NHS Digital Data Release Register - reformatted

Market And Opinion Research International Limited projects

M&ORI Ltd Study of ASCS and SACE datasets for the Representativeness of Adult Social Care Surveys

30 data files in total were disseminated unsafely (information about files used safely is missing for TRE/"system access" projects).

M&ORI Ltd Study of ASCS and SACE datasets for the Representativeness of Adult Social Care Surveys — DARS-NIC-349486-Y3C3L

Type of data: information not disclosed for TRE projects

Opt outs honoured: No - data flow is not identifiable, Anonymised - ICO Code Compliant, No (Does not include the flow of confidential data)

Legal basis: Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 s261(2)(b)(ii)

Purposes: Yes (Commercial)

Sensitive: Sensitive

When:DSA runs 2020-05-11 — 2021-05-10 2020.08 — 2020.08.

Access method: One-Off

Data-controller type: MARKET AND OPINION RESEARCH INTERNATIONAL LIMITED

Sublicensing allowed: No

Datasets:

Personal Social Services Adult Social Care Survey
Personal Social Services Survey of Adult Carers
Personal Social Services - Adult Social Care Survey (ASCS)
Personal Social Services - Survey of Adult Carers in England (SACE)

Objectives:

The Personal Social Services Adult Social Care Survey (ASCS) and the Personal Social Services Survey of Adult Carers in England (SACE) are important sources for the Adult Social Care Outcomes Framework (ASCOF) outcome measures, used heavily by the Department for Health and Social Care (DHSC) to inform policy development and service delivery.

The Personal Social Services Adult Social Care Survey (ASCS) is an annual national survey, conducted by Councils with Adult Social Services Responsibilities (CASSRs). It gathers information from a sample of services users; aged 18 and over; in receipt of long-term support services funded or managed by social services. The ASCS seeks to learn more about how effectively services are helping service users to live safely and independently in their own homes, and the impact that these services have on their quality of life. The ASCS data is collected via postal survey and it has been ongoing since the financial year 2010/11.

The Personal Social Services Survey of Adult Carers (SACE) is a biennial survey which was last conducted in 2018/19. It is conducted by CASSRs, and covers carers aged 18 or over, caring for a person aged 18 or over, who have been assessed or reviewed by social services (or where the cared-for person has received respite or another form of carer support) during the 12 months prior to the sample being extracted. The SACE seeks carers’ opinions on a number of topics that are considered to be indicative of a balanced life alongside their caring role. The SACE data is collected via postal survey and it has been ongoing since the financial year 2009/10 (main stage only since 2012).

Questions have been raised over the representative quality of the users who take part in the ASCS, owing to low response and because only those who have mental capacity are eligible to take part. The survey also currently excludes short–term users which represents a gap in user experience. Previous work has identified a range of barriers to taking part in the surveys. Furthermore, whilst the surveys have been running for some time, they have not taken advantage of the range of innovations that have occurred in survey methodology, design and delivery in recent years. Research is required to review the representative quality of both surveys and to identify ways in which the views of under-represented groups could be captured by local authorities.

This request for data is being made by Market and Opinion Research International Limited (M&ORI Ltd) as part of a research project funded by the National Institute for Health Research (NIHR) on behalf of the Department of Health and Social Care [Ref. NIHR200692]. The aim of the research is to explore and understand which groups are under-represented in the Personal Social Services Adult Social Care Survey (ASCS) and the Personal Social Services Survey of Adult Carers in England (SACE) data, identify potential causes, the impact these have on survey estimates, and suggested approaches for how this under-representation can be addressed.

The NIHR research funding started in May 2019 and runs to July 2020 so NHS Digital data are required in order to carry out the analysis as part of this project. M&ORI Ltd have discussed the possibility of an extension with DHSC so that a complete analysis can be conducted should the data be received too late to complete by July 2020.

