The CSRI is used to comprehensively record the support and services received by participants in research studies. This includes any item that has resource implications, such as accommodation and any on-site care, employment status and income, formal health and social care services used, and unpaid care. From this individual level data, we can create an aggregate picture for the study population. Below we use tables from various studies to illustrate how data collected using the CSRI can be used to highlight a range of important statistics.
Statistics related to rates and intensity of service use
- Rates of service use of individual services
- Mean intensity of service use
- Rates of use of accommodation over time
Statistics related to costs calculated
- Unit costs
- Mean costs for individual services
- Mean cost by service category
- Contributions of cost categories to total costs
- Comparison of costs by category across studies
First, the CSRI allows the service use consequences of a new policy or intervention to be described, or the changes in employment or income status to be detailed. Analysis of CSRI service use data could reveal changes in the level of use of high cost nursing or residential care, hospital inpatient admissions, or accident and emergency attendances. Perhaps the use of community nursing or general practitioners increases as a result of the new intervention, or there is an increased burden on unpaid carers. The changes may be small in an evaluation involving, say, 100 participants, but should the policy or intervention be made available to a wider population, commissioners and providers need to be aware of the impact on the full range of services so they can plan for a change in the level of provision.
Rates of receipt of different services are an important factor to consider. As an example of how the CSRI may be used to calculate these, we turn to a study into the costs and effectiveness of two psychosocial treatments for personality disorder by Beecham et al. This study investigated how rates of receipt for different service categories differed among the study groups. This is shown for the baseline group in table 1.
Table 1: Service use at baseline for the year after treatment within the Beecham et al. study
| Baseline: N (%) group using | |||
|---|---|---|---|
| One-Stage N(=32) | Step-Down (N=29) | TAU (N=47) | |
| Hospital services | |||
| Psychiatric inpatient | 5 (16%) | 2 (7%) | 14 (30%) |
| Non-psychiatric inpatient | 2 (6%) | 5 (17%) | 13 (28%) |
| Non-inpatient hospital services | 30 (84%) | 26 (90%) | 42 (89%) |
| Legal services | |||
| Police | 1 (3%) | 4 (14%) | 15 (32%) |
| Lawyer | 4 (13%) | 7 (24%) | 19 (40%) |
| Mental health services | |||
| Psychiatrist | 14 (44%) | 10 (34%) | 11 (23%) |
| Psychologist | 4 (13%) | 5 (17%) | 2 (4%) |
| Community psychiatric nurse | 7 (22%) | 10 (34%) | 28 (60%) |
| Private psychotherapist | 8 (25%) | 7 (24%) | 2 (4%) |
| Other counselling services | 0 | 5 (17%) | 0 |
| Community-based services | |||
| GP | 29 (91%) | 27 (93%) | 47 (100%) |
| Social Worker | 6 (19%) | 7 (24%) | 21 (45%) |
| Education classes | 5 (16%) | 7 (24%) | 11 (23%) |
| Employment services | 0 | 2 (7%) | 8 (17%) |
| Voluntary services | 7 (22%) | 3 (10%) | 19 (40%) |
The average number of contacts can reveal important differences. This was reported for each service in a study investigating costs and outcomes related to relapse in schizophrenia conducted by Almond et al. This is shown in table 2.
Table 2: Mean intensity of service receipt by trial category
| Service | Non-relapse (n=68) | Relapse (n=77) |
|---|---|---|
| Mean usage | Mean usage | |
| In-patient care (days) | 0 | 57.8 |
| Out-patient | ||
| Psychiatric visits | 1.4 | 2.1 |
| Other | 0.1 | 0.3 |
| Day hospital (visits) | 2.3 | 2.1 |
| Community mental health centre (visits) | 2.4 | 1.4 |
| Day care centre (visits) | 5.9 | 0.9 |
| Group therapy | 0.4 | 0.1 |
| Sheltered workshop | 1.1 | 0 |
| Specialist education | 2.9 | 0 |
| Other (not specified) | 0.6 | 0 |
| Visits by: | ||
| Psychiatrist | 2.5 | 2.3 |
| Psychologist | 0 | 0 |
| General practitioner | 1.8 | 1.6 |
| District nurse | 0.1 | 0 |
| Community psychiatric nurse | 12.6 | 5.2 |
| Social worker | 0.1 | 0.4 |
| Occupational therapist | 0 | 0.8 |
| Home help/care worker | 0.4 | 0.6 |
Table 3 describes the accommodation placements for people who left long-stay hospitals to live in community-based accommodation in the mid-1980s under the Care in the Community demonstration programme, having been followed-up twelve years later. The definition of each accommodation type was derived using data for this study and the economic evaluation of the NETRHA re-provision policy.
