19.9 Clinical audit in palliative medicine

Quality and value for money in health care is now centre stage in many countries. Whatever the funding system, most developed countries place a high priority on ‘consumer’ choice and accountability. Scandals related to poor care continue in many countries, higher public expectations, increased health-care spending, and the move towards quality service in many public and private companies have all heightened the desire to introduce quality assurance, clinical audit and evaluation into clinical care (Shaw, 1980a–1980e; Selman and Harding, 2010; Eastman and Martin, 2012) The move from palliative and hospice care from being a concern of primarily the voluntary and charitable sector, to the greater involvement of public and, increasingly, private sector services bring with it a growing need to continuously monitor quality through clinical audits. Audits allow clinicians, managers, and governments to standardize practice and support ‘best’ possible practice Shaw, (1988, 1989, 1990). Audit data also can be the foundation for future research—testing methods, identifying potential recruits, or better understanding the focus of research (Currow et al., 2010; Evans et al., 2013; Higginson et al., 2013).

Although many terms in the quality dictionary are new, many of the ideas, such as clinical review, are not. The history of quality improvement and audit can be traced back as far as 1700 bc in Egypt when King Hammurabi (the King who proposed the code of ‘eye for an eye’) introduced penalties for different degrees of ‘surgical incompetence’ (Rosser, 1985). This included a proposal that: ‘if a doctor operates on a patient and the patient dies, the doctor’s hand will be cut off’ (Rosser, 1985). In 1518, the Charter of the Royal College of Physicians included: ‘to uphold the standards of medicine both for their own honour and public benefit’. The General Nursing Council (GNC) was established in 1919 to ‘monitor nurses’ practice and conduct’. Florence Nightingale was one of the first clinicians to insist on measuring the outcome of care for her patients, to evaluate treatment. Grand rounds, postgraduate lectures, and clinical presentations all contributed to the review of medical and nursing performance.

References in the medical and nursing literature to quality assurance and clinical audit began in the late 1970s. They continue to rise, although the terms have changed over time (Fig. 19.9.1), including terms such as continuous quality improvement, clinical governance, standards, and preferred practices Shaw, (1980a, 1988, 1991; Higginson, 1994). The new emphasis on clinical audit, quality improvement, and governance has brought four main differences:

Palliative care arose out of a desire to improve the quality of care for patients with advancing disease and their families. This was based on evidence that care for people with progressive illness, and care for the imminently dying, was lacking (Higginson and Evans, 2010). Many of the early evaluations of palliative services compared hospice, home, and hospital care (Higginson et al., 2003) and some sought to influence those working in oncology and other professions (Buckingham et al., 1976; Higginson, 1993, 1994; Ling, 2000; Lee et al., 2001). Later, palliative care often led the way in developing ways to examine the quality of care generally and its core precepts were seen to align with public preferences. A recent survey of 9344 citizens across Europe found that respondents highlighted (a) a need for improved quality of end-of-life and palliative care, and access to this care for patients and families and (b) the recognition of the importance of death and dying and holistic care including comfort and support (Daveson et al., 2014).

Palliative and hospice care have grown rapidly such that there are services in all countries of the globe. This growth has been accompanied by the development of many tools and measures to assess and audit the quality of palliative care. The use of these tools is becoming more common and national and international guides for quality are now available (Dudgeon, 2007; Ferrell, 2007). There is now evidence demonstrating the effectiveness of palliative care and hospice services, based on empirical studies in different sites and systematic literature reviews (Higginson et al., 2003; Higginson and Evans, 2010; Gomes et al., 2013a). Although the importance of this work is undeniable, the lack of comparison of different models of care has been criticized (Higginson and Evans, 2010). Practice still varies from one country to another, from one part of a country to another, and from one service to another, even in simple aspects such as staffing levels and mix within a hospice or home care team, the catchment populations, the operational policies and the throughput (Higginson and Evans, 2010). Even today, we need to know which models of care work best and for which types of problems palliative care is most effective. Those providing and those who resource palliative services need to know which interventions and in what combination work best, for what kinds of patients and families, and in which type of localities. Research is important in discovering some of these aspects, but equally audit or quality assessment, is important to identify what works best, for whom and when.

Hospices and palliative services bring new therapies, such as new treatments for symptoms, support and counselling services, or complementary and alternative medicine therapies. These new therapies must be evaluated and audited to determine if and in whom these are useful. Otherwise hospices’ resources and the patient’s time will be wasted. There is a great danger on concentrating on only current concerns without reviewing previous failings and using those findings to plan improved care in the future.

