Data Quality Statement

Purpose

This data quality statement section details the steps taken to improve the quality of local authorities data collected through the Adult Social Care Workforce Data Set (ASC-WDS) as at September 2019. It highlights current data quality issues and summarises the implications for results already published in previous years.

This data quality section has used data collected by the NMDS-SC / ASC-WDS for the past eight years (from 2011). The ASC-WDS is managed by Skills for Care (SfC) on behalf of the Department of Health and Social Care (DHSC) and has been collecting information about social care providers and their staff since 2006. NHS Digital quality assures the data and publishes this annual publication, which has been prepared in conjunction with SfC.

Before 2011 the data source for this publication was NHS Information Centre’s (a predecessor organisation to NHS Digital) Social Services Staffing Collection (SSDS001). Following consultation across the social care sector it was decided that the NMDS-SC would replace the SSDS001 as the main source of adult social care staffing data from local authorities. 72 responses were received including 61 from local authorities, as part of social care collection consultations. Further information on the consultation can be found here.

The publication covers all staff employed, both directly and indirectly, by adult social services departments in local authorities as at 30^th September 2019. It does not cover staff employed wholly by the independent sector (private and voluntary) or by children’s social services departments (published separately by the Department for Education).

Each local authority was required to complete a defined subset of fields in the ASC-WDS during the period 9th September to 11th October 2019.

The ASC-WDS is split into two main sections: the establishment or team section and the individual worker level section.

For the establishment or team section each local authority split their workforce into teams, meaning each Care Quality Commission (CQC) registered location should be its own team. No further strict guidance was given due to the diverse range of business models operated by local authorities. For each establishment or team, local authorities completed;

details of the services provided,
the types of users of social care services that are cared for,
staffing numbers,
starters in the past twelve months,
leavers in the past twelve months,
vacancies on completion date,
service capacity and utilisation data.

For the individual worker level section, local authorities provided information regarding each worker’s demographics (e.g. gender, date of birth and ethnicity) and employment details (e.g. pay, working hours, sickness and employment status). From 2014, local authorities were also required to provide data regarding the qualifications held by their workers. Local authorities are required to complete worker level information on at least 90% of their workforce.

Local authorities are required to complete a number of mandatory fields in the ASC-WDS. In addition to the mandatory fields, some local authorities also completed other data items collected by the ASC-WDS. Some of these data items are used in this publication where the coverage was judged to be sufficient. These non-mandatory fields were:

Data derived from National Insurance Numbers in order to calculate headcount
Average hours and zero-hours contract (to assist with calculating Whole Time Equivalent Pay)

Relevance

The degree to which the statistical product meets user needs in both coverage and content

The information in this publication is provided from all 151 local authorities in England (prior to April 2019 the total number of local authorities submitting data was 152). The data are used by central government to monitor the impact of social care policy and by local government to assess performance in relation to their peers. The data are also available for use by researchers looking at local authority performance and by service users and the public to hold local authorities and the government to account. The data have also been used previously by the CQC for their Annual Performance Assessment (APA). Skills for Care will use the information in the publication to inform their work with local authorities and wider work on estimating the size and structure of the adult social care sector.

Accuracy and Reliability

The proximity between an estimate and the unknown true value

Action taken to safeguard data quality

Several methods of data validation were carried out. These data validations included validations at source, online reporting, post validations and follow up validations. These validations increase the quality and accuracy of the data reported and are summarised below.

Validation at source

As local authorities submitted data, a range of validation checks were carried out at source by the ASC-WDS collection system to prevent inaccurate data from being submitted.

These validations included:

Establishment name, address and contact details must be completed.
National Insurance numbers must be in the correct format.
Workers must be between 14 and 100 years old.
Hourly rates of pay less than £2.50 or annual salary less than £500 cannot be entered.
Sickness days must be between 0 and 365 (to the nearest half day).
Contracted hours per week cannot exceed 75 hours.

