EURONHEED Methodological Guide V3 - unirioja.es · 2 ACKNOWLEDGEMENTS This is the third draft of the European Network of Health Economic Evaluation Databases (EURONHEED) methodological

EURONHEED Coordination centre : http://www.euronheed.org CES – 7, rue de Citeaux – 75012 Paris – France E-mail : [email protected] Tel : +33-(0)1 43 45 75 65 – Fax : +33-(0)1 43 45 75 67

DELIVERABLE 6:DELIVERABLE 6:DELIVERABLE 6:DELIVERABLE 6:

METHODOLOGICAL GUIDEMETHODOLOGICAL GUIDEMETHODOLOGICAL GUIDEMETHODOLOGICAL GUIDE

EURONHEEDEURONHEEDEURONHEEDEURONHEED QLQLQLQLG7G7G7G7----CTCTCTCT----2002200220022002----02379023790237902379

OctoberOctoberOctoberOctober 2004 2004 2004 2004

Third versionThird versionThird versionThird version

Project Financed byProject Financed byProject Financed byProject Financed by the European Commissionthe European Commissionthe European Commissionthe European Commission

Project inProject inProject inProject in collaboration with :collaboration with :collaboration with :collaboration with :

2

ACKNOWLEDGEMENTSACKNOWLEDGEMENTSACKNOWLEDGEMENTSACKNOWLEDGEMENTS

This is the third draft of the European Network of Health Economic Evaluation Databases (EURONHEED) methodological guide. It is based mainly on CRD Report 6 (2nd Edition) (NHS Centre for Reviews and Dissemination, 2001), which was edited by Boyka Stoykova and John Nixon, and the CODECS Methodological Guide (Collège des Economistes de la Santé, 2000), compiled by Gérard de Pouvourville, Emile Lévy, Francis Fagnani, Robert Launois, Catherine Buron, Cécile Charasse, Stéphanie Boulenger, Emmanuelle Brun and Philippe Ulmann. Contributors to specific chapters of the original documents are retained, along with new contributors as appropriate. The harmonisation of these two guides has been undertaken for the EURONHEED project by Stéphanie Boulenger, Philippe Ulmann, Emile Lévy, John Nixon and Julie Glanville, Mike Drummond, Stephen Rice, Dawn Craig. It is anticipated that revised versions of this document will be issued over the life of the EURONHEED project as methodology and practice are refined. © 2004 Collège des Economistes de la Santé, Paris, and Centre for Reviews and Dissemination, University of York for EURONHEED.

3

Table of contentsTable of contentsTable of contentsTable of contents 1111---- Guidance for writing EURONHEED abstracts Guidance for writing EURONHEED abstracts Guidance for writing EURONHEED abstracts Guidance for writing EURONHEED abstracts

5

A) General Information and effectiveness evidence Boyka Stoykova, Elisabeth Fenwick, David Smith, Jos Kleijnen

1. Subject of Study 6 2. Key Elements of Study 6 3. Details about Clinical Evidence 8

B) Economic Analysis and EURONHEED Commentary John Nixon, Anne Mason, Helen Weatherly, Francis Pang, Sally Thompson, Michael Drummond, Gérard de Pouvourville, Emile Lévy, Franci Fagnani, Robert Launois, Catherine Burin, Cécile Charasse

4. Economic Analysis 12 5. Results 17 6. Critical Commentaries 19 7. Implications of the study 21 List of AppendicesList of AppendicesList of AppendicesList of Appendices

Appendix 1. Explanatory Notes 22 Appendix 2. The Commentary Step-by-step 34 Appendix 3. NHS EED and CODECS Example Abstracts 43 BibliographyBibliographyBibliographyBibliography

54

4

LIST OF ABBREVIATIONSLIST OF ABBREVIATIONSLIST OF ABBREVIATIONSLIST OF ABBREVIATIONS CBA Cost-benefit analysis CCA Cost-consequences analysis CEA Cost-effectiveness analysis CHE Centre for Health Economics CI Confidence interval CINAHL Cumulative Index of Nursing and Allied Health Literature CMA Cost-minimisation analysis CODECS COnnaissances et Décision en ÉConomie de la Santé CRD NHS Centre for Reviews and Dissemination CRM Cost, review and methodology studies CUA Cost-utility analysis DARE Database of Abstracts of Reviews of Effectiveness DH Department of Health EQ-5D Euroqol EURONHEED European Network of Health Economic Evaluation Databases HYE Healthy Years Equivalent ITT Intention to treat NHS National Health Service NHS EED National Health Service Economic Evaluation Database NICE National Institute for Clinical Excellence NNT Number needed to treat QAG Quality Assurance Group QALY Quality adjusted life year RCT Randomised controlled trial RD Risk difference RR Relative risk SD Standard deviation WTP Willingness to Pay

5

I. GUIDANCE FOR WRITING EURONHEED ABSTRACTS. I. GUIDANCE FOR WRITING EURONHEED ABSTRACTS. I. GUIDANCE FOR WRITING EURONHEED ABSTRACTS. I. GUIDANCE FOR WRITING EURONHEED ABSTRACTS. EURONHEED will make accessible information on published economic evaluations in a user-friendly format. The aim is to ensure that its contents are of maximum use to its potential audience of health care professionals, managers, policy makers, researchers and academics. In order to appreciate the potential value of economic evaluations, it is important to critically appraise each study and not just to accept the author’s conclusions. The critical analysis of each study is undertaken systematically, following the guidelines in this section. They are based on existing criteria for judging the quality of economic evaluations of health technologies. The aim of a EURONHEED abstract is to provide a structured summary of the study and to facilitate an understanding of the methods used as well as the assessment of its quality; it also allows comparison across studies and highlights any features of special interest. Commentary fields contain summaries of key features which affect the usefulness of the evaluation and give an independent critical view on the study’s conclusions and implications. The structure of the abstract allows separate consideration of the quality of the evidence of clinical effectiveness and the economic component. It also distinguishes between economic evaluations carried out alongside a single effectiveness study, those based on a review or synthesis of more than one study, and those in which the estimates of effectiveness are based on opinion. Where the estimates of effectiveness are based on a study published elsewhere, and insufficient details are provided in the report of the economic evaluation, the relevant paper can be retrieved and examined where necessary. The quality of the effectiveness study, review or model is assessed as well as the relevance of the costs to the treatments. The abstract structure has been developed to be very general and therefore usable with different studies.

How To Use The Guidance How To Use The Guidance How To Use The Guidance How To Use The Guidance

The guidance follows the structure of the abstract and is presented under the same headings and subheadings. A detailed description is given about what should be included in each field. Further guidance is available in the footnotes. All terms marked with an asterisk (*) are briefly explained in Appendix 1, where some further reading is suggested in the “useful references” list. Abstractors/users should distinguish between two types of information to be included in an abstract (marked with the corresponding bullet): � MandatoryMandatoryMandatoryMandatory - relevant to all studies and, therefore, comments should be made even if the information

is missing or not explicitly stated in the paper; � Additional informatioAdditional informatioAdditional informatioAdditional informationnnn – relevant to some studies only; comments should be made if missing but

considered relevant.

6

1.1.1.1. SUBJECT OF STUDYSUBJECT OF STUDYSUBJECT OF STUDYSUBJECT OF STUDY

1.1. Health technology*1.1. Health technology*1.1. Health technology*1.1. Health technology* � Describe the intervention(s) studied by providing relevant details. If no health technology is explicitly stated as the comparator, provide a description of all technologies. 1.2. Disease 1.2. Disease 1.2. Disease 1.2. Disease This is a broad disease category derived from National Library of Medicine Subject Headings and is assigned by information staff. For more details see Appendix 1. 1.3. Type 1.3. Type 1.3. Type 1.3. Type of intervention. of intervention. of intervention. of intervention. � Report the intervention type in broad terms, using one or more of the following phrases:

- Primary prevention*; - Secondary prevention*; - Screening*; - Diagnosis*; - Treatment*; - Rehabilitation*; - Palliative care*;

- Health professional training; - Other (Other public health policies, Information and education in health*, Integrated care*, techniques and equipments, specify).

1.4. Hypothesis/study question1.4. Hypothesis/study question1.4. Hypothesis/study question1.4. Hypothesis/study question � Summarise the general objective of the study and specify the hypothesis/question(s) posed. � State whether any alternative health technology was explicitly stated as a comparator. � Report relevant details about the comparator(s). � Report the justification given for the choice of the alternative technology. � Report the perspective adopted for the economic analysis as stated by the authors.1

2. KEY ELEMENTS OF STUDY2. KEY ELEMENTS OF STUDY2. KEY ELEMENTS OF STUDY2. KEY ELEMENTS OF STUDY

2.1. Economic study type2.1. Economic study type2.1. Economic study type2.1. Economic study type � To define the study type, choose one or more of the following2:

1�� !��"��2�#��$��

7

Cost-effectiveness analysis3*; Cost-utility analysis*; Cost-benefit analysis*.