Clinical audits have been slow to materialize in some settings. Antipathy to the approach is based on various arguments, such as:

None of these arguments can be supported by evidence. Numerous audits and quality assessments of palliative care have helped to improve care (Charlton et al., 2002; Cooper and Hewison, 2002; Galvin, 2002; Higginson and Bruera, 2002; Ackroyd, 2003; Burt et al., 2004; Galbraith and Booth, 2004; Hunt et al., 2004; Newbury and Hatherell, 2004; Bausewein et al., 2005; Brandt et al., 2005; Currow and Abernethy, 2005; Harding and Leam, 2005; Morita et al., 2005; Veerbeek et al., 2006; Godfrey and Poole, 2007; Kaasa et al., 2007; Le and Ashby, 2007; Payne, 2007; Amigo et al., 2008; Currow et al., 2008a, 2008b; Deane et al., 2008; Eisenchlas et al., 2008; Miller and Hopkinson, 2008; Morita et al., 2008; Tsai et al., 2008; Eagar et al., 2010; Harding et al., 2010; Nadimi and Currow, 2011; Benitez-Rosario et al., 2012).

Audit has been recommended in national strategies (Kaasa et al., 2007) and has had major impact in the development of hospice and palliative care in much of Europe, Canada and Australia. The results have been used in the following ways:

Audit or clinical governance can help most patients and families receiving palliative care, because it looks at routine practice, rather than at a few ‘special’ cases. Quite apart from mistakes, suboptimal care may be due to professional or administrative problems which tend to escape anecdotal case reviews. Audit also is important for education and training, because the structured review allows analysis, comparison, and evaluation of individual performance; it promotes adherence to local clinical policies and offers opportunity for publication of results. Educational programmes can be constructed to meet the demonstrated needs of individuals or groups. Increasingly, audit, clinical governance, or quality review are required for the recognition of training posts and for the revalidation of doctors. Royal Colleges and Faculties increasingly seek evidence of formally organized review and could withdraw recognition from departments that do not provide this.

Audit is important for those who resource palliative care, such as commissioners, primary care trusts, and health-care insurance agencies, because it provides tangible evidence that the service is seeking the most effective use of existing clinical resources and aims to improve the quality of care. This is increasingly important when competing for health-care contracts (Clark et al., 1995). Requirements for audit and the implementation of research findings may well be included in such contracts (Clark et al., 1995).

The costs of not auditing are as important as the benefits of auditing (Shaw, 1980a–1980e; Higginson, 1994). In the absence of review, there is a risk that inappropriate treatment may be undertaken, uncontrolled symptoms may cause admission to hospice or hospital, or delay discharge, or unnecessary services may be delivered. All this activity wastes the patients’ and families’ time and resources, as well as staff time and resources.

Clinical audit takes time and resources (Shaw, 1980a–1980e; Higginson, 1994). These should not be underestimated. Indeed, it may be impossible to undertake audit if a service is struggling, with insufficient numbers of staff to give even basic care such as washing and administering drugs. Perhaps the audit here is to show simply that the service cannot even measure and assess a patient’s needs properly. Indeed, simple assessments, such as whether follow-up occurs, can be useful in assessing practice (Selman and Harding, 2010). The resources needed for audit and quality assessment can include:

Computers are not necessary for audit, but they are increasingly becoming part of routine health care. A number of computer programmes are available for palliative care and other services, many from commercial companies. The audit measures can be included within these programmes, and increasingly application in clinical practice is aided by staff or patients using mobile computing devices. However, if the audit is small or in its early stages, too rigid use of computers by inexperienced staff can be a hindrance, because the need to update the computer delays the evolution of the audit.

The costs of audit mean that it is important to ensure that the audit itself is as effective as possible. What is the purpose of collected audit data if the changes are not acted on? Mechanisms to review the audit and to ensure it is effective are discussed in a later section.

Quality, as defined in many dictionaries, pertains to ‘degree of excellence’ or ‘general excellence’. But what constitutes excellence in the context of a service, who defines it, and what components should be addressed? For example, should ‘excellence’ be limited to clinical skills, or should it encompass broader aspects, such as whether the service reaches all those in need? Surely, when measuring quality of a service all the features and characteristics that bear on its ability to satisfy the stated or implied needs of the users of that service should be assessed. Quality in health care is a multidimensional concept, and as such, a multidimensional approach to measurement and assessment of that quality is required Donabedian, (1966, 1976, 1981a, 1997, 2005).

