Do virtual renal clinics improve access to kidney care? an impact evaluation of a virtual clinic in east London

BACKGROUND Evidence from the UK national chronic kidney disease (CKD) audit, identifies deficits in the identification and management of CKD within primary care. Aligning the requirement of GPs for a responsive nephrology service, with the capacity of renal services and the need to prioritise patients with progressive disease requires a re-think of traditional models of care. Utilising the health data in the primary care electronic health record (EHR) to bridge the primary secondary divide is one way forward. METHODS We describe a novel community kidney service based in the renal department at Barts Health NHS Trust and four clinical commissioning groups (CCGs) in east London. An impact evaluation of the changes in service delivery used quantitative data from the virtual CKD clinic and from the primary care electronic health records (EHR) of 166 participating practices. Survey and interview data from health professionals were used to explore changes to working practices. RESULTS Analysis of the virtual clinic data shows a rapid rise in referrals. The majority (>80%) do not require a traditional face to face appointment but can be managed with advice to the referring clinician. The wait for a nephrology opinion fell from 64 to 5 days.

but can be managed with advice to the referring clinician. The wait for a nephrology opinion fell from 64 to 5 days.
Early identification of people with CKD in primary care, particularly among those with risk factors such as diabetes and hypertension, enables proactive management of blood pressure, cardiovascular risk and lifestyle factors with referral to specialist services where there is diagnostic uncertainty or evidence of progressive disease (2). There is some evidence that progression of CKD can be delayed by reduction of blood pressure (3) (4), 3 and the high rates of cardiovascular risk associated with CKD can be reduced by blood pressure control and the use of statins (5).
Currently almost 70% of health and social care budgets are directed towards the care of people with long term conditions. (6) The NHS Long Term Plan, released in 2019, envisages efficiencies in the management of chronic diseases and major changes to the delivery of hospital outpatient care which is described as outdated and unsustainable. It endorses digitally-enabled primary and outpatient care, which 'will go mainstream across the NHS', and 'will free up significant medical and nursing time'. (7) A number of UK studies describe a variety of virtual renal clinics which include alternatives to face to face consultations. Harnett  demonstrate the cost saving compared to routine clinic attendance. (10) Other approaches which use structured test monitoring independent of clinic attendance include eGFR graph surveillance by laboratory staff, this is specifically designed to identify those with progressive CKD and encourage onward referral. (11) In contrast with these schemes, which are run from hospital clinics, the east London community service includes dashboard data on every GP registered patient with biochemical evidence of CKD (3)(4)(5), not only those referred into renal clinics. It has an emphasis on upstream CKD management in primary care (blood pressure control and statin prescribing) with benefits for reducing the risk of cardiovascular disease associated with a declining eGFR. (12) All general nephrology referrals from GPs are assessed in the virtual clinic. The clinic aims to support the management of less-complex CKD within the 4 framework of primary care management of long-term conditions, by providing timely advice, but restricting traditional outpatient clinic follow up to the small number of progressive cases which require more intensive specialist management.
Aims a) To describe the development of a virtual CKD clinic set within a community kidney service which integrates data across primary and secondary care, based on the concept of a learning health system -in which the data from every patient encounter is used for system development and better practice. (13) b) To evaluate the impact of the virtual CKD clinic on timely access to specialist advice, and on satisfaction with changes to service delivery by primary care clinicians and renal specialists.

Study design and setting
This observational study was set in east London primary care and the Renal Unit at Barts Health NHS Trust between 2015-2018. Barts Health NHS Trust is the sole tertiary renal provider for North-East London, reporting a high incident need for renal replacement services, with over 30% of patients with new end stage renal disease commencing dialysis in an unplanned manner, compared to 15.6% across the UK as a whole. (14) All 130 GP practices in the three contiguous inner east London clinical commissioning groups (CCGs) co-terminus with their London boroughs of City and Hackney, Newham and Tower Hamlets (total population 800,000) were involved in the first stage of the service change during 2016, with 36 practices in Waltham Forest CCG joining in 2017. All practices use Egton Medical Information Systems (EMIS Web) for the patient electronic health record. In the 2011 UK Census, almost half of the population in each of these CCGs was recorded to be of non-white ethnic origin (15), and the English indices of deprivation 2015 show that all three inner east London localities fall in the lowest decile for social deprivation in England (16).

Theoretical stance
Many of the key strategies for change management described by Kotter (17) were reflected in planning the design and implementation of this programme. These include: building the case for change and forming a guiding coalition which includes both clinicians and managers, empowering others to act on the programme by the provision of education, comparative performance data and quality improvement (QI) tools, creating early wins for the programme and consolidating the new approach into work as usual to ensure sustainability.

