UK Essays Services:

Free Help:

Economic Haemodialysis Costs

Appropriate estimation of staff costs for economic evaluations: A case study in haemodialysis

The substantial resources devoted to publicly funded health care mean that costs matter. Reasons for costing are numerous and include resource allocation, budgeting, and service planning and therefore it is unsurprising that there are different approaches to define and estimate 'costs'. However, the nature of health care presents many challenges to costing since many resources are shared across multiple patients, interventions and services.

The thesis examines methods to cost health care for economic purposes. Costing must balance a number of competing objectives in relation to data quality and feasibility. The thesis assesses the implication of these objectives on the choice both between costing approaches (top-down and bottom-up and between methods to quantify and cost staff inputs for different patients. It considers the impact of patient heterogeneity on nursing costs for chronic haemodialysis, a life-saving but costly treatment for end-stage renal failure that is used as a case study for the empirical work.

Background to costing

This section describes the background to the thesis in terms of four key aspects of costing: the purposes for costing, application of economics in costing health care, challenges due to the definition of health care outputs, and costing approaches.

Purposes for costing

As highlighted above, there are numerous reasons to undertake costing in health care, which include:

decision-making (e.g. how to allocate resources, to set prices or tariffs for reimbursement, and to plan new services or changes to existing services);

monitoring and control (e.g. budgeting, activity/cost analysis, cash flow analysis, benchmarking and assessment of efficiency);

valuation (e.g. for financial accountability (financial performance and standing), and to value resource used for programmes, services, or illnesses / diseases).

Costing may follow either accounting or economic principles. Accountancy is based on financial costs, the money actually spent on resources. Therefore, items that are donated and unpaid (informal) care have no value in accountancy terms. In contrast, costing using economic principles is based on the 'opportunity cost' of items. This is the value of the benefits that would be obtained from using the same resources for their next best alternative (regardless of whether the resources were purchased). Therefore, items that are donated and unpaid (informal) care have an 'economic' value, for example the replacement cost and the carer's lost earnings. For example, the opportunity cost of creating a new nurse specialist role might be the value of either nursing time no longer available for other tasks or patients, or the other therapies such as drugs etc. that could have been bought. The practicalities of valuing opportunity costs are revisited later (section 2.2.4).

Application of economics in costing health care

The application of economic principles in costing is potentially useful in many areas of health care decision-making, monitoring and control, and valuation. Economics involves the study of how resources and incentives affect choices about the production, distribution, and consumption of goods and services. Resources do not refer to money, but the resources used to produce the goods or service, namely labour (e.g. nurses, doctors), capital (buildings, property, equipment, etc.) and materials (e.g. drugs, consumables). Two concepts are central to economics. First - the concept of 'scarcity' - there are never sufficient resources available to meet all potential demands and so choices must be made. Second, the choice to use resources in one way precludes their use for other alternatives (i.e. the concept of opportunity cost).

A further key concept is that of the market, a mechanism that allows exchange of information about demand and supply between potential buyers (consumers) and sellers (producers) and facilitates the exchange of goods or services. The demand from consumers relates to their value of, or willingness to pay for, the product. Supply from producers relates to the costs to produce the product and overall price charged for the product affect the quantity supplied. In theory, a perfect market (free from regulation) and based on individuals' pursuit of self-interest would enable goods and services to be valued and priced, and would be efficient and maximise benefits to society. Crucially, in a perfect market, the prices of transactions would reflect the opportunity cost because the individual buyers and sellers would have traded-off other possible uses for their resources. Furthermore, the market would determine the most efficient use of resources.

Whilst in reality no markets work perfectly, the characteristics of health care lead to collapse of all the basic assumptions underlying the market approach (Donaldson et al 2005). Furthermore, the nature of health care renders it susceptible to government intervention and regulation to achieve greater health benefits to society on efficiency, equity or moral grounds. For example, uncertainty about the need for health care and the associated financial risks mean that payment for health care is rarely solely direct from patients. Instead, payment is commonly by a third party (government or insurer) with funding from taxation, compulsory social or voluntary private health insurance. As a result, the third party, such as a Primary Care Trust, interrupts the link between the consumer and producer (health care provider) and whilst people using health care may pay a contribution (called co-payment), they do not bear the full cost. This, in turn potentially distorts 'prices' because they may no longer reflect the amount consumers would be willing to pay or because health care providers may lack awareness about their production costs.