2.2. Study population2.2. Study population2.2. Study population2.2. Study population � Outline the main characteristics of the patient population from which the study sample was

drawn; e.g. age, sex, health status (e.g. duration and severity of disease or co-morbidity), socio-economic status, etc. Briefly summarise any inclusion/exclusion criteria reported by the authors.

� State the target population of the health technology if provided in the paper. This may differ

from the actual study population and therefore affect the generalisability of the results for the intervention technology4.

2.3. Setting2.3. Setting2.3. Setting2.3. Setting � Specify the setting of the study using one of the following terms: Outpatient care*, Inpatient

care*, Home care*, Community care*, Institutional care. � If you find it is relevant and appropriate, specify the level of health care using one of the

following terms: Primary*, Secondary* and Tertiary care*. � Quote the country or specific place in which the economic study was carried out (e.g.

Sheffield, UK). 2.4. Dates to which data relate. 2.4. Dates to which data relate. 2.4. Dates to which data relate. 2.4. Dates to which data relate. Report the years during which the data were collected, for: � The effectiveness analysis5. � The resources used, such as equipment, manpower, medicines, etc. (if different for different

components or interventions, report separately). � The prices used. 2.5. Source 2.5. Source 2.5. Source 2.5. Source of effectiveness dataof effectiveness dataof effectiveness dataof effectiveness data � State whether the evidence/estimate for final outcomes was one or more6

of the following: (a) Derived from a single study – complete fields 3A (b) Based on a review/synthesis or ad hoc use of previously published studies - complete

fields 3B (c) Estimates of effectiveness based on experts’ opinion and/or authors’ assumptions -

complete fields 3C 2.6 Modelling2.6 Modelling2.6 Modelling2.6 Modelling* 3� #��%��!��!��&��

4�'��(��)*��+*��%��!��5�#�� 6�#��

8

� State whether a model was used to estimate benefits and/or costs; � State the purpose of the model; � State the type of model used or enter “type not specified”. 2.7 Link between effectiveness and cost data 2.7 Link between effectiveness and cost data 2.7 Link between effectiveness and cost data 2.7 Link between effectiveness and cost data (To be completed only if a single study has been used to derive effectiveness). � State whether the costing was undertaken on the same patient sample as that used in the

effectiveness study. � State whether the costing (i.e. the collection of resource data) was undertaken prospectively

(alongside the effectiveness study) or retrospectively (after the effectiveness results were known).

3. DETAILS ABOUT CLINICAL EVIDENCE3. DETAILS ABOUT CLINICAL EVIDENCE3. DETAILS ABOUT CLINICAL EVIDENCE3. DETAILS ABOUT CLINICAL EVIDENCE7777

3A.3A.3A.3A. Single study Single study Single study Single study

3A.1. Study sample3A.1. Study sample3A.1. Study sample3A.1. Study sample � State whether the sample size8 was determined in the planning phase of the study to assure a

certain power*, or if power calculations* were retrospectively performed based on the existing sample size (provide the relevant details).

� Summarise the method of sample selection. � State whether there is evidence that the initial study sample is appropriate for the clinical

study question.9 � Report the trial/study size. Distinguish between the number of subjects (e.g. patients, doctors,

health care provider units) overall, in the intervention group(s), and in the control group(s). � Report the percentage of subjects who refused to participate. � Report the percentage of subjects excluded and the reasons for the exclusions from the initial

sample. 3A.2. Study design* 3A.2. Study design* 3A.2. Study design* 3A.2. Study design* � Define the type of study. For example:

7� �#�� '�� ,-�.�/��0��,-�1�#��0��,2�.�%��0��,2�1�3��0��8� �� 4� �� #��9�5��$��6��

9

Randomised controlled trial; Nonrandomised trial with concurrent controls; Cohort study; Case-control study;

Study with historical controls; Before-and-after study in one group of patients; Case series; Other (specify).

� State whether the study was single or multi centred. Give details (e.g. number of centres or

number of sites). � If an experimental study, report the subject allocation method, choosing for example:

Unit of randomisation*; Method of random allocation*, including concealment of allocation; Block randomisation*; Stratified allocation*; Haphazard or other non-random allocation; Other (specify).

� Report the duration of follow-up of the groups. � Quote the loss to follow-up* (report the reasons for withdrawals if provided).

For example in an experimental study report: % overall; % in intervention group; % in control group.

� Specify the blinding method* for assessment of outcomes, if any. For example: Patient; Clinician; Assessor of patients’ outcome.

3A.3. Analysis of effectiveness 3A.3. Analysis of effectiveness 3A.3. Analysis of effectiveness 3A.3. Analysis of effectiveness � For experimental studies state whether the analysis of the clinical study was based on: 1)

intention to treat*; or 2) treatment completers only*; or 3) not stated. If the study is the report of an observational investigation report whether all the patients included in the study were accounted for in the analysis.

� Report the primary health outcomes used in the analysis (e.g. CHD deaths, quality of life,

side-effects)10, and specify any particular instruments used to evaluate these data (e.g. quality of life questionnaire).

� State whether, at analysis, groups were shown to be comparable in terms of age, sex and

prognostic features or if there was adjustment for confounding variables. List, if any, the confounding variables*.

3A.4. Effectiveness results 3A.4. Effectiveness results 3A.4. Effectiveness results 3A.4. Effectiveness results � Summarise the results of the trial or study, including the side effects or adverse effects (give

quantitative results when reported). � Quote the 95% confidence interval* and p-values* of primary outcomes, using the following

format:

10�#�� !��3��!��!��,-�.�/��0��,-�1�#��0��,2�.�%��0��,2�1�3��0��

10

(95% CI: 1.23 – 4.56; p=0.001), or (p<0.001), or (range: 12 – 34) where appropriate. 3A.5. Clinical conclusions 3A.5. Clinical conclusions 3A.5. Clinical conclusions 3A.5. Clinical conclusions � Summarise the clinical conclusions derived from the evaluation of the intervention.

3B. Review/synthesis or ad hoc use of previously published studies. 3B. Review/synthesis or ad hoc use of previously published studies. 3B. Review/synthesis or ad hoc use of previously published studies. 3B. Review/synthesis or ad hoc use of previously published studies.

3B.1. Outcomes assessed in the review/Parameters used in the model3B.1. Outcomes assessed in the review/Parameters used in the model3B.1. Outcomes assessed in the review/Parameters used in the model3B.1. Outcomes assessed in the review/Parameters used in the model � Report the outcomes (parameters) chosen by the author and determined by the assessments

made in the primary studies. Outcomes may include mortality, morbidity, quality of life, adverse events, sensitivity, and specificity of tests, etc.

� For model studies report all effectiveness, epidemiological and demographic variables used to populate the model.

3B.2. Study3B.2. Study3B.2. Study3B.2. Study designs and other criteria for inclusion in the review designs and other criteria for inclusion in the review designs and other criteria for inclusion in the review designs and other criteria for inclusion in the review � List the particular study designs included in the review as described by the author (e.g.

RCTs). � State any other inclusion or exclusion criteria, such as age, disease state, outcome

measurement, duration of follow up, measure of quality of studies.11 3B.3. Sources searched to identify primary studies 3B.3. Sources searched to identify primary studies 3B.3. Sources searched to identify primary studies 3B.3. Sources searched to identify primary studies � List the sources searched by the author(s)12. Example: MEDLINE, other databases,

unpublished data, journal references, personal communication. 3B.4. C3B.4. C3B.4. C3B.4. Criteria used to ensure the validity of primary studies riteria used to ensure the validity of primary studies riteria used to ensure the validity of primary studies riteria used to ensure the validity of primary studies � Specify the criteria used to assess the validity of the primary studies. For example:

concealment of treatment allocation; blind assessment; low drop out rates, etc. 3B.5. Methods used to judge rele3B.5. Methods used to judge rele3B.5. Methods used to judge rele3B.5. Methods used to judge relevance and validity, and for extracting datavance and validity, and for extracting datavance and validity, and for extracting datavance and validity, and for extracting data13131313 � State how relevance was judged, taking into consideration the process of selecting the papers

for inclusion, how many authors were involved and how they worked on the selection? (E.g. blinded, independent, how disagreements were resolved14?).

� Specify the judgement criteria applied by the author(s) for assessing the validity of the

primary studies. How many reviewers were involved and how did they work on the assessment?

