Southampton City Council Fair Cost of Care Report – Home Care October 2022
Market engagement and response
Southampton City Council has worked closely with Hampshire Care Association (HCA), the local representative body for care providers throughout this exercise as well as sharing information with providers about the national support offer from Care England.
The council began raising local awareness about the exercise via three separate webinars held in early May 2022. All local home care and care homes were invited to attend to hear a presentation about the exercise, which included an independent view from HCA and an opportunity to ask questions. Representatives from five different home care providers attended these sessions. The presentation was also subsequently sent out to all local home care providers.
The home care exercise began on 16 May with all 26 local home care providers asked to take part and submit a response. The LGA commissioned ARCC tool was used for this exercise. Weekly drop-in sessions were made available for providers to ask questions of commissioners and a dedicated email address was also provided although very few questions were asked via these routes. As the initial deadline of 10Ju ne neared, individual contact was made with every provider to check if they had any questions or needed any other support. The council received considerable feedback from providers indicating that the complexity of the exercise meant that they wouldn’t be able to achieve this deadline, therefore an extension was given until early July. A further extension was additionally given to several providers who were able to give a specific date when the exercise would be completed. During this extended period, weekly individual contact was made with each provider to offer support. The outcome of the exercise was that 14 out of the 26 providers (54%) submitted a response however upon initial validation two responses were not sufficiently completed to use all of their data even after clarification discussions with these providers. Therefore 12 full responses have been used in the analysis and two partial responses. The providers who made a submission are considered to be broadly representative of the whole market as the proportion of small and large providers is similar to the overall market composition. This is also the case for the proportion of providers which have been established in the city for many years compared to those which are relatively new. The main reasons given by providers who did not complete the exercise were lack of time (even after further extensions were provided) and not having all of the information needed to accurately complete the tool.
Statistical analysis and data quality considerations
The council conducted an initial validation of the responses to sense check against data already held, for example the number of care hours commissioned by the council. Alongside this, LaingBuisson were commissioned to support the council with a more detailed analysis and validation of the data. This involved checking the returns for sense and consistency before attempting direct contact with each provider to resolve ambiguities, particularly around the time period for which the data was applicable (2021/22 or April 2022). Their analysis also involved identifying data outliers and requesting clarification or evidence where necessary.
The council’s approach to analysis of the fair cost of care provider returns and instructions to LaingBuisson has been to include as much data as possible with the minimum of adjustments. Providers spent considerable time and effort in calculating and submitting responses, so it is only reasonable for them to expect that we utilise the maximum amount of valid data possible. This approach means that there have not been significant amendments to the data submitted and the council considers the fair cost of care figures are reflective of the data submitted.
The following statistical analysis and approaches have been used:
- The council has calculated the fair cost of care figure by summing the median of each individual cost line. This approach was taken (as opposed to taking the median of the sub- totals or the overall totals) because it is felt that 14 responses is sufficient to enable calculation of representative line by line medians. This approach also means that the minimum amount of data has been excluded (see below description of the outliers analysis)
- The council worked with LaingBuisson to agree an appropriate statistical approach to identifying and excluding data outliers. This approach was to calculate the Double Median Absolute Deviation (DMAD) for each cost line. Please see the end of this report for a detailed explanation of this approach.
- The analysis excluded zero cost lines except where it is reasonable to expect that a provider could have zero cost on a particular line and still operate safely & effectively. For example, a zero line for CQC registration costs cannot be correct as every home care provider must be CQC registered so in this case any zero lines were excluded from the median calculations. The cost lines in which zero figures were included were:
| Cost line | Reason for including zero costs |
|---|---|
| Additional non-contact pay | This is a catch-all category for any non-contact pay related costs and providers don’t necessarily have to have any costs in this line to be viable |
| Travel cost (parking / vehicle lease) | This is separate from staff travel costs and relates to broader travel related costs, none of which are essential for viability |
| Assistive technology | Although the council is keen to work with providers to make best use of assistive technology, it is not yet a necessity for viability |
| Central / head office costs | For smaller providers without a head office these types of costs would not be incurred |
| Other additional costs | This is a catch-all category for any costs not accounted for in any other cost line and so providers could reasonably provide a zero cost here if everything had already been covered. |
- The fair cost of care data was collected in 2022/23 financial year and providers were asked to ensure it was from this current year. Providers were asked to do this by either giving 22/23 actual costs or uplifting 21/22 data by an amount which could be evidenced and justified. For all 14 responses, these instructions were followed and it was not necessary for the council to apply any of its own uplift calculations or to adapt those from providers.
