Fifty-Third Report - Covid 19: supporting the vulnerable during lockdown

DHSC and NHS Digital took too long to identify all clinically extremely vulnerable people. Individuals were not formally eligible for the central support of food boxes and medicines delivery offered through the shielding programme until they were on the Shielded Patients List. NHS Digital used national hospital and GP data to identify clinically vulnerable people. However, it took over six weeks for the number of people on the Shielded Patients List to stabilise at 2.2 million people, with 900,000 people added between 18 April and 7 May. The time taken to add people was because of the need to work with GP IT system providers to design, build and gather GP data, which were on different systems to the readily available national data, and then to complete the next necessary step of GPs and hospitals using their clinical judgement to add and remove people from the list. NHS Digital believes that faster access to data in GP records would help. It also suggests that government invest in the digitisation of hospital records, noting that primary care data has been digitised, and is now a richer source of information than hospital data. Recommendation: Within six months, DHSC and NHS Digital should set out a detailed plan on how they will improve access to and join-up NHS data systems to ensure quick and secure access to all patient records.

Huge local variation strongly suggests that GPs were inconsistent when judging who was clinically extremely vulnerable and should therefore be advised to shield and be eligible for support. As well as NHS Digital using national data to identify clinically vulnerable people, GPs and hospital doctors were quite sensibly asked to review those listed, and, using their clinical judgement, add or remove people. The 6 Covid 19: supporting the vulnerable during lockdown list grew from 1.3 million to 2.2 million largely as a result of GPs adding people. However, the extent to which it grew varied hugely in different areas, with increases in those being added between 12 April and 15 May ranging from 15% to 352% by local authority. DHSC acknowledges that clinicians took different approaches to adding people. DHSC and NHS Digital believe they did everything possible to ensure consistency, where they identified possible over-inclusion or over-exclusion, they have worked with NHS England’s clinical directors to challenge some of the differences. DHSC tells us that NHSE&I and NHS Digital consider that ultimately additions were a decision for local clinicians. DHSC has also provided us with details of NHS Digital’s analysis of the variation as of 11 February 2021. Based on this analysis, DHSC concludes that the level of variation in how local clinicians added people to the list is acceptable. However, and despite the best intentions of all involved, it is not credible to assert that the same criteria and judgements were applied consistently in all parts of the country when the extent of local variation in numbers added was so vast. Recommendation: Within six months, DHSC and NHS Digital should provide to the Committee a detailed explanation for the local variation in growth for the shielded patient list between April and May 2020 including the extent it was due to appropriate clinical judgements and identify lessons for how to support a consistent clinical approach in future.

Government chose a centrally-directed system to support clinically vulnerable people as it did not have confidence all local authorities and supermarkets could meet people’s needs, particularly for food. MHCLG spoke with some local authorities and supermarkets early on to assess their capacity, but could not do a full assessment of local authority capacity to support the most vulnerable because of the urgency of the task. Instead, it used the information it had available to have a centrally-directed supply of food boxes which cost £200 million, as this was likely to guarantee a supply of food to every part of England, particularly given its concerns about shortages in supermarkets. However, some local authorities had queried why government chose a centrally-directed rather than a local system of support, particularly for food, and felt that they would have provided better quality support. Starting in April, as confidence grew in the supply chain and as it developed its understanding of local authority capacity, MHCLG moved to a locally-led model which was in place by summer 2020. This model focused on access to supermarket deliveries and having local authorities offer food to suit the needs of the local population where needed. MHCLG calculates that it has provided local authorities some £4.6 billion in un-ring-fenced funding in 2020–21 to help with COVID costs. Recommendation: MHCLG should ensure that local authorities will continue to have the capacity and resilience to support the needs of clinically extremely vulnerable people, particularly given the significant increase of people advised to shield in February 2021 – from 2.2 million to 3.9 million people.

