Uses of Management Accounting Information for Benchmarking in NHS TrustsThis article presents findings from the first phase of a longitudinal study on implementation of benchmarking in three large acute National Health Service (NHS) trusts in the eastern region of England. Benchmarking of performance uses Healthcare Resource Groups (HRGs) (NCMO, 1997, p. 5) as the basis for compiling the National Reference Cost Index (NRCI) (NHSME, 2000), which ranks all hospitals in terms of their cost position relative to the national average. The findings concentrate on the attitudes of professional groups towards preparation and uses of this management accounting information for performance measurement and comparison between NHS trusts. It considers: •The management accounting challenges in building up the HRG costing information. •Accountants’ views on the validity of the costing information. •The different approaches to benchmarking adopted in the three trusts. •Contrasts in the views of accountants, managers and clinicians towards benchmarking. Role of HRGs in Performance Measurement The 1989 White Paper Caring for Patients instituted a split between the role of district health authorities as purchasers of services and hospitals as providers of services governed by contractual arrangements. An internal market was created and hospitals began competing, rather than collaborating, with each other. Resulting problems included high transaction costs and dysfunctional behaviour. These were recognized when the newly-elected Labour government in 1997 abandoned the internal market, and heralded a new era of co-operation and sharing of information and learning when it published The New NHS White Paper (DoH, 1997). The proposals under the new framework introduced a period of radical change. The fundamental theme linking the objectives of the White Paper was the concern to reduce variations in performance. Publication of the Acheson Report (1998) on inequalities in health emphasised this theme. A quality framework for the new NHS was proposed in A First Class Service (DoH, 1998), which set out how variations in quality of care would be tackled. Initiatives intended to ensure consistent, highquality patient care included the setting of national standards in National Service Frameworks (NSFs), and the setting up of the National Institute of Clinical Excellence (NICE), to provide guidance to clinicians on the most clinically- and cost-effective treatments and drugs. The government claimed that the new NHS would give equal importance to cost and quality of health services and would make use of both groups of performance measures (NHSME, 1999). These major initiatives were coupled with a huge increase in funding for the NHS. Spending on the NHS in cash terms reached £90 billion in 2007/08 (Audit Commission, 2007). In the context of the massive increase in funding instituted by the Labour government, and criticised by some on the basis that extra funds had previously failed to produce commensurate increases in performance, it was necessary to demonstrate that this time the funding was being put to good use, and value for money was being achieved. The government Pinar Guven-Uslu is a lecturer in accounting at Norwich Business School, University of East Anglia. Lynne Conrad is a principal lecturer in accounting at Portsmouth Business School, University of Portsmouth. Uses of Management Accounting Information for Benchmarking in NHS Trusts Pinar Guven-Uslu and Lynne Conrad This article investigates the implementation of benchmarking in three large acute NHS trusts. The findings concentrate on the attitudes of professional groups towards the preparation and use of management accounting information for performance measurement and comparison in NHS trusts. The problems revealed in developing appropriate costing information in this organizational context suggest difficulties lie ahead as more far-reaching organizational and financial change permeates the NHS. Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 240 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA announced the publication of the National Reference Cost Index (NRCI), based on a compilation of HRG costs reported by all hospitals in England, weighted by casemix. The result was a league table, which ranked the performance of hospitals based on reported costs of procedures categorized by HRG. More far-reaching change was announced when details of the planned payment by results method of financing hospitals were published in Reforming NHS Financial Flows (DoH, 2002). This method uses the NRCI to determine the average cost of each HRG, and then pays hospitals that average sum according to the actual activity carried out for each HRG. This marks a major change in the funding mechanism, which was previously based on block sums paid by purchasers to providers, according to agreed planned levels of service. HRG costs are crucial and central to funding in the NHS, and this article looks at the extent to which experience of HRGs to date suggests that they are fit for their very important purpose. Study Design Our case study used multiple research tools including preliminary interviews, questionnaire responses, follow-up interviews, investigation of archival materials and informal observation in the trusts. The basic methodology was to conduct intensive analysis of a relatively few cases, rather than a more superficial analysis of a larger number. Perceptions of four groups of professionals working in three large acute trusts in the same region were analysed. As these trusts were geographically close and working in partnership under the control of the same health authority, it was considered that they would provide an appropriate context for a comparative study of benchmarking practices. At the first stage of data analysis, the data were separated according to three data collection methods for professional groups and trusts, and then subdivided according to specified aspects of preparation and uses