< Previous | Contents | Next >

EFFECTIVE AND COST-EFFECTIVE MEASURES TO REDUCE ALCOHOL MISUSE IN SCOTLAND: A LITERATURE REVIEW

CHAPTER THREE REVIEW OF COST-EFFECTIVENESS STUDIES: METHODS

Introduction

3.1 The cost-effectiveness literature is smaller than the effectiveness literature and of variable quality. Several other reviews of alcohol services have been undertaken previously. Godfrey (1994) reviews the literature prior to 1994, dividing studies into cost-offset, cost-benefit and cost-effectiveness. More recently, French (2000) provides an updated review. French covers a wide range of interventions, although the categorisation of studies demonstrates the majority to be merely costing studies without reference to patient outcomes. In addition, French identified a number of studies relating to general substance abuse treatment, as opposed to specific alcohol treatment. A systematic review of studies published in the period 1990 — 2000 has been undertaken for this report. It is considered that any earlier studies that might be identified would be of limited relevance. The search strategy is outlined below.

Methods

Search Strategy

3.2 Systematic searches of electronic databases have been undertaken to identify any study that has undertaken some form of economic evaluation even if this is not complete. Studies where a passing reference to costs or cost-effectiveness is given without any supporting data have been excluded. The databases searched were Medline, Embase, DARE (B system), NHS EED (B system), HTA (B system), Ongoing Reviews d/b (T system), National Research Register, Cochrane Library, HEED, EconLit, Social Science Citation Index, Science Citation Index, Cinahl, British Nursing Index, PsychInfo, SIGLE, HMIC, AMED, PAIS, ASSIA Plus, EconBase (WWW), and HDA Evidencebase (WWW)

3.3 The search for cost-effectiveness literature for the economic analysis searched 22 electronic databases. The systematic search strategy is shown in Annex 1 and the databases are shown in table 3.1. A total of 2,303 references were found. The abstracts of the articles were checked by hand and studies with economic data were closely scrutinised. A Microsoft Access database was constructed to abstract data from studies including details of both costs and consequences of alcohol interventions. A key selection criterion was to limit the analysis to studies that related to specific alcohol treatments or ranges of treatments, since these studies would permit costs to be derived for the treatments in question.

Table 3.1 List of databases and date range searched

Quality Assessment

3.4 All the identified studies were subject to a rigorous individual critique to assess how much confidence could be placed on the specific results. This involved the use of the general checklist devised by Drummond et al (1997). The extent to which studies had encompassed all relevant costs and consequences was reviewed using the more specific item checklist outlined in chart 3.1.

Assessing Cost-effectiveness

3.5 Economic evaluation involves building upon effectiveness information to assess both the costs of delivering the different policies or interventions and also assessing a wide range of consequences. The steps in any economic evaluation involve identifying, measuring, valuing and then comparing the costs and consequences of two or more alternatives. There are a number of different variables involved in the final result and variations in results could be due to a number of factors. Also, local conditions can influence the value of costs and consequences, especially between countries. This problem can be overcome if data on units of resource use are reported. In this case, relevant local data on costs can be applied. Where available in the review, these units have been recorded in the database as well as the financial cost estimates.

3.6 The application of economic evaluation techniques also involves researchers making a number of assumptions and, generally, individual studies undertake a range of sensitivity analysis to test the robustness of their findings to changes in these assumptions. Synthesising evidence on cost-effectiveness is not as straightforward as for effectiveness reviews nor are there well-developed techniques. There do exist, however, a number of checklists to assess the quality of individual studies.

3.7 In practice there are very few good quality economic evaluations that have been undertaken. The largest literature is for face to face treatments. It may be expected that these studies would be of a similar quality to those found in other health care areas. Unfortunately, many have omitted major costs or consequences. The evidence that can be drawn from such studies is, therefore, of a very different quality from that which can be taken from a well-conducted systematic review. In general, the lessons drawn illustrate some of the issues that will impact on cost-effectiveness rather than lead to any ranking between interventions.