� Explain the process of extracting the data from the included studies. For example: Single reviewer; two independent reviewers; whether blind to source, author, journal etc., and how discrepancies were resolved 11��&��$��(��7�8��8��,(��0��#��(��,9(��8��6��12�3��13�/��(��$��14��'��

11

3B.6. Number of primary studies included3B.6. Number of primary studies included3B.6. Number of primary studies included3B.6. Number of primary studies included � Report the number of primary studies included in the review, and the types of studies (if

more than one type). 3B.7. Method of combination of primary studies3B.7. Method of combination of primary studies3B.7. Method of combination of primary studies3B.7. Method of combination of primary studies � If studies are combined, specify the methods used to combine the results of the individual

primary studies. For example: Meta-analysis*; Narrative method15; 3B.8. Investigation of differences between primary studies 3B.8. Investigation of differences between primary studies 3B.8. Investigation of differences between primary studies 3B.8. Investigation of differences between primary studies � State whether the author investigated the differences between the primary studies, focusing

attention for example on the differences between participants, exposures or interventions, outcome measures, study designs, etc.16

� State whether the author provided an explanation of any differences between individual

studies, including the results of statistical tests of homogeneity17 (e.g. significant, not

significant, not done, not applicable). � State whether the author investigated how these differences affect the estimate of the

effectiveness of the technology. 3B.9 Results of the review/Values used for model parameters3B.9 Results of the review/Values used for model parameters3B.9 Results of the review/Values used for model parameters3B.9 Results of the review/Values used for model parameters � Summarise the authors’ findings from combining primary studies, including actual data,

summary estimates, such as odds ratios and relative risks (both for discrete variables); or weighted mean differences or standardised mean differences (effect sizes) for continuous outcomes18; quote the 95% confidence interval* and p-values* of primary outcomes, using the following format where appropriate: (95% CI: 1.23 – 4.56; p=0.001), or (p<0.001), or (range: 12 – 34) • For modelling, where authors have extensively reported model parameter values abstractors

should use their judgement to limit reporting by selecting the most informative data (for example, reporting data for a given patient group, reporting minimum and maximum values, etc.).

3C. Estimates of effectiveness based on opinion. 3C. Estimates of effectiveness based on opinion. 3C. Estimates of effectiveness based on opinion. 3C. Estimates of effectiveness based on opinion.

3C.1. Methods used to derive estimates 3C.1. Methods used to derive estimates 3C.1. Methods used to derive estimates 3C.1. Methods used to derive estimates of effectiveness/model parametersof effectiveness/model parametersof effectiveness/model parametersof effectiveness/model parameters � Report the methods used to derive estimates of effectiveness (model parameters), e.g.

consensus, experts’ opinion, authors’ assumptions. 3C.2. Estimates of effectiveness and key assumptions3C.2. Estimates of effectiveness and key assumptions3C.2. Estimates of effectiveness and key assumptions3C.2. Estimates of effectiveness and key assumptions � Report the estimates of effectiveness (model parameters) of health technology and key

assumptions used in producing them.

15�:��

16�3��(��(��&��17��'��!��;��6��<��18��#��&�� /��(�� (��;��<'3��

12

4. ECONOMIC ANALYSIS4. ECONOMIC ANALYSIS4. ECONOMIC ANALYSIS4. ECONOMIC ANALYSIS

4.1. Measure of health benefits used in the economic analysis4.1. Measure of health benefits used in the economic analysis4.1. Measure of health benefits used in the economic analysis4.1. Measure of health benefits used in the economic analysis � Specify which health benefit measure was used in the economic analysis, using for example: Heart attacks avoided;

Lives Saved; Life Years gained; Quality Adjusted Life Years (QALYs)*; Healthy-Years Equivalent (HYEs)*; Monetary benefits (in CBA only)19.

If no summary health benefit is used in the economic analysis and clinical outcomes are left disaggregated, report that a cost-consequences analysis (CCA) was performed20.

If the authors have demonstrated that the clinical effectiveness results are equal the researchers may have actually conducted a cost-minimisation analysis. In both cases comment appropriately21.

� Report the basic method of valuation of health states or interventions22. For example it can be

one or more of:

Generic valuation matrix, (e.g. EQ-5D, Rosser, HUI etc.); Direct measurement; Author's assumption; Willingness to pay* (in CBA);

Human capital approach* (in CBA)

� In the case of direct measurement report whose values, and how many (i.e. number of people used to elicit values) values were used to assess the health states. For example, use one or more of the following:

Author; Clinician; Patient; Relative;

Society; Other health worker; Literature.

� Specify, if stated, when they were valued. For example: at baseline, at 1 month, at three

months from the intervention.

19�3��=��%='��20�#��!��&�� >�� 9��2�?�� :�� 3��2�1��21�9&��7��#��!��&��!��&��#�� 6�� !��#��2�?�� 2�1��22��@��(��

13

� Specify how they were valued. For example: postal or telephone survey, interview, etc. � Report the valuation tool used. For example:

Standard gamble*; Time trade-off*; Conjoint analysis*; Other (specify).

4.2. Direct costs*4.2. Direct costs*4.2. Direct costs*4.2. Direct costs* Provide details of methods used, distinguishing, where possible, between the prices/unit costs and quantities of the resources*: � State whether resource quantities and costs were reported separately23. � Report whose direct costs were included in the analysis24. For example:

Hospital Hospital Hospital Hospital This includes a hospital or other institutions providing outpatient and in-patient services. It might be private or public. Costs may relate to surgical and diagnostic procedures, nursing care, acquisition and administration of drugs, hotel, etc.

PatientPatientPatientPatient aaaand/ornd/ornd/ornd/or

RelativesRelativesRelativesRelatives

These are non-health care sector costs borne by the patients and/or their families for visitation, travel, or goods and services purchased in the market by the consumers, e.g. therapeutic appliances, optical equipment, hearing devices. The lost income from work is not included in this category since this is considered part of indirect costs.

Health Health Health Health serviceserviceserviceservice

In this case a broad range of resources for providing health care services has been considered in the study. These may relate, for example, to more than one of the following: provider institution for outpatient and in-patient services (e.g. hospital, tertiary care centre), general practice, non-hospital laboratories costs, dentist services and any other ambulatory care, nursing home costs, acquisition and administration of drugs outside the hospital, professional health services provided by a single physician (e.g. specialists, consultants) outside the hospital; any other medical costs, (such as those related to the implementation of vaccination programmes).

Other Other Other Other agencies agencies agencies agencies

This option refers to the case where the study analysed the costs of agencies providing social care (non-medical direct costs), such as residential care, voluntary services, and social services.

Others Others Others Others (define) (define) (define) (define)

For example, other societal costs different from the patients/relative costs previously described may have been included in the economic evaluation.

� List all direct costs included in the cost analysis. � State the source of the direct cost data and whether a model was employed to extrapolate to a

longer time-frame or another setting25. � State whether the estimation of the prices or unit costs/charges26 was one or more of the

following: i) a guess; ii) based on actual data (for example on one unit of analysis or a

23�#��%��3��+��24�;�� $�� ,9��0��25��(��!��!��,A��0��

14

sampling technique or on published information from other studies or standard prices or tariffs such as Medicare, social security reimbursements); iii) derived using modelling techniques.

� State, whether discounting* was relevant (when costs were incurred over a 2 year period or

longer), whether costs were discounted and if justification for the chosen rate was given. � State whether the study reported marginal*, incremental* or average* costs. � Specify the date(s) to which the price data refer. � State whether costs were adjusted for inflation and report the method used. � State whether the author established that some cost components were excluded because they

were common to both alternatives (differential costing) or were too small to change the overall result of the study.

� State whether any adjustments were made to the observed costs to correct, for example, for

protocol-driven costs or learning effects for new technologies. France France France France Origin of the data: 1- Public Hospital: For public hospital costs it is recommended to use, if possible, analytical

accountability data presented per DRG (PMSI database) by using the corresponding value in ISA points or the complete average costs per DRG, available in the « Etude Nationale de Coût » (they are divided into 14 components which can be grouped in 4 headings: medical cost/medico-technical cost/ lodging, management and logistic costs/capacity costs). Until this process becomes widely applied to all domains in the public and private sector, billing information and/or conventional tariff data may continue to be used.

2- 2- Primary Care: From the point of view of the National Sickness Fund (Assurance Maladie)point of view of the National Sickness Fund (Assurance Maladie)point of view of the National Sickness Fund (Assurance Maladie)point of view of the National Sickness Fund (Assurance Maladie), direct costs are, by convention, allocated on the basis of money reimbursed. Allocation of values to procedures is normally obtained by applying the coefficients of the Nomenclature Générale des Actes Professionnels (Relative Value Scale) and tariffs for each key letter, which are published by public bodies, in addition to other tariffs and prices used (drugs, prostheses etc), which are adjusted to take into account the specific reimbursement figure for the procedure (this practice is valid for direct outpatient costs: laboratory tests and diagnostic investigations, medical consulting visits and outpatient appointments, general practitioners and specialists, nursing and physiotherapy procedures etc. It is conventional to apply the factors used by the Régime Général (Sickness Fund for person on salary) in calculating reimbursement rates. From the point of view From the point of view From the point of view From the point of view of the patientof the patientof the patientof the patient, other adjustments are usually made based on the frequency and size of fee over-expenditure (practitioners with the right to exceed budgetary expenditure or who are practising in sector II, in receipt of private fees) and additional insurance coverage. In most cases, information about additional insurance coverage is not available in sufficient detail and the payment attributed to individuals generally includes all that part which is not reimbursed by the compulsory National Health Insurance. From From From From a societal perspective, a societal perspective, a societal perspective, a societal perspective, the value applied should ideally be expressed in terms of opportunity costs. This value is often, however, obtained by adding the reimbursable costs to the costs borne by the patient, including additional coverage.

