The role of charitable funding in the provision of public services: the case of the English and Welsh National Health Service

The role of charity in the provision of public services is of substantial academic and practitioner interest, and charitable initiative within the English and Welsh National Health Service (NHS) has recently received considerable attention. This study provides rich insights into the role that NHS-linked charities present themselves as playing within the NHS. The dataset analysed is a novel construction of 3,250 detailed expenditure lines from 676 sets of charity accounts. Qualitative content analysis of itemised descriptions of expenditure allows us to explore how these charities portray their activities. We distinguish between expenditures that can be framed as supplementary to government funding (such as amenities and comforts) and items that suggest charitable effort is substituting for government support (such as funding for clinical equipment). We also consider the claims being made through these representations, and suggest that the distinctiveness of the charity and NHS spheres are currently under question. We argue that, through their representational practices, charities are both shaping and blurring the expected roles of government and charity. Acceptance of the benefits that charitable initiative does provide, in terms of innovation, pluralism and participation, must be tempered with the realisation that charitable funds are playing a role in service provision that is not guided by clear policy, and that this has the potential to widen existing inequalities within a key public service.


Introduction
At the time of its inception in 1948, the UK's National Health Service (NHS) was conceived as a state-funded and centrally-planned health service, drawing its funding from taxation, rather than through the prior patchwork of local initiatives Unauthenticated | Downloaded 09/15/23 03:55 AM UTC in which access to healthcare was often a function of the 'caprice of private charity' (Bevan, 1946).Nevertheless, charitable funds have always played a role in the NHS; some income is derived from inherited assets, while health authorities have never been prevented from accepting charitable legacies or donations.The role of charity is not trivial, with annual charitable income peaking at the equivalent of around £700 million in 2020 prices in the early 2000s, while more recent estimates place the figure at approximately £450 million. 1There are very substantial variations in charitable funds available to NHS trusts: several entities have budgets in the tens of millions which, in some cases, are equivalent to several per cent of the annual budgets of the Trust itself (Bowles et al, 2023).This role has led to ongoing questions about the extent to which charitable funds could or should be used to provide goods or services which are properly thought of as being part of a comprehensive statutory health service.
In an early (1953) legal case (quoted in Meakin, 1998: 27), the court held that it was legitimate to permit charitable support of the NHS, on the basis that the state would never be able to meet all conceivable health needs.That judgment left unexplored, however, the challenges in determining what ought to be supported by the state, and what could or should be met by charity.The implication was that charity would provide resources over and beyond those offered by the state for the core services of the welfare system.However, at an early stage, it was recognised that defining the border between 'core' and 'additional' services was likely to pose challenges.
In fact, there does not appear to have been a systematic effort to demarcate this border.Less than 25 years into the NHS's existence, the NHS Reorganisation Act (1973) eschewed any attempt to do so, stating that charitable funds could be used for 'all or any purposes relating to the health service'.The Health Services Act (1980), which enabled health authorities to engage in fundraising, similarly indicated that charitable monies could be used 'in providing or improving any services or any facilities or accommodation which is or are… to be provided as part of the health service or to assist them in connection with their functions with respect to research'.
In the 1990s, the Charity Commission for England and Wales (CCEW) did distinguish between the responsibility of government, and the possibilities for charity involvement, in the delivery of public services, advising that the 'relief or substitution of statutory funds or services was not an appropriate use of charitable funds, although charities might legitimately use their funds to supplement or enhance statutory provision' (Paines, 2007: 505).However, since 2004 the CCEW has 'concede [d] that if it is charitable for a charity to contract to provide public services even if these are the statutory responsibility of Government, then so too must it be charitable for grants to be made to fund such services' (Paines, 2007: 513;HFMA, 2017).Currently, therefore, guidance from the CCEW states that 'there are few legally defined boundaries for services that must only be provided and funded directly by the public sector.It is therefore legitimate for charities to consider becoming involved in the delivery of such services' (CCEW, 2012).The Department of Health (2012) has also asserted that 'charitable funding provided by NHS Charities can in effect replace exchequer funds'.
Therefore, there is no extant legislative justification for the argument that charities should focus their activities on supplementing rather than substituting for state provision.Nevertheless, this argument remains alive in the discourse surrounding charities in the NHS.Fundraising for the NHS had a very high profile in the early stages of the COVID-19 pandemic, prompting discussion about how the substantial funds so raised might be spent.Rimmer (2020) states (without reference) that 'money donated to NHS charities cannot be spent on direct patient care'.Ainsworth (2020) similarly argued that 'money raised for the NHS can't be spent on PPE [personal protective equipment]'.Wade et al (2022) suggested that charitable funds could not 'be directed toward essential "core" services of the NHS', apparently believing that essential NHS functions always remain a state responsibility.The expenditure of donated funds by NHS-linked charities was also frequently the subject of mainstream news discourse during this period (for example, the disparaging headline 'Well-wishers donate more than £100m to NHS charity… but funds will be spent on iPads', Daily Mail, 2020), yet reporting is rarely able to extend beyond cherry-picked anecdotes of what the donations have been able to provide.For example, a 2022 article investigating the expenditure of the bounty generated by centenarian fundraiser Captain Sir Tom Moore (The Independent, 2022) relied on individual press reports and quotations from NHS staff to give readers an overview of how the donations were spent.The confusion surrounding this issue, and its high salience and visibility -particularly as we approach the 75th anniversary of the NHS in 2023 -has prompted this article.
This study draws on data relating to the years 2014-2020.Despite political turmoil and the COVID-19 pandemic, during this period there was a degree of continuity in policy approaches to the voluntary sector.Against the background of public sector funding cuts and welfare transformation programmes (Taylor-Gooby, 2012;Dayson, 2017), the government has pursued policies designed to increase choice and promote competition in public service delivery (Naumann and Crouch, 2020), including encouragement of voluntary sector provision, while also devolving power and espousing localism.However, that constellation of policies has been criticised for enabling central government to evade responsibility for the consequences of its actions (Fitzpatrick et al, 2020: 547, 556).This includes continued contracting out and commissioning of public services to be provided by the voluntary sector.
The present policy context, therefore, is a lack of guidance as to what should be provided by government and what can be provided by charity within the NHS, as well as a government approach that promotes the role of charities as service providers.This article seeks to further understanding of how, within this environment, charities within the NHS have defined their role.Drawing on qualitative analysis of over 3,000 lines of expenditure-related accounts data from 676 sets of annual accounts from 340 separate charities, we argue that these charities choose to represent themselves in ways that are consistent with both supplementing and also being substitutive of state provision of healthcare.Their role is shown to be not just providing 'additional' items, but also goods that are essential to the provision of a comprehensive health service, including basic equipment and furniture.This, we argue, suggests a need for more evidence-based discussion around the appropriate role of charity in providing services within the NHS, and of how, paraphrasing Stewart et al (2022), such charitable action 'co-exist[s] with commitment to dutiful, means-based funding of healthcare via taxation'.
This discussion also has wider significance.NHS-linked charities are one example of a set of charities that provide funding and support to the provision of 'social' goods and services that are primarily the responsibility of the government in England and Wales.Similar types of questions to those posed here have been raised in consideration of parent-teacher associations (PTAs) by Body et al (2017) and Body and Hogg (2022), while the implications for the funding of the public school system in the USA of large-scale fundraising initiatives are explored by Reich (2006).Consideration of these NHS-linked charities therefore has wider policy, practical and theoretical interest and resonance.By presenting a rich, robust and in-depth analysis of NHSlinked charity expenditure reporting, we offer a detailed consideration of how these charities position their role.