MHCLG and DHSC do not know whether 800,000 clinically extremely vulnerable people slipped through the net and missed out on much needed support. DHSC explains that it took a ‘multi-channel’ approach to engaging with those affected. Through this approach, it focused first on sending letters, then an email, then calls from the contact centre, which was established at a cost of £18.4 million. 1.8 million Covid 19: supporting the vulnerable during lockdown 7 people did not register their needs or respond when contacted by letter, so their details were passed to the contact centre for follow-up. However, the contact centre was unable to get in touch with around 800,000 vulnerable people, despite apparently making hundreds of thousands of calls every day. It took central government one month to pass the details of these people to local authorities, so local authorities could check if they needed help. Crucially, MHCLG has no knowledge of whether local authorities then managed to reach any or all of these people. Recommendation: MHCLG should urgently update the Committee on whether it has now successfully confirmed the support needs of all vulnerable people, including the additional 1.7 million people advised to shield in February 2021.

Missing or inaccurate telephone numbers in NHS patient records undermined government’s efforts to contact 375,000 people. The contact centre relied on telephone numbers in NHS patient records when calling people to check their needs. Over 20% of the 1.8 million telephone numbers passed to the contact centre from NHS records, for roughly 375,000 people, were missing or found to be incorrect, with the consequence that when the contact centre needed to rely on phone numbers too many were not right and so people could not be contacted to check they were well and getting what they needed. DHSC argues that NHS records are only as good as the information patients provide and explains that its preference is to contact people by letter first, as addresses are the highest-quality contact records. DHSC notes that it also relies on GPs to make sure that the necessary contact is made with clinically extremely vulnerable people. DHSC is trying to improve contact information by asking those affected to ensure their GP records are up to date. Recommendation: DHSC and NHS Digital should ensure that different NHS bodies can securely source the most up to date, reliable and complete patient records, including contact details. It should update the Committee on its plan to achieve this progress within six months. 8 Covid 19: supporting the vulnerable during lockdown 1 Identifying and supporting vulnerable people

On the basis of a report by the Comptroller and Auditor General, we took evidence from the Department of Health & Social Care (DHSC), Ministry of Housing, Communities & Local Government (MHCLG), Department for Environment, Food & Rural Affairs (Defra), and NHS Digital about protecting and supporting clinically extremely vulnerable people during lockdown.1

As its understanding of the disease has grown, DHSC has developed a new risk assessment tool, QCovid, to identify people at risk based on wider factors which make them at more risk from COVID-19. DHSC described the tool as having technical, clinical and academic elements.10 QCovid identifies people who have combined risk factors which put them at enhanced risk, including personal characteristics, such as age, ethnicity and body mass index.11 DHSC told us it considered it was ‘pretty good going’ to take 10 months to develop the tool and that it would have been difficult to develop it more quickly.12 DHSC used this tool to identify an additional 1.7 million Clinically extremely vulnerable people in February 2021.13

DHSC told us it recognised that advising people to stay inside and away from society does have risks as well as benefits. Of those surveyed, some 36% reported worsening mental health and wellbeing while shielding.14 Charities also told us of the impact of lockdown on people not categorised as clinically extremely vulnerable people. They reported how the over 70s and the blind and partially sighted, who were not advised to shield, and therefore not eligible for support through the Programme, struggled to access food. The Royal National Institute for the Blind wrote that the government’s ‘one size fits all’ approach left many blind and partially sighted people behind.15 Identifying all clinically extremely vulnerable people

The list of medical conditions that the chief medical officers developed to define clinically extremely vulnerable people was shared with NHS Digital on 18 March 2020. DHSC tasked NHS Digital to use patient data to identify those affected and create a list of people to be advised to shield (the shielded patient list – or the list).16 NHS Digital held or had easy access to some patient data but GP data were held within GP IT systems, which were not immediately accessible to NHS Digital.17