The focus of this data sharing application is to fulfil the secondary analysis task, which involves identifying the under-represented groups. The first stage of assessing the representative quality of the samples for the surveys will be to quantify the impact of the exclusion of sample members, described above, who would otherwise have been eligible for the survey. In other words, to assess how representative the sample selected is. M&ORI Ltd would expect a major reason for exclusion to be users who lack the mental capacity to consent to the survey, and the proportion excluded could be estimated. M&ORI Ltd therefore need data about the characteristics of the eligible population at the time of sampling as well as the characteristics of those invited to take part (whether responding or not). The second stage is to explore whether those who respond are representative of the issued sample.

This secondary analysis element is part of the whole project funded under NIHR200692. Other elements include methods review, literature review, online survey of local authorities, interviews with stakeholders, local authorities and service users and carers. These elements are already in progress and the secondary analysis will not result in any decisions about who to interview or how to carry out those stages. The secondary analysis is the main way to identify under-represented groups. The other elements allow M&ORI Ltd to explore the reasons and how to address the issue. All elements of this work are being carried out at M&ORI Ltd. M&ORI Ltd are collaborating with Kings College London on the literature review and plans for interviews but Kings College London are not involved in the secondary analysis and will not have access to any data.

Although each phase of the work is linked in having a common purpose, the data requested from NHS Digital will not be used for deciding which local authorities to interview or survey as that has already been done or will have been done by the time the data from NHS Digital are received. The data from NHS Digital will be pseudonymised record level and will not be used to identify individuals. Although M&ORI Ltd will be speaking to individual service users and carers it will not be possible for M&ORI Ltd to link anything the service users tell them with individual information within the data M&ORI Ltd have received from NHS Digital. M&ORI Ltd will have no means to know whether they are interviewing anyone whose data is included in the ASCS or SACE data M&ORI Ltd already hold.

M&ORI Ltd require unsuppressed pseudonymised data from the ASCS and SACE with more detailed information on ethnicity, age and religion and primary reason for support than is available in the public data. The reason M&ORI Ltd need unsuppressed data is that they cannot explore under-representation unless they know the characteristics of the eligible population, the issued sample and the responding sample. If data are suppressed then M&ORI Ltd cannot fully understand the under-representation of particular groups. The reason M&ORI Ltd needs data on ethnicity, detailed age bands, religion and primary reason for support is that the public data only shows 18-64 and 65+ and White British/ non-White British categorisations. There are likely to be very different patterns in response for subgroups within the working age and older adults and within the broad Non-White British category. Ethnic disparities in response cannot be explored without having more detailed information and thus research questions raised by DHSC could not be answered without these fields.

M&ORI Ltd recognise that when unsuppressed data are provided with local authority identifiers, more detailed ethnic group than non-white and more detailed age than 65+ the risk that an individual could be identified from the data is increased. However, M&ORI Ltd will not have access to any other information with which they could link the data and so identify who individuals are. Although theoretically M&ORI Ltd could match information given by individuals in the service user and carer interviews (20 interviews in total being conducted under another area of the same project) with the data in the survey results on the basis of their answers or demographic characteristics, in practice this would be impossible because the data will have been collected a year apart and these two elements of the project are separate so M&ORI Ltd would not collect details which would specifically enable them to do this. M&ORI Ltd will store the information about individuals they interview in a separate location from the data from NHS Digital.

DATA MINIMISATION
M&ORI Ltd are requesting data from England only, which is the full coverage of the surveys. The analysis requires a sufficiently large sample size for subgroup analysis. In addition, the response patterns and ethnic, age and support reason mix vary between local authorities. M&ORI Ltd need data from across all local authorities as well as the Councils with Adult Social Services Responsibilities (CASSR) identifier to fully understand the issue of under-representation.

The fields M&ORI Ltd are requesting from ASCS and SACE need to be for both responding and non-responding cases where available (for non-responders this will be information held by LA, not survey answers). NHS Digital's Adult Social Care Statistics team have confirmed they can provide information on non-responders.