The present study: two types of NHS-linked charity
This research explores two different groups of charity that have direct relationships with NHS bodies: 'Friends of ' organisations, and the group of organisations commonly known as 'NHS Charities'.
'Friends of ' (including Leagues of Friends) organisations are charities which take part in volunteering and fundraising for the hospitals to which they are attached (Millward, forthcoming).According to Cooney (1960: 263), these organisations were founded to provide 'personal service to patients, supplementing the efforts of the paid staffs' and 'to supply hospitals with equipment not likely to come from the budgeting of the authorities or to enable them to obtain equipment sooner than would otherwise be possible'.These charities have a long history.Many developed out of groups that had been attached to pre-NHS voluntary hospitals, and build on the institutional legacy of the pre-1948 voluntary hospital network (Millward, forthcoming).Ellis Paine et al (2019) find that Leagues of Friends support community hospitals by providing financial, human, practical and intellectual resources.These resources were found to enhance community hospital services, patient experience and staff morale, as well as volunteer wellbeing (Ellis Paine et al, 2019).Subsequently, Millward (forthcoming) demonstrates that many 'Friends of ' charities have, over the years, focused on providing a modest level of funding for capital expenditure.The budgets of the larger such entities may cover a 'significant proportion of capital investments such as sunrooms or entertainment halls', while the smaller charities might focus on 'purchasing a single, relatively large item each year'.The limited extant literature, therefore, provides a picture of 'Friends of ' charities as being focused on both patient and staff wellbeing, with a particular focus of activity being relatively small-scale purchase of equipment.
The group of organisations known as NHS Charities have a shorter organisational history.While, as described earlier, charitable monies have always been held by NHS organisations, these charities have (largely) only been formally managed as separate charitable organisations since the 1990s, after regulatory changes put in place by the Charities Acts of 1992/1993.Between the 1990s and 2020, subsequent Acts (including the 2006, 2011 and 2016 Charities Acts, and the 2016 NHS (Charitable Trusts etc) Act) have reformed the regulation and governance status of charities, and this subset of charities specifically.This has led to different definitions of the term NHS Charities, with the CCEW stating that NHS Charities refers only to those organisations that have an NHS body as a corporate Trustee.This includes most such charities that are directly associated with an NHS body, but not those that have an independent governance status: as of 2022, 25 of these charities had moved to independent status (NHS England and NHS Improvement, 2022).While small in number, some of the largest such charities are (now) independent charities.These charities nevertheless do often maintain some governance links with their NHS body, with guidance from the Department of Health and NHS Charities Together (2020: 8) advising that the constitution of the NHS Charity should allow for at least one trustee to be appointed either from or by the NHS body.This means, for example, that, as of January 2023, Leeds Cares (the Leeds teaching hospitals charity) has two trustees who also serve on the NHS Trust Board (including the Chair of the Trust Board).Cwplus (the charity that supports the Chelsea and Westminster NHS Foundation Trust) similarly has two trustees who also serve on the Trust's Board, while the Alder Hey Children's Charity has one.However, Great Ormond Street Hospital Children's Charity, the NHS charity with the largest annual income, has none.NHS Charities Together, the membership organisation for NHS Charities, refers to all such organisations -both independent and those with corporate trustee status -as NHS Charities, and this convention has also been followed within academic research, such as Stewart and Dodworth (2021: 2).This article also follows this practice, and describes all such organisations as NHS Charities.
NHS Charities Together itself states that the priority of these organisations is to provide 'the extra support needed for staff, patients and communities' (NHS Charities Together, 2020;2022).Their 2020 Annual Report describes the focus of member NHS Charities expenditure as 'supporting research and development… brightening up hospital environments, and donating state-of-the-art technologies and equipment' (NHS Charities Together, 2020: 13).
In summary, therefore, the literature suggests that the primary role of 'Friends of ' charities was to improve the experience of staff and patients, often expressing their 'friendship' (Millward, forthcoming) through the purchase of equipment.Likewise, existing literature suggests that NHS Charities have historically been associated with capital expenditure (buildings and equipment) (Fitzherbert and Giles, 1989;Fitzherbert, 1992;Lattimer and Holly, 1992;Lattimer et al, 1996) and, more recently, as focusing on staff and patient welfare and wellbeing, staff welfare, research, andperhaps more occasionally -the purchase of equipment.
Finally, while these charities are 'NHS-linked', they are also able to move beyond supporting their associated NHS body, to fund both other NHS bodies and private charities, as is widely allowed for by their charitable purposes.For example, the charitable objects of Leagues of Friends may be to 'relieve patients and former patients' (emphasis added) of their associated NHS hospital -as with, for example, the Friends of University Hospital Lewisham -or to 'provide a link between the hospital… and the community' (Friends of the Hospital of St Cross), or to support and relieve 'outpatients and former patients of the hospital and other persons in the local community' such as by 'supporting other organisations that support and relieve out-patients and former patients of the hospital' (The Hawkhurst Community Hospital League of Friends).The purposes of many NHS Charities similarly enable them to move beyond supporting their associated NHS body.The Gateshead Health NHS Foundation Trust Charitable Fund, and the Bolton NHS Charitable Fund for example, have as their objects as being 'for any charitable purpose or purposes relating to the National Health Service' (drawing on prior standard model NHS charity governance documents).Cwplus, has similarly broad objects, namely 'any charitable purpose or purpose relating to the general or any specific purposes of Chelsea and Westminster Hospital NHS Foundation Trust or the purposes of the health service'.Finally, the Maudsley Charity (associated with South London and Maudsley NHS Foundation Trust) includes within its objects, very broadly, the 'relief of sickness and the preservation of the health and social welfare of people living in the United Kingdom'.
Despite the policy context outlined here, the growth and use of NHS charitable funds has received relatively little detailed research attention.Existing work provides overviews of the aggregate data, highlights some organisations with particularly large charitable funds, and identifies ways in which the presence of significant charitable resources poses operational challenges for NHS authorities (Fitzherbert and Giles, 1989;Williams, 1989;Lattimer and Holly, 1992;Lattimer et al, 1996;Mohan and Gorsky, 2001;Pharoah and Mocroft, 2001;Exworthy and Lafond, 2021).
Some authors -notably Fitzherbert, Lattimer and co-authors, and Williamsstrongly argued that charity was substituting for statutory funding, with particular regard to items that might reasonably be regarded as 'core' to the business of the service.An implied distinction was being drawn between the general categories of amenities and comforts ('nice to have' items that contribute to the working and therapeutic environment) that are often assumed to be the raison d'être of NHS fundraising, and the core ('need to have') items essential to delivering healthcare, responsibility for which was felt to be the preserve of the government.Often the arguments were made in relation to substantial capital projects or significant items of capital expenditure (Williams, 1989).More recent work by Stewart and Dodworth (2021) draws on research by New Philanthropy Capital (2019) which similarly asserts that the distinction between 'core' and 'non-core' services and funding is still being maintained by charitable organisations which fund '"add-ons" to patient care (such as arts in health) and staff development'.However, Stewart and Dodworth (2021: 2) suggest that these organisations also '[o]ccasionally… purchase medical equipment for which there would be no business case by a needs-based definition of the local population'.