Another consequence of 'market failure' is that since the market does not decide how resource should be used. This is a reason for the development of economic evaluations that provide information to assist decision makers to use resources efficiently and effectively. These evaluations offer an explicit approach to compare the expected costs and benefits associated with different treatments, interventions or technologies such as devices, or service strategies. Efficiency is important in two ways - allocative and technical efficiency. Allocative efficiency addresses how to maximise benefit from available resources by deciding whether to allocate resources to a given objective (amongst competing objectives) and if so at what level. For example, if and how (what level) should NHS resources be allocated to provide cancer care, fertility treatments, cosmetic surgery, etc.?. Technical efficiency concerns how to meet a given objective by maximisation of output from given resources, or minimisation of costs for a given output. For example, given a decision dialyse patients with renal failure, should they receive haemodialysis at hospital or home?

Various bodies commission and potentially use economic evaluations (typically cost-per-QALY or cost-effectiveness analyses). For example, the Department of Health's Health Technology Assessment Programme commissions evaluations and the National Institute for Health and Clinical Excellence (NICE) commissions technology assessments and produces guidance to the NHS in England and Wales. In each decision context, service planners weigh up economic evaluation information alongside a number of other factors including equity issues about the fairness of funding and distribution of health care resources.

A further application of economics principles is programme budgeting and marginal analysis (PBMA). This examines the use of existing resources (programmed budgeting) and assesses the possible impact on costs and benefits of changes (marginal analysis). It thereby offers a pragmatic approach to funding new services or altering the scale of existing ones within limited resources and, unlike most economic evaluations, acknowledge which budgets gain or lose. The technique has been widely used despite the major challenges of organisational behaviour and the interactions of stakeholders (Mitton and Donaldson 2004).

Challenges due to the definition of health care outputs or products

The nature of health care presents several key challenges for costing. Most industries produce a number of readily identifiable products to sell. This is not so in health care, which serves multiple purposes. For example, Fetter (1991), suggested that hospitals have two separate production functions:

to produce standard or intermediate outputs - goods or services such as meals, laboratory tests, x-rays or inpatient days, and

to diagnose and treat patients (which involves use of the various intermediate products), the main overall function and product. For example, a hospital may treat patients across all age groups, and according to numerous patient categories, e.g. emergency / elective admissions, medical / surgical, or provide general care with no specific treatment.

However, patients are unique due to variation on numerous characteristics: physical, mental, social, and clinical diagnosis, procedures or illness severity. This variation, or patient heterogeneity, is important because it makes it inherently difficult to define the 'products' of health care. Moreover, due to the complex interplay between the production processes and patients, definition of final 'patient-products' in terms of a 'market' is difficult because there are no uniform (homogenous) products for which there can be exact quality control guidelines.

Despite the problems of definition, it has been recognised that variation between overall products was important in understanding cost differences both between hospitals and between patients within a hospital (Fetter 1989). Consequently, there have been various attempts to develop suitable classifications of health care 'case mix' - products or patients - with a manageable number of definitions. The development of the Diagnosis Related Groups (DRGs) was an important attempt to provide a classification to both examine and control expenditure, and to make reimbursement fairer (Fetter 1989, Fetter 1991). An equivalent classification, Healthcare Resource Groups (HRGs), was developed in the UK for similar reasons.

The nature of health care also poses problems to costing in terms of linking resource use inputs to 'products'. The complexity of the production process means it is difficult to determine how to attribute multiple inputs (staff, services and departments) across shared 'products'. Information may be gathered on resource use or unit costs at various levels of aggregation (e.g. diagnostic tests, bed-days, by procedure). However, in costing, although a fundamental requirement, it is often difficult to find (or collect) the resource use data in units that match those of the unit costs. A key challenge to costing, especially in terms of staff inputs, is patient heterogeneity that is associated with different resource use implications. For example, whilst a degree of variation is expected, some patients may require consistently more or less staff input that might be delivered by multiple staff (who in turn, each typically care for multiple patients).