of management accounting information. Twentyfive in-depth interviews were undertaken. A semi-structured interview schedule was developed which contained a common core section and also an issue specific section. The interviews were fully transcribed and subjected to content analysis. Forty questionnaires were collected from the respondents and were crosscompared between trusts and professional groups for statistical differences. Emergent themes were identified for both trusts and groups separately and cross-compared around the specified aspects of management accounting information preparation and uses to investigate the interrelations between them. The questionnaire included Likert scale questions to measure attitudes of professional towards costing information, and in order to statistically analyse their perceptions Kolmogorov-Smirnov non-parametric tests were considered most appropriate, given the relatively small sample size (Siegel and Castellan, 1988). Thus the emergent themes of qualitative data were supported with statistical significance test results. Relevant Recent Literature A number of issues were raised by this investigation. First, the issue of standard costs being employed in a very different arena from usual. Developed in the context of mass production of identical units, they are now being employed in hospitals, where no two patients are identical. Jones (1999) considers the problems associated with the use of costed HRGs as standard costs, and suggests that HRGs lack the sophistication of standard costs established in manufacturing settings in a precise engineering manner, and represent an attempt to apply average costs when ‘there is no such thing as an average patient’. Llewelyn and Northcott (2005) develop this insight further, explaining how the costing initiative in the NHS has led to the construction of the ‘average hospital’, where the primary objective becomes ‘being average’ without any clear notion of whether that average is good or bad. This is a major consequence of the failure to link cost and quality measures. Being average avoids attracting too much attention, and Smith (1993) suggests that the media attention that comes with publication of public sector league tables means that public service organizations aim for convergence, so that they do not stand out as exceptional in either good or bad performance. Furthermore, these standard costs are now being used not only for internal performance management purposes, but to determine the flow of funds to hospitals. Second, HRGs are held up as an example of benchmarking. Consideration of the principles of benchmarking reveals considerable differences between the normal private sector practices used in implementing benchmarking and those being adopted in the NHS. Particular issues are the contrast between the confidentiality ethos surrounding benchmarking in the private sector and the requirement for openness in the public sector. Another contrast is the ‘no-blame’ culture Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 241 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA surrounding benchmarking in the private sector—the objective is continuous improvement. In the public sector, by contrast, the publication of league tables is being used to ‘name and shame’ poor performers (Northcott and Llewellyn, 2003). However, the tensions between the openness required by public accountability and the ‘no-blame’ culture associated with continuous improvement are not explored. Co-operation between medical, accounting and managerial staff is an important aspect of successful implementation of benchmarking in the health sector, and this study investigates the perceptions of different professional groups towards benchmarking, recognizing the tensions found to exist between these groups in previous literature (Broadbent et al., 1992; Jones and Dewing, 1997). Third, it could be argued that benchmarking is a misnomer, and that the developments discussed here more closely resemble yardstick competition. Dawson et al. (2001) draw parallels with the use of the NRCI and the use of yardstick competition in utility industries such as water where there is a lack of competitive pressure. Yardstick competition tries to encourage efficiencies in these industries by measuring their costs relative to each other, and using that information to inform regulation. Using yardstick competition enables a regulator to force firms facing different markets effectively to compete on cost, by relating the utility’s price to the costs of firms identical to it. Shleifer (1985) likens Medicare, a prospective payment system for reimbursement of healthcare costs in the US, to yardstick competition. Patients are divided into 500 diagnostic related groups (DRGs), assigned to a group based on a physician’s diagnosis, and Medicare pays a fixed fee per patient. The fee paid is the average of costs of treating patients who fall into a particular group taken across comparable hospitals over the previous year. When introduced, the system was found to decrease typical length of stay for a patient possibly because payments, unlike costs, do not increase with length of stay. This system is imperfect because it does not fully account for heterogeneity (i.e. does not adjust for severity of illness, or equivalent duration of stay), and highlights the problem of moral hazard as doctors seek to beat the system. This article adds to the existing literature by investigating the quality of the standard costs used for performance measurement in the NHS, by ascertaining whether existing practice is more akin to benchmarking or yardstick competition, and thus evaluating whether use of the NRCI is likely to contribute towards the provision of high-quality, costeffective patient care. Determination of HRG Costs HRGs are intended to have clinical coherence so that both managers and clinicians can use them as a common language. Two fundamental processes are used in the definition and preparation of HRGs: •A