Moving on from 'core' versus 'comforts': supplementing, substituting or sidestepping government provision
In seeking to explore the role and contribution of these charitable actors, we turn to the academic literature that considers rationales for the existence of charitable organisations and for the role of charity in the provision of public services.This literature can be dated back at least to Weisbrod (1975), who argued that a core rationale for voluntary provision was the failure of majoritarian democratic societies to respond to the diverse needs of heterogeneous populations, and to James ' (1986; 1987; 1989) 'supply-side' theory -arguing that the primary motivating factor for the existence of charities is individuals who establish charities in order to achieve a specific (often personal) objective.Building on these, Salamon's (1987) theory argues that, rather than charities being the response to government failure, instead it is government action that responds to 'voluntary sector failure' (Salamon, 1987: 39).Billis and Glennerster (1998) alternatively argue that charities can have a comparative advantage in 'restricted but important areas of human service provision' compared to private and public sector organisations.These theories as developed by Weisbrod, James, Salamon, Billis and Glennerster and others, are still being engaged with and tested to provide novel insights into the role of charities in the provision of public services, as exemplified by Zuhlke's (2022) and Flanigan's (2022) critical appraisals of heterogeneity theory.
Similarly, there is an extensive literature that examines partnership between government, the private, and voluntary sectors in order to 'make, manage and deliver public policy' (Pill and Guarneros-Meza, 2018: 410; see, for example, Carmel and Harlock, 2008;McMullin et al, 2021).There is also a substantial body of work that considers these relationships through the lens of hybridity (Alcock, 2010;Billis, 2010), as well as substantial exploration of the role of funding in shaping these relationships (see for example, Lipsky and Smith, 1989;Binder, 2007;Young, 2007;Wilsker and Young, 2010).
These bodies of knowledge primarily consider operational charities that directly provide goods and services.The charities that are the focus of this study are, however, more akin to grant-makers than to operational charities: they provide funding either to directly pay for goods and services for their associated NHS body or other organisation, and/or give grants to these NHS bodies or other partners.This article therefore seeks to explore these charities using a theoretical framework developed by Toepler and Abramson (2021) -drawing on Young (2000), Young and Casey (2016) and Gidron et al (1992) -to consider the relationship between government and private foundations in the US.This work is more appropriate to considerations of charities as funders, and develops a typology of relationships between government and foundations as being either collaborative (substitutive), parallel (innovative or supplementary), or adversarial.
According to Toepler and Abramson (2021: 224-225), collaborative relationships are ones in which foundations 'fund government goods that are principally a government financial responsibility'; while in the parallel pattern, foundations will either foster 'innovation', or will support needs deemed to be outside the remit of (and therefore supplementary to) government action.In the adversarial pattern, foundations will 'pursue structural changes… particularly on behalf of groups without voice in the political process'.
This study explores how NHS-linked charities report their expenditure within their annual accounts, as a means of understanding the role these charities present themselves as playing within the NHS.Annual accounts (alongside their associated annual reports) are 'seen as [charities'] most important publicly available communication' by 'a range of stakeholders' (Connolly and Hyndman, 2013;Hyndman and McConville, 2018: 138).We recognise documents such as annual accounts as a constructed account (Coffey, 2014: 377) of an organisation's expenditure practices, reflecting both internal organisational practice and active decision-making (Yasmin and Haniffa, 2017: 82).They provide not just accounting and accountability information, but are also an important way through which organisations seek to manage their identity, gain legitimacy with external actors (Dhanani, 2019;Cordery and McConville, 2022), and develop the 'dialogic codifications' (Manetti et al, 2021) that both represent and shape understandings of charity's role and position within society.
The detail given in charity accounts therefore reflects choices made as to how to classify, record and present their financial position.This study explores these choices, to understand how these NHS-linked charities represent their activities, and to explore the 'identity projected by [these] actions' (Hoffman and Weiss, 2008: 281).The expression of different types of expenditure has a specific social and symbolic value, that links to the role and identity that these charities are actively choosing to present for themselves.
Drawing on this framing thus predicts that NHS-linked charities' role within the funding of NHS services will project one (or more) of three roles: • in the collaborative pattern, charities in the NHS will describe their expenditure activities in ways that suggest their role is to substitute for government service provision by providing funding to meet the clinical needs of the health service, including 'filling the gaps' (Calabrese and Ely, 2022) in government service provision, and thereby enabling or allowing government to reduce the public provision of such goods and services; • in the parallel pattern, these charities will describe expenditure in ways that are consistent with an approach that complements and/or supplements public provision (see Paines, 2007) by innovating, taking risks (Toepler and Abramson, 2021), or providing additional non-essentials focused on welfare and wellbeing ('comforts'); or • these charities may position themselves as supporting work that takes place outside or alongside the NHS, sidestepping government provision, such as by giving to other, private organisations.While the theoretical approach of Toepler and Abramson (2021) describes the 'third' role of foundations as being adversarial or conflictual, with foundations seeking to pursue structural changes in society, prior work -such as by Dunn (2007) -suggests that charities such as those studied here do not take a political role.In the context studied here we suggest an alternative framing of this third category, to consider whether these charities work outside the NHS.
Use of such a typology allows for a more expansive understanding of the possible roles of government and charity in the provision of NHS services than the 'core versus comforts' lens provides.In the parallel or supplementary pattern, charities' role is not only about providing comforts, but also about innovating -such as through researchand taking risks that the state and market are 'structurally unlikely' to take (Reich, 2018: 197).The collaborative framing accepts that defining 'core' health needs is extremely challenging, but argues that charities are filling gaps in government service provision if and when they are providing funding to meet clinical needs.This framing therefore nevertheless still suggests an acceptance that there are both essential and non-essential services provided within a health service.Such a claim is debatable and contentious -particularly given the recognition that, for example, supporting staff morale and wellbeing is key to a well-functioning, efficient and ethically-managed workforce (Johnson et al, 2018;Ahmed, 2019).This article does not seek to argue that supporting such activities is not the responsibility of the government.However, we also recognise that -even in a well-funded and politically-prioritised public service -there will be limits to government provision.We therefore argue that, at minimum, if charities are paying for basic clinical goods and services, this suggests that the role these charities seek to represent for themselves includes substituting for government provision.Funding of welfare and wellbeing may suggest that these charities see their role as complementary to government provision.
The next section describes the methods used in this study, and is followed by the findings and discussion, and conclusion.