Table 3: Accommodation for people with learning disabilities and people with mental health problems one and 12 years after leaving long-stay hospitals
| People with learning disabilities (n=103) | People with mental health problems (n=75) | |||
|---|---|---|---|---|
| Accommodation type: | ||||
| Residential/nursing home | ||||
| Hostel | ||||
| Staffed group home | ||||
| Unstaffed group home | ||||
| Adult foster /sup. lodgings | ||||
| Sheltered housing | ||||
| Independent living | ||||
| Unclassified | ||||
| Hospital | ||||
Accommodation types were defined as follows:
- Residential/nursing homes provide six places or more with continuous staff cover by day and waking staff at night.
- Hostels provide six places or more with continuous or intermediate staff cover by day and sleeping-in or on-call cover at night.
- Sheltered housing provides individual living units within a larger complex which are rented by clients and some day and night staff cover is available.
- Staffed group homes provide two to five places with continuous or intermediate staff cover by day and any form of night cover.
- Unstaffed group homes have two to five places with ad hoc or no day staff cover and on-call or no staff cover at night.
- Foster placements have intermediate day support and on-call support at night where clients have moved in with an established household and in supported lodgings clients move into an established household with ad hoc day staff cover and on-call night staff support.
- Independent living arrangements cover single or group tenancies in domestic housing, including living with relatives or spouse, where there is ad hoc or no day staff cover and no staff cover at night.
- Hospital placement is where clients have been readmitted from community placements.
Another key way to use the data recorded on the CSRI is to estimate the associated costs. The CSRI identifies which service has been used and the intensity of use: how many contacts there have been in the specified period, and the duration of these contacts. The unit cost of each service – per hour, per contact, per week etc. – can then be multiplied by each person’s duration of use of that service and these figures can then be totalled. A useful source of nationally applicable unit costs for around 150 health and social care services in England is The Unit Costs of Health and Social care. This compendium is produced annually by the PSSRU, and can be found here.
In the next few tables, we show examples where costs are considered, derived from individual level service use data collected on the study’s CSRI. They are identified by service type (hospital inpatient care, GP, social worker, etc) and by service category, which summarises the cost burden to specific agencies and organisations.
Unit costs of each service captured within the CSRI may be described, as was the case in a study investigating the cost-effectiveness of Sertraline and Mirtazapine for people with depression and dementia, conducted by Romeo et al. This allows readers to gauge the transferability of research findings into different contexts. This is shown in table 4.
Table 4: Unit costs (2009/10 prices) used within the Romeo et al. study
| Service | Unit cost (£) |
|---|---|
| In-patient (bed days) | 299 |
| Day hospital (attendance) | 50–205 |
| Out-patient (appointment) | 21–165 |
| Accident and emergency (attendance) | 37–97 |
| General practitioner (per surgery consultation) | 28 |
| Geriatrician (minute) | 1.83 |
| Nurse (minute) | 0.43–0.52 |
| Occupational therapist (minute) | 0.65 |
| Community psychiatrist (minute) | 1.83 |
| Counsellor (minute) | 0.57 |
| Psychologist (minute) | 1.2 |
| Chiropodist (contact) | 0.37 |
| Social worker (minute) | 0.67 |
| Care manager (minute) | 0.82 |
| Home care worker/care attendant (minute) | 0.35 |
| Sitting scheme (minute) | 0.45 |
| Self-help group (minute) | 0.57 |
| Meals on wheels (meal) | 4.8 |
| Dentist (minute) | 2.9 |
| Optician (minute) | 0.48 |
| Day care (day) | 42–66 |
| Lunch club (meal) | 7 |
| Social club (session) | 5 |
Unit costs and service receipt data collected using the CSRI are used together to calculate costs of service receipt, which are typically a key component of economic evaluations. An example of this would be a study regarding Multiple Sclerosis (MS) in the UK, conducted by McCrone et al. which included a more extensive form of the following table:
Table 5: Inpatient use and contacts with professionals and associated costs (£) in 6 months prior to survey [mean (Standard Deviation)]
| Costs for any reason (entire sample) | MS-related costs (entire sample) | Non-MS-related costs (entire sample) | |
|---|---|---|---|
| Neurology outpatient | 115 (181) | 105 (172) | 8 (64) |
| Other outpatient | 105 (282) | 52 (242) | 49 (159) |
| Day hospital | 11 (57) | 9 (53) | 1 (11) |
| Nursing/residential home | 81 (795) | 68 (724) | 13 (292) |
| Neurology inpatient | 138 (1477) | 109 (1392) | 18 (379) |
| Intensive care unit | 74 (1320) | 11 (284) | 48 (1157) |
| Other inpatient | 310 (2176) | 134 (1540) | 160 (1543) |
| General practitioner | 58 (78) | 35 (72) | 21 (40) |
| Physiotherapist | 63 (183) | 57 (186)c | 4 (30) |
| Social worker | 24 (77) | 17 (74) | 2 (20) |
| Practice nurse | 8 (37) | 4 (36) | 2 (11) |
| District nurse | 66 (371) | 53 (323) | 12 (183) |
| Speech therapist | 2 (10) | 1 (10) | <1 (1) |
| Home help | 183 (899) | 162 (914) | 9 (137) |
| Acupuncturist | 10 (59) | 7 (53) | 1 (19) |
| Homeopath | 2 (16) | 1 (15) | <1 (2) |
| Herbalist | 3 (63) | 2 (62) | <1 (6) |
| Aromatherapy | 7 (41) | 4 (37) | 1 (11) |
| Reflexology | 15 (65) | 11 (60) | 1 (21) |
As well as reporting costs for each individual service, subtotals for each cost category (hospital, community social care, medication etc.) are typically calculated. We provide as an example a study investigating the cost-effectiveness of a cognitive stimulation therapy for people with dementia by Knapp et al. , as shown in table 6.