Although all of the above efforts have enabled a more representative figure to be calculated, the council has not been able to fully verify each cost line, nor to validate that the overall figures from providers represent actual costs incurred. As part of the exercise the council asked providers to state how much of their care provision was purchased by self-funders. There was a correlation between a provider having more self-funders and submitting a higher rate however in most cases these self- funder rates were not high enough to explain how the FCC rate was so much higher than is currently paid by the council. An example of this discrepancy is given below, it does not use actual figures from any specific provider to avoid any commercial sensitivity issues but is representative of a common issue:
- Provider A submitted a FCC response totalling £25 per hour
- The council currently pays £20 per hour to this provider and commissions 75% of all care hours they deliver
- The provider stated that they charge self-funders £7 per hour more to self-funders (who represent 25% of all care hours delivered)
- A weighted average using the above figures gives a rate of £21.75 per hour. This is significantly less than the £25 per hour FCC figure and would mean that the provider was losing £3.25 on average for each care hour provided. The logical conclusion to this is that the provider cannot be currently viable, however that is evidently not true as they are currently providing services.
This type of situation arose for most provider responses, and it was not possible to obtain a satisfiable explanation for the discrepancy. There are likely to be a range of reasons for this and it is not necessarily the case that providers have made any deliberate attempt to artificially increase the figures. This uncertainty means that the council has concerns about the validity of the overall figures and is minded to put less weight on them than other contextual factors considered when setting fees. These other factors include the rates currently paid, the resultant impact on service quality and market ability to meet current levels of demand for home care.
Return on Operations
For home care providers, the council’s approach to deciding on a fair Return on Operations is to use the median average of 7.7% from the exercise.
Fair cost of care summary tables
The full line by line breakdown of the fair cost of care figures including the first/third quartiles and count of observations can be found at the end of this report.
The summary figures for home care are below and based on full responses from 12 providers and two partial responses.
| Rate | Median | Lower quartile | Upper quartile |
|---|---|---|---|
| FCC Rates | £23.49 | £21.06 | £32.91 |
The table below shows the lower quartile/median/upper quartile of number of appointments per service per week by visit length.
| Visit length | Median | Lower quartile | Upper quartile |
|---|---|---|---|
| 15 minutes | 79 | 59 | 82 |
| 30 minutes | 304 | 87 | 551 |
| 45 minutes | 66 | 6 | 151 |
| 60 minutes | 105 | 46 | 243 |
The table below shows the cost per visit for each of 15, 30, 45 and 60 minute visits after accounting for relative travel time.
| Visit length | Median cost |
|---|---|
| 15 minutes | £6.63 |
| 30 minutes | £12.25 |
| 45 minutes | £17.87 |
| 60 minutes | £23.49 |
Accounting for inflation
These Fair Cost of Care figures will be uplifted to account for inflation each April by considering a range of factors including the percentage increase from National Living Wage as well as the rate of inflation. Other factors which will be taken into account include affordability, approvals in line with the council’s budget setting process and any other factors which might have a material effect on the cost of care (both inflationary or deflationary).
Template format of exercise
The LGA commissioned ARCC tool was used to collect fair cost of care data from home care providers. In addition to this the council also asked providers to complete a short survey detailing the number of individuals self-funding their care and the hours associated with this. This information was used to better understand the self-funder market and inform planning for the care cap and change to charging thresholds.
Explanation of Double Median Absolute Deviation (DMAD) – outlier statistical analysis
Median Absolute Deviation (MAD) is calculated by finding the absolute difference between each validated data point and the validated sample median, and then calculating the median of these absolute differences.
Statistical analysis showed that the data suffered from a highly asymmetric distribution across all categories therefore using a singular Median Absolute Deviation value would mean disregarding the asymmetry in the distribution and produce unreliable results. For this reason, we opted for an enhanced method called “Double MAD”.