Data acquisition and assembly
This article focuses on understanding how NHS-linked charities represent their activities.To do so, the research draws on a purposive sample of available annual reports and accounts for financial years ending in 2015, 2016, 2019and 2020. As Morgan (2011: 215) has noted, charity accounts data has limitations in terms of both availability and completeness.For example, while each set of accounts normally covers a 12-month period, there is no requirement on charities to follow a specific accounting year, and charities have up to ten months to file their accounts.In addition, accounting guidelines allow substantial room for charities to represent their expenditure in different ways.This means that there is a challenge of both time lag and comparability when using these accounts for research purposes.Inconsistences in reporting practices limit the potential of extant computational social science approaches (the potential of which for nonprofit studies is summarised by Ma et al, 2021) to robustly analyse large datasets from these accounts.
This research therefore uses a manual qualitative coding process.The focus of data collection was thus on sampling for range and diversity, aiming for resonance (Tracy, 2010) rather than representation (Small, 2009) or full coverage.This is an approach widely used in qualitative studies of nonprofit and philanthropic organisations as a way of extracting a study sample from a sector with porous and ill-defined boundaries (Waters et al, 2021;Bekkers, 2022).The article seeks transferability through generating analytic and conceptual understandings (Smith, 2017) that have relevance for wider questions about the role of charities in the provision of public services.
The dataset that we analyse in this article is a novel construction of 3,250 detailed expenditure lines from 676 sets of accounts from 340 NHS-linked charities (130 NHS Charities -just under half of the total number -and 210 Friends of charities) that are registered with the CCEW.These NHS-linked charities were identified in a number of ways: the membership list of NHS Charities Together; prior lists of NHS Charities held by the authors; 2 and keyword searches of the CCEW register of charities to identify relevant 'Friends of ' charities according to standard terms commonly found in their governing documents. 3The sample is restricted to organisations that report annual incomes or expenditures greater than £25,000 because, above this threshold, accounts are made available via the CCEW website.
The focus of this study is specifically on the detail of expenditure included within either the Statement of Financial Activities (SoFA 4 ) or the notes to the Accounts 5 related to expenditure included within this corpus of annual reports and accounts.Examples of relevant sections of the SoFA and notes to the accounts related to expenditure for two different sets of accounts used in this research are given in Figures 1 and 2 respectively.This also shows the varying nature of the detail given.
For the corpus of annual reports and accounts, each individual line of relevant accounting information was manually extracted and included within the dataset.We do this by employing a specialist data entry firm.As shown earlier, there are, of course, variations between organisations in the extent to which they provide such granular information, and the variability in the location of itemised expenditure records within the accounts meant that an automatic extraction technique was not suitable.Through this process of data extraction, data was classified by the specialist data entry firm as either 'Charitable Expenditure' or 'Grant Expenditure'.The final database held 4,205 relevant observations of itemised expenditure within these two classifications.