This is how the publicly available data are presented. M&ORI Ltd also need a summary of the eligible population against which they can compare the issued and responding sample. This will be gender, age, PSR and support setting (cross tabulated from 2015-16 onwards and single variable frequencies for 2014-15) by local authority.

M&ORI Ltd have taken advice from NHS Digital's Adult Social Care Statistics team and are requesting only data from 2014-15 onwards. Data requested:

Adult Social Care Survey (ASCS): 2014-15, 2016-17, 2017-18, 2018-19
Survey of Adult Carers in England (SACE): 2014-15, 2016-17, 2018-19

M&ORI Ltd require multiple years rather than the most recent year due to the fact that survey response has been declining over time and M&ORI Ltd need to explore the extent to which under-representation has been made worse as response declines, and to question whether declining response is found equally among all groups.

Not all financial variables have been requested; only those directly relevant to the analysis in order to look at whether service users funded in certain ways are under-represented. Only three of these fields have been selected.

COMPARABLE DATA
In respect of the data sources for the three groups outlined (those responding[1], non-responders[2] and those eligible for the survey[3]) the ASCS and SACE the data on age group, gender, support setting, primary support reason are drawn from the administrative data held by local authorities which form the basis of the sample. This means that the data source on these characteristics are the same for all groups [1,2, and 3] as it is the information held by the local authorities. Data on ethnicity is also administrative data but only available for groups 1 and 2. Age, gender and ethnicity are only asked in the survey if the information is missing or the local authority believes it to be unreliable.

The data source for group 2 is the ASCS and SACE sample, which include the administrative data held about them. M&ORI Ltd have requested in this agreement these variables for group 2 which will have the same format as the same variable for the respondents (group 1).

The data source for group 3 is the administrative data held about the eligible population; data held in the SALT returns (although the data covers a slightly different period, it should be still helpful to estimate totals of service users). The published SALT return data has quite broad groups (e.g. 18-64/ 65+) but M&ORI Ltd have also requested in this agreement summary data about the eligible population using more detailed bandings. Ethnicity is not available for the eligible population so the analysis will need to be confined to age, gender, primary support reason and support setting. M&ORI Ltd have been working with NHS Digital’s business team preparing the data and they are aware of the need to provide summary data on these characteristics on the eligible population in a way which is comparable to the data provided for the issued sample. The data will be summary data but based on the same administrative data held by councils so the source variables are the same as the group 1 and 2.

M&ORI Ltd are therefore confident that the data on these key variables come from the same source and are comparable for groups 1, 2 and 3.

In addition to using the eligible population data from NHS Digital, for the carer population it will be possible for M&ORI Ltd to compare the profile of the issued and responding sample with data on carers from other sources. Both the Health Survey for England (HSE) and the Labour Force Survey (LFS) collect information on caring responsibilities. This gives a wider population of carers, not just those in touch with the local authority and so this analysis will serve a slightly different purpose, to explore the extent to which the issued and responding sample are representative of a wider group of carers. The HSE provides information on age, gender and ethnicity as well as the nature of support provided. LFS provides information on age, gender and ethnicity. M&ORI Ltd will ensure that tey prepare the data in such a way that they are as comparable as possible (ensuring age bands are consistent, adjusting for any differences in the gender question related to the inclusion of other categories in some but not all data sources). The ethnicity question on HSE and LFS is the same as that on the ASCS and SACE survey when asked. If this differs from the administrative data on ethnicity held by local authorities M&ORI Ltd will manage and group the data before analysis to ensure M&ORI Ltd are comparing comparable groups.

M&ORI Ltd can analyse the publicly available data but this tells us very little about the nature of the representation since the age, ethnicity and support reason variables are too broad and it lacks other information such as wider health information and financial support. Using the broad data from the ASCS and SACE provided publicly does not give enough granularity to fully explore the issue of under-representation in the survey as only 18-64/ 65+ and White/ BAME categories are available. Tabular data from NHS Digital is insufficient as M&ORI Ltd need to explore how people with different combinations of characteristics are under-represented (e.g. particular age group, ethnicity, support combinations, funding arrangements). M&ORI Ltd also need to understand groups excluded from the samples which is why they need information on the eligible population. For example, M&ORI Ltd need to understand the extent of the exclusion of those with memory and cognitive support needs resulting from the rules about excluding those who lack mental capacity from the survey sample.