NHS Digital created the first iteration of the list of some 900,000 people within two days using readily accessible data sources—hospital, maternity and prescribed medicines data. By 12 April 2020, three weeks after shielding began, a further 420,000 people had 7 Q 17; C&AG’s report para 2.3 8 Qq 13, 17 9 Q 17 10 Qq 3–4, 13–14 11 C&AG’s Report, Figure 14. 12 Qq 14–15. 13 https://www.gov.uk/government/news/new-technology-to-help-identify-those-at-high-risk-from-covid-19 14 Q93; C&AG’s report, para 4.4 15 Royal National Institute for the Blind; Independent Age. 16 C&AG’s Report, para 11 17 Q 21; C&AG’s Report, para 11 10 Covid 19: supporting the vulnerable during lockdown been added using GP data, bringing the total to 1.3 million people.18 We asked why it took until the 12 April 2020 to access GP data. NHS Digital told us that it took three weeks working with GP IT systems providers to design, build and gather the GP data. NHS Digital told us that this was quick, as work of this scale would usually take four to six months.19

As well as NHS Digital using patient data to add people to the shielding list, GPs and hospital doctors were asked to review the list and use their clinical judgement to add or remove people. GP and hospital doctors’ additions brought the total to 1.8 million by 18 April and then 2.2 million by 7 May.20 As people were added, NHS England & NHS Improvement (NHSE&I) sent them letters advising them to shield and of their eligibility for support.21 We received written submissions from charities which reported delays in people receiving these letters, potentially putting them at risk of infection, and causing distress and delays in accessing support.22 Overall, 900,000 people were added to the list between 18 April and 7 May.23

DHSC acknowledged that there are advantages with NHS data systems—such as having large amounts of data—and disadvantages, for example challenges in connecting and using legacy systems.24 We asked NHS Digital what would help to identify patients earlier. NHS Digital told us this would require a technical solution, faster access to GP data would help, while ensuring that the general practice which collects the data is comfortable with how it will be used. NHS Digital also suggested that government invest in the digitisation of hospital records, noting that primary care data has been digitised, and is now a richer source of information than hospital data.25 Local variation in the extent to which people were added to the Shielded Patient List

NHSE&I asked GPs and hospital doctors to add or remove people from the list, based on their clinical judgement, and as their patients’ conditions or treatments changed over time.26 However, the extent to which the list grew between 12 April and 15 May 2020 varied hugely in different areas, with increases in the list ranging from 15% in Carlisle to 352% in Hounslow, with an average increase across local authorities of 73%. This was not a small number of authorities with variance from the average 73%: 33 authorities saw their list sizes more than double between these dates, whereas 17 saw their list sizes increase by less than a third over the same period.27

NHSE&I was not responsible for managing any local variations and did not challenge local clinical decisions.28 DHSC has told us that NHSE&I and NHS Digital considered that ultimately additions were a decision for local clinicians. It noted that the approach to local additions was endorsed by the UK Chief Medical Officer who provided guidance on 18 Q 21; C&AG’s Report, paras 12, 2.5 19 Qq 21–22 20 Q 21; C&AG’s Report, para 2.6 21 C&AG’s Report paras 2.5, 2.10 22 SVL0002 - Written Evidence submitted by Asthma UK and the British Lung Foundation p.3 23 C&AG’s Report, para 2.6 24 Qq 23–24 25 Qq 21–22, 97–98 26 Qq 21, 23, 25; C&AG’s Report, para 2.6 27 C&AG’s Report, Figure 8, analysis of underlying data 28 C&AG’s Report, para 2.9, Figure 8 Covid 19: supporting the vulnerable during lockdown 11 the shielded patient list. DHSC also explained to us how NHSE&I and NHS Digital took steps to try to ensure this process was consistently applied across England. For example, it told us how on 12 April NHS Digital noticed that additions by GPs showed wider than expected variation and identified that GPs were adding patients in bulk using computer searches rather than assessing individual patients. In response, NHSE&I told GPs that there should be no automated process used to compliment or supplement individual clinical identification.29

NHS Digital told us that for the people whom it had identified and added centrally to the list, in line with the clinical criteria set by the chief medical officers, there is very little variation by area.30 DHSC acknowledged that it has seen variation in regions, and in local areas clinicians would have adjusted their approaches when adding people. DHSC explained that in a few cases it tried to understand whether variation reflected a genuine underlying illness, or different thresholds for adding people to the list.31