A range of cross validation checks between variables was also carried out:

An employee cannot have started in their current job before they started working in the sector or before their date of birth.
An employee must have been at least 14 years old when they started work in the sector or in their main job.
Employees flagged as ‘full time’ must have at least 24 contracted hours and workers flagged as ‘part time’ must have fewer than 40 contracted hours.
Cannot select zero-hours contract for a worker if they have entered a value greater than zero for contracted hours. Zero-hours contract workers must have additional hours completed (can be zero).
The total number of starters, leavers or vacancies cannot be more than three times the total number of staff.
Full time equivalent salaries (calculated by the system using hours and salary or hourly rate) must fall with acceptable ranges which are calculated separately by job role group.

These checks prevented incorrect data from being entered at source. Further manual data quality checks were also carried out and details of these checks are outlined in the following sections.

ASC-WDS online reporting

During the submission period each local authority was given access to a live report on the ASC-WDS system which detailed what they had submitted and highlighted any missing data or data quality issues with their return. Local authorities were encouraged to use this publication to highlight any problems or gaps in their return and then resubmit data as necessary.

Post validations

After each local authority finalised their data return, they then independently confirmed the total number of jobs at their local authority via the confirmation form. This process of confirming total job numbers was used to identify whether or not the data supplied in the ASC-WDS related to the whole local authority or whether the local authority provided a partial ASC-WDS return. For example, a local authority with 11 establishments or teams might only complete data for 10 of them. Without the confirmation form the number of workers for the missing establishment or team would not have been captured.

If the number of jobs at the local authority had increased or decreased by 5% or more from the previous year then the local authority was asked, via a free text box in the confirmation form, to provide details of the main reason(s) for these changes including any structural changes that have affected the total number of jobs or types of jobs within the local authority. Further detail can be found in the accompanying data quality tables.

Further validations were carried out on starters and leavers data collected on the confirmation form for each local authority for the 2014 collection onwards. Where this starters and leavers validation showed a discrepancy of 50 or more jobs, SfC contacted the local authority for further explanations as to the reasons for the discrepancy if they hadn’t already supplied this information. The additional validation work has provided further transparency regarding the movement of staff in and out of local authority adult social services departments.

SfC also performed manual data quality checks to ensure there were no issues that had not been resolved by the local authority. If data quality issues were found at this stage, within the submission period, the local authority was informed and asked to resubmit their data.

Follow up validations

The vast majority of data quality issues were picked up throughout the submission period by the checks and processes described above. After this period, if any issues were found within a submission, the associated local authorities were contacted to investigate. Depending on the nature of the problem, these local authorities were asked to resubmit their data, or the data were manually altered based on the advice received.

Coverage

In 2019 all 151 local authorities provided an ASC-WDS return, however the coverage for each individual data item varied. The number of local authorities contributing to the analysis of each data item is specified under each table or chart throughout the publication and is summarised in Table 1 below.

The figures in Table 1 show the number of local authorities contributing to the analysis of each metric. A local authority is only excluded from a metric when all of the related records supplied by the local authority cannot be included. For the majority of measures, the number of local authorities submitting data that can be used within analysis has increased since 2012.

Table 1: Number of local authorities included in each metric, 2019

Metric	2012	2013	2014	2015	2016	2017	2018	2019
Total Jobs	152	152	152	152	152	152	152	151
Headcount¹	98	122	105	99	103	100	102	97
Whole Time Equivalent	152	152	152	152	152	152	152	151
Total Starters and Leavers²	-	-	151	152	152	152	152	151
Starters³	128	141	147	151	151	152	152	151
Leavers³	128	141	147	150	152	151	151	151
Vacancies	97	122	121	133	145	149	151	150
Employment Status	152	152	152	152	152	152	152	151
Capacity and Utilisation⁴	124	131	129	126	122	135	136	136
Gender	152	152	152	152	152	152	152	151
Age	152	152	152	152	152	152	152	151
Ethnicity	152	152	152	152	152	152	152	151
Whole Time Equivalent Pay	152	152	152	152	152	152	152	151
Sickness Days⁵	149	141	146	146	147	149	151	149
Qualifications⁶	-	-	142	152	152	-	-	-
Social Care Qualification⁶	-	-	-	-	-	152	152	151
Zero hours contract	-	-	-	-	-	-	150	150

¹These fields were not specified as mandatory

²Only collected via the confirmation form from 2014 onwards

³Starters and leavers by job role were separated in 2015 to allow local authorities to be included if they had completed only one of the two data items

⁴In 2017, local authorities submitting capacity and utilisation data for any service have been included. Previously, only local authorities submitting capacity and utilisation data for one of the four services reported on were included.