quantitative process, based on the analysis of national data using length of stay as a proxy for resource use, to identify optimal groupings from a statistical viewpoint. •A qualitative process, based on the consideration of these statistical optimal groups by clinical groups, to validate their clinical homogeneity. HRGs group together treatments that are clinically similar, consume similar quantities of resources and are likely to be similar in cost. Costing for HRGs involves building up costs from care profiles for specific clinical procedures. Each costed HRG comprises a weighted average of these grouped procedure costs. HRG costing then divides total costs attributed to an HRG by the total activity occurring within that HRG. The basic unit of activity is the Finished Consultant Episode (FCE), i.e. the treatment provided for a particular patient by a specific consultant. One admission does not necessarily constitute only one FCE, as some patients may be treated by several consultants, and some hospitals classify this as several FCEs. Clearly, the higher the activity level in terms of FCEs, the lower the average cost per HRG. NHS costs are calculated on a full absorption basis, meaning that all costs of health treatment and support services should be absorbed into the appropriate speciality, service or programme. The cost allocation process requires the management accountant to analyse costs of hospital services according to their behaviour, distinguishing between direct, indirect and overhead costs as well as fixed, variable and semi-variable costs. This is not a straightforward process, and ideally requires a good understanding not only of cost structures and behaviour in different services but also of clinical procedures and practices for different treatments. Similarly, the process would be facilitated if clinicians were introduced to costing methodology and gained an enhanced understanding of the accountant’s role (Jones, 1999). The NHS Costing Manual (NHSME, 2001) Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 242 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA provides some guidance to be followed by all NHS trusts, but states that: ‘Coding procedures for health services are so complicated that the current detailed methodology is only able to make the cost information meaningful comparatively but not absolutely’. NHS trusts cost as follows. Direct costs are attributed directly to the generating services, specialities or programmes, and then indirect and overhead costs are allocated across these using a two-stage allocation and apportionment process. First, these costs are traced to cost centres, termed ‘support services’ (catering, laundry etc.). They are classified as fixed, variable or semi-variable and allocated to costing pools, termed ‘patient treatment services’ (wards, pharmacy etc.). In the next stage, service costs are apportioned into specialities (general surgery, paediatrics, etc.) using appropriate ‘cost drivers’. For example, for pathology services, the cost driver is number of tests for specialities; for ward-based costs, the cost driver is number of bed days (or length of stay); for theatre-based costs, the cost driver is theatre hours. Then the NHS manual defines ‘highlevel control totals’ where allocated and apportioned indirect and overhead costs are summed up with direct costs. These are grouped into four main categories: in-patient elective, in-patient non-elective, day case and outpatient. Each group is constituted of several weighted average HRG costs. These weighted average costs are used to set up resource profiles for each HRG which include a list of all resources used and their associated costs. HRG costing divides the total cost attributed to an HRG by the total activity occurring within that HRG. The empirical research investigates the use of HRGs as the building blocks of comparative cost data for benchmarking and performance improvement. Empirical Findings This section discusses the empirical findings of the case study under three broad headings. First, a detailed analysis of costing systems in hospitals is afforded by investigating accountants’ views of cost allocation problems and their consequences for the quality of the comparative cost data being used for benchmarking purposes. Second, different approaches to benchmarking adopted in the three units studied are discussed. Third, the perceptions of the four professional groups involved in the study towards the usefulness of HRGs as management accounting information are compared. Accountants’ Perceptions of Costing Practices Perceptions of accountants in relation to management accounting information were grouped into three categories: •Consistency between the NHS manual and actual costing exercises. •Problems associated with overhead allocation and apportionment. •Perceptions of data quality and ways to improve it. Responses to the first three questions in table 1 indicate that accountants perceive that the NHS manual provides consistent guidelines which are followed in their trusts. Results for the last question, however, indicate that classifications of variable costs are considered to be unrealistic. According to interviewees the definitions of direct, indirect and overhead costs in the manual were not very clear, sometimes very subjective and often confusing, providing an opportunity for cost shifting between activities. Another technical problem related to the treatment of semi-fixed and stepped costs, where areas of guidance were mentioned as being ‘still ambiguous’: I do not think that the manual is prescriptive enough to dictate all costs. It is unclear how they Table 1. Accountants’ perceptions of the NHS costing manual. Variables Unit A Unit B Unit C Overall score HRG costings are consistent with NHS manual 6 7 6 6 NHS costing Manual provides consistent guidelines 6 6 4.5 6 Variable costs in my trust are classified in accordance with the manual 6 6 6 6 The manual variable cost classifications are inconsistent with reality in my trust 3 5 3 3 Notes: Median scores derived from Likert scale 1–7, from strong disagreement (1) to strong agreement (7). All range scores were 1, 2 or 3. Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 243 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA want you to actually calculate the costs. So each trust will end up deciding for itself what the manual meant. The more detail you get into, the more it is going to be different for each trust (director of finance, unit C). Overhead Allocation and Apportionment Overhead apportionment was considered one of the most important technical difficulties by seven of ten respondents. Four suggested that current practice impeded an assessment of why particular areas may be expensive or inexpensive, and made comparisons between units very difficult. The overhead apportionment for each unit based on the costing manual guidelines should ideally be based on the same or very similar procedures. The scores for the first question in table 2 on subjectivity in overhead apportionment indicate mixed responses, although according to nine of ten respondents the bases of apportionment were being improved. Shifting costs between HRG and non-HRG activities was also directly related to bases of apportionment. The evidence suggested that the majority of respondents believed that there was scope for cost shifting, and a significant number of accountants believed that this would be used to address problems of HRG costs above national average. Others did not support this idea, however. The deputy director of finance of unit A commented: Cost shifting cannot be the only and major issue resulting in average cost difference but it should be considered when investigating cost differences. Data Quality Overall, the findings suggested that there were inconsistencies between the NHS costing manual and cost allocation practices of units. The classification of costs as fixed, variable and semi-variable, and allocation and apportionment of overhead costs to HRGs were not standardized. This limited the usefulness of the cost database for fair comparison and was considered a major reason for decreased data quality and validity for benchmarking purposes. On the other hand, accountants argued strongly that HRGs had potential uses for benchmarking length of stay data for certain treatments and also for determination of cost drivers for certain specialties. Medical HRGs were also considered to be inadequately developed. Feeder systems were described as not sufficiently sensitive to reflect the differences in drugs, nursing dependency etc., which are usually the main cost drivers in medical HRGs. It was also mentioned that within an HRG there was scope for significantly different mixes of procedures, which may involve very different costs. The casemix within a particular HRG can have a significant influence on the cost and length of stay data. Clinical variations in techniques used also affected HRG costing data and increased the inconsistencies between organizations. Different Approaches to Benchmarking In unit A, a newly-appointed benchmarking team was responsible for leading and implementing benchmarking in the organization. The deputy director of finance who led that team was concerned about integrating non-financial measures and comparisons into benchmarking exercises. He emphasised that comparison of financial data without reference to qualitative data was of limited use. In unit B, a management accountant was one of the key people leading change through benchmarking in the organization. He was interested in using qualitative and quantitative performance measures and was leading benchmarking processes where both measures could be used for improvement. Table 2. Accountants’ perceptions of overhead allocation practices. Variables Unit A Unit B Unit C Overall score Overhead apportionment for costed HRGs are not subjective 5* 4 2.5 3.5* The bases of overhead apportionment for costed HRGs are constantly being improved 5 6* 5 5* There is scope for shifting costs between costed HRGs and non-HRG activities 5 5 4.5 5 HRG costs above national average may be addressed by cost shifting 4 5* 4.5 4.5* Notes: Median scores derived from Likert scale 1–7, from strong disagreement (1) to strong agreement (7). *Indicates a high range value 4, 5 or 6. Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 244 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA In unit C, benchmarking was considered a departmental issue rather than an organizational one, but implementation of clinical governance in the unit was an important item on the management agenda. The CEO and the director of finance were more involved with using benchmarking exercises to compare qualitative data rather than financial data, for the purpose of controlling and improving quality of service rather than working on cost efficiency. Thus issues of clinical governance were considered to be more important in unit C, and costing comparisons for specialties were not the primary objective. The CEO commented: There is no work going on about reference costs other than Audit Commission reports. Therefore we are not leading any effort about cost efficiency through benchmarking. If there is not an incentive where we are paid on benchmarked costs, then there will not be any major effort to use these costs for benchmarking. The empirical evidence from these three units highlights the variations in how the term ‘benchmarking’ meant different things in different organizations. In units A and B, where primary responsibility for benchmarking was placed with the finance team, the importance of benchmarking of costs was clear and was given priority. The importance of integrating cost and quality measures was recognized; albeit the difficulties of achieving this remained unresolved. In unit C, in contrast, benchmarking of costs was being disregarded completely, in the absence of clear sanctions for doing so. Management Accounting Information for Benchmarking For