSRI also includes an expanded carer section. One section, which has only recently been considered, regards estimating the time spent by carers and other costs associated with assisting the care recipient in accessing services such as day care.
Generic
The ‘generic’ mental health CSRI was developed in response to need for economic evaluations to be compared across interventions for similar mental health problems. To make such comparisons, not only must the outcome measure employed be similar, but so too should the cost measure include the same scope of services.
This version of the CSRI is UK-specific, although developed from work for an international project with resulting careful consideration of services and support that would be available across the UK for people with severe mental health problems. The service options listed were considered to be a good compromise between limiting the CSRI to a ‘core’ group of services, and sufficient differentiation within service types (for example, accommodation or hospital wards) to pick up the major cost differences. Questions on levels of police contact and medication use complete the CSRI. This CSRI ask about service use during the previous three months and incorporates a coding system for easier data-entry.
The CSRI can be downloaded here.
Arthritis study CSRI
This trial evaluated a primary care self-management programme for people with arthritis. The CSRI format was adapted to support self-completion and used when participants were recruited at primary care practices, and then sent by post to participants four and twelve months later. Participants completed the CSRI, found , alongside other outcome measures so the initial patient identifiers were omitted.
To fit better with the order of questions in the booklet, this version of the CSRI started with questions on service use, focusing on services that people with arthritis were most likely to use. Each question clearly asks for resource use information ‘….because of your arthritis’. The study took both a health and social care and a societal perspective, so questions were included on social security benefit receipt, out-of-pocket expenditure and unpaid care, as well as time off work. CSRI completion rates were in line with those for the outcome questionnaires, suggesting the format was acceptable to participants. Of the 812 people recruited to the trial, 94% completed the CSRI at baseline (95% for the outcome measures), 77% at four months (80%), and 73% at twelve months (76%).
A copy of the paper reporting on the economic evaluation can be downloaded here.
Homestart CSRI
This interview-based CSRI was developed for use in an evaluation of the Homestart service for young families under stress. The CSRI for this study covers the usual fields: background data; employment and income; use of education, health and social care services (yes/no, frequency and average duration); and unpaid care supports. Most questions related to the previous three months, and the CSRI was supported by a comprehensive list of service options and the interviewers’ knowledge of local services.
Homestart offered trained volunteer support to families with children less than five years of age. This service context meant that there were some formatting changes to allow the focus to shift from an individual to the family. For this CSRI, it was important to record service use for all family members, rather than an individual. Moreover, as the study was undertaken in Northern Ireland and in South-East England, we also needed to take into account the availability of country-specific services as well as local variations. The resource use tables, therefore, are not populated with service lists but were left blank, and space is provided to record systematically the services used by all the children in the family as well as parents. Shared services – perhaps the mother and a child seeing the GP at the same time – could also be recorded. While this made data coding and data entry more time-consuming, it allowed the interviewers more flexibility in administering the CSRI during the face-to-face interview.
Within an interview lasting about two hours that covered all the research questionnaires, the CSRI typically took between five and fifteen minutes to complete, and thus was only a small additional time burden on parents. Generally, parents had little problem answering the questions, and many kept personal records of the services they used so were confident about making an accurate estimate of frequency and duration.
The CSRI can be found here.
More information about this CSRI and the resulting service use and cost data can be found in:
‘Sleed, M., Beecham, J., Knapp, M., McCauley, C. and McCurry, N. (2006) Assessing services, supports and costs for young families under stress, Child: Health, Care and Development, 32, 1, 101-110.’
The summary and full report from the JRF-funded Homestart study can be found here.
Economic cost of severe antisocial behaviour in children study
The individuals of interest within this study were children aged 3-8 years with severe antisocial behaviour as a persistent problem. The abstract for this paper can be found here.
The CSRI was completed by the main carer of the child, and can be found here.
The first section covers demographics. In this instance they were collected as part of the CSRI; however, in many cases these are collected separately.
Under ‘employment and income’, which refers to the main carer, the carer is asked a number of questions which may assist with calculating an estimate cost of lost employment due to caring. In particular, question 14c asks for monthly wage. Other CSRIs (that are administered in the form of an interview instead) avoid asking for wage or salary information due to this being a sensitive question for a number of individuals, and typically result in a lower response rate. Instead, they request a job description and hours of employment lost- and refer to average wage data by job category to calculate income estimates instead.
The CSRI also has a section (‘school support or special school’) which is tailored entirely towards capturing service receipt related to children. Within the final section (health service use), question 27 is tailored towards collecting children’s data; however, the other questions are typically also asked within CSRIs aiming to capture data on adults. Since the CSRI covers a broad range of services that many different people would receive, as shown here tailoring a CSRI to a population group typically involves adding specific services to the pre-existing list as opposed to removing questions.
Chronic pain CSRI
This study investigated the costs of chronic pain in adolescents (aged between 11 and 18). The CSRI was specifically adapted for this study, and was completed by 52 families. The abstract for the paper can be found here and the CSRI can be downloaded here.
The sections related to hospital service receipt are quite extensive compared to CSRIs used in studies involving different population groups (for example, people with dementia).
We can also see that in question 7 respondents can answer yes OR no to the receipt of each service. This can be helpful in identifying missing values in the dataset that would be compiled from answers to these questionnaires (if the respondent does not have an opportunity to specifically state that they did not use a service researchers, would have to judge if the blank space they received at the end meant that they did not use it or if this data is missing).
This CSRI also does not ask for data related to the receipt of benefits. In some instances this data is not used and therefore does not need to be collected. One example of this is if a societal perspective is adopted: benefits are payment transfers and therefore would not be considered a cost.