26�@��%��3��+��3��!�!��

15

4.3. In4.3. In4.3. In4.3. Indirect costs* (Productivity losses and intangibles). direct costs* (Productivity losses and intangibles). direct costs* (Productivity losses and intangibles). direct costs* (Productivity losses and intangibles). Provide details of the methods used (include discounting) distinguishing, where possible, between costs and quantities. � State whether a rationale for the inclusion or exclusion of productivity costs was given. � Specify which quantities/costs were measured. For example, productivity gains/losses. � Report the source of quantity/cost data. For example: national income statistics (for average

incomes), employer’s records, social security tariffs, patients’ records etc. � State whether quantities and costs were analysed separately. � Report the quantity/cost boundary. For example: Patient; relative; Others (define). � State whether the estimation of the quantities was one or more of the following: i) a guess; ii)

based on actual data (for example based on one unit of analysis or a sampling technique or on published information from other studies; iii) derived using modelling studies, based on the friction cost method*.

� State whether the estimation of the costs was one or more of the following: i) a guess; ii)

based on actual data (for example on one unit of analysis or a sampling technique or on published information from other studies or standard prices; iii) derived using modelling studies, based on the friction cost method. . . .

� Specify when the quantity of resources was measured (a range of years should be quoted if

necessary). For example: trial date. � State, whether discounting* was relevant (when costs were incurred over a 2 year period or

longer), whether costs were discounted and if a justification for the chosen rate was given. � Specify the date(s) of the price data. � State whether costs were adjusted for inflation and report the method used. 4.4. Currency*4.4. Currency*4.4. Currency*4.4. Currency* � Quote the currency as reported in original study. For example: US $, Yen. (A list of preferred currency abbreviations is available in Appendix 1) � State whether any conversions were undertaken with dates if given. 4.5. Statistical analysis of quantities/costs. 4.5. Statistical analysis of quantities/costs. 4.5. Statistical analysis of quantities/costs. 4.5. Statistical analysis of quantities/costs. The authors may have undertaken statistical analysis on resource use or total cost data to test for statistical significance of the results (i.e. whether they have treated the data stochastically). Additionally descriptive statistics* may have been provided, or alternatively the cost data may have been treated only in a deterministic* manner.

16

The abstractor should determine how the resource use and cost data were handled and make appropriate comments according to the following: � If the resource use and/or costs were treated as point estimates (i.e. the data were

deterministic*), the abstractor should comment appropriately. � If descriptive statistics* were given, provide methodology details (e.g. the authors provided

mean/median values and standard deviations). � If resource use and/or costs were treated in a stochastic* manner, report the statistical test

used and levels of significance. � Report the parameters tested and whether the tests* used were justified by the authors given

the distribution of the data. � Note whether the study was powered to detect this difference27. . . . 4.6. Sensitivity analysis*4.6. Sensitivity analysis*4.6. Sensitivity analysis*4.6. Sensitivity analysis* When sampled data are not available it is not possible to deal with uncertainty using classical statistical techniques, and other methods, such as sensitivity analysis, are required. � State whether a sensitivity analysis of benefits/costs was carried out and specify the parameters

(e.g. discount rate, estimates of effectiveness or cost data, etc.). � Specify which areas of uncertainty* were investigated. Choose one or more of the following:

variability in data; generalisability of results*; extrapolation from primary data source to make the results more comprehensive*; analytical methods*; other (specify).

� Specify the method or rationale used to determine the ranges over which variables were

tested28.... � Specify the method used. Choose one or more of the following:

One-way simple sensitivity analysis*; Multi-way simple sensitivity analysis*; Threshold analysis*; Analysis of extremes*; Probabilistic sensitivity analysis*; Other (specify).

27�#��6�� %�� B�� 4��28� �� (�� '��(��5��

17

5. RESULTS 5. RESULTS 5. RESULTS 5. RESULTS

In the following subfields quote the results, if any, including figures for quantities/costs, discounted and not discounted29. 5.1. Estimated benefits used in the economic analysis. 5.1. Estimated benefits used in the economic analysis. 5.1. Estimated benefits used in the economic analysis. 5.1. Estimated benefits used in the economic analysis. Where no summary benefit measure was determined by the authors (in the case of cost-consequences and possibly cost-minimisation analyses) the abstractor need only state ‘the reader is referred to the effectiveness results reported previously’. � Report (incremental) benefits using for example:

(incremental) lives saved; (incremental) life-years gained and discount rate; (incremental) quality adjusted life years gained and discount rate. Include p values and/or confidence intervals, if any.

� Report the duration of benefits from the intervention to the patient used in the economic

analysis (e.g. length of follow-up as in clinical study, ten years, life time). State whether extrapolated beyond the trial/observation time.

� Report the duration of comparator benefits to the patient used in the economic analysis (e.g.

length of follow-up as in clinical study, ten years, life time). State whether extrapolated beyond the trial/observation time.

� State whether the side-effects of treatment were considered in the economic analysis, or if

they were not relevant. 5.2. Cost results 5.2. Cost results 5.2. Cost results 5.2. Cost results � Report the total intervention cost and discount rate, in the original currency. � Report the total comparator cost and discount rate, in the original currency. � Report the results of any statistical analysis of costs or confidence intervals. � State whether the costs of adverse effects or knock on costs were dealt with in the costing, or

were not relevant. � Report, the result of any currency conversion. � Report any incremental quantities/costs (discounted and not discounted figures). � Report the duration of intervention quantities/costs used in the economic analysis. � Report the duration of comparator quantities/costs used in the economic analysis.

29�#��

18

5.3. Synthesis of costs and benefits. 5.3. Synthesis of costs and benefits. 5.3. Synthesis of costs and benefits. 5.3. Synthesis of costs and benefits. � Specify how the estimated benefits and costs were combined30.

For example:

Cost/Life Saved and discount rate; Cost/Life-Years gained and discount rates of cost and benefits; Cost/QALY and discount rates of costs and benefits; Net benefit in cost-benefit analysis; Cost-effectiveness acceptability curve*; Incremental net benefit*; Others (define); or Not combined31.

� State whether an incremental analysis was performed32. � Report the summary findings, quoting currency33 and year. For example: (incremental)

Cost/Life Saved and discount rates; (incremental) Cost/Life-Year gained and cost/benefit discount rates; (incremental) Cost/QALY gained and cost/benefit discount rates.

� Report the range of the incremental costs/outcomes ratio and the discount rates. For example:

incremental Cost/Life Saved and discount rate referring to: baseline and discount rate of costs; lowest value; highest value.

� Report any important differences in the cost-effectiveness of sub-populations. � List the sensitive parameters, quoting any influential parameters discovered and figures. � Summarise the author's comments about the sensitive parameters and their range of variation. � Report the statistical testing results for the sampled data. 5.4. Author's conclusions 5.4. Author's conclusions 5.4. Author's conclusions 5.4. Author's conclusions � Summarise the author's conclusions.

30�#��

31�� (�� !��&�� #��7� '��A��(��&��!��32��!��(��'��(��!��6��5�� !��33�5��

19

6. CRITICAL6. CRITICAL6. CRITICAL6. CRITICAL COMMENTA COMMENTA COMMENTA COMMENTARIES: Methodological DiscussionRIES: Methodological DiscussionRIES: Methodological DiscussionRIES: Methodological Discussion

The commentary should reflect a good understanding of the paper and its implications. It should present a general overview of the study’s strengths and weaknesses, highlighting any relevant concerns about the reliability of the findings. The authors of the paper may offer their own critique in the discussion section of their paper, which should be acknowledged in the relevant sub-fields of the Critical commentaries34.... Each sub-field of the commentary should contain suitable comments, based on these instructions and the flow diagrams in Appendix 2. Repetition of details reported in the main abstract should be supported by suitable comments35. The diagrams and instructions are intended to identify key issues for comment: they are intended to be a guide for critical thought and comment. Comments are, therefore, unlikely to be simple yes or no answers, but should be based on an assessment of the paper and the abstractor’s view of the extent to which the paper meets EURONHEED quality criteria36.

6.1. Choice of comparator6.1. Choice of comparator6.1. Choice of comparator6.1. Choice of comparator The choice of comparator in an economic evaluation influences the interpretation of the findings. Generally, the comparator(s) should reflect standard practice (or placebo in clinical trials) in the authors’ setting, but this may affect the generalisability of the results to other settings37. The authors, however, may or may not provide an explicit justification38

for their choice of comparator(s).