accountants, benchmarking was mainly about comparative cost measures, and they were generally more positive than the other groups about the potential of HRG costing information: HRGs form the only database, which could help indicate areas where further investigation might be worthwhile to improve clinical practice. This could result in not only cost savings, but also more effective patient treatments (director of finance, unit B). Nevertheless, use of HRGs for benchmarking purposes was impeded by technical difficulties in accounting procedures and cost allocation methods. As a result, respondents mentioned that in several benchmarking exercises they ended up using their time and efforts inefficiently: We started benchmarking some of our services with partner units in our region. We selected certain specialities where we thought we were expensive. As a result of several days of work, we understood that the difference was because of an extra charge for cleaning staff. We could not get the benefit we were expecting from it. That was because of the different practices in overhead allocation (deputy director of finance, unit A). When clinicians realized the level of detail in which clinical practice could be tracked and recorded, this led to a degree of defensiveness and rejection of the validity of the information: The major weakness is the information or the data they use to compare our practice. We are often being compared based on poorly collected, largely meaningless information (consultant physician, unit A). You have to have reliable data to compare. Let’s say we have longer length of stay in our department but our department might be dealing with more complicated patients. Every patient is different (clinical director, unit B). Doubts about the validity of HRG costing information were further exacerbated by difficulties in HRG coding procedures. This is an area where reaching a common understanding between clinicians and accountants is crucial, but this has not yet been achieved. The coding process begins with the completion of the medical records by the clinicians and is only completed when the person who codes this information applies the episode to a very specific health resource, the HRG. This frequently entails an examination of the medical records in some detail and may involve consultation with the clinician to clarify aspects of the treatment: The medical records often run into several pages and may involve the management accountant in the application of a good deal of judgement to translate the often brief description of clinicians into an appropriate HRG (management accountant, unit A). Central managers endeavoured to promote greater understanding of coding procedures and problems among clinicians but this was not Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 245 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA easy: I know that it is quite difficult to put what we are doing into HRGs. I worked closely with accountants when they were first introduced. But I know that most of the clinicians do not use HRG information and do not know anything about coding (clinical director, unit B). In the main, clinicians’ understanding of, and approach to, benchmarking was more qualitative and significantly different to the accountants’ more quantitative approach of benchmarking service costs. For clinicians, benchmarking usually meant learning from scientifically-proven best practice, as a service manager in the pathology department of unit B explained: I am not aware of any benchmarking activity within the unit. But our department has its own benchmarking policy and external benchmarking partners. These are other pathology departments around the country. This is a private initiative, which started …before the preparation of reference cost tables. HRGs were expected to give greater visibility to the financial implications of clinical decisions, and thus facilitate control and comparison by managers of their unit’s clinical resource usage, potentially by influencing decisions on medical/surgical procedures and/ or length of stay. However, there remained areas where cost allocation procedures were inadequate for a fair and efficient comparison, resulting in different views among accountants, managers and clinicians on the effectiveness of these systems of monitoring and control. Implications and Concluding Discussion This study reveals the lack of integration of cost and quality aspects of performance in the development of costed HRGs as the basis for the NRCI. The academic literature indicates that cost efficiency and quality improvement should be considered simultaneously in benchmarking health services. However, such an approach was not observed in the studied units. The approach adopted towards performance measurement in the NHS is more characteristic of yardstick competition than of benchmarking as commonly understood in the private sector. Different units and professional groups attributed relatively different importance to cost efficiency compared to quality improvement, and thus attitudes towards uses of management accounting information for benchmarking purposes varied. In the public domain NHS trusts were compared according to their average costs and were expected to benchmark accordingly. The deficiencies in data quality revealed by this investigation, especially in relation to costing practices, casts doubt on the reliability of the average costs which provide the standard of acceptable achievement, and which hospitals are ‘benchmarked’ against. The data of this study implied that the usefulness of the comparative cost database for benchmarking was very limited because the complexity of cost allocation and behavioural implications associated with costs, obscured cost transparency and impeded the standardization of cost allocation procedures. There was a clear need for improvement of data quality for benchmarking purposes. The limitations of standard costing (Drury, 2005) result in attention being focused on