⁵Sickness days methodology changed in 2018 and is now based on directly employed (permanent or temporary) workers. Figures for previous years have been updated.

⁶In 2017, the fields used to collect a worker's qualification data changed. Therefore 2017 onwards may not be directly comparable with previous years

Source: Skills for Care - ASC-WDS, Confirmation form, establishment / team and worker sections, 2019

Whilst Table 1 above showed the number of local authorities included in each metric, Table 2 below shows the proportion of workers.

Table 2: Percentage of workers included in each metric, 2019

Metric	2012	2013	2014	2015	2016	2017	2018	2019
Total Jobs	99%	100%	100%	100%	100%	100%	100%	100%
Headcount¹	81%	84%	82%	82%	80%	80%	81%	82%
Whole Time Equivalent	90%	92%	93%	94%	95%	94%	94%	98%
Total Starters and Leavers²	-	-	100%	100%	100%	100%	100%	100%
Starters³	90%	95%	98%	100%	100%	100%	100%	100%
Leavers³	90%	95%	98%	99%	100%	100%	100%	100%
Vacancies⁴	56%	72%	77%	85%	93%	95%	99%	99%
Employment Status	99%	100%	100%	100%	100%	100%	100%	100%
Gender	99%	99%	99%	98%	98%	98%	98%	98%
Age	99%	99%	98%	98%	98%	98%	98%	98%
Ethnicity	92%	92%	90%	89%	89%	88%	85%	84%
Whole Time Equivalent Pay⁴	90%	92%	93%	94%	95%	94%	94%	98%
Sickness Days⁵	-	-	-	-	-	-	98%	99%
Qualifications⁶	-	-	86%	84%	88%	-	-	-
Social Care Qualification6	-	-	-	-	-	82%	84%	79%

¹These fields were not specified as mandatory

²Only collected via the confirmation form from 2014 onwards

³Starters and leavers by job role were separated in 2015 to allow local authorities to be included if they had completed only one of the two data items

⁴The 2016 NMDS-SC publication understated the 2016 vacancy figure and overstated the whole time equivalent pay figures for this table. It has been corrected in this publication.

⁵Sickness days methodology changed in 2018 and is now based on directly employed (permanent or temporary) workers.

⁶In 2017, the fields used to collect a worker's qualification data changed. Therefore 2017 may not be directly comparable with previous years

Source: Skills for Care - ASC-WDS, Confirmation form, establishment / team and worker sections, 2019

The percentage of workers included in each metric has remained similar to 2019. There has therefore been little need for estimations within this reporting period, apart from the headcount calculations (this is discussed in detail later in this section).

For a more detailed summary of the individual completion rate of each local authority based on records submitted, please see Table 3 of the Data Quality tables accompanying this publication.

Data quality filters

Headcount calculations

For the headcount calculation, only local authorities with National Insurance (NI) numbers and date of births completed for more than 95% of their jobs were included in the analysis. The England headcount is therefore an estimate, based on the data from these local authorities.

This requirement is more stringent than for other measures in the publication because low coverage can cause underestimation of the number of people working multiple jobs. As coverage increases, workers with multiple jobs are more likely to be detected in the analysis.

In 2019, 97 local authorities met this criteria and so were able to be included in the headcount calculations.

For information on the local authorities that were included, please see Table 2 of the Data Quality tables accompanying this publication.

Whole Time Equivalent (WTE)

Whole time equivalent is used to measure the size of the workforce in terms of contracted hours worked as opposed to counts of jobs or people.

As a result, if the full time and part time distribution of the workforce were to change, this measure would still be comparable between years whereas the jobs and people measures may be misleading.

Workers on a zero-hours contract were included in this calculation if their working hours had been completed.

Whole Time Equivalents (WTEs) =

Contracted hours / 37

98% of all workers were included within WTE counts. All local authorities were included in the national WTE totals.