� See Figure [1] of Appendix 2 and record suitable comments. 6.2. Validity of estimate of effectiveness.6.2. Validity of estimate of effectiveness.6.2. Validity of estimate of effectiveness.6.2. Validity of estimate of effectiveness. Comment according to the source of effectiveness data. In the case of a type A study, the main focus will be the internal validity of the study’s design39. In the case of a type B study, the main concern will be the quality of the review, the quality of retrieved studies, and the methods used to combine results from each source40. In the case of a type C study, comments should centre on the reliability of the estimates41. Where effectiveness data are derived from more than one source, complete all the relevant effectiveness fields. Figure [2] contains a loop to facilitate this process.

� See Figure [2] of Appendix 2 and select the relevant study type/s:

34�C��(��!�� 35�C��(�� 36�C��(��&��1�� D��#�� 37�C��(��6��

38�'��$��'��(��39�#��7��&��6��4��40�9��#��C��&��35��7��766��6��&��41��/��(��E��$��

20

� Type A Type A Type A Type A –––– see Figures [2] and [3] of Appendix 2 and comment on the internal validity of

the study and its impact on the effectiveness results. � Type B Type B Type B Type B –––– see Figures [2] and [4] of Appendix 2 and comment on the quality of the review

and process of determining estimates of effectiveness results and model parameters. � Type C Type C Type C Type C –––– see Figures [2] and [5] of Appendix 2 and make comments on the validity of

the estimates and assumptions used. Figure [5] contains a loop to facilitate this process. 6.3. Validity of estimate of health benefit. 6.3. Validity of estimate of health benefit. 6.3. Validity of estimate of health benefit. 6.3. Validity of estimate of health benefit. Many studies report both effectiveness and benefit measures, for example unit reduction in cholesterol (effectiveness) and QALYs (health benefit). In this case it is also necessary to report the strengths and weaknesses of the methods used to derive health benefits. Other studies, however, report only effectiveness or only benefits (or do not differentiate between them). The abstractor will need to determine the approach adopted and report accordingly.

� Comment on the methods used to derive that measure according to the bottom half of

Figure [2] of Appendix 2.

6.4. Validity of estimate of costs. 6.4. Validity of estimate of costs. 6.4. Validity of estimate of costs. 6.4. Validity of estimate of costs. Cost is the product of two elements: the quantity of resources consumed and the unit cost of the resources. Here the abstractor can make comments on one or more of the following: the chosen perspective, methodology issues, data collection issues, data analysis issues, and presentation issues.

� Refer to Figure [6] in Appendix 2 for an overview overview overview overview of the costing process and make

appropriate comments. The inclusion of relevant costs needs to be considered at two levels:

(1) The inclusion of certain categories of cost: these are determined by the boundary adopted in the economic analysis (e.g. direct healthcare costs; indirect costs etc.).

(2) The inclusion of specific costs within these categories: these will vary according to the particular study.

� Refer to Figure [7] in Appendix 2 for issues related to the quantity of resource usequantity of resource usequantity of resource usequantity of resource use.

Decide on the source(s) of the data on quantities and comment on whether sensitivity and/or statistical analyses were used appropriately, taking into account the ranges used. Figure [7] contains a loop to facilitate this process.

� Refer to Figure [8] in Appendix 2 for issues related to prices prices prices prices (unit costs). Figure [8]

contains a loop to allow for multiple price sources. If prices were not taken from the authors’ setting, comment on whether a sensitivity analysis of prices was conducted.

If prices were taken from the authors’ setting, comment on whether charges were used to proxy costs. Irrespective of the source(s) of prices, comment on whether the following were relevant, but omitted: discounting, price date and currency conversion.

6.5. Other Issues 6.5. Other Issues 6.5. Other Issues 6.5. Other Issues Issues to consider include: were the findings compared with the findings of other relevant studies; was the generalisability of the results to other settings (with focus on the NHS) assessed;

21

did the authors present results selectively; were the authors’ conclusions justified, given the limitations of the study (these are often addressed by the authors in the discussion). Report any further limitations of the study, as acknowledged by the authors, and not already covered in the commentary.

� Refer to Figure [9] in Appendix 2 and make appropriate comments. 6.6. Publications of related interest 6.6. Publications of related interest 6.6. Publications of related interest 6.6. Publications of related interest Publications presenting the clinical trial on which is based the economic evaluation and/or the preliminary results and/or the methodology used in the study

7. IMPLICATIONS OF THE STUDY7. IMPLICATIONS OF THE STUDY7. IMPLICATIONS OF THE STUDY7. IMPLICATIONS OF THE STUDY

7.1 According to the findings of the paper 7.1 According to the findings of the paper 7.1 According to the findings of the paper 7.1 According to the findings of the paper The implications of the study should be summarised. If the authors’ implications are summarised this should be made clear by a leading phrase such as “The authors suggest……”. Alternatively, abstractors with relevant experience can also add their own assessment of the implications of the study. If the implications claimed by the author do not seem warranted based on the study the abstractor should comment on this. � Report any recommendations made by the authors regarding changes in policy or practice. � Report any recommendations made by the authors regarding the need for further research42. � If the authors give no clear implications, the abstractor should only comment if the

implications can be clearly inferred. 7.2 According to the EURONHEED expert [Where experts are employed by centres] 7.2 According to the EURONHEED expert [Where experts are employed by centres] 7.2 According to the EURONHEED expert [Where experts are employed by centres] 7.2 According to the EURONHEED expert [Where experts are employed by centres] Identifies the main implications of the study in terms of clinical practice, health policy or research43. Specifies if the study has an impact on equity (distribution).

42�C��(��6��

43�@ ��(��#��#��(��&�� 6�� %�� $��

22

APPENDIX 1.APPENDIX 1.APPENDIX 1.APPENDIX 1.

EXPLANATORY NOTESEXPLANATORY NOTESEXPLANATORY NOTESEXPLANATORY NOTES

1.1.1.1. Subject Of StudySubject Of StudySubject Of StudySubject Of Study � Health TechnologyHealth TechnologyHealth TechnologyHealth Technology covers any method used by those working in health services to promote

health, prevent and treat disease and improve rehabilitation and long-term care. "Technologies" in this context are not confined to new drugs or pieces of sophisticated equipment.

� Medical subject headings:Medical subject headings:Medical subject headings:Medical subject headings:

�� ,�� $�� :�� B�� A�� A�� 3��$�� ;�� A3;�� #��(��

��A��3��$��;��,��0��(��A3;!��(��A9/B#:9�� F3'�:�� B�� A�� A3;�� A3;� �� $�� 9�� /��7�

�Bacterial infections and mycoses Mental disorders Behavior and behavior mechanisms Virus diseases Parasitic diseases Neoplasms Musculoskeletal diseases Digestive system diseases Stomatognathic diseases Respiratory tract diseases Otorhinolaryngologic diseases Nervous system diseases Eye diseases Urologic and male genital diseases Female genital diseases and pregnancy complications Cardiovascular diseases Hemic and lymphatic diseases Neonatal diseases and abnormalities

Skin and connective tissue diseases Nutritional and metabolic diseases Endocrine diseases Immunologic diseases Disorders of environmental origin Pathological conditions, signs and symptoms Therapeutics Investigative techniques Equipment and supplies Surgical procedures, operative Health care: Residence characteristics Health care: Socio-economic factors Health care: Health promotion Health care: Emergency medical services Health care: Health policy Health care: Patient care management Health care: Primary health care Other organisational issues of health care

To determine under which term(s) or in which Disease Field(s) falls a certain disease, you can go the MeSH website: http://www.nlm.nih.gov/mesh/MBrowser.html.

� Primary preventionPrimary preventionPrimary preventionPrimary prevention - prevention of disease or mental disorders in susceptible individuals or populations through promotion of health, including mental health, and specific protection, as in immunisation, as distinguished from the prevention of complications or after-effects of existing disease.

23

� Secondary preventionSecondary preventionSecondary preventionSecondary prevention involves identification and treatment of persons who have already developed a disease, in order to prevent worsening or repeat events (e.g. aspirin to prevent the recurrence of heart attacks).

� Palliative care Palliative care Palliative care Palliative care ---- treatment aimed at relieving symptoms and pain rather than affecting a cure. � Screening Screening Screening Screening –––– the examination of people with no symptoms, to detect unsuspected disease. � Diagnosis Diagnosis Diagnosis Diagnosis ---- t t t the determination of the nature of a case of disease.

� Rehabilitation Rehabilitation Rehabilitation Rehabilitation ---- t t t the return of function after illness or injury, often with the assistance of

specialised medical professionals. � Information and education in health Information and education in health Information and education in health Information and education in health ---- programmes aimed at improving the patient's

knowledge of his disease in order to prevent morbid events and delay mortality (e.g. education programs for diabetic patients).

� Integrated care Integrated care Integrated care Integrated care ---- an approach used to manage all aspects of health care including primary

care (see next page for definition), acute care and long-term care.