Table 6: Mean costs for service categories by treatment allocation
| Intervention group ( n =91) | Control group ( n =70) | Difference between intervention and control groups | |||
|---|---|---|---|---|---|
| Mean (s.d.) | Mean (s.d.) | Bootstrap mean difference | Bootstrap 95% confidence interval | P | |
| Baseline | |||||
| Residential care | 334.93 (91) | 331.63 (101) | 3.3 | -27 to 34 | 0.829 |
| Hospital services | 44.75 (159) | 18.95 (54) | 25.8 | -6 to 65 | 0.152 |
| Day services | 21.47 (55) | 23.16 (67) | -1.69 | -22 to 17 | 0.861 |
| Community services | 13.37 (42) | 11.12 (30) | 2.25 | -9 to 14 | 0.704 |
| Medication | 9.20 (21) | 9.13 (15) | 0.07 | -5 to 6 | 0.980 |
| Other accommodation | 0 (0) | 1.2 (10) | -1.2 | -4 to 0 | 0.321 |
| Total | 423.72 (178) | 395.19 (110) | 28.53 | -14 to 74 | 0.241 |
| Follow-up | |||||
| Residential care | 334.93 (91) | 331.63 (101) | 3.3 | -26 to 35 | 0.829 |
| Hospital services | 29.82 (116) | 4.51 (36) | 25.31 | -2 to 54 | 0.051 |
| Day services | 8.66 (33) | 16.32 (44) | -7.66 | -20 to 4 | 0.226 |
| Community services | 11.47 (25) | 7.04 (16) | 4.43 | -2 to 11 | 0.204 |
| Medication | 3.99 (13) | 6.84 (16) | -2.85 | -7 to 2 | 0.219 |
| Other accommodation | 0 (0) | 1.37 (11) | -1.37 | -5 to 0 | 0.321 |
| Intervention | 24.92 (13) | — | 24.92 | — | — |
| Total | 413.80 (151) | 368.61 (111) | 45.18 | 5 to 86 | 0.037 |
Having calculated costs for individual services and service categories, the identification of primary cost drivers or high cost areas is generally of interest to readers of research findings. The CSRI enables researchers to investigate these questions, as demonstrated by Knapp and Beecham in their 1993 ‘Reduced list costings’ paper investigating where the bulk of costs occur, with the paper abstract available here. Table 7 demonstrates these figures, as found through administering the CSRI.
Table 7: Reduced list costs (£ per week, 1992 prices) estimated by accommodation type: psychiatric reprovision (NETRHA- North East Thames Regional Health Authority)
| Percentage of full costs accounted for by: | |||||
|---|---|---|---|---|---|
| Accommodation type: | Full cost | Accommodation | Top 5 | Top 10 | Sample N |
| Residential/nursing home | 597 | 91 | 97.4 | 98.9 | 164 |
| Hostel | 462 | 85.1 | 96.4 | 98.8 | 60 |
| Sheltered housing | 168 | 72.9 | 94.7 | 99.2 | 6 |
| Staffed group home | 445 | 83.3 | 96.6 | 98.7 | 41 |
| Unstaffed group home | 371 | 64.2 | 92.5 | 98.9 | 27 |
| Adult foster placement | 357 | 69.8 | 96.1 | 99.2 | 11 |
| Independent living | 240 | 63.2 | 87.7 | 95.9 | 32 |
| Full sample | 488 | 83.3 | 94.4 | 98 | 341 |
Comparisons of the relative contribution of different service categories to total costs across different studies have also been completed using the CSRI. As part of a project funded by the Alzheimer’s Society to estimate the cost of dementia in the UK, the researchers investigated the distribution of costs across service categories found in three different studies which collected costs using a CSRI . These studies included Cognitive Stimulation Therapy (CST), Maintenance Cognitive Stimulation Therapy (MCST), and Sertraline or mirtazapine for depression in dementia (SADD). Results are shown in table 8.
Table 8: Variations in distribution of costs of care between services (%) across studies, excluding unpaid care
| Short-stay accommodation | |||||
| Hospital services | |||||
| Community health care | |||||
| Community social care | |||||
| Adaptations and equipment | |||||
| Day services | |||||
| Medication | |||||
| (Total) |
–>Home