Starters and leavers by job role

Only local authorities with at least one starter were included in the starter figures and only local authorities with at least one leaver were included in the turnover figures. Local authorities with less than 15 employees are included as it is possible that these local authorities could have no starters or leavers during the year.

The ASC-WDS collects starters and leavers data for establishments or teams that were operating at September 2019. The confirmation form (introduced for the 2014 collection) collects starters and leavers data for establishments and teams that operated at any point over the year to September 2019. These data were also validated, at local authority level, against the change in the number of jobs between years to maximise data quality. The total confirmation form figures are used to calculate the overall starters and leavers rates.

These numbers are published alongside the submitted totals through ASC-WDS and can be seen in Table 3 of the Reference Tables.

Validations are carried out to compare the starters and leavers figures supplied in the confirmation form and in the ASC-WDS. The majority of differences can be explained by the ASC-WDS not gathering the starters or leavers figures from teams which were no longer operating at the time of submission.

All 151 local authorities provided their total number of starters and leavers for the year on their confirmation form and through ASC-WDS. Details of the local authorities included in calculations relating to starters and leavers can be seen within Table 3 of the Data Quality tables accompanying this publication.

Vacancy rates

Only establishments within local authorities where the ‘total vacancies’ equalled the sum of the number of vacancies recorded by job role are included within the vacancy rate calculations. The establishments with a mismatch in these areas would indicate potential data quality issues and are therefore excluded from vacancy calculations.

Local authorities with a missing ‘total vacancies’ figure are excluded from the overall vacancy calculations. In 2019, one local authority was excluded from the vacancy rate calculations.

Table 2 of the Data Quality tables accompanying this publication contains a list of local authorities which contributed towards the vacancy rate analysis.

Sickness

Local authorities where more than 90% or less than 10% of directly employed (permanent and temporary) workers had zero sickness days in the past 12 months are excluded from the analysis.

In 2019, 149 of 151 local authorities were included within the sickness calculations.

Table 2 of the Data Quality tables accompanying this publication contains a list of local authorities which contributed towards the sickness analysis.

Qualifications

In 2017, the fields used to collect a worker's qualification data changed slightly. In previous years, the local authority was required to supply every individual qualification that a worker had achieved from a pre-defined list of around 110 qualifications. Based on these individual responses, SfC calculated the highest level of relevant social care qualification the worker had achieved. This question was simplified in 2017, whereby local authorities were first required to select whether each worker had achieved a relevant social care qualification and, if so, the highest level of qualification from a pre-defined list ranging from Entry level to Level 8 or above (based on the governments definition of what each qualification level means.

Table 2 shows that this change did not seem to affect the amount of qualification data submitted to the ASC-WDS.

In 2019, all 151 local authorities provided social care qualification data for at least 90% of their workers.

Capacity and utilisation

136 local authorities are included within the capacity and utilisation data. Please note that these three selected main services are a subset of a larger list of 72 services. Local authorities are included in the capacity and utilisation data if one or more of the 72 services had completed capacity data, therefore it is possible that a local authority can be included in this analysis but not offer any of the three services shown in this publication.

The main three services are;

Care homes with nursing
Care only homes
Day care

Table 2 of the Data Quality tables accompanying this publication contains a list of which local authorities are included in the capacity and utilisation section of the publication.

Estimation

This year, all 151 local authorities in England provided an ASC-WDS return. However, the coverage for each individual data item varied. In previous iterations of this publication estimations were used to estimate a full national and regional figure if data items were missing, since 2016 this was deemed unnecessary for all metrics (except headcount) due to increased level of completeness across data items and an overall improvement in the quality of data submitted.

Levels of completeness for each data item by local authority can be found within Tables 2 and 3 of the Data Quality tables accompanying this publication.

Headcount (number of people working in adult social services departments) calculation methodology

Headcount makes the distinction between the number of local authority employed adult social services jobs and the number of people working in those jobs.

The ASC-WDS was designed to make this distinction by creating a unique reference number for each worker using a worker’s NI and their date of birth. If the same NI and date of birth combination appears more than once in the dataset it indicates that the same person has more than one job. Typically, this occurs when a person has two part-time jobs in different teams within the same local authority.