2. Key Elements o2. Key Elements o2. Key Elements o2. Key Elements of Studyf Studyf Studyf Study � CostCostCostCost----effectiveness analysiseffectiveness analysiseffectiveness analysiseffectiveness analysis is a type of economic evaluation where the health outcomes are

expressed in natural units (e.g. life-years gained, deaths avoided, heart attacks avoided, cases detected). Studies with multiple outcome measures (cost-consequences analyses) are also included in this category. In addition, cost-minimisation analyses, where the alternatives show identical effectiveness, are also considered as a subtype of cost-effectiveness analysis.

� In CostCostCostCost----utiutiutiutility analysislity analysislity analysislity analysis the benefits are measured in utility units or utility-weighted life-years

(such as QALYs). The final result is expressed as a cost/utility ratio for each technology compared.

� CostCostCostCost----benefit analysisbenefit analysisbenefit analysisbenefit analysis measures costs and benefits in monetary terms, and the final result is

expressed as a net monetary gain (or loss) or as a cost/benefit ratio, using approaches such as ‘willingness-to-pay’ or ‘human capital’. Studies are not regarded here as cost-benefit analyses if the savings of resources are considered as benefits (outcomes), even if the authors define them as cost-benefits analyses. However, such studies might report an estimate of an effectiveness outcome (e.g. number of cases of disease prevented) and might instead be classified as cost-effectiveness analyses.

� Outpatient (or ambulatory care)

- Core definition: Medical and paramedical services delivered to patients who are not formally admitted to the facility (physician’s private office, hospital outpatient centre or ambulatory-care centre) and do not stay overnight (OECD, 2000b).

- Supplementary definition: Services provided in a physician’s office, clinic, or other ambulatory setting (Getzen, 1997).

� Inpatient

24

- Core definition: A patient who is formally admitted (or “hospitalized”) to an institution for treatment and/or care and stays for a minimum of one night in the hospital or other institution providing inpatient care (OECD, 2000b).

- Supplementary definition: A patient who has gone through the full admission procedure and is occupying a bed in a hospital inpatient department (NAHAT, 1991).

� Home careHome careHome careHome care

- Core definition: Medical and paramedical services delivered to patients at home (OECD, 2000b).

- Supplementary definition: Included are obstetric services at home, home dialysis, telematic services, etc.; excluded are the consumption of medical goods (pharmaceuticals, other medical goods) dispensed to outpatients as part of private household consumption (Source: OECD, 2000b).

� Community care (social care)Community care (social care)Community care (social care)Community care (social care):

- Core definition: Services related to long-term inpatient care plus community care services, such as day care centres and social services for the chronically ill, the elderly and other groups with special needs such as the mentally ill, mentally handicapped and the physically handicapped. (Source: European Observatory on Health Care Systems, 2001).

� Institutional care

- Core definition: Institutional care is similar to community care (social) (see definition above) but applies to individuals who are long-term residents of an institution but are not classified as in-patients.

� Primary health care

- Core definition: The first level contact with people taking action to improve health in a community (WHO, 1998a).

- Notes: In a system with a gatekeeper, all initial (non-emergency) consultations with doctors, nurses or other health staff are termed primary health care, as opposed to secondary health care or referral services. In systems with direct access to specialists, the distinction is usually based on facilities, with polyclinics, for example, providing primary care and hospitals secondary care (Witter, 1997; Getzen, 1997).

- Supplementary definition: The usual point at which an individual enters the healthcare system. It encompasses the routine care of individuals with common health problems and chronic illnesses that can be managed in the home or through periodic visits to a community health centre. Providers of primary care include health professionals who staff community health centres, out-patient departments, general practitioner units and health units attached to schools and workplaces (Kielhorn A., Graf von der Schulenburg J.-M., 2000).

� Secondary health care

- Core definition: Specialized ambulatory medical services and commonplace hospital care (outpatient and inpatient services). Access is often via referral from primary health care services (European Observatory on Health Care Systems, 2000).

- Notes: Does not include highly specialized, technical inpatient medical services (which is tertiary health care).

- Supplementary definition: The provision of specialised medical services (either diagnosis or treatment) by a specialist or hospital, usually on referral by a primary-care physician. It also includes emergency treatment and critical care (Kielhorn A., Graf von der Schulenburg J.-M., 2000).

25

� Tertiary health care

- Core definition: Refers to medical and related services of high complexity and usually high cost (WHO, 1998a ).

- Supplementary definition: This is treatment of individuals and families with complex or complicated health needs. Providers are health professionals who specialise in a particular clinical area and provide services on a regional or national basis, e.g. psychiatric hospitals. Entry to care is usually by referral from either the primary or secondary level (Kielhorn A., Graf von der Schulenburg J.-M., 2000).

� A model model model model involves the synthesis of inputs from various sources in order to calculate an

estimate of costs and/or benefits. � The purpose of the modelpurpose of the modelpurpose of the modelpurpose of the model could be:

� to extrapolate the progression of clinical outcomes � to transform final outcomes from intermediate measures � to examine the relation between inputs and outputs in production function models to

estimate or apportion resource use � to use data from a variety of sources to undertake a decision analysis � to use evidence from trials, or systematic reviews of trials, to reflect what might happen in

a different clinical setting or population � To include any relevant differences in resource provision not measured in the trials and to

exclude "protocol-driven" resource provision � Common types of modelstypes of modelstypes of modelstypes of models employed include:

� Decision tree modelsDecision tree modelsDecision tree modelsDecision tree models - determine the likely outcomes and/or costs associated with each of various treatment options according to probabilities of occurrence to determine the option which best meets the objective of the decision-maker. For example for a decision tree assessing outcomes only the option which gives the lowest death rate or the shortest stay in hospital. The main limitation of decision trees is that they are not suited to represent multiple outcome events that recur over time.

� StateStateStateState----transition modelstransition modelstransition modelstransition models - allocate and subsequently reallocate members of a population

into one of several categories, or states (representing different health states, events or costs which happen to the patient). Transitions occur from one state to another at defined time intervals (defined as cycle length) this is often 1 year, but can be shorter for rapidly progressive diseases. The time horizon of the model specifies the number of cycles which are modelled. This is usually until there is no difference between the treatment groups. A special type of state-transition model in which the transition probabilities depend only on the current state, and not on the full previous history of the patient is called a Markov model. A specific model where the probabilities associated with movement between states are constant over time is called a Markov Chain.

3. Details About Clinical Evidence3. Details About Clinical Evidence3. Details About Clinical Evidence3. Details About Clinical Evidence � Study populationStudy populationStudy populationStudy population is a set of individuals with given characteristics. In clinical trials a samplesamplesamplesample

of individuals from larger populations can only be observed. Therefore, it is important to

26

have a representative sample to avoid bias. Also, samples are sometimes drawn from a population, which is more restricted than the one about which conclusions are drawn. Findings from such studies can onlyonlyonlyonly apply to the population from which the sample originates.

� Sample size determinationSample size determinationSample size determinationSample size determination: The size of a study determines the magnitude of the p-value and

thus the likelihood that any observed difference will attain statistical significance. SamSamSamSample size ple size ple size ple size calculationscalculationscalculationscalculations are used in the planning phase of a study in order to determine the number of subjects needed to assure a given probability of detecting a statistically significant effect of a given magnitude if it does exist. However, if there is a limited number of subjects to be included in the study, power , power , power , power calculationscalculationscalculationscalculations can be used to establish the likelihood of a statistically significant effect of a given magnitude to be identified among this group of subjects if it is really there. A number of formulae can be used, depending on the study design, the particular research question being addressed, as well as the type of data to be collected. Another major factor influencing the power of the study to detect a true difference between treatment groups is compliance. Noncompliance in any participant makes the intervention and comparison group alike, which results in decreasing the ability of the trial to detect any true differences between groups.

Since sample size calculations based on detecting differences in economic outcomes are not common, sample sizes are often based on the clinical outcomes only.

� Study designs: Study designs: Study designs: Study designs: A particular research question may be addressed using various design

strategies. The choice of the study design is influenced by the research question, logistic considerations of time and resources, and also by previous research and current gaps in knowledge. Defining the design of a study is a complex and not always straightforward process given the existence of variations among textbook definitions. Although authors often state what study design they have adopted, for the purpose of the abstract their classification might need to be amended to allow consistency within the database. The following are the most common study designs: � DesDesDesDescriptivecriptivecriptivecriptive – describing patterns of disease occurrence in relation to variables such as

person, place and time, without an explicit comparison group: � populations (correlational studiescorrelational studiescorrelational studiescorrelational studies); � individuals:

- case reportscase reportscase reportscase reports (of a single patient) - case seriescase seriescase seriescase series (of a group of patients) - crosscrosscrosscross----sectional surveyssectional surveyssectional surveyssectional surveys – exposure and disease status are assessed simultaneously

among individuals in a well-defined population;

� Analytic studiesAnalytic studiesAnalytic studiesAnalytic studies – use explicit comparison group, which allows testing of epidemiologic hypotheses:

� observatiobservatiobservatiobservational studiesonal studiesonal studiesonal studies – the investigator observes the natural course of events:

- casecasecasecase----control studiescontrol studiescontrol studiescontrol studies – the investigator selects the individuals based on whether or

not they have the disease and then determines their previous exposure.