Patient heterogeneity is a particular challenge to costing changes that involve removing a defined sub-group of patients from a heterogeneous ‘usual care' group. A key issue is how to separate the staff costs within the ‘usual care' patients in order to ensure a like-for-like comparison. Costing the new service or setting is usually relatively straightforward because patients are well defined and more homogeneous in their care needs.

Costing approaches

Costing entails three steps; i) identification of resource use; ii) measurement of resource use in meaningful units; and iii) valuation by multiplying the measured resource use by a monetary value, the unit cost. Due to the nature of the approach, in top-down costing (discussed below) the second and third steps may merge.

The health economics literature describes two broad approaches to costing - top-down and bottom-up costing (Drummond et al 2005). Top-down or gross costing involves use of aggregated financial and activity data to breakdown expenditures into the cost estimate of interest. This requires the researcher to make assumptions about how to apportion costs to the activity. For example, the cost per bed-day could be the total ward costs (staff, equipment, drugs, consumables, etc.) divided over the occupied patient bed-days. Bottom-up or micro costing involves collecting or estimating resource use data on all the component parts that contribute to the activity and deriving a unit cost for each. These component costs are then added together to derive the cost for the activity.

Advantages of top-down costing are that it usually offers a more global impression and is cheaper and quicker to obtain. In contrast, typically bottom-up costing is thought to offer increased precision and richness of data, and allow greater insight into activities and costs (Drummond et al (2005).

A key issue for the thesis was to examine the choice between the two approaches, identifying the trade-offs and examining both costing advice and empirical evidence about the impact of choosing one approach over the other.

The importance of staff costs - background to the thesis

Staff are a major health care expenditure as salaries constitute approximately 60% of NHS spending (2003, http://www.performance.doh.gov.uk/HPSSS/TBL_E3.HTM). The dominant staff group is nurses whose salaries accounted for 70% of staff expenditure (Department of Health 2000). Therefore, staff inputs are likely to be important cost consideration in numerous types of evaluation, including where:

staff are the major part of the intervention e.g. psychological interventions such as counselling patients;

there are differences between the comparators in terms of staff time inputs

to administer the intervention;

as a result of the intervention, i.e. where patient outcomes have an impact on staff contacts or length of stay;

there is substitution of roles, services or settings since these may result in different usage patterns for staff time or skills.

Challenges experienced in trying to measure staff inputs in a previous research project, the Renal Satellite Evaluation study (RSU study, Roderick et al 2005), was the motivation for the thesis. This evaluation compared local with more central service provision and is described in detail later (section 4.5.2). In summary, the study aimed to evaluate the costs and outcomes of haemodialysis in two types of setting - renal satellite units (RSUs) and their parent main renal units (MRUs). Whilst care involved a range of staff including doctors, technicians and clerical workers, nurses provided the major routine input. RSU patients were relatively homogeneous whilst MRU patients were heterogeneous since some were sicker or more unstable than those at RSUs. At that time, it was not possible to resolve how to allocate the cost of nurses' time to the different types of patient and so the evaluation approximated by averaging costs across all patients. The impact of not costing on a like-for-like basis was not clear. In theory, the averaging should have inflated the costs of the sub-group of patients at the MRU and therefore introduced bias in favour of treating patients at a RSU. However, it was not clear whether the potential differences in resource use between the patient groups were sufficient to affect costs in a meaningful way. The thesis examines this and the following other questions that arose. In similar circumstances, do researchers i) measure staff time and if so, how, or 2) average costs across patients, and iii) acknowledge the potential confounding due to patient heterogeneity?