We asked DHSC if it had created a postcode lottery of support, and if people with certain conditions in some areas, would have different support to people in other areas with the same conditions.32 DHSC told us that it did everything possible to ensure that it had consistent application of the policy. NHS Digital and DHSC explained that where they saw areas that had high or low numbers of people being added, they had worked with NHS clinical directors to challenge these areas and followed some up directly.33 As a result of such work, DHSC was confident that it had not identified any systematic differences in approach, and that were there was variation, there was no indication that guidance had not been consistently applied. DHSC told us that it also worked with the Royal College of General Practitioners, who had training modules to try and ensure that there was a consistent understanding and approach in adding people to the list.34 DHSC also provided us details of NHS Digital’s analysis of the variation of 11 February 2021 which DHSC considered showed few local areas outside the normal expected range. Based on this analysis, DHSC concluded the level of variation in how local clinicians added people to the list to be acceptable. It contended that local variation in the number of people identified and added to the list could be explained by demographic variations in the English population and the “inevitable difference” in decision making arising from local clinical judgement.35 A centrally-directed system to support clinically extremely vulnerable people

Government quickly needed to ensure that those shielding had reliable access to food, medicines and care. It chose a national system of support run by central government.36 MHCLG considered that a centralised offer was more likely to guarantee delivery of food boxes in every part of England at the start of the pandemic, when there was real concern about food shortages in supermarkets.37 MHCLG consulted with a small number of local 29 Letter from DHSC to the Committee 10 March 2021 30 Qq 27, 32, 46 31 Qq 25–26, 28, 47 32 Qq 29–30 33 Qq 25, 28 34 Qq 30, 31 35 Letter from DHSC to the Committee 10 March 2021 36 C&AG’s Report, para 1.5 37 Q 73 12 Covid 19: supporting the vulnerable during lockdown authorities as to the best way to support people shielding, but acknowledged that it had not done a full assessment of local authority capacity in the way it would for a business-as- usual programme, and had made a judgment based on the evidence available at the time.38 MHCLG told us that it engaged with local authorities, and some reported they would not have been able to provide the food delivery service in the early months of the pandemic.39 However, the NAO reported that some local authorities queried why government had chosen a centrally directed rather than a local system of support, particularly for food, and some authorities felt that they would have provided better quality support.40 MHCLG was confident that it had made the right decision to have a national system to provide food boxes, rather than a local one.41

Defra was chosen to lead on providing food to people shielding because, according to MHCLG, it had the expertise and relationships with the food industry.42 Defra consulted with supermarkets and wholesalers, to understand their capacity to provide people with food, nationally and quickly. Defra told us that it was very clear supermarkets did not have the capacity, and so it had used wholesalers direct.43 Food box deliveries started five days after the start of shielding, on 27 March, and went to 510,486 clinically extremely vulnerable people from then to 1 August, when the programme was paused. Overall, Defra spent £200.2 million on the food support service up to 1 August 2020.44 MHCLG told us that it had asked local authorities to provide those shielding with ‘supplementary food’ for people with dietary requirements.45

In August 2020, the government conducted an early lessons learned review of the programme which noted that, should shielding be needed again, a local support model could improve flexibility and potentially be more cost-effective.46 MHCLG told us how it started to move to a locally-led model over summer 2020 as it gained confidence in the food supply chain and in the capacity of local authorities.47

This locally-led model focused on providing eligible people with priority access to book supermarket deliveries, rather than Government providing standard food parcels.48 Defra told us that almost everybody that signed up for food box support was matched and prioritised by a supermarket, and that these half a million people have continued to be prioritised by the supermarkets over the last months and for the foreseeable future. MHCLG observed that as the pandemic has progressed, fewer people have needed support through this programme, and felt confident that a locally led model, reinforced with the new registration system built over the summer, has proven an effective way of delivering the programme. MHCLG also reported that it now has good information about what 38 Q 74 39 Q 80 40 C&AG’s Report, para 3.22 41 Q 80 42 Q 78; C&AG’s Report, para 1.7 43 Q 74; C&AG’s Report, para 3.7 44 C&AG’s Report, paras 19, 21, 27, Figure 1 45 Q 80 46 C&AG’s Report, para 4.10 47 Q 75 48 Qq 74, 75; C&AG’s Report, Figure 14 Covid 19: supporting the vulnerable during lockdown 13 local authority activity is on the ground, in terms of the delivery of the programme.49 MHCLG calculates that is has provided local authorities some £4.6 billion in un-ring- fenced funding in 2020–21 to help with COVID costs.50 49 Qq 74–76 50 Q 83 14 Covid 19: supporting the vulnerable during lockdown 2 Communicating with vulnerable people Difficulties in gaining assurance that people’s needs were met