M&ORI Ltd cannot fully meet the project objectives they have set without access to the data requested in the application.

DATA CONTROLLER AND PROCESSOR
Market and Opinion Research International Limited (M&ORI Ltd) is a specialist research agency, commonly known as “Ipsos MORI”. Ipsos MORI is part of the Ipsos worldwide group of companies (the parent headquarters located in France), and a member of the Market Research Society.

Ipsos MORI is formed of a number of legal entities including Market and Opinion Research International Limited (M&ORI Ltd) and Ipsos MORI UK Limited. It operates as a single business. The National Institute for Health Research (NIHR) funded project which this data request forms part of is being carried out by the Public Affairs business which sits within the M&ORI Ltd legal entity (which is the legal entity through which all our work for the public sector is delivered), not Ipsos MORI UK Ltd.

For the purpose of this Data Sharing Agreement:
- The Data Controller and Data Processor will be known as Market and Opinion Research International Limited (M&ORI Ltd).
- Ipsos MORI UK Limited will not be involved in the work described within this agreement.
- All staff involved in processing NHS Digital data under this agreement will be employees of M&ORI Ltd (employment contracts are with M&ORI Ltd).
- All NHS Digital data under this agreement will not be processed, stored or sent outside of the UK.

Kings College London (KCL) is involved in the wider project through the literature review and advising on research with ‘hard to reach groups’, but KCL are not dictating the method of data processing and will not have access to the data. Therefore KCL are not listed as a Data Controller or Data Processor. KCL's only involvement in the secondary analysis will be reviewing the summary outputs - with no record-level data - in order to assist in the development of implications and recommendations.

The research has been funded by National Institute for Health Research (NIHR) on behalf of the Department of Health and Social Care (DHSC). NIHR have commissioned M&ORI Ltd to carry out this analysis. DHSC and NIHR will not have access to the data. Only M&ORI Ltd will store and have access to these data.

DHSC and NIHR will receive a report of findings which does not identify individuals. Any data outputs will contain only aggregated data suppressed according to the data suppression rules.

LEGAL BASIS FOR PROCESSING
Data processing is conducted under GDPR Article 6(1)(f) Legitimate interests and M&ORI Ltd shared a copy of its Legitimate Interests Assessment with NHS Digital, who have assessed it against the ICO’s checklist and are content that the requirements are met.

Health data is Special Category data to be processed under GDPR Article 9(2)(j) (Archiving, research and statistics (with a basis in law)). The data are being used purely for statistical purposes in order to understand under-representation on the ASCS and SACE and explore the impact. This will assist DHSC and NIHR in improving the representative quality of these surveys. The results of this will be made available in the public domain (in a way which could never identify individuals).

If an individual has a unique combination of age, ethnicity, gender, primary support need within a local authority then an individual can be identified as being unique or one of a very small number with that combination in the data. M&ORI Ltd will know their unique combination of characteristics in the issued sample but they will not know whether they are unique in the local authority since not all eligible long term service users are invited to take part. Furthermore, the M&ORI Ltd research team will not have access to any other information which would allow them to identify who that person is. M&ORI Ltd will have no access to any information such as name, date of birth, or contact details of any kind.

There is no harm to the public by dissemination of the data from NHS Digital or resulting analysis output reports. There will be an advantage to the public as the findings will identify which groups are under-represented in surveys carried out by local authorities about the experiences of social care service users and carers. This will assist NHS Digital in addressing this so that in future the surveys better represent these groups. If the surveys improve the extent to which they represent the populations of interest, local authorities will be better able to assess the effectiveness of their services and plan improvements.