Data cleaning
Having established a database of itemised expenditure, the data was then cleaned using OpenRefine.This also enabled data familiarisation, in preparation for the development and testing of the coding framework.OpenRefine is an open-source data tool that enables the cleaning of messy data as studied here.OpenRefine's 'cluster' function was used to remove spelling errors, leading and tailing whitespace, and other typographical errors, enabling the identification of groups with similar names.The function was used at its lowest power to ensure no over-clustering, which could lead to miscoding.As an example, 85 different database rows with different spellings of 'patient's welfare and amenities' were adjusted to have the same format and spelling of 'Patients' welfare and amenities'.This process also enabled the identification and removal of accounts relating to a small number of charities (17 in total) that did not meet our definition of NHSlinked charities, for example charities supporting hospitals based overseas.This resulted in a database of 4,181 different lines.

Data analysis
We then began a process of qualitative content analysis.Content analysis is widely used to classify and interpret financial records and narrative reports in studies of charitable activity, and in accounting research more generally (Steenkamp and Northcott, 2007;Cordery and McConville, 2022).The form of content analysis used in this study is similar to that referred to by Hsieh and Shannon (2005) as 'summative content analysis' in that the researchers both quantified the occurrence of key themes and explored the context in which they are used.The coding framework used for the analysis was abductively constructed: deductively through drawing on guidance issued by relevant membership bodies (HFMA, NHS Charities Together, NHS England, and general guidance found in the charities' SORP (statement of recommended practice)), and inductively through the data cleaning and familiarisation approach described earlier.This process of data cleaning using OpenRefine also enabled the identification of 'clusters' of lines, where data with the same row description appears multiple times: for the 4,181 different lines in the database, there were in total 1,810 different row descriptions.Of these, 83 appeared five or more times.The draft coding scheme was tested through a process of keyword coding of the 1,630 lines with these 83 row descriptions, to ensure its suitability for capturing the broad range of expenditures by organisations in our population.This process also reinforced the appropriateness of our choice of a manual, rather than machine learning, approach.While a machine learning approach may have some success in classifying data with pre-assigned codes and inductively identifying emergent codes, the itemised expenditure descriptions often refer to expenditure which requires further contextualisation.For example, significant outlay on discrete items of clinical equipment may be documented as a brand or model name which is only identifiable after further desktop research.Similarly, the accurate coding of grant expenditures and spending linked to specific campaigns required coders to read the narrative text within the annual report and explore promotional material to ascertain the applicable code.Therefore, the majority of data within this dataset -2,575 rows -was manually coded by two raters as described next.
The final structure for the coding framework contained eight thematic codes, including a code for 'Organisational' costs.While these were coded at the point of data collection due to the substantial presence of such costs within this data, this data is excluded from subsequent analysis, as the focus of this article is on those costs that these charities spend externally.By these organisational costs, 931 lines of expenditure were accounted for, meaning that the final dataset discussed refers to 3,250 lines of data from 676 sets of annual reports and accounts.These codes are then considered in the light of the theoretical consideration of substitution, supplementary and sidestepping, in the Findings and discussion section.The coding process was undertaken by two of this articles' authors, who each initially coded a proportion of the 2,575 itemised expenditure records.To ensure internal validity, the coders shared a detailed coding framework and guidance and met regularly to discuss individual items of expenditure that were unclear.After completion of data collection by both coders, a random sample of 10 per cent of the manually-coded items was second-coded.Agreement between the two raters was 83.7 per cent, with a kappa test result of 0.798 (p<0.001),indicating a high degree of inter-rater reliability.

