Understanding health and care expenditure by setting – who matters to whom?

Objective To assess service use and associated expenditure across a range of care settings in one local authority in London, United Kingdom. Methods An analysis of linked electronic health and council records of adults living in the borough of Barking and Dagenham, east London, for the financial year 2016/17. Unit costs were applied to individual service use to provide expenditure at an individual and population level for five settings of care. Population and expenditure volumes were compared for 32 possible combinations of service use. Results The total expenditure for the cohort (114,393 residents) for 2016/17 was £180.1 million. Almost half (47%) of total expenditure was incurred by community care, social care and mental health services, with hospital care and primary care incurring, respectively, 35% (£63.3 m) and 18% (£32.6 m). The two most common combinations in terms of total population volume and expenditure were primary and hospital care, and primary, hospital and community care. Primary care was present in all combinations. Mental health service use accounted for just over a tenth of all expenditure in the borough, but using mental health services substantially increased mean expenditure per patient. Conclusions A whole system perspective across all settings of care improves understanding of service user patterns. Setting-level analysis remains important, particularly for mental health users.


Ethnic Group
The The analysis for this paper used cost-weighted utilisation by setting and overall. The expenditure was estimated from activity data, with different methods used for the different settings of care. The total expenditure for the financial year was calculated for each individual by aggregating expenditure across five settings of care: hospital, primary care, community, mental health and social care.

Hospital
Activity was collated across four domains of hospital care: Emergency department attendances, Elective inpatient stays (including day cases), Non-elective inpatient stays (including people admitted but staying for less than 24 hours) and Outpatient attendances.
The dataset included the Healthcare Resource Group (HRG) national tariff assigned to each unit of activity, using the HRG grouper and national tariff in place for the 2016/17 financial year(1). The costs used were reflective of the true cost to the commissioner of the activity. The tariff for each emergency department attendance varied depending on the type of emergency department (consultant-led emergency departments; consultant-led mono-specialty services; other types of minor injury departments; and NHS walk-in centres), whether the patient was admitted or not, and whether they arrived at the emergency department by ambulance. These adjustments were made by the HRG grouper and the national tariff for individual level activity reflected these adjustments. The national tariff for elective inpatient stays varies according to the specialty of the department, the type of admission, diagnostic tests and procedures performed during the inpatient stay and the overall length of stay for the patient. Non-elective cases are assigned different tariffs for short stays (less than two days) and long stays (two days or longer). Outpatient attendance tariffs alter depending on the specialty of the department and agreed thresholds of new to follow up ratios. There was activity that had no cost attached to it, for example outpatient visits that exceeded the tariff trim point. This activity was therefore not reflected in the total hospital expenditure Primary care Our activity data from primary care records provided the monthly count of attendances with a General Practitioner (GP), with a non-GP, and the number of prescriptions for each individual. Unit costs from the 2016/17 Unit Cost Health and Social Care were used for GP attendances (2). The report calculates costs for GP services by comparing salary, overheads, and other costs for the practice to perform the activities taking into account how long each activity takes. We used the 2016/17 unit cost of £38 per visit, which includes direct staff costs, qualification and training, for a surgery consultation with a GP. This assumes attendances are on average 9.22 minutes in duration.
The activity data provided a count of the number of attendances an individual had per month with a "Non-GP". Non-GP attendances may be with a practice nurse, pharmacist or health care assistant. Unit costs from the 2016/17 Unit Cost Health and Social Care for GP practice nurses were used, which was £42 per hour with qualifications. It was assumed that a nurse would see four people per hour, providing a unit cost of £10.50 per visit, and as with the GP attendances no distinction was made between face to face and telephone consultations. The local prescription data provided total prescription costs per practice and total number of prescriptions per practice for 2016/2017. This allowed for a unit cost per prescription to be calculated for each practice and applied to calculate prescription costs for each patient within that practice.

Community services
Activity data was extracted from the RiO system of the local community provider, NELFT. This provided information on the referral source and primary reason for referral, the cost centre the activity was assigned to (to facilitate matching of activity to unit costs), the type of consultation (new or follow up), location of the care contact (e.g. clinical setting or individuals home), whether it was in person or a telephone consultation. Each of these impacted the unit cost of the activity.
Data from the patient level information and costing system from NELFT was used to assign unit costs for each component of activity. This was calculated by the trust by distributing the budget within each department across the activity of the department. Unit level costs were then provided, by department for each type of activity (new or follow up, in person or by phone, location of care contact).

Mental health
As with community services, activity data was extracted from the RiO system of the local community provider, NELFT. This provided information on the referral source and primary reason for referral, the cost centre the activity was assigned to (to facilitate matching of activity to unit costs to the appropriate department), the type of consultation (new or follow up), location of the care contact (e.g. clinical setting or individuals home), whether it was in person or a telephone consultation. Each of these impacted the unit cost of the activity. Data from the patient level information and costing system from NELFT was used to assign unit costs for each component of activity.

Social care
Local authority social care costs were obtained from council data which lists the billed cost for each care package per week for each care recipient. This provided the granularity on in year changes to packages and the resultant change in package costs. Weekly packages included crisis intervention, home care, supported living placements, residential and nursing home placements. Data on self-