Numerous authors have identified some of the dimensions that should be included when assessing the quality of a service. For example, Maxwell identifies the following as dimensions of care that need to be measured when monitoring quality Maxwell, (1984, 1986, 1992):

Black agreed with all these, but added a further dimension: humanity, a dimension which would be particularly important for palliative care services (Black, 1990).

So how do we determine whether we are offering a good quality service if we have to consider all these dimensions? One method is to conduct a programme of review and improvement of the service, a process that incorporates clinical audit, quality assurance and quality improvement.

Donabedian and others have translated the assessment of the quality of health care into:

This model is built on manufacturing industry and, despite limitations discussed below, it has value in that it defines the steps in which health and social care are delivered.

Structural aspects influence the process of care so that its quality can be either diminished or enhanced. Similarly, changes in the process of care, including variations in its quality, will influence the output, and in turn, the effect of care on health status and outcomes. Thus, there are functional relationships (Fig. 19.9.2).

Structure is easiest to measure because its elements are the most stable and identifiable. However, it is an indirect measure of the quality of care and its value depends on the nature of its influence on care. Structure is relevant to quality in that it increases or decreases the probability of a good performance. Process and output are closer to changes in the health status of individuals and their measurement reflects the most immediately discernible attributes of care activities. However, they are only valuable as a measure once the elements of process are known to have a clear relationship with the desired changes in health status. Outcome reflects the true change in health status, and thus is the most relevant for patients and society. However, it is difficult to eliminate other causes for change, such as prior care or external events. A useful approach is to focus on the difference between the desired outcome and the actual outcome. Services can then identify whether or not their goals are being achieved and investigate any failings.

Organizational audits usually assess the structure and process of care, whereas clinical audits often measure the process and outcome of care. Although structure is easiest to measure it is furthest from influencing change in the patient and family. Outcome is most difficult to measure but is of direct relevance to the patient and family. Standards of structure or process are most useful when these are of proven effectiveness, or if there is an overwhelming consensus that these are desirable. Structure, process, and outcome measures which have been either used or advocated to assess palliative care are shown in Table 19.9.1.

Stewart et al. have extended the model, subdividing the ‘structure’ elements into (1) personal and social environment of the patient and family and (2) the structure aspects of care (Stewart et al., 1999). They also subdivided the outcomes into (a) satisfaction and (b) quality and length of life. These subdivisions are consistent with the earlier model of Donabedian, and those proposed for palliative care audit in the United Kingdom. However, Stewart et al.’s differentiation of the personal and social environment is a useful demonstration of how audit results may be different because of different settings, or with a different ‘case mix’ where patients and families have differing conditions or problems. Casarett et al. have built further on this model, proposing a minimum data set (Casarett et al., 2006).

Over the last 30 years, it has been interesting to observe how different terms have come in and out of vogue to describe the evolving approaches to assessing and ensuring the quality of health care. There are also differences among countries. Somewhat disappointingly, the literature in one country tends to ignore audit and quality improvement initiatives in another, thereby reducing the chance to learn. In the United Kingdom and many other European countries, clinical audit (as opposed to financial audit, concerned only with financial matters), became the initial term of choice. Later this was accompanied by quality assurance (which tended to include a programme of several audits, all working together). Such a programme was similar to the development of total (or continuous) quality improvement in the United States and many other countries, which incorporated a cycle called Plan, Do, Study, Act. Some of the widely accepted definitions are shown in Table 19.9.2.

There are elements common to all types of clinical audit or quality review. These include systematic assessment in a cycle that includes the monitoring of care, review of results, and change to improve practice (Fig. 19.9.3). The cycle is repeated and the process should become part of the organization’s culture. Practice changes usually are made to be consistent with standards of care that are set locally or nationally, but it is possible to enter the audit cycle at any point; standards can be developed after practices are observed and change is made based on the results.

Clinical audit is the systematic critical analysis of the quality of clinical care (Higginson, 1994). It may focus on the procedures used for diagnosis and treatment, the use of resources, or the resulting outcome and quality of life for the patient. It can be prospective, where the standards and measures are agreed at the start and are recorded on patients and families during their care, or retrospective, which looks back at the care of patients using either the clinical notes and extracting the information or by asking families. Although early forms of audit involved only single professions—for example, medical or nursing audit (Table 19.9.2)— it is now widely accepted that audit in palliative care should be multiprofessional, to reflect the multiprofessional nature of care.

