Data quality is measured on submission of annual data by local authorities, and processes are followed to try and improve quality of data submitted. The Data Quality Statement gives further information, and presented below is an overview of the key data quality issues impacting on the 2020-21 DoLS data.
Summary measures indicate that the data submitted was valid and complete to a high degree, with the final returns yielding a very high national completeness and validity score (99.85%). The annex tables that accompany this publication also includes some data integrity checks. There are 11 data integrity checks that are carried out across each record in each local authority return. In 2020-21 there were approximately 15,500 data cells with validation breaches, which is comparable to previous years.
Missing Data
Due to a serious cyber-attack, Hackney Council (called London Borough of Hackney in the data tables) has been unable to submit 2020-21 DoLS data to NHS Digital.
In order to present England level statistics that can be compared to previous years, NHS Digital have calculated a range of estimates based on different scenarios to account for the missing data. These can be summarised as:
- add 2019-20 Hackney data to 2020-21 totals
- uplift 2019-20 Hackney data by a factor equal to the national percentage increase between 2019-20 and 2020-21, and add to 2020-21 totals
- uplift 2019-20 Hackney data by a factor equal to the regional percentage increase between 2019-20 and 2020-21, and add to 2020-21 totals
- uplift 2019-20 Hackney data by a factor equal to the nearest neighbour* percentage increase between 2019-20 and 2020-21, and add to 2020-21 totals
These estimates have been applied to the key measures only, giving a range of:
- between 256,580 and 256,610 total DoLS applications received
- between 245,995 and 246,045 total DoLS applications completed
- between 119,735 and 119,745 total DoLS applications not completed
Because these ranges of estimates are very narrow we have chosen to only present the first scenario (add 2019-20 Hackney data to 2020-21 totals) in the web page Key Facts and Key Measures data table.
For all other DoLS detailed data tables or the Adult Social Care Analytical Hub, the missing data has not been included and therefore England totals do not match those presented in the Key Facts or Key Measures. Where age-standardised population rates have been presented, the numerator (activity) does not include Hackney but the denominator (population) does include Hackney in order to maintain the same approach to standardisation. This means care must be taken when comparing figures over time.
The detailed data tables display London Borough of Hackney data as [x], denoting missing data.
* nearest neighbour is a statistical model developed by the Chartered Institute of Public Finance and Accountancy to facilitate benchmarking of local authorities.
Not completed applications
As in previous years, NHS Digital has used the data available to calculate the estimated volumes of applications not completed at year end. An estimated number of applications not completed can be created by taking last year’s reported number of applications not completed, adding the number of applications received and then subtracting the number of applications completed.
This calculation produces an estimated figure different to the equivalent figure reported by local authorities. NHS Digital has worked with local authorities to try to understand the reasons for this. From the local authorities who provided explanations for their variation in previous years one common theme was that this variation could be explained by the figures from the previous reporting period being higher than they should have been. This was due to several reasons, such as the previous return including applications that should have had a status of Not Granted but the application was still showing as in progress, or data quality issues caused by migrating data from older reporting systems, or through duplicate recording of applications. This year the pandemic is also cited as having an impact on local authorities’ ability to complete assessments and also to resolve inconsistencies in the data.
Variance in active authorisations
The DoLS data collection should include all applications that were active at any point in the year. During the analysis of the data, it was noticed that a number of local authorities had once again not submitted data for authorisations that had been started in the previous year but had finished in the current year. During the data validation process, many local authorities resolved this issue and included the missing authorisations for their final data return. This year some local authorities reported to NHS Digital that the variance was a result of the pandemic, and that the cases should have been closed with a retrospective date before April 2020 but the local authorities were not notified of the change in circumstances until later. This situation will always exist to some degree regardless of the pandemic and can help to explain some of the variance.
As a result, 17 local authorities are showing significantly lower numbers of active authorisations on 1 April 2020, the opening date of the 2020-21 collection, compared with the previous day, 31 March 2020, the closing date of 2019-20 collection. Having these data allows us to accurately measure the actual duration of granted authorisations and also measure the number of authorisations in place throughout the year. NHS Digital will continue to provide the guidance document and encourage all local authorities to review this document each year to ensure they are including the correct records.
Incomplete Actual End Dates
During the data validation process, NHS Digital identify where the Planned End Date of the authorisation was during the current reporting period however the Actual End Date had been left blank, indicating the unlikely scenario that the authorisation was still in place. This year NHS Digital embedded this check into the automated data validation tool used by local authorities. This has reduced the number of these cases by half from approximately 3,300 last year to 1,200 this year. In some cases local authorities told us this is a deliberate approach to allow ongoing monitoring via case management systems. The issue has the effect of inflating the figure reported in Table 6 in the Time Frames data tables, which is the number of authorisations in place on 31st March 2021.
Start Date of Authorisation recorded earlier than Application Sign-off Date
There were approximately 5,100 authorisations across 16 local authorities where the Start Date of Authorisation had been recorded earlier than the Application Sign-off Date. Some local authorities provided the explanation that this was an accurate reflection of local practice.
Some local authorities gave comments in their data return to explain errors, changes or issues with their data. NHS Digital would like to thank these local authorities for their transparency. The table below summarises specific key issues explained by local authorities that are not covered by the general comments above.
Local Authority |
Table Affected - reference |
Table Affected - details |
Local Authority comment |
114 - Stockton on Tees |
Applications Data Tables – Table 1 |
Comparison of Applications Received between current and previous years |
Stockton-on-Tees advised that although the data return indicates there has been an increase in applications this year, this is not actually the case and there has been a slight decrease. This is because this year’s return correctly includes applications where no assessments were completed, whereas they have not been included in previous years. |
114 - Stockton on Tees |
Timeframes Data Tables - Table 1 |
Longest and mean duration from receiving to both completing and last assessment |
Stockton advised us that there was an input error on one of their applications, which recorded the date received as January 2020 when actually it was received in January 2021. This means that the mean duration between application and authorisation (and also last assessment) is reported as being greater than it was for the reporting period. |
205 - Doncaster |
All |
All |
Doncaster advised us that due to a change of case management system a partial return has been submitted covering April-Dec 2020. |
209 - Bradford |
Demographics Data Tables – Table 1 |
Count of Ethnicity and Sexual Orientation recorded as ‘Not Known’ |
Bradford Council advised they are aware that a gap remains with some demographic information (also highlighted last year). Efforts continue to be made to redress this. |
310 - Stockport |
Applications Data Tables – Table 9 |
Not Completed applications |
Due to a change in client data management system Stockport have identified an error in the number of not completed applications. The correct figure is approximately double of the reported figure. |
705 – Hammersmith & Fulham |
Applications Data Tables – Table 4, Table 5
Timeframes Data Tables – Table 1, Table 2 |
Completed applications, Not Granted applications |
There is likely to be an underreporting of completed applications where they were not granted, due to incomplete application sign-off dates as a number of applications were withdrawn before assessment and no date was recorded. |
719 - Brent |
Applications Data Tables – Table 9 |
Not Completed applications – comparison between current and previous year |
Brent advised there was an over-reporting on last year’s data return in relation to not completed applications. This year’s return now shows the correct position. |