Expected Benefits:

Local authorities and NHS Digital can use the findings from the ASCS and SACE review to benchmark against England and regional results, to assess the extent to which service users are satisfied with the services they receive and the outcomes of services for their quality of life. These are used to inform policy and service developments and assess the effectiveness of existing services. Their ability to do this relies on the quality of the statistics. If certain groups are under-represented in the surveys then this may affect the estimates and result in misleading findings. This research project for which M&ORI Ltd are making the data request is important because it will lead to an understanding of under-representation and the impact on survey estimates. This will allow NHS Digital and local authorities to put in measures to address this – through weighting, revised protocols or improved efficiency in the use of existing tools. If this results in better data which informs service design and delivery, service users and carers will benefit.

The ASCS is carried out annually and SACE every two years. Service users and carers are asked to complete the surveys and local authorities spend a considerable amount of time and resource on running them. Taxpayers’ money is spent on collecting and analysing these data. It is therefore in the public interest that research is carried out into how well the surveys represent the populations of interest and that this is made public. Only through dissemination will it be possible to address the causes of under-representation, since NHS Digital and Local Authorities need to understand the issues before they can address them.

The resultant report for DHSC, written by M&ORI Ltd, will also be available to NHS Digital and local authorities. It will include recommendations for changes to the survey method, advice on weighting and recommendations about how to ensure existing protocols are followed more closely which will enable NHS Digital and local authorities to make improvements.

Other elements of the project are important in understanding issues such as the protocols for the surveys, how they are administered in practice, challenges faced by local authorities, best practice for gathering feedback from hard to reach groups and the experiences of service users and carers who complete the survey. However, the most important element of the project is the secondary data analysis since only through analysis of the enhanced data will it be possible to understand which groups are under-represented through the rules for exclusions and through differential response. Without this analysis, the information on which groups are under-represented will be very broad (e.g. two wide age groups and white/non-white) or based on analysis from a decade ago.

M&ORI Ltd will take responsibility for disseminating the results and recommendations as described under outputs. DHSC and NHS Digital will then have opportunity to use these findings. This is just one of several reviews (e.g. ASCOF Refresh) being undertaken by M&ORI Ltd who will ensure their findings and recommendations from these different streams of research will be brought together. DHSC will then be able to consider the implications as part of their wider consideration of statistics about social care.

The findings will feed into future ASCS and SACE data collection. It is likely that any changes would not be implemented until the 2022-23 surveys since DHSC and NHS Digital will need time to consider the recommendations, make provision for any changes and update their guidance. Any change to the survey will affect all 151 local authorities which carry out the survey as well as those who are invited to take part in the research (over 200,000 for the ASCS).

The results from the 2022-23 surveys will be published in June 2023 (Carers) and October 2023 (Service Users) and at this point local authorities will benefit from any improvements to the representative quality of the survey data which have resulted from this research.

Outputs:

The expected analysis will be written up into a report for consumption by DHSC to help DHSC assess the representative quality of the survey. The report will also be available for analysis by NHS Digital. M&ORI Ltd will also submit an article to a peer reviewed journal, in collaboration with Kings College London (who will only have access to the summary outputs, not the data). This paper will include findings from all elements of the project, not just the secondary analysis.

It is anticipated that M&ORI Ltd will give a presentation to DHSC and NHS Digital and they would also offer to present the findings at the NHS Digital hosted Social Services User Surveys Group (SSUSG) meeting. Membership includes Councils with Adult Social Services Responsibilities (CASSRs) (information specialists agreed by the LA Lead in association with the ADSS Standards and Performance Committee); NHS Information Centre for health and social care (social care and health colleagues); Department of Health (DH) (policy and analytical); Care Quality Commission (CQC) (formerly Commission for Social Care Inspection (CSCI)); Personal Social Services Research Unit (PSSRU) and relevant groups and service users (though at present there are no service user members). M&ORI Ltd would also offer to present to Association of Directors of Adult Social Care (ADASS) and would request the opportunity to present at the School for Social Care Research annual conference.