The Kings Fund Centre in London originally described organizational audit as the developmental and voluntary stage towards accreditation (Higginson, 1994). Accreditation schemes exist in many countries and usually operate nationally. The audit focuses on the organization and compares specific aspects against agreed standards. These standards usually consider only the structure and process of care, and may try to identify administrative delays, uncoordinated care, poor environment or sign-posting, poor staff training, etc. Organizational standards need to be straightforward enough to be monitored by an external surveyor.

Although there are various definitions for quality assurance, a widely accepted one is the ‘definition of standards, the measurement of their achievement and the mechanisms to improve performance’. Clinical audit lies within the frame of quality assurance. Audits may involve the review of the quality of a local clinical practice on a regular basis, for example, through internal ‘peer review’ by practising clinicians.

Total quality management is a recent term, which has been defined as a strategy to get an organization working to its maximum effectiveness and efficiency. It has been facilitated by closer working relationships between clinicians and managers (Higginson, 1994), builds on the other definitions, and switches the focus from quality practised within professionals to quality within the whole organization. Thus, clinical audit would lie within a total quality management programme. It also introduces the concept of managing the quality process, such as cataloguing reports of local quality initiatives (Higginson, 1994) and using managers to ensure that improvements in quality occur.

Clinical governance has developed recently because of concerns about variations in clinical practice, particularly assessment, diagnosis, and prescribing. This is of increasing concern in modern medicine because of the rapid development of diagnostic investigations and treatments that may have potentially dangerous effects if not carried out properly. Clinical governance programmes assess what doctors (and most recently nurses) do, and examines both their training and continuing professional development, and all components of their practice. It aims to ensure and develop the practice of all clinicians. Participation in clinical governance is already required of many health-care services, and will soon be required to ensure the revalidation of doctors in many countries (Ferrell et al., 2007).

The different approaches to audit and the assessment of quality can be categorized according to two axes. The first is defined by the identity of the auditor. This could be clinicians, managers, or an organization. The second is defined by the object of the audit, whether it focuses on the care of an individual or a few patients, or on the whole organization or population. Common forms of audit according to these two axes–internal versus external and individual versus organization/population—are shown in Fig. 19.9.4. The type of audit should be determined by the setting and goals. Those focused on improving specific practices often are more oriented towards internal and individual assessment. They are more likely to be educational and voluntary. Those that are external and focused on the overall clinical quality of care are more likely to rely on organizational or environmental standards, determine whether professionals are employing proven high quality treatments, examine structural elements such as staff mix, and note whether quality activities such as a clinical audit programme is in place. These types of audits are more likely to be mandatory and to provide funding or accreditation (see Fig. 19.9.5).

Clinical audits and other quality activities should be actively embraced by those working in palliative care. Specific steps should be taken to implement the audit processes:

The definitions of palliative care and palliative medicine provide good guidance on the goals of palliative care, and the appropriate outcomes that might be measured in clinical audit. These include aspects of pain and symptom control, improving the quality of life for the patient, relieving fears and anxiety, and caring for the family members or carers. Therefore, in analysing the goals or standards of care, many have audited their effectiveness in improving outcomes such as control of symptoms, their effect on a patient’s quality of life or psychological well-being, or the patient’s or family members’ satisfaction with care. A European-wide survey of public attitudes recently found that pain, breathlessness, and burden to others were top concerns that needed to be addressed near the end of life (Bausewein et al., 2013).

Specialists in palliative care often focus on end-of-life care and those undertaking clinical audits may choose to evaluate specific concerns encountered in the care of the dying. Kellehear described the features of a modern ‘good death’ as ‘awareness of dying, social adjustments and personal preparations, public preparations (legal, financial, religious, funeral, medical), work or activities reduced, and farewells’ Kellehear, (2000, 2001, 2009) Singer et al. identified five domains of quality end-of-life care: receiving adequate pain and symptom management, avoiding inappropriate prolongation of dying, achieving a sense of control, relieving burden, and strengthening relationships with loved ones Singer et al., (1999, 2000). Payne et al. identified similar concerns from family members (Payne et al., 1999). These definitions might suggest audit of aspects of patient and family awareness of the illness or their planning of personal and public preparations, and such aspects have been included in some measures, such as the Support Team Assessment Schedule (Higginson et al., 1992; Bausewein et al., 2011a).