Findings and discussion
As described, 3,250 lines of text associated with 3,250 lines of accounts data from 676 sets of accounts were analysed and coded.These are described in Table 1.Table 1 also quantitatively summarises the presence of these codes.Each line was coded into only one theme.Overall, we find that these charities present their role as both supplementary to and substitutive of government service provision, and both types of charities include both types of roles.However, neither charity yet shows clear efforts to sidestep government provision.These themes are now explored in more detail.

Supplementing government provision: education, welfare and amenities, and research
As outlined earlier, this article's framing of government-grantmaker relationship theory predicts that the charities studied here may supplement public provision (see Paines, 2007) by innovating, taking risks (Toepler and Abramson, 2021), or providing additional non-essentials focused on welfare and wellbeing.As demonstrated in Table 1, 1,130 lines of expenditure data analysed here are linked with activities that may fit within this definition.This quantitative data summarises the qualitative way in which these charities seek to represent these activities.When describing their expenditure on welfare and amenities, these charities often simply categorise their work as providing 'Patients' welfare and amenities' (accounting for 168 lines), 'Staff welfare' (36 lines) or 'Staff welfare and amenities' (93 lines).The more detailed descriptions, however, more clearly demonstrate the types of expenditure noted by these charities.This includes substantial spending on flowers (including 'Mother's Day posies'), newspapers, television reports, toiletry packs ('for patients' emergency use'), toys for play areas, cushions, wall art and books.Christmas activities -including decorations, parties, lunches and presentsare also a strong and recurrent theme within their activities.Many of these costs therefore clearly fall into the realm of 'comforts': improving the environment for staff, patients and the wider community, but without a clear clinical aim.
It is much less clear, however, that 'education' is supplementary rather than a clinical necessity.Education has been included within this thematic code, because this is how these costs are often represented within the accounts.A number of lines, for example, refer to 'Staff welfare, education and amenities', 'Staff welfare, training and development', 'Education and welfare', 'Patient education and welfare' and 'Patient education, welfare and amenities'.However, education -for both patients and staffis a key aspect of both wider health and wellbeing, and of staff professional and clinical development.These charities' incorporation of education within the category of welfare and amenities may however create an impression that this spending should be seen as 'additional', rather than essential to a well-managed workforce, and to a comprehensive preventative health programme.This may suggest a further, and more active, blurring of the lines between substitutive and supplementary spending by these organisations.
Research can be associated with innovation in healthcare, an activity that is often framed as a key virtue of charitable or philanthropic funding (Reich, 2018: 197).The charities within this study, however, provide minimal detail as to the research activities for which this funding is paying.Of the 120 lines of research funding, 98 are simply described as 'Medical research', 'Research', 'Research grants', 'Research and innovation' or 'Research and development'.Only three charities provide more detail on their research spending: these include a 'Sleep study' funded by the Friends of the Royal Marsden, and The League of Friends to the Robert Jones and Agnes Hunt Orthopaedic Hospital providing funding for a research grant into an 'Os-Stretch Device'.Overall, however, these charities resist specifically defining their research spending.
As outlined earlier, extant literature, as well as NHS Charities Together themselves, suggests that the focus of NHS Charities is '"add-ons" to patient care' (Stewart and Dodworth, 2021: 2) and 'brightening up hospital environments' (NHS Charities Together, 2020: 13).This is supported by the data found here: over a third (34.8 per cent) of lines explored within this study are related to education, welfare, amenities and research.However, as noted, the line between 'additional' and 'clinical' servicesif seen as the difference between 'essential' and 'non-essential' services -is not clear, particularly with regards to support for education, and charities themselves are blurring these boundaries in the way in which they represent some of this expenditure.