Description of the east London community kidney service
The system wide changes to the delivery of the community kidney service had three components described below. This report focuses on the evaluation of the virtual CKD clinic.
1) The virtual CKD clinic supports electronic-only referrals from GPs for general nephrology advice into a weekly hospital clinic serving each CCG. Service development included the introduction of the EMIS Web platform to the renal unit, and sign up by all practices to data sharing agreements to allow nephrologists to view the complete primary care electronic health record (EHR), with informed patient consent. This facilitates review of eGFR plots over time, proteinuria and all recorded investigations, examinations, medication history, co-morbidities, hospitalisations and other specialist in-and out-patient documents (https://www.qmul.ac.uk/blizard/ceg/renal-health-service/ ). Each participating CCG agreed additional pilot funding to initiate these changes to the renal service. The ambition was to fund the service as a block contract based on the previous years' general nephrology activity. The NHS Long Term Plan increasingly commissions for whole pathway care rather than itemised episodes of care. The per CCG contract for the virtual system was priced at an annually reviewable, fixed tariff for all activity, including education, developing and delivering dashboards, and practice facilitation. Hence the Renal Unit carried the risk of growth in appointments and activity above baseline, with CCGs holding the risk of the service contracting traditional outpatient activity and GPs providing more extended management in primary care settings. In addition each CCG developed customised local enhanced services (activity additional to the GP core contract (18)) with financial incentives to promote best CKD management (treating blood pressure to target, use of statins for secondary CVD prevention and monitoring CKD progression.) Quarterly dashboards identified the number of patients with evidence of CKD, changes in practice performance in CKD coding and management, and were available to practices, commissioners and the renal department. The other elements of the community service included: 2) A package of IT tools which support practices to identify patients requiring diagnostic coding, improvements to blood pressure and cardiovascular management, and monthly practice alerts to identify patients with a falling estimated glomerular filtration rate

Data sources for evaluation of the virtual CKD clinic
Data on referrals, appointment numbers, cost and type (whether virtual, traditional general nephrology outpatient first or follow up attendance) and wait time were collected from the care records system (CRS) at Barts Health NHS Trust. This was supplemented by nephrology department data on transfers between virtual and traditional appointments, and on renal follow up of patients in the virtual clinics.
Anonymised data on practice coding and primary care management were collected on a 8 quarterly basis through EMIS Web and collated into practice and CCG level dashboards.
Questionnaire survey data from Tower Hamlets GPs (the pilot locality for the virtual clinic service) was collected soon after the clinic went live. This data was enriched with interviews with GPs recruited from Tower Hamlets practices, and all three nephrologists involved at the front line before the service had become 'work as normal'. These individual interviews with seven GPs, three nephrologist and one CEG facilitator were recorded and transcribed and athematic analysis using the Framework approach was The dashboard showing variation in CKD coding rates and primary care management of CKD across the four CCGs is shown in Table 1. The majority of practices engaged with the IT tools, and within the first year CKD coding rates improved, with the lowest coding CCG improving performance by 50%. (23)

Referrals to the virtual CKD clinic
From the start of the service all routine general nephrology referrals from GPs were processed through the virtual clinic. GPs were encouraged to refer anyone they would previously have sent to out-patients, and had local guidance which conformed to the 2014 NICE CKD guidelines.
(2) The 'falling eGFR' trigger tool, run monthly in practices, also identified cases to be considered for referral, on average 10% of trigger tool cases were referred to the virtual clinic.
In the twelve months prior to April 2015 the average annual referral rate to general nephrology outpatient clinics was 0.8/1,000 GP registered population. By the second year of the service (2018) the average, annual referral rate was 2.5/1,000 registered patients as shown in the funnel plot ( Figure 1). This graph shows that 15% of practices fell outside the upper control limit for referrals, and four practices with a list size >9,000 made no referrals during the year.

Clinic data
The average waiting time from GP referral to a first outpatient appointment in 2015 was 64 days. When the virtual clinic started the average time between GP referral and virtual clinic assessment fell to 4-6 days. The nephrology opinion can bae viewed in the GP record on the day it is written, and a clinic notification is sent electronically to the practice within a few days. Across the whole service nephrologists arrange an outpatient face-to-face review following just 12% of virtual appointments. Over 40% are discharged back to the GP, with up to 50% being tagged for a further specialist review in the virtual clinic (see Table 2)

*lowest and highest figures across the four CCGs
It was also possible to measure the 'hidden work' associated with virtual clinics by observing the repeated virtual reviews done by nephrologists. More than 40% of initial referrals had a second virtual review, and 30% of these had a third review (Figure 3). The repeated review of virtual referrals was often linked to requests to GPs to arrange further investigations to facilitate a more complete assessment. This virtual review work made up approximately 50% of a virtual clinic session, and alongside the early surge in new 11 referrals contributed to a perception of overload by nephrologists. This work was not transparently captured by routine hospital recording systems.