27

- cohort studiescohort studiescohort studiescohort studies - retrospective or prospective “follow-up” studies44 – groups of individuals are defined on the basis of presence or absence of exposure to a suspected risk factor for a disease.

� intervention studiesintervention studiesintervention studiesintervention studies45454545 – clinical trials – the investigator allocates the intervention:

- randomised controlled trialrandomised controlled trialrandomised controlled trialrandomised controlled trial – participants are randomly allocated to study

groups; - nonnonnonnon----randomised controlled trial randomised controlled trial randomised controlled trial randomised controlled trial –––– “quasi “quasi “quasi “quasi----experiment”experiment”experiment”experiment”–––– non-random

allocation to study groups. The control group can be concurrentconcurrentconcurrentconcurrent or historicalhistoricalhistoricalhistorical. Figure 1 represents an algoFigure 1 represents an algoFigure 1 represents an algoFigure 1 represents an algorithmrithmrithmrithm46464646 which can be used in classifying study design. which can be used in classifying study design. which can be used in classifying study design. which can be used in classifying study design.

44�,5��0��,��0��(��45�#��(��!��

46� /�� !�� "�� %�� >�� 3��!!�� C�� %�� >�� 3��'��G��>��A��.***�G��?H�?�3��7�12!-1��

Comparison group(s) NoUncontrolled trial (descriptive study)e.g. case series

Yes

Exposure andoutcome

determined at thesame time

Yes Cross-sectional

No

More than onegroup studied?

Yes

Measurementsmade duringintervention?

Yes

NoNo

Time series

Before and after study

Investigatorsassign exposure? Yes

No

Randomisation?

No

No

Non-randomisedcontrolled trial

Yes

Randomisedcontrolled trial

Groups definedby?

Exposure

OutcomeCase-control

Cohort design?

Yes

Other designs withconcurrent controls (e.g.

time series study withcomparison group

Prospectivecohort study

Retrospectivecohort study

Prospective?Yes

No

28

� Methods of randomisation:Methods of randomisation:Methods of randomisation:Methods of randomisation: Participants are randomly allocated if they have the same chance of being assigned to any of the study groups. Groups are balanced at baseline if their characteristics are similar. Randomisation can be achieved using a variety of procedures: flipping a coin, rolling a die, random number tables, computer generated random numbers, sealed envelopes, etc. Block randomisationBlock randomisationBlock randomisationBlock randomisation is used to keep the number of participants as close as possible. Within each block the same number of participants are allocated to the study groups. Stratified randomisationStratified randomisationStratified randomisationStratified randomisation is used to keep characteristics of the participants as similar as possible across groups. Examples of units of randomisation include: patients, physicians, nurses, groups of people – hospitals, families, geographical areas (also known as cluster randomisationcluster randomisationcluster randomisationcluster randomisation), etc.

� Blinding:Blinding:Blinding:Blinding: RCTs where all participants are aware of the interventions are referred to as openopenopenopen

(e.g. trials on surgical interventions). Blinding is used in RCTs as a methodological strategy to control for biases. In a singlesinglesinglesingle----blindblindblindblind RCT one group of individuals (participants or investigators) involved in the trial is not aware of the identity of the intervention given to each participant. Double-blinding refers to trials in which two groups of individuals involved in the trial (usually participants and investigators, assessing the outcomes) do not know the identity of the intervention given to each participant. Triple-blind and quadruple-blind trials are rare and involve blinding of three or four groups respectively, e.g. participants, data analysts, investigators giving the intervention, investigators, assessing the outcomes, etc.

� Loss to follow upLoss to follow upLoss to follow upLoss to follow up describes the participants who were initially included in the study but did

not complete the observation period. It is also known as withdrawal or drop out rate. � A confounding variableconfounding variableconfounding variableconfounding variable is correlated with the outcome and the probability of receiving a

treatment or exposure, so biasing estimates of the association between treatment and outcome.

� Handling of missing dataHandling of missing dataHandling of missing dataHandling of missing data: The way missing data are handled in a trial is important as it can

be a source of bias if inappropriate. The analysis is called “intention to treatintention to treatintention to treatintention to treat” if all study participants are included in the analysis as part of the groups they were initially allocated regardless of whether or not they completed the study. Sensitivity analysis should be performed to account for the uncertainties around the data. In contrast, in the “treatment completerstreatment completerstreatment completerstreatment completers” analysis, only participants who completed the observation period are included and are analysed according to the treatment they actually received. This may introduce bias due to non-random drop out.

� The confidence intervalconfidence intervalconfidence intervalconfidence interval is the plausible range of values above and below the point estimate of a treatment effect, that is likely to include the true value of the treatment effect. A confidence interval of 95% means that there is a probability of 95% that the confidence interval calculated from a specific study includes the true value of a treatment effect. If the interval includes a null treatment effect, the null hypothesis cannot be rejected.

� In hypothesis testing, the pppp----valuevaluevaluevalue is the probability that an observed difference or a greater

difference between the control and the intervention groups was due to chance alone if the

29

null hypothesis (no difference) is true. Conventionally, if “p” is less than 0.01 or 0.05, the null hypothesis is rejected.

� MetaMetaMetaMeta----analysisanalysisanalysisanalysis is a method of analysis which combines results from independent studies using

statistical techniques. The results are weighted according to the sample size or variability, and are sometimes weighted by study quality. The aims are mainly: to obtain a pooled estimate of an overall "true treatment” effect and/or to obtain an improved estimate of the outcome of a specific intervention which has been analysed in a class of studies. It is also known as “data synthesis” or “quantitative overview”.

4.4.4.4. Economic AnalysisEconomic AnalysisEconomic AnalysisEconomic Analysis � Quality Adjusted LifQuality Adjusted LifQuality Adjusted LifQuality Adjusted Life Yearse Yearse Yearse Years measures the duration of life discounted by some fraction

between 0 and 1 which estimates the quality of life during the period. It represents the equivalence of being alive for a certain proportion of a year in perfect health. It is also known as “well years”, “well life expectancy”, “health life years” and “years of healthy life”.

� HealthyHealthyHealthyHealthy----Years EquivalentYears EquivalentYears EquivalentYears Equivalent is a measure of health-related quality of life. It is the hypothetical

number of years spent in perfect health which could be considered equivalent to the actual number of years spent in a defined imperfect health state. It differs from a QALY because not only is it based on the individual’s preferences for the duration of life, but also on the individual’s preferences for the states of health. QALYs and HYEs are used to measure the benefits in cost-utility analysis, where single or multiple effects are not common to both alternatives.

� In willingness to paywillingness to paywillingness to paywillingness to pay a target group such as the general public or patients is asked how

much, in monetary terms, it is willing to give up (from surplus income) to acquire the benefits of the intervention. In the human capital approachhuman capital approachhuman capital approachhuman capital approach the method involves estimating the value of a human life by, for example, calculating the lost income of an individual due to his/her condition without the benefits of the intervention.

� Standard gambleStandard gambleStandard gambleStandard gamble is a method to measure utility directly. It is based on the fundamental von

Neumann-Morgenstern axioms of expected utility theory. The subject is asked to make a trade-off between the certainty to have a chronic disease for a certain period of time, and a gamble. The gamble has the following alternatives: good health for the same period and death.

� In time tradetime tradetime tradetime trade----offoffoffoff, the utility value is measured by finding the point at which the respondent

is indifferent between two health states for different lengths of time. For chronic states, the choice is between the index health state for time t and perfect health for a shorter time, both followed by death. For temporary states, the choice is between the index health state for time t and a worse health state for a shorter time, both followed by the same specified outcome.

� Conjoint analysisConjoint analysisConjoint analysisConjoint analysis is gaining widespread use in health care and has been applied successfully in several areas, including eliciting patients' and the community's preferences in the delivery of health services; establishing consultants' preferences in priority setting; developing outcome measures; determining optimal treatments for patients; evaluating alternatives within randomised controlled trials; and establishing patients' preferences in the doctor-patient relationship.

30

� Direct costsDirect costsDirect costsDirect costs include the value of all resources consumed in the provision of a health care

intervention – the acquisition, administration and monitoring costs – as well as current and future consequences of that intervention such as the treatment of adverse effects. Direct costs can be divided into health care costs which include the acquisition costs of pharmaceuiticals and other necessary medical supplies, the costs of laboratory and radiology tests, the labour costs of health care personnel and institutional overhead costs, for example support services, heating, electricity, maintenance, administration and capital costs; and non-health care costs which include formal patient care costs, volunteer informal care costs, childcare costs, and the cost of transportation to and from treatment centres.