Substituting for government provision: equipment, and buildings and furniture
As outlined earlier, Cooney (1960) and Millward (forthcoming), suggest that the provision of equipment is a key part of the 'Friends of ' charities' activities.NHS Charities Together also states that their member charities spend on 'state-of-theart technologies and equipment' (2020).This research demonstrates that these charities also see their role as providing equipment for their associated NHS body: 30.4 per cent of the 3,205 lines coded are spending on either equipment, buildings or furniture.
Where detail is given, however, evidence suggests this funding is not being spent on advanced equipment, or trailblazing clinical developments.The expenditure described within this corpus of data includes funding for basic hospital equipment, such as vital signs monitors, laryngoscopes, bladder scanners, dental drills, pulse oximeters, treadmills and weighing machines.Larger pieces of equipment are also included -including x-ray machines and defibrillators -as well as mattresses, high/ low beds, drug cupboards and a range of different chairs (ranging from waiting or consulting room chairs, to 'Clinical treatment chairs for [a] Community Ulcer clinic').There are examples of more 'high-tech' clinical spending: review of the 2019 Annual Report of Leeds Cares, an independent NHS charity that spent £3.3 million on equipment costs in that year, shows that this charity spent £1.8 million of spending listed as 'Equipment' to improve patient care for those with brain tumours, including neurosurgical microscopes and transcranial doppler ultrasound machines.However, these last have been utilised for such purposes for at least 20 years (Meyer et al, 2001).
Overall, this equipment, building and furniture expenditure is presented as 'filling the gaps' for public provision, rather than on supplying advanced technology.This may, in Cooney's (1960) words enable NHS trusts 'to obtain equipment sooner than would otherwise be possible'.The concern here would be that the availability of charitable funding works against the grain of declared public priorities for resource allocation.A further possibility -though perhaps one more likely to be apparent in the context of large fundraising for capital schemes -is that charitable funding would enable government to reduce spending on what would otherwise be seen as vital elements of comprehensive healthcare.

Sidestepping government provision: grants to other organisations
This research also suggests, however, that very few of these charities have moved beyond their role in supporting their associated NHS body.This code was only generated from the data 37 times, and in a number of instances spending remained within the NHS, but was given to a different NHS body.
There is, however, evidence of charities also working outside the NHS, with grants and donations being given to other non-NHS-linked charities (including Action for Children, the London Community Foundation, the Samaritans, and Cancer Research) and -in one case -to another NHS charity.Funding was also given to churches and parochial church councils (PCCs), to hospices, to local authorities and directly to communities.As noted earlier, such spending is clearly in line with the mission and purposes of these organisations.The wording of their charitable objects in fact seems to provide NHS Charities with scope to use their funds as they see fit, without reference to a boundary dispute about the respective responsibilities of charity and state.

The unknown: miscellaneous and unspecified grants
As shown in Table 1, just over a third of the lines within the data (1,128 lines: 34.8 per cent of the total corpus of text analysed) were coded either as 'Miscellaneous/ other' or as 'Unspecific grants for associated NHS body'.
Lines were coded as Miscellaneous/other either: 1) because this is how such spending is recorded in the accounts of the charities themselves (of the 902 'miscellaneous' lines, 153 are simply called 'other' or 'other charitable activities' in the original data; 71 are originally termed 'miscellaneous'; and a further 45 are called 'charitable activities' or 'charitable expenditure'); 2) when insufficient information has been given (such as 'direct purchases for hospital'); or 3) for spending that cannot be otherwise categorised (this includes 'groups', 'heritage' and even one line simply termed 'Jackie').The Unspecified grants theme similarly gives no detail as to the ultimate use of these monies.A lack of completeness in charity annual reports has been previously noted (see Dhanani and Connolly, 2015).Ho et al (2021) also highlight the importance of drawing attention to the 'silences' within such documents.While this article cannot determine what falls within these 'other' and 'unspecified' categories, we can draw attention to this absence, and to note that there are substantial elements of their work that these charities have chosen not to make visible in their accounts.Of course, we recognise that there are trade-offs for organisations: what may appear as a lack of completeness and detail to some, may simply be viewed by charity accountants and trustees as a pragmatic acceptance that there are limits to how many items of expenditure can reasonably be recorded.Nevertheless, substantial proportions of expenditure remain relatively opaque.

The importance of role representation
An organisation's role refers to the functions and activities that individuals or organisations undertake, as well as the behavioural characteristics of that person or organisation (Biddle, 1979).Roles are framed by social context: while structuralists argue that the social context defines the characteristics of a role, interactionists recognise an actor's agency in 'role-making', albeit shaped by the societal structure in which that actor is situated (Callero, 1994).As noted earlier, an organisation's communications and discourse shape their role and practice (Maier and Meyer, 2011), and the way organisations choose to represent themselves is a substantial part of constructing and understanding their role.
This article argues that the charities studied here choose to represent themselves in ways that are consistent with being both supplementary and substitutive of state provision of healthcare.Their role is shown to be not just providing 'comforts', but also goods that are essential to a comprehensive healthcare provision, including basic equipment and furniture.Furthermore, the representation of both 'education' and 'research' within these accounts may contribute to a discourse in which these activities are seen as additional rather than essential to a well-managed and up-todate healthcare service.