In order to estimate the national level headcount, a ratio of the number of jobs to number of people in each service area is obtained from the sample of included local authorities. This ratio is then applied to the national level number of total confirmed jobs.

Headcount =

Total number of jobs in service area / Average number of jobs per person in service area

The local authorities included within the overall headcounts are included within Table 2 of the Data Quality tables accompanying this publication.

Job confirmation numbers (total jobs)

In 2019 local authorities were asked, via the confirmation form, to provide reasons where there was more than a 5% difference in the number of jobs compared to the previous year. In addition to providing reasons for changes in total jobs, local authorities were also asked for any other comments they wanted to make, including any structural changes that might not have affected the total number of jobs or the types of jobs within the local authority (please refer to the data quality tables – reason for change sheet).

Timeliness and punctuality

Timeliness refers to the time gap between publication and the reference period. Punctuality refers to the gap between planned and actual publication dates

The data relates to September 2019 and the lag from the end of September 2019 to the publication of these data (27 February 2020) is under five months. These five months includes a month required by data suppliers to finalise their data for submission.

The date of the publication this year is similar to the date of the previous publication in 2019.

This publication has been released in line with the pre-announced publication date and is therefore deemed to be punctual.

Accessibility and clarity

Accessibility is the ease with which users are able to access the data, also reflecting the format in which the data are available and the availability of supporting information. Clarity refers to the quality and sufficiency of the metadata, illustrations and accompanying advice

The accompanying Reference Tables contain the aggregated data that are used to create all the tables and charts within the main publication report and a CSV file contains the number of jobs by main job role, main service group, job role group, service provision and employment status as well as starter, leaver and vacancy totals, and worker characteristics (demographic, full-time employment status, sickness and qualifications) for all local authorities who provided data.

Links to this and other NHS Digital adult social care publications are available.

There are no restrictions to access the published data. The data are rounded (see below for details).

Rounding

All figures have been rounded to the nearest 100 in the publication, excluding figures relating to gender, ethnicity and age which have been rounded to the nearest 5. All percentages have also been rounded to 1 decimal point.

Some additional steps are taken to exclude data from certain calculations where small numbers may lead to data being interpreted incorrectly.

When discussing pay data, time series information is only shown for job roles with 2000 or more workers in 2019.

Only local authorities with at least one starter or leaver are included in the respective starter and leaver rate calculation. Local authorities with less than 15 employees are included in both of these rates as it is not unlikely that these local authorities could have no starters or leavers during the year. Local authorities with a missing vacancy total are also excluded from the calculation vacancy rates.

Local authority data may be excluded from other calculations due to data not being submitted or not being of sufficient quality, as determined through the validation checks carried out on the data. Where this has occurred, these local authorities are identified via a footnote on the relevant reference table.

Coherence and comparability

Coherence is the degree to which data that are derived from different sources or methods, but refer to the same topic, are similar.

Comparability is the degree to which data can be compared over time and domain

Coherence

SfC publish annual adult social care workforce estimates and other reports on the adult social care sector and workforce which include the adult local authority workforce. Wherever possible the data will be taken directly from this publication. There may be slight differences for historic data in these publications however as SfC may revise data based on new information received.

SfC also make local authority data available on the ASC-WDS website as dashboards and on the data.gov website as a CSV file. These data may refer to different time periods to this publication and may not include some of the calculations and adjustments made within this publication. As such the data from these sources may not exactly match the data published in this publication.

There are no other national sources of local authority staffing data.

Comparability

NMDS-SC 2014, 2015, 2016, 2017, 2018 and ASC-WDS 2019

The vast majority of data from 2014 onwards are comparable, for detail prior to this date please see the 2014 data quality report.

All 151 local authorities have provided data and worker level data coverage has remained similar to 2018.

Local authorities were only required to complete a partial worker record for agency staff. As a result, it came to light that several local authorities had not included agency staff on their previous returns. In 2016 no local authorities have specifically mentioned not including agency staff, though some have mentioned that they had data quality issues which may have affected their return.