Costs and consequences relevant to assessing the cost-effectiveness of alcohol interventions.

3.8 Alcohol misuse impacts on the individual drinker, their families, their communities and the wider society. In economic terms, drinking is associated with a range of private and external (third party) costs. For this study we have assessed all (social) costs and consequences where possible. Also, different policies have different resource inputs and costs and a range of consequences. Fiscal and legislative approaches are imposed on drinkers, not voluntarily undertaken in the same way as those seeking help through treatment. Some policies impact (and impose costs) on all drinkers whether or not they are causing problems for themselves or others. A checklist of the cost and consequences was given in chart 3.1.

Chart 3.1 Checklist of potential costs and consequences of alcohol interventions

3.9 Not all interventions involve all of these costs and consequences. In particular, a media campaign or schools education campaign may involve a high level of direct resource inputs. However, any change in drinking patterns that resulted from any intervention would be undertaken voluntarily and therefore not involve any loss in consumption benefits. Tax policies are not resource costly and do not involve large implementation costs but may involve such lost consumption benefits. A change in legislation will involve some costs but is unlikely to be effective unless it is enforced and this can involve larger resource outlays.

A taxonomy of economic evaluations.

Cost-offset studies or cost analysis.

3.10 Many economic studies in the alcohol field have restricted their attention to the question of whether treatment costs are offset by savings in future alcohol related medical care and other external costs of alcohol. The implicit alternative in such studies is the counterfactual that the individuals concerned would not have received any specialist care. This is a very realistic alternative in many countries, for example, only 1 to 5 per cent of dependent drinkers receive any specialist help in the UK in any one year (Godfrey, 1997). These studies, however, set a higher hurdle for alcohol treatments compared to other health care interventions. Also the value to the individual of treatment is excluded. Holder (1998) concludes that these studies do indicate that alcohol treatment is likely to be cost beneficial, even with limited consequences included. However, the results are influenced by the size of savings in treatment costs and these will tend to be much higher in the US than in the UK, as both the volume of service use and the cost per item of service are higher.

Cost-effectiveness analysis.

3.11 Cost-effectiveness analysis refers strictly to evaluations where the main individual outcome is measured in a single, specific way, for example, reduction in alcohol consumption, abstinent days or reduction in a problem index. The approach is of less value when a range of outcomes has to be considered. Direct comparisons across different studies are only possible to the extent that the same outcome measures have been used. The term, cost-effectiveness analysis, is often used more generically in the literature to refer to economic evaluation.

Cost-utility analysis.

3.12 Cost-utility studies have been particularly favoured in health care evaluations (Drummond et al, 1997). These studies use specific health related quality of life measures, where values exist, of the benefits perceived from individuals moving from one health state to another. These data are combined with estimates of the time period for which the health benefits will last to give quality adjusted life years. These measures allow the comparison of life saving with life enhancing interventions. While a number of studies are currently underway using these measures alongside alcohol specific outcomes, their usefulness in evaluating alcohol interventions is unknown. Their benefit more generally is that the relative cost-effectiveness of interventions in very different areas can be compared using a measure with clear equity weights; one QALY has the same value for all people. This contrasts to the human capital measure of valuing life, often used in social cost studies, where higher earners have more value than those with low earning potential.

Cost-benefit analysis.

3.13 Cost-benefit analysis refers to evaluations where all the effects are measured in money terms. This allows direct comparison with the costs of each intervention and the net worth of each of the alternatives can be estimated. Such studies are rarely carried out in full, and the methods for determining monetary values for health effects are still subject to some controversy.

Interpreting the available literature on cost-effectiveness.

3.14 The limited available literature has been reviewed and the results are reported within each chapter. In addition, a commentary on the relevance of the available results to the Scottish context is provided. The second stage of the economic analysis has been to use the results of the effectiveness reviews to attempt some outline modelling for brief interventions and relapse prevention, using Scottish costing on the resources needed for the intervention and the potential consequences. These results are reported in chapters 8 and 10.

< Previous | Contents | Next >