For both SSUSG and ADASS, the outputs will contain summary statistics. There will be no record level data sharing. All data outputs will be presented with strict adherence to the Disclosure Policies for ASCS and SACE. M&ORI Ltd would also take care to avoid identifying individual local authorities in what they write up and share publicly. M&ORI Ltd will not present any tables at a level of detail which would require suppression.

The funding for this research is conditional on providing outputs to DHSC. In addition to the report and peer reviewed article which have been agreed on M&ORI Ltd have also offered a toolkit to assist in putting the findings about the survey's representative nature and quality into practice. DHSC will decide on whether this is needed once they have seen the findings.

During the research M&ORI Ltd have involved NHS Digital, DHSC, local authorities, academics, a service user and carer in the advisory group. M&ORI Ltd's primary research has involved interviews with stakeholders and local authorities and M&ORI Ltd will share the report with those who M&ORI Ltd have involved in the research and who would like to receive the outputs. The report of the findings will be made publicly available on the M&ORI Ltd website. The journal will be published in an open access journal.

M&ORI Ltd's research partner, Kings College London, will host a dissemination seminar (subject to DHSC funding approval). At this seminar M&ORI Ltd would present findings from all elements of the research including the secondary analysis. Members of the research team have previously presented to the School for Social Care Research conference and to an ADASS meeting. This project has ADASS approval and M&ORI Ltd have kept in touch with ADASS during the research. Once M&ORI Ltd have findings, M&ORI Ltd would discuss whether there would be an opportunity to present to them. M&ORI Ltd also have previous links with the Local Government Association (LGA) from previous research projects and would also discuss the findings of this project with LGA.

All outputs of this research will be published with open access. Detailed analysis and tables, will not be published. The pseudonymised record level data will not be shared or made available outside the research team at M&ORI Ltd, and will be deleted 6 months after the end of the project so there will be no long-term data usage rights.

The final report is due in June and will be signed off in July 2020. M&ORI Ltd will still be able to include the findings in this final report even if analysis is not ready by the April draft report as they can work on the analysis while DHSC reviews the report containing other aspects of the research. M&ORI Ltd have also raised the possibility of an extension with DHSC to ensure they can use any data they are provided with through NHS Digital. Once this data sharing agreement is finalised DHSC will consider approving the extension.

Throughout M&ORI Ltd’s research, they have involved or engaged service users and carers and organisations representing under-represented groups through:
1. an advisory group with one user and two carer members. Its aim has been to ensure that M&ORI Ltd's research addresses salient issues, produces findings which stand up to scrutiny and develop recommendations which are appropriate and will engage those responsible for making changes. The service user is in the older age group. The two carers are from BAME groups.
2. a separate consultation with the user and carer members of the advisory group on the design of topic guides, and participant information.
3. In-depth interviews carried out with a range of stakeholders working for organisations representing under- or over-represented groups including BAME groups, LD groups, people suffering from mental health issues, etc.
4. fieldwork with service users and carers so they can share their views and experiences of barriers and facilitators to participating in the ASCS and SACE. The fieldwork was due to take place in March but had to be paused due to the Covid-19 outbreak. M&ORI Ltd are now discussing with DHSC whether the fieldwork could go ahead but be carried out remotely rather than face to face.

In addition, M&ORI Ltd are currently working with DHSC on an engagement plan to share the findings. One of the findings from other aspects of the research is the importance of engaging with under-represented and seldom heard communities and therefore sharing the findings and discussing the implications will be very important. Until M&ORI Ltd have done the analysis, M&ORI Ltd cannot be sure which groups are under-represented so will finalise their engagement plans once the analysis has been done. As mentioned above, M&ORI Ltd already have contacts with organisations representing or working with groups M&ORI Ltd believe may be under-represented, through their stakeholder interviews. M&ORI Ltd would ensure they invite them and other organisations working with under-represented groups M&ORI Ltd identify from the analysis to share in and discuss the findings of the research. Because of current uncertainty and the likely prolonged nature of the Covid-19 outbreak, we will consider sharing M&ORI Ltd's findings through a webinar or online workshop- rather than a face to face dissemination event as originally planned.