The premise underlying this method are similar to the classic version, with the only difference being the calculation of two Median Absolute Deviations: (1) the median absolute deviation from the median of all points less than or equal to the median and (2) the median absolute deviation from the median of all points greater than or equal to the median. This allows us to set pertinent outlier thresholds taking into account skewness in the data sample. Finally, for each cost line, we have defined as an outlier any data point which is more than 2 X DMAD above or below the median. All such outliers have been excluded from the calculation of median costs.
We have made one exception to the general outlier exclusion rule described above. It relates to the treatment of outliers in the Total Business Costs line. These back-office costs were highly variable and presented several anomalies including staff doubling up as care workers and as back-office staff members, leading to possible double counting, and back-office staff being used to support other business lines, leading to possible overstatement of costs. We also noted that toolkit submissions for back-office costs stood out as being substantially higher than the benchmark cost for ‘Running the business’ within the Home Care Association’s pro forma minimum cost structure. The balance of evidence is that many of the toolkit submissions did overstate Total Business Costs. Consequently, for this cost category, the outlier exclusion method has been customised to “Median – 3 X DMAD” for the lower boundary and “Median + 1 X DMAD” for the upper boundary. This has the effect of restricting the acceptable range of values above the median. The reasoning for adjusting the lower threshold is that the standard “Median – 2 x DMAD” would mean figures being excluded which were similar to the accepted values from the Homecare Association benchmark of £3.02 per hour for back-office staff.
Full table from Annex A, section 3
Including count of observations, first and third quartile
| Cost of care exercise results | Median | 1st Quartile | 3rd Quartile | Count of observations |
|---|---|---|---|---|
| Total careworker costs | £16.00 | £15.43 | £20.26 | - |
| Direct care | £11.17 | £10.82 | £11.99 | 13 |
| Travel time | £0.31 | £0.92 | £1.93 | 14 |
| Mileage | £0.61 | £0.40 | £1.04 | 12 |
| PPE | £0.29 | £0.25 | £0.72 | 13 |
| Training (staff time) | £0.35 | £0.20 | £0.58 | 13 |
| Holiday | £1.60 | £1.52 | £1.66 | 10 |
| Additional non contact pay costs | £0.00 | £0.00 | £0.22 | 14 |
| Sickness/maternity and paternity pay | £0.30 | £0.25 | £0.53 | 12 |
| Notice/suspension pay | £0.05 | £0.04 | £0.11 | 4 |
| NI (direct care hours) | £0.93 | £0.73 | £1.06 | 11 |
| Pension (direct care hours) | £0.39 | £0.30 | £0.42 | 12 |
| Total business costs | £5.81 | £4.12 | £10.30 | - |
| Back office staff | £4.17 | £3.20 | £6.77 | 11 |
| Travel costs (parking/vehicle lease et cetera) | £0.00 | £0.00 | £0.02 | 14 |
| Rent/rates/utilities | £0.60 | £0.33 | £0.96 | 13 |
| Recruitment/DBS | £0.11 | £0.08 | £0.44 | 12 |
| Training (third party) | £0.09 | £0.03 | £0.20 | 12 |
| IT (hardware, software CRM, ECM) | £0.22 | £0.12 | £0.37 | 12 |
| Telephony | £0.07 | £0.06 | £0.10 | 9 |
| Stationery/postage | £0.03 | £0.01 | £0.06 | 12 |
| Insurance | £0.11 | £0.08 | £0.15 | 11 |
| Legal/finance/professional fees | £0.12 | £0.06 | £0.17 | 11 |
| Marketing | £0.11 | £0.04 | £0.24 | 13 |
| Audit and compliance | £0.04 | £0.01 | £0.13 | 11 |
| Uniforms and other consumables | £0.04 | £0.02 | £0.06 | 9 |
| Assistive technology | £0.00 | £0.00 | £0.05 | 14 |
| Central/head office recharges | £0.00 | £0.00 | £0.24 | 14 |
| Other overheads | £0.00 | £0.00 | £0.23 | 14 |
| CQC fees | £0.10 | £0.08 | £0.11 | 10 |
| Total Return on Operations | £1.68 | £1.51 | £2.35 | - |
| TOTAL | £23.49 | £21.06 | £32.91 | - |