Qualifications data were required data items for the first time in 2014. In 2015 it was made clear that “no qualifications held” should be interpreted literally and those workers with qualifications but without any relevant to social care should be recorded as ‘any other qualification’. Qualification coverage increased from 84% in 2015 to 89% in 2016 (See table 2 for time series of qualification inclusion data ).

In 2017, the fields used to collect a workers qualification data changed. Therefore, the qualification data used from 2017 and onwards publication may not be directly comparable with previous year’s qualification data. However, this change did not seem to affect the amount of qualification data submitted to the ASC-WDS.

In 2014 and 2015, pay and ethnicity data were weighted to ensure any regional bias in the missing data did not affect the national totals and that reliable trends could be produced. Due to the overall coverage in both metrics from 2016, it was deemed unnecessary to apply weighting to these metrics. This should not affect comparability.

In 2019 the number of local authorities completing enough NI numbers to be included in the headcount figures decreased to 97 local authorities from 102 in 2018, 100 in 2017, 103 in 2016, 100 in 2015, 105 in 2014 and 122 local authorities in 2013. Through estimation, efforts have been made to ensure the figures are comparable between 2013 and 2019, however, they should be treated with some caution due to the relatively small sample sizes.

2019 saw the continuation of the methodology comprised in 2015 for calculation of WTE pay. Previously, only workers with hours worked were included in the calculations. Further investigation revealed the base used for these figures could be increased by allowing workers with hourly rates recorded but without hours data recorded to be included.

Comparing the results from the new methodology to the previous one revealed that this change did not substantially impact the final results at a national level and as a result the figures between years can be compared. No retrospective adjustments to previously published figures were required.

Trade-offs between output quality components

Trade-offs are the extent to which different aspects of quality are balanced against each other

Each reference table is based on the local authorities that pass the inclusion criteria for that reference table. Some of the tables and charts are therefore based on fewer local authorities or workers than others. The number of local authorities on which each analysis is based is given as a footnote below each reference table. The majority of analyses have not been adjusted for differing completion levels so there is an assumption that those local authorities that did provide data are a good representation of all.

Assessment of user needs and perceptions

The processes for finding out about users and uses, and their views on the statistical products

SfC run a quarterly ASC-WDS data user group which is attended by representatives from all sectors including local authorities. NHS Digital consult on data collections and run user feedback surveys on a regular basis.

User feedback on the format and content of the publication is also invited via a web form on the NHS Digital website.

Confidentiality, transparency and security

The procedures and policy used to ensure sound confidentiality, security and transparent practices

The Health and Social Care Act 2012 states that NHS Digital’s statutory data quality role is to assess the extent to which the data it collects meets applicable published standards and to publish the results of the assessments. In addition, NHS Digital may give advice or guidance on data quality relating to the collection, analysis, publication or other dissemination of data and information. NHS Digital is not responsible for the quality of the data it collects. That responsibility lies with the organisations producing the data. However, Skills for Care who manage the collection help data producers improve the quality of their data by sharing the results of the data quality assessments it undertakes and by providing advice and guidance when asked to do so.

Please see links below to relevant NHS Digital policies:

The data contained in this publication are National Statistics. The code of practice for official statistics is adhered to from collecting the data to publishing.

Statistical Governance Policy

Freedom of Information Process

Statement of Compliance with Pre-Release Order

Feedback

This publication is free to access via the NHS Digital website and the majority of users will access the publication without being known to the NHS Digital.

It is therefore important to try to understand how these users are using the statistics and also to gain feedback on how we can make the data more useful to them. We welcome feedback from publication users, ideally covering the following points:

How did you find out about this publication?

How useful did you find the content in this publication?

What type of organisation do you work for?

What did you use the publication for?

What information was the most useful?

Were you happy with the data quality?

To help us improve our publications, what changes would you like to see (for instance content or timing)?

Would you like to take part in future consultations on our publications?

Comments on this publication would be welcomed. Any questions concerning any data in this publication, or requests for further information, should be addressed using details on the contact us page

Internet explorer is no longer supported

Personal Social Services: Staff of Social Services Departments, England 2019

Page contents