a possibly spurious average. These standards and the limitations identified in relation to costing practices give rise to similar problems in the UK in relation to moral hazard as identified by Shleifer (1985) in the USA. This article explains how cost-shifting can take place (‘coding up’ of patients to more serious illnesses to ensure higher payments) and how FCEs, the main measure of activity level, can be manipulated. Achieving excellence requires co-operation and collaboration between accountants, managers and clinicians in setting high standards for excellent and cost-effective patient care. Early experiences of costed HRGs have alienated medical staff, who are found in this study to lack faith in their validity, and to ignore them wherever feasible. On the other hand, as the empirical evidence demonstrates, HRGs for quality were used by clinicians in a manner more in keeping with the usual ethos of benchmarking. Clinicians were more inclined to use qualitative data for benchmarking purposes and referred to evidence based medicine (EBM) for clinical improvement. They were also inclined to protect clinical autonomy and showed resistance to direct managerial or financial control or comparison of their practices. Failure to bring clinicians on board is very important, as accurate coding of activity to HRGs is crucially dependent on expertise in understanding the medical treatment which has taken place. This study suggests that coding is being undertaken by accountants lacking in the necessary understanding of medical terminology. Widely reported variations in costs Downloaded by [University of Liverpool] at 13:19 26 December 2012 PUBLIC MONEY & MANAGEMENT AUGUST 2008 246 © 2008 THE AUTHORS JOURNAL COMPILATION © 2008 CIPFA per HRG reported by different hospitals have destroyed their credibility among clinicians. This issue is key to the lack of enthusiasm demonstrated by clinicians towards HRGs for costing, in contrast to their recognition of their value for enhancing quality. Despite the difficulties in the compilation of a national database of costs for HRGs, the perceptions of all groups of professionals were positive concerning its use for benchmarking purposes in the future. Numerous initiatives recognize the need to complement assessment of financial performance with a range of qualitative indicators. These include the Balanced Scorecard approach coupled with the award of star ratings to hospitals (DoH, 1999, 2001), subsequently replaced by the Annual Health Check (Healthcare Commission, 2005). Nevertheless, recent pronouncements also make it likely that with the implementation of payment by results hospitals which cannot provide certain services at average cost will be required to close down these services. As Jones (1999, p. 11) remarked presciently: The compilation of a national database of costs for HRGs…will extend the power of purchasers as they make detailed decisions concerning the sources from which to procure services. It will also highlight efficiency/inefficiency and, by implication, could be used as a starting point when considering the redistribution of resources across the nation. This could lead to the restriction of the range of services provided by individual acute units, or even the closure of units, thus restricting patient choice as to the location for treatment. Issues such as the convenience of patients or visiting relatives could become secondary to receiving treatment more speedily, at lower cost, or in a centre of excellence. The announcement of an initiative to close numerous accident and emergency departments in the UK is illustrative of many reports of rationalization and reduction of services in UK hospitals: Yesterday, the prime minister attempted to persuade people that having to travel further for medical help in emergencies would be good for them. The apparently bizarre logic of this argument rested on the premise that highly specialised super-regional (i.e. not local) centres of excellence are the best place to treat lifethreatening events…Many accident and emergency departments would be closed as fullscale emergency facilities and replaced by minor injury units (Daily Telegraph, 2006). Implementation of payment by results may be viewed as the reintroduction of the internal market into the NHS, with added power as it involves greater proportions of hospital budgets and stronger financial incentives. ¦ References Acheson Report (1998), Independent Inquiry into Inequalities and Health (The Stationery Office, London). Audit Commission (2007), A Prescription for Partnership (London). Broadbent, J., Laughlin, R. and Shearn, D. (1992), Recent financial and administrative changes in general practice: an unhealthy intrusion into medical autonomy. Financial Accountability and Management, 8, 2, pp. 129– 148. Daily Telegraph (2006), Labour’s lethal logic (6 December). Dawson, D., Goddard, M. and Street, A. (2001), Improving performance in public hospitals: a role for comparative costs? Health Policy, 57, 3, pp. 235–248. DoH (1997), The New NHS: Modern and Dependable, Cm 3807 (The Stationery Office, London). DoH (1998), A First Class Service: Quality in the New NHS (London). DoH (2002), Reforming NHS Financial Flows: Introducing Payment by Results (London). Drury, C. (2004), Management and Cost Accounting (Thomson, London). Healthcare Commission (2005), Assessment for Improvement: The Annual Health Check—Measuring What Matters (London). Jones, C. S. (1999), Developing financial accountability in British acute hospitals. Financial Accountability and Management, 15, 1, pp. 1–20. Jones, C. S. and Dewing, I. P. (1997), The attitudes of NHS clinicians and medical managers towards changes in accounting controls. Financial Accountability and Management, 13, 3, pp. 261– 280. Llewellyn, S. and Northcott, D. (2005), The average hospital. Accounting, Organizations and Society, 30, 6, pp. 555–583. NCMO (National Casemix Office) (1997), Uses and Users of HRGs (NHS Executive, Winchester). Northcott, D. and Lle