As highlighted above, the need to measure staff inputs is potentially important in many comparisons and role substitution is an obvious example. In ‘nursing', due to various factors, there has been much interest in potential substitution or extension of roles, new roles, generic workers, and the provision of specialist vs. generalist care (Royal College of Nursing 2003b). A Department of Health programme has been looking at workforce capacity (numbers and skills) to deliver the NHS agenda, and improve productivity by redesign of services, new ways of working and greater efficiency (http://www.dh.gov.uk/en/Policyandguidance/Humanresourcesandtraining/Modernisingworkforceplanninghome/index.htm). Local initiatives have included Workforce Development Confederations to plan and develop the health care workforce. There has been emphasis on moving work from doctors to other health care professionals and from health care professionals to the support workers, partly to decrease doctors' working hours in response to the European Union Working Time Directive (The Working Time Regulations 1998). Moreover, changes in nursing training and the instigation of supernumerary status for students necessitated the use of support workers. Whilst some changes have been borne out of necessity, other role substitutions have been proposed to improve skill mix efficiency or to find cheaper ways of working (Gibbs et al 1991). In addition, other evaluations of care delivery have been motivated out of the desire to expand the quality of services by changes in staff roles and / or settings (e.g. specialist rather than generalist care).

Given such interest and the potential budgetary impact, it is therefore surprising that economic evidence is generally lacking. Richardson et al (1998) found few cost-effectiveness studies on doctor-nurse substitution and the evidence was mostly dated (1970s and 1980s) and from the US. They found that many studies were poorly designed, had a small sample size and used single sites that limited the generalisability of the results. Systematic reviews by Horrocks et al (2002) and Laurant et al (2004) comparing nurse practitioners and doctors in primary care found similar methodological limitations to the published studies. In general, there appeared to be no differences in health outcomes, but patient satisfaction was greater for nurse practitioner care. However, nurses' consultation times were longer, they recalled patients more frequently, and requested more investigations. The effect on costs was variable, although few studies provided cost data, and often these were underpowered and had used different approaches. Moreover, other reviews have challenged the assumption that nurses are a cheap alternative to doctors (Spilsbury and Meyer 2001, Watts et al 2001).

Substitutions of services or settings are also situations that may result in different usage patterns for staff time or skills. For example, the ‘NHS Plan' outlined increased investment in intermediate care facilitates to promote early discharge or prevent admission (Department of Health 2000a). The NHS Confederation (2006) made the case for the continuing reduction in hospital beds, increased care in the community and in some cases care from specialists rather than the local hospital. There is an extensive literature on various intermediate care schemes for early discharge or admission prevention. These embrace ‘hospital at home', early supported discharge, rapid response teams, community rehabilitation teams, day hospital and day care centres, community hospitals, nurse-led units, and care homes (Roe 2005). A Cochrane review of hospital at home schemes by Shepperd and Iliffe (2005) found that patient outcomes were generally similar. They concluded that whilst patient satisfaction may be higher for hospital at home, the carers' burden was greater and there was little evidence of cost savings to the health service although relatively few included the cost impact. Overall, Little (2005) concluded that there remained a shortage of evidence on the economic benefits of intermediate care and such schemes do not always save money.

Measuring staff inputs

Measurement of staff inputs presents several important and potentially inter-related considerations. Staff inputs are multi-faceted and encompass not just time, but the personnel involved (i.e. nurses, health care assistants, doctors, etc.), grade mix (within a professional group), salary level, and skill mix (i.e. variations in experience or proficiency). A further complication is the potential interaction between staff time inputs, the quality of care and the effect on patient outcomes. Measurement of staff inputs can either focus on staff (individual tasks or aggregated categories e.g. administering drugs) or focus on the patient (e.g. care for a dialysis session).

Staff-patient interactions have inherently flexible or ill-defined boundaries, which often pose challenges to the consistent and accurate categorisation and measurement of staff activity. However, some staff-patient interactions are better defined than others are. For example, if consultations are booked into fixed appointment times, the interaction may over or under-run the allotted time, but the scope for variability is likely to be more restricted than, for example, a staff-patient interaction on a ward. Furthermore, given the availability of an appointment schedule, consultation times may be monitored relatively easily, especially if they involve a single member of staff at a known location. Indeed, depending on the purpose for costing, measurement of some interactions may be not worth the research effort.

Other, potentially important aspects of care for specific patients occur when the patient is not actually present. Examples of these 'indirect' care activities include writing reports and letters of referral, phone calls, and arranging transport or post-discharge support for the patient. Without the presence of the patient, it is difficult for an observer to link such care activities to specific patients. Moreover, many indirect care tasks may not be readily distinguishable from other general administrative activities that are not for any specific patient, but necessary for the service to function.