Survey and interview data
During the first year of the service a questionnaire was sent to all 68 GPs in Tower Hamlets who had used the virtual clinic. There were 28 (41%) responses, with 86% of responders reporting that it was very or quite easy to use the service and 96% being happy with the referral advice they received from the nephrologist (Figure 4). GPs reported that most patients were satisfied with the service although one quarter reported no feedback from patients. The overall value of the new kidney service was rated as 5/5 by 60% of respondents.

Benefits for patients
Every GP interviewed said that all patients had readily consented to their records being shared, with many expressing surprise that this was not already happening: "I think the system is great, and keeping people out of hospital is clearly a good thing." Some GPs described how patients were now being referred when they had not been in the past: "It is useful to get a bit of advice. In the past I would probably have not done anything to be honest….as they (the patients in the nursing home) were not fit enough to go up to hospital" The timeliness of the referral was also considered important for patients.

"Having the nephrologists seeing people within one week is a great benefit."
Finally, a number of GPs spoke about how the new system was educating them in managing CKD in the future. Nephrologists reported that in the old system, some patients were referred but did not actually need to be seen, often there was no referral letter and up to half the time there were incomplete notes in clinic. Other challenges included the duplication of tests, not knowing the medication list, transport or language difficulties.
"What I was not doing was anything meaningful." The referral process is now easier with quicker response times. The ability to see the full record, including all tests and correspondence allows a more in-depth case-review. From The key messages are that patients are content to share their primary care record with nephrologists, so that management advice can be obtained without needing a visit to the hospital. The service provides timely advice back to GPs, who value the improved relationship with the nephrologists. One nephrologist concluded by saying: "There's a lot of kidney disease out there that we did not know anything about."

Main findings
Over a three year period this project developed a complex intervention to improve primary care management of CKD and provide timely access to specialist advice.
The introduction of this unique virtual CKD clinic, based on sharing full access to the primary care record, was followed by rapid take up of the service by local GPs across all four participating CCGs. Improved access to specialist advice also included disadvantaged groups, such as care home residents, for whom traditional OPD attendance poses most difficulties. Clinic barriers to effective assessment, such as lost notes, transport delays, language barriers and patient non-attendance simply disappeared. Time from GP referral to nephrology advice visible in the GP record fell from 64 to 6 days. Surveys identified high rates of satisfaction from GPs with ease of clinic use, the value of timely specialist advice and increased confidence in managing CKD. Nephrologists valued seeing the entire patient record, particularly the eGFR graph, but were more affected by changes to 14 traditional working practices and the loss of patient contact. Virtual assessment took an somewhat less time than a traditional first outpatient appointment. Less than 15% of referrals required a subsequent face-to-face appointment, however 40-50% of referrals had at least one virtual follow up before discharge back to the GP. The service change was highly successful in providing an expansion in the capacity to assess patients with kidney disease and provided rapid access to traditional outpatient services for the small numbers who needed specialist investigation and follow up.

Strengths and limitations
A strength of this study is the application of this complex service change to whole health economies, rather than to selected practices. The first year of the project involved three CCGs which already had a well-developed working relationship with the Clinical Effectiveness Group and the primary care data management, practice comparisons and facilitation services they offer (24). Historically the clinicians and managers in these CCGs have been early adopters of clinical change of value to patients and the health economy.
The project was slower to engage with the fourth CCG (Waltham Forest) where there was less experience of data sharing and system wide quality improvement work.
This was a pragmatic programme evaluation, recognising variation in the way the intervention was implemented in each of the four CCGs. Identifying contextual differences between CCGs in clinical leadership, in approach to implementing and incentivising change in their constituent practices is an essential consideration for the success of scaling up such interventions. Differences in context will help shape local perceptions of the value of change. In turn this may affect the implementation process and contribute to differences in speed of diffusion across differing geographical areas. We also acknowledge that a longer evaluation period is needed to fully understand the impact of the change, over time and against prior practice.

Implications for practice and future research
Changes to the traditional patterns of delivering care will involve a complex interaction between making the most effective use of data within the electronic health record (EHR) All patient-level data are anonymised, and only aggregated patient data are reported in this study. All GPs in the participating east London practices provide written consent to the use of their anonymised patient data for research and development for patient benefit.
to Improve Care in England. Report of the National Advisory Group on Health Information Technology in England. National Advisory Group on Health Information Technology in England; 2016. Figure 1 Annual (   Appendices.pdf