Processing:

M&ORI Ltd are requesting personal data including age, gender, types of service used, financial arrangements, and responses to survey questions about outcomes and experience of service use. M&ORI Ltd are requesting special category data including information about health and disability and ethnicity.

The data being requested is Adult Social Care Survey data and Survey of Adult Carers in England data for these years:

Adult Social Care Survey (ASCS): 2014-15, 2015-16, 2016-17, 2017-18, 2018-19 (5 data periods)
Survey of Adult Carers in England (SACE): 2014-15, 2016-17, 2018-19 (3 data periods)

The special category data requested:
• personal data revealing racial or ethnic origin;
• personal data revealing religious or philosophical beliefs;
• data concerning health;

The ASCS and SACE datasets cannot be linked, as they do not contain identifiers.

The pseudonymised record level data would flow from NHS Digital to M&ORI Ltd via a Secure File Transfer Service (SEFT) in one dissemination of data.

The data flowing from NHS Digital to M&ORI Ltd are the requested variables from the Adult Social Care Survey and Survey of Adult Carers in England for the years specified above. M&ORI Ltd are requesting pseudonymised record level data.

The data will be stored in a secure folder. The data may be manipulated to create derived variables necessary for analysis. Data analysis will take place using SPSS or Stata. M&ORI Ltd will work on a version which has data banded into derived variables where possible.

M&ORI Ltd will compare the NHS Digital publicly available 'issued cases' sample (appropriately weighted) against the profile of the population (i.e. the subset of fields requested in this agreement) for all issued cases to identify any characteristics that are associated with not being selected for the survey – in other words to assess the implications of any coverage error. M&ORI Ltd would use multivariate regression models to explore the main drivers of not responding to the survey once invited. The team will also use Chi-square automatic interaction detection (CHAID) analysis to identify groups with combinations of characteristics which are more or less likely to respond. M&ORI Ltd have used some of these approaches on the public data and would extend this to allow more granular analysis on the data they have requested.

This analysis will be repeated for multiple survey years to add power to the analyses, and also to explore how patterns of differential response and exclusion from the sample have shifted over time. They would then use predicted estimates from the models to explore how the under-representation of certain groups from the issued and responding samples impacts on the resulting survey estimates. The data would be destroyed using Blancco file shredding software and a record of destruction would be kept, with a Certificate of Data Destruction issued to NHS Digital when requested.

Substantive employees of M&ORI Ltd may access the secure folders on which the data are held when working remotely. They will log on using Forticlient using their M&ORI Ltd encrypted computer. The data will never be stored on a local device. The data will only be stored and processed in a secure folder with access restricted to authorised named individuals. Remote workers are only permitted to work within the defined territory of use set out in the agreement.

The data will be stored at M&ORI Ltd. There will be no attempt to re-identify individuals. There will be no attempt to link the data at an individual level. M&ORI Ltd may attach local authority level data such as region of the local authority to the data for the analysis but this will be limited to additional information about the local authority, not about the individual or about any groups of individuals within the local authority.

All data held by M&ORI Ltd is backed up on a regular basis to a second local disk based storage system under the control of a Backup solution, Iron Mountain UK Ltd. Iron Mountain UK Ltd do not access data held under this agreement as they only supply the building. Therefore, any access to the data held under this agreement would be considered a breach of the agreement. This includes granting of access to the database[s] containing the data

All organisations party to this agreement must comply with the Data Sharing Framework Contract, including requirements on the use (and purposes of that use) by "Personnel" (as defined within the Data Sharing Framework Contract i.e.: employees, agents and contractors of the Data Recipient who may have access to that data).