In many countries, the role of palliative care in supporting, advising and educating other professions is stressed. For example, family practitioners and hospital clinicians may identify educational needs for symptom control and patient and family support. Therefore, another form of audit could examine the educational and supportive role of palliative care services.

Poor communication is a frequent cause of stress for patients and families. Doctors and nurses need to communicate well with patients and their families rather than withdrawing or appearing hurried or abrupt in their manner. Communication also is needed among professional staff caring for the patient and family member, to ensure liaison and to prevent duplication or delay. Usually patients and family members want to be involved in decision making, and patients want their family members to make decisions when they are not able to, although this wish is not universal Daveson et al., (2013, 2014). These aspects are also goals suitable for audit.

Total pain has been described as including physical, emotional, social, and spiritual components. Although the earlier discussions have covered the emotional and physical aspects of palliative care, it is also important to consider audit of the social and spiritual aspects. Social aspects might include whether the patient and family have sufficient practical and financial support to remain at home, or whether they have been cared for in their place of choice, usually home (Gomes et al., 2013b). Spiritual audit might consider whether patients are able to raise aspects of spiritual concern to them and find a mechanism to relieve problems, or whether patients are in any form of spiritual crisis. Guidelines for spiritual practice have been developed, and several of the more recent measures that can be used for audit include spiritual aspects, measuring items such as peace and feeling good about self Selman et al., (2010, 2011).

There are also simple goals, such as place of death, or meeting other patient or family preferences, or access to services by certain patient groups (such as those from ethnic minorities, or non-cancer conditions) that might be suitable for audit. In auditing the processes and outcomes associated with these goals, clinicians must recognize and accommodate variation. For example, although around 70% of patients would prefer to die at home, increasingly patients choose hospice as a second choice, wishes are very individual, are little understood in some sections of society and are altered by experience (Gomes et al., 2013b). Thus this indicator, like other process indicators, needs to be reviewed cautiously.

How can these goals be turned into standards? Any goal or standard which is set must be measurable, sufficiently challenging but achievable. It would be unrealistic to set a goal or standard that all patients would be free from pain, but it would be reasonable to set a goal of what proportion of patients’ pains might be controlled and in what circumstances pain might be uncontrolled. A baseline could be established from current practice. Standards of action when pain is uncontrolled could be audited. Such an approach has been used in developing audit programmes in many countries, including Africa (Harding et al., 2010; Selman and Harding, 2010).

When standards are ill-defined, achievement of goals can be compared across services. This approach has been used successfully by the Australian Palliative Care Outcomes Collaborative, which has compared data on the assessment of symptoms and concerns using the Palliative care Outcome Scale (Eagar et al., 2010).

For all of the above aspects, measurement tools or indicators are needed. There are now a number of these available in palliative care which have been validated to different degrees. The Association for Palliative Medicine of the United Kingdom established a working party to review these and subsequently published a ‘which tool’ guide (Clinical Effectiveness Working Group et al., 2002). The group carried out a MEDLINE search to identify potential measures and to check whether they had been used in more than one centre. Then the group provided a preliminary guide as to whether they would recommend the tools in research or clinical audit palliative care.

More recently, a Europe-wide project PRISMA surveyed clinicians internationally on outcome measures used in clinical audit, clinical practice and research, and examined the barriers to using outcomes in routine clinical practice and for audit (Bausewein et al., 2011b; Harding et al., 2011). Key findings were that:

Some measures used in clinical audit have been validated in many languages and settings. Table 19.9.3 shows examples of these. The two most common multidimensional measures used in clinical audit, the ESAS (Bruera et al., 1991; Amigo et al., 2008; Richardson and Jones, 2009; Bergh et al., 2011; Carvajal et al., 2011; Selby et al., 2011; Watanabe et al., 2012a; Hui and Bruera, 2013) and the POS (see Box 19.9.1 for an example of one item in this measure, pain) (Hearn and Higginson, 1999; Bausewein et al., 2005; Brandt et al., 2005; Stevens et al., 2005; Siegert et al., 2010; Higginson et al., 2012) (see also <http://www.pos-pal.org>) have both had extensive research into their validity, reliability, and factor structures, as well as translations and cultural adaptations in many languages. This includes determining factor structures, clinically important differences, and use for screening for depression and other symptoms (Siegert et al., 2010; Bergh et al., 2011; Selby et al., 2011; Zimmermann et al., 2011; Harding et al., 2013; Watanabe et al., 2012b; Hui and Bruera, 2013; Saleem et al., 2013). POS builds on an earlier widely used measure, the Support Team Assessment Schedule (STAS) (Box 19.9.2) but was designed for use by patients as well as staff and families.