Conclusion
When the place of charitable support for NHS organisations was first discussed at a high level in the Ministry of Health, the concern was with the presentational issues that would arise if members of hospital authorities were seen to be publicly supporting fundraising efforts, when the official position was that the financial position of hospitals under the NHS was 'free from doubt', since institutions were 'in no way dependent on voluntary financial help' to meet their running costs (Ministry of Health, 1948).Discussion of these questions was framed by a presumption that fundraising would be for projects on a scale large enough to distort revenue allocations.However, with NHS charitable funding largely flatlining during the 1950s, there does not seem to have been any pressure to devise a clear policy to demarcate the state-charity boundary in the provision of NHS services, and no further measures were implemented to do so.Important Acts governing the reform of the NHS in the 1970s and 1980s made no reference to the matter, implying that charitable funds could be used for any purpose relevant to the NHS (Mohan and Gorsky, 2001: 91-96).
A clear legal definition of 'core' was thus never settled.However, as we argued in the introduction, there seems to exist in media coverage and commentaries a perception that charitable funds in the NHS may not be used to support 'core' elements of health service provision.That view informed critical writings in the 1990s and coverage of the resurgence of fundraising in 2020.As shown in the introduction, however, there is no legislative or policy guidance that determines what should be provided by government and what can be provided by charity within the NHS.The purpose of this article has been to investigate what can be learned from how charities operating in this space present what they do through the medium of their accounts.It is a complex situation.We should also note some of the limitations of the data we have: nearly one-fifth of the data by value could be assigned only to a 'miscellaneous' category, because of the lack of detail available in notes to the accounts of the sampled charities.This has echoes of a criticism made by Holly (1996: 100) about the lack of detail in the accounts of NHS Charities; her inference appeared to be that this vagueness hid the usage of charitable funds for core NHS services.Nevertheless, the combined value of the items in our database is some £250 million and we are aware of no other study that has explored spending at the interface between charitable fundraising and public provision on such a scale or in such detail.
If the role of charity is one of supplementing statutory provision, the accounts show that charitable support encompasses a wide range of expenditures on facilities and services which support staff wellbeing, amenities and comforts, education and training.Much of the activity here improves the working and therapeutic environment for hospital users and staff, but has no obvious clinical benefit.That being said, one might take the view that support for staff wellbeing would be expected of an enlightened employer, and the same could be said for education and training.Framed in these terms, the case could be made that such items constitute expenditure on what is the core business of the NHS, and thus that some of them are substitutive.As to research, it is clear that significant contributions are being made, given that the mean size of expenditures here is over £300,000; however, without more detail on the nature of the research, it is difficult to say much about how innovative it is or the extent to which it supplements other sources of research funding.
The question of substitution for state funding is also raised by spending on equipment, buildings and furniture.The public presentation of such expenditures is one of supporting state of the art provision; from our data and analysis, the reality seems more prosaic, with lengthy lists of pieces of equipment which have been routinely provided in NHS hospitals for many years, and can justifiably be regarded as part of core clinical provision.
Finally, on the issue of 'sidestepping', from this dataset we find little evidence of charities actually moving beyond their original role of supporting the NHS body with which they are associated, though our sense from other work is that this is moving onto the agenda.
What then are the wider conclusions here?The idea of a rigid border between the domains of responsibility of charitable and statutory funding may have exercised campaigners against what they saw as the disruption of needs-based resource allocation in the NHS by the vagaries of charitable fundraising, but there has never been such a border in practice.If anything, the substantial dataset constructed for this article demonstrates the difficulties of demarcating it.But nor does the data bear out the claims of supporters of charity that they are enabling the spread of cutting-edge technology and innovation through the NHS; much of what is being purchased might once have been regarded as cutting-edge but now, rather as Fitzherbert and Giles (1989) argued, looks like the routine purchase of standardised equipment.
While this article has not sought to draw a normative conclusion as to if and where there should be a boundary between government responsibility and charity opportunity in the provision of healthcare services within the NHS, we argue that this research suggests that the discourse around this currently blurred boundary needs to recognise the extant evidence and policy gaps.Drawing on Boswell's (2018) understanding of boundary problems, the distinctiveness of the charity and NHS spheres are currently under question, and through their representational practices, charities are both shaping and blurring these lines.Given that charity has lots to offer when it is enabled to support, advocate and work closely with communities, but is also beset by well-documented challenges of insufficiency and unequal distribution of resources, we contend that this has the potential to contribute to growing inequalities in the provision of NHS services.
It has long been argued that policies towards the NHS saw the state's role as a "co-ordinative, almost entrepreneurial" one and that this might portend a mix of public, private and voluntary services for health, with a consequent (and unacknowledged) increase in inequalities (Davies, 1987: 315).Acceptance of the benefits that charitable initiative does provide, in terms of innovation, pluralism and participation, must be tempered with the realisation that charitable funds are playing a role in service provision that is not guided by clear policy.Against the background of substantial variations in the level of charitable support available to individual NHS trusts, a potential future scenario is one in which elements of health service provision will come, ultimately, to depend to a greater degree on charitable initiative.

Notes
1 The difference in part reflects the fact that until the establishment of NHS Foundation Trusts from 2003, returns on charitable fundraising were compiled centrally by the Ministry of Health and its successors.After 2003, NHS providers ('Foundation Trusts') were no longer required to provide financial returns, and estimates therefore rely on aggregation of the resources of charities thought to be associated with NHS bodies; they may therefore not be comprehensive. 2These were supplied directly by the Charity Commission which had identified several hundred such organisations as part of a review of NHS Charities during the 1990s, and flagged them in their databases subsequently. 3The hundreds of hospital Leagues of Friends were supplied with standardised terms for use in declaring their charitable objects when registering with the Commission. 4If no additional detail is given in the notes to the accounts. 5Notes to the accounts explain the accounting policies, provide detail about income and expenditure, and provide other relevant information.

Funding
We gratefully acknowledge the support of the Wellcome Trust (Grant 219901/Z/19/Z; 'Border Crossings: Charity and voluntarism in Britain's mixed economy of healthcare since 1948).

Figure
Figure 1: Example A -expenditure detail included in a charity's SoFA Statement of Financial Activities for the year ended 31 March 2020 Unrestricted Funds Unauthenticated | Downloaded 09/15/23 03:55 AM UTC