Government used a range of ways to engage with clinically vulnerable people, to advise them to shield and how to register to access support.51 Government wanted all affected to register whether they needed support or not. NHSE&I and DHSC were initially responsible for advising people to shield, and began sending letters and texts from 23 March. The Government Digital Service developed a website and an automated telephone helpline for people to register for support.52 Government also commissioned a contact centre through the Department for Work & Pensions (DWP) to call around 1.8 million people who had not yet registered using the website or automated helpline, despite having been sent letters. The contact centre cost a total of £18.4 million. However, the contact centre was unable to get in touch with around 800,000 people despite MHCLG’s assessment that the centre was making hundreds of thousands of calls every day.53

DHSC and MHCLG explained that for the 800,000 vulnerable people that the contact centre could not reach, their contact details were passed to local authorities, as it was thought local authorities might be better placed to contact these people and identify their need for support.54 These details were given by the contact centre to local authorities starting from 28 April, over one month after the start of shielding. MHCLG told us that local authorities had, before 28 April, received the full shielded patient list and the details of those who had registered for support by then and that they had started contacting people before receiving the details of the 800,000 ‘uncontactables’.55

We asked MHCLG how many extra people took up the offer of support after being contacted by local authorities. MHCLG confirmed that it did not collect this information and explained that it is difficult to disentangle how many people registered support needs via contact with local authorities from those who registered from central government contact. MHCLG has no knowledge of whether local authorities reached the 800,000 people that the contact centre could not.56 Missing or inaccurate telephone numbers

The DWP contact centre relied on telephone numbers in NHS patient records to call clinically vulnerable people who had not yet registered their needs. In some 375,000 cases out of the 800,000 people that the contact centre could not get hold of, or over 20% of the 1.8 million people the centre attempted to contact, it was because of missing or inaccurate phone numbers. While government knew that a proportion of telephone numbers in NHS records were missing or inaccurate, the Programme agreed to use telephone numbers from NHS records to follow-up hard-copy letters.57 51 Qq 60–64 52 C&AG’s Report, paras 2.10, 3.2 53 Q 65; C&AG’s Report, paras 17, 3.4–3.5 54 Qq 64–65 55 Q 68; C&AG’s Report, para 17 56 Qq 66–69 57 C&AG’s report, para 2.8, 3.6, 3.24 Covid 19: supporting the vulnerable during lockdown 15

MHCLG considered that it established a communication strategy that took all reasonable steps to reach people.58 DHSC explained how it used a ‘multi-channel’ approach to communicate with clinically vulnerable people: its preference was to contact people using letters first as it considered letters used the highest-quality contact records, followed by email, and then a telephone call. DHSC highlighted that it also relied on GPs to make sure that the necessary contact was made.59 We asked DHSC why such a large proportion of vulnerable people have incomplete patient records. DHSC responded that it is difficult to make sure phone numbers are up to date, as a ‘surprisingly large number’ of people change their phone numbers quite often. It said that NHS records are ‘only as good as what patients provide’.60 The NAO also reported how local authorities struggled with inaccurate contact data which created additional work and potentially delayed getting support to those who needed it.61

DHSC told us it is trying to improve contact information by asking clinically extremely vulnerable people to ensure their GP records are up to date and will continue to update the records as patients improve their record keeping with their doctor. DHSC also noted that increasing numbers of patients are adding their email addresses to their GP records.62 58 Q 65 59 Qq 56, 60–62, 64 60 Qq 56, 61 61 C&AG’s report, para 3.24. 62 Q 60 16 Covid 19: supporting the vulnerable during lockdown