An alternative to multidimensional assessment is audit of single symptoms or other problems, for example, constipation, communication, or access to care. This is called topic review. Other approaches are also possible, although seldom used. Box 19.9.3 reviews these.

In contrast to the use of these measures to perform internal audits of individual patient care issues, other strategies often are used for organizational audits or peer review systems. These have tended to employ indicators that can be used in inspections of services, and concentrate on the structure and process of care, such as:

Ideally, the audit cycle should include a period to implement change to improve quality based on the findings of the audit. Data should be discussed with the whole team to determine what changes might be appropriate. Decisions also may involve a review of the literature or consulting with other individuals to determine whether they have solutions to particular problems. For example, in London we identified that our control of breathlessness in the last weeks of life was not reaching our targets: this was the most severe symptom in patients (Higginson and McCarthy, 1989). To decide what change was needed we had to consult chest physicians, physiotherapists, and other colleagues about what treatment might be appropriate in the home, review the literature on effective treatment, and examine the possible causes of breathlessness. This led us to substantially change our practice and it also has probably led to further research into the management of breathlessness.

In larger specialist palliative care services, difficulties in implementing the change have been described (Higginson et al., 1996). These include the following:

The change management process is a vital component of clinical audit but is often neglected. Continuing education is important and can be targeted to ensure that those to be educated feel part of the teaching or learning process. A systematic review by Antunes et al. shows how the process of introducing routine outcome measurement into practice requires leadership, facilitation, education, and ongoing support, including attitude change in staff (Antunes et al., 2014).

It is appropriate to ask whether audit is yet relevant in developing as well as developed countries? Palliative care programmes in these countries have relatively few health-care professionals, and these have varied levels of training. In addition, some patients and families have high levels of illiteracy, making self-completion of assessment forms impossible. The burden of illness and symptom distress in patients dying of AIDS, tuberculosis, malaria, and even cancer-related syndromes may be different, just as belief systems vary, and therefore assessment tools may need to be adjusted in these populations (Higginson and Bruera, 2002).

Nonetheless, the professionals who deliver health care in the developing world document the results of interventions and use them to guide future direction in a manner not different than professionals in developed countries. Symptom control, information needs, and physical, emotional, spiritual, and social support for patients and families are important, just as in Europe, Australia, and North America (Selman et al., 2009). It can be argued that the limited resources make it even more imperative to ensure that, as the programmes evolve, they are capable of serving the largest possible number of patients and families in the most effective possible way. Audit is one way to minimize the risk of failure, and to learn, at an early stage, about potential problems and to identify success. A good example of a successful adaptation are the programmes of clinical audit now well established in Africa, with culturally adapted and validated forms of existing measures, in particular the African APCA POS (see Fig. 19.9.6). The tool is used to aid assessment, conduct clinical audit and in research (Harding et al., 2010). Considerable progress has now been made in Africa with audits using this measure to identify successes and shortcomings and ways to improve care (Harding et al., 2013).

Audit approaches and methods are now well advanced in palliative care, especially in clinical audit. We have a choice of possible and already tried methods and measures which we can adapt for our own needs, rather than having to undertake much of the development ourselves. Practical tools for clinical audit may use multidimensional measures, such as the Support Team Assessment Schedule, the new shorter Palliative care Outcome Scale, or the Edmonton Symptom Assessment Scale, or specific measures for topic audits. Clinical audits which use satisfaction surveys or surveys after bereavement are probably more costly, but are still possible. Apart from completing the audit cycle, clinical audit can look to developing clinical protocols for treatment, or algorithms to predict patient problems and the need for specialized care. Organizational audit and peer review is also now well established, and different systems are available.

Audit, or the various alternative terms that describe this assessment of the quality of care, is here to stay, and is now widely accepted. But it requires resources, and so it must be sure to benefit patients and families, be kept as simple and efficient as possible, and have a strong educational component. Further work is needed to evaluate the impact of different audit approaches and methods on improving care, so that we know which approach is most cost-effective. In addition, there is a need to develop and test methods of audit in developing countries. If palliative approaches extend backwards to include patients earlier in care, rather than those just near to death, then the audit could become a means for clinical dialogue and education between specialties. Palliative medicine could take the lead in encouraging this, promoting methods among their medical and surgical colleagues and presenting their own results.