A gap in the health economics literature is advice on data collection methods for resource use and costs, and in particular, there is little practical guidance on how to cost staff inputs that are shared (Adam et al 2003b). There are various possible techniques to measure staff time including observation methods of ‘time and motion' and ‘work sampling', self-recording, retrospective self-reporting and ‘expert opinion'. A key aim of the thesis was to evaluate the pros and cons of these methods to measure staff inputs, particularly at the patient-level, and to review relevant sources to determine how staff inputs have been quantified for economic purposes.

Aims and objectives of thesis

The previous sections have outlined the background to the thesis and introduced key concepts. Based on review of literature and other sources, methods and issues were selected for testing in the empirical work. Haemodialysis was selected for the case study as it provided an opportunity to address the questions that arose from the RSU study. Moreover, haemodialysis is an expensive treatment for which the NHS is under continuing pressure to increase provision.

The aim of the thesis is to examine methods to quantify and cost staff inputs for specific patients for economic purposes. The empirical work uses chronic haemodialysis as a case study.

The objectives of the thesis are:

To give an overview of costing and review costing approaches for economic purposes.

To evaluate methods to measure staff time per patient.

To assess the impact of patient heterogeneity on nursing costs for chronic haemodialysis.

Research questions

What can be learned about the feasibility and data quality of barcode scanning to self-record nursing time per patient?

Is it feasible to use work sampling to measure nursing time per patient?

In chronic haemodialysis, is the nursing time per patient statistically or economically different between:

i) patients who are eligible and ineligible for RSU care, and

ii) patients of different dependency?

Outline of methods used

Health economics recent literature

Outline of thesis

Chapter 2 provides the background to costing, introducing key costing concepts and terms, and gives an overview of important considerations for costing. The chapter reviews literature on costing guidance and research in general and specifically in relation to both the top-down and bottom-up costing approaches and the classification of patient heterogeneity.

Chapter 3 concerns the measurement of staff time. The chapter examines what can be learned from background and specific literature about the best method to measure staff time per patient. The chapter discusses the importance of possible categories of staff time since the thesis focuses on patient's care time not how nurses use their time in general or for specific tasks. It outlines methods available to measure staff time and compares their advantages and disadvantages. Lastly, through targeted literature reviews, the chapter discusses key findings from the literature about application of the measurement techniques in economic evaluations and health care more generally.

Chapter 4 explains the background to why haemodialysis offered a useful case study for the empirical research. It outlines the clinical and epidemiological aspects of renal failure, available treatment options and service provision, and discusses policy initiatives to address the challenges faced by the NHS to meet increasing demand for haemodialysis. Finally, the chapter reviews literature on economic evaluations in relation to lessons that can be learned about costing haemodialysis, where heterogeneity in resource use appears to be important.

Chapter 5 draws together findings from the previous chapters and provides the rationale for the methods selected for the empirical work. Subsequently, it presents the aims and research questions addressed by the empirical work and remainder of the thesis.

Chapter 6 describes the methods used in empirical research. The overall aim was to measure and evaluate both the data collection methods and outcome (nursing time per patient) in a single main renal unit for patients who were either eligible or ineligible for RSU care. Data collection included:

Measurement of nursing time using barcode scanning and work sampling.

Descriptive information on data collection feasibility issues including a nurse acceptability questionnaire and research costs.

Classification of patient heterogeneity by eligibility for RSU care and other tools.

The chapter describes the rationale for and objectives of the initial pilot phase (the SUHT and Totton studies) that aimed to ensure that it was feasible to collect relevant time and dependency data. The chapter also describes the setting for the main data collection (Portsmouth study) and data collection methods, tools, and statistical analyses used.

Chapters 7 and 8 are the results chapters. Chapter 7 presents the descriptive results and information from the pilot studies (SUHT and Totton). Chapter 8 gives an overview of the data collection, presents the results from the main data collection (Portsmouth study), and concludes with a summary of the empirical results.

Finally, Chapter 9 discusses the results and reviews the themes addressed by the thesis before presenting the overall conclusions.

We provide a professional essay writing service that thousands of our customers use as an effective way of improving their grades, improving their research and saving them lots of time.