� DiscountingDiscountingDiscountingDiscounting is important because costs and benefits available today have a higher value than costs and benefits occurring in the future. Discounting future health benefits reflects individual positive time preference but is more controversial. The rate of discounting for costs is typically 2 to 6% per annum – at publication of this report the UK Treasury recommends 6% (1-1.5% for benefits), while the Washington Panel recommends 3%.

� Resource quantitiesResource quantitiesResource quantitiesResource quantities can be measured at different levels of precision. The most precise level, microcosting is where resource use is identified at a detailed level and a unit cost attached to each resource. For example, staff time administering drugs could be measured and valued by a staff cost per hour. Gross-costing involves a more aggregated approach using cost estimates for units of input and outputs that are large relative to the intervention being analysed. For example, the number of hospital days could be measured and valued by the average cost of a hospital day.

� Incremental costIncremental costIncremental costIncremental cost and marginal costmarginal costmarginal costmarginal cost are two different concepts, although the terms are often used interchangeably. Incremental cost is the difference in costs found between two technologies (i.e. intervention and comparator). Marginal cost is the difference in cost due to the expansion or contraction of a programme (e.g. increase of one day in the length of stay). They coincide only when the expanded (or contracted) programme is considered as the alternative option. Average costsAverage costsAverage costsAverage costs are calculated by dividing the total costs for the intervention or comparators(s) by the number of patients receiving it.

� Indirect costsIndirect costsIndirect costsIndirect costs are the costs of lost productivity due to a disease, for example the time lost from work. They are typically calculated from the gross earnings (including employment overheads and benefits) of those in employment. Imputed data may be derived for those not in employment (e.g. homemakers) using average earnings, the cost of replacing that individual with a paid worker or the opportunity cost of the production they could have achieved if they were not at home.

� Friction costsFriction costsFriction costsFriction costs estimate productivity losses based on the amount of time organisations need to restore the initial production level and are arguably more realistic in assessing productivity losses due to periods of absence on behalf of the patient.

� CurrencyCurrencyCurrencyCurrency Some common abbreviations used with the database:

'�� '��I� :��J�� :JI�

=�� =�� "�� /A�

%�� %��I� 3��'�� 5�

/��K�� /�� 3�� 39K�

31

'�� '��I� :��J�� :JI�

/�� /�� 3�� 3��

C�� C�� FK�� L�

#�� B� F3�� I�

� Where patient level data on costs are available (e.g. through clinical trials or observational

studies) these data are said to be “stochastic” in that they are random variables subject to probability estimates. Tests of significance for this type of data can be undertaken through e.g. Student’s t test, Mann-Whitney U test. Also, confidence intervals may be placed around the means or standard errors to indicate the precision of cost estimates.

� Types of “descriptive statisticsdescriptive statisticsdescriptive statisticsdescriptive statistics” include measures of central tendency (e.g. mean (average) or median) and measures of variability / dispersion (e.g. standard deviations or the inter-quartile range) and the shape of the distribution (e.g. whether data are skewed). Typically cost data are highly skewed with a few patients incurring very high costs and in this case the median is the best measure of central tendency. However, within economic analyses, mean costs should be estimated so that total intervention costs may be calculated by multiplying the mean by the number of patients receiving the intervention.

� The term “deterministic costsdeterministic costsdeterministic costsdeterministic costs” refers to the costs that were included in the analysis as point estimates only. If costs were treated as deterministic, sensitivity analysis (section 4.6) should be employed to identify critical areas of uncertainty associated with estimates.

� Only parametric testsparametric testsparametric testsparametric tests assume data are normally distributed. If data are skewed non non non non----

parametric statistical testsparametric statistical testsparametric statistical testsparametric statistical tests might be undertaken. Alternatively parametric tests may be applied following e.g. bootstrapping or log-transformation.

� Sensitivity analysisSensitivity analysisSensitivity analysisSensitivity analysis is a method testing the robustness of the analysis: the deterministic variables for which there is uncertainty (e.g. effectiveness data, discount rate, costs, etc.) are varied over a range and the related changes in the final results are analysed.

� Areas of uncertaintyAreas of uncertaintyAreas of uncertaintyAreas of uncertainty which can be investigated in sensitivity analyses:

1. Generalisability of resultsGeneralisability of resultsGeneralisability of resultsGeneralisability of results: in this case the uncertainty may concern for example the validity of the effectiveness and cost results for different groups of patients, the applicability of the cost-effectiveness observed in a trial to a routine clinical practice, the possibility of getting similar consequences in different settings (hospitals, regions, countries).

2. ExtrapolatioExtrapolatioExtrapolatioExtrapolation from primary data sourcen from primary data sourcen from primary data sourcen from primary data source to make the results more comprehensive: The extrapolation may concern for example: a) the extension of cost-effectiveness results based on intermediate clinical end-points to a model based on final clinical end-points; b) the prolongation of the time horizon of a study.

3. Analytical methodsAnalytical methodsAnalytical methodsAnalytical methods: There exists a certain degree of disagreement about the methods and the instruments of analysis used in the economic evaluation studies (for example the inclusion of indirect costs, the valuation of the consequences of the interventions, etc.).

32

� OneOneOneOne----way simple sensitivity analysisway simple sensitivity analysisway simple sensitivity analysisway simple sensitivity analysis: By this method each parameter is varied individually in order to isolate the consequences of each parameter on the results of the study.

� MultiMultiMultiMulti----way simple sensitivitway simple sensitivitway simple sensitivitway simple sensitivity analysisy analysisy analysisy analysis: By this method two or more parameter are varied at

the same time and the overall effect on the results is evaluated (this is sometimes called "scenario analysis"). Simple sensitivity analysis is generally useful in investigating all the areas of uncertainty previously described.

� Threshold analysisThreshold analysisThreshold analysisThreshold analysis: This analysis identifies the critical values of the parameters above or

below which the results of a study vary. This method is usually used jointly with the simple sensitivity analysis, and it is generally useful to investigate all the areas of uncertainty described, with the exception of the uncertainty surrounding the analytical methods.

� Analysis of extremesAnalysis of extremesAnalysis of extremesAnalysis of extremes: By this method, a base-case analysis based on the best estimates of

inputs is considered. Therefore alternative analyses based on extreme estimates variability and extrapolation from primary data source, and may be useful in addressing the issue of the generalisability. ("pessimistic" and "optimistic") are generated. This analysis, which is also known as worst/best case analysis, is generally useful to deal with uncertainty in data.

� Probabilistic sensitivity analysis:Probabilistic sensitivity analysis:Probabilistic sensitivity analysis:Probabilistic sensitivity analysis: This method of analysis attributes distributions of

probabilities to the uncertain variables which are incorporated into evaluation models based on decision analytical techniques (e.g. Monte-Carlo simulation). This method can only be used to deal with uncertainty in data input.

5. Results5. Results5. Results5. Results

� CostCostCostCost----effectiveness acceptability curveseffectiveness acceptability curveseffectiveness acceptability curveseffectiveness acceptability curves assist the decision-making process by depicting the ceiling costs per extra unit of effect values on the x axis against the probability of these values being achieved on the y axis. � The incremental net benefit approachThe incremental net benefit approachThe incremental net benefit approachThe incremental net benefit approach is new in the field of economic evaluation. It uses a societal (or decision-makers’) value, depicted by 1/g or λ, to specify the value below which the incremental cost-effectiveness ratio must be, to produce a net benefit. ��

�

/��AC��E =��=��3��"B��"@ ��A��9��9��;��%��>��.��:��M��7�E(��F��>��?NN)��"��A5��3��G9��5��B=��@��A%��%��!��:��M��7�E(��F��>��?NN+��5�� KG� �� "�� 3�� A�� 9�� B��!5�� >��>��?NNH��=��G5��>��3"��A��/��A��?NH1��17-?N!2H��

33

Sonnenberg FA, Beck JR. Markov models in medical decision making: a practical guide. Medical Decision Making 1993;13:322-38. Briggs AH, Gray AM. Handling uncertainty when performing economic evaluation of healthcare interventions. Health Technology Assessment 1999;3(2). Briggs AH, Sculpher M. An introduction to markov modelling for economic evaluation. Pharmacoeconomics 1998; 4:397-409. Detsky AS, Naglie G, Krahn MD, Naimark D, Redelmeier DA. Primer on medical decision analysis - getting started. Medical Decision Making 1997; 17:123-5. Detsky AS, Naglie G, Krahn MD, Redelmeier DA, Naimark D. Primer on medical decision analysis - Building a tree. Medical Decision Making 1997; 17:126-35. Krahn MD, Naglie G, Naimark D, Redelmeier DA, Detsky AS. Primer on medical decision analysis - Analysing the model and interpreting the results. Medical Decision Making 1997; 17:142-51. Naglie G, Krahn MD, Naimark D, Redelmeier DA, Detsky AS. Primer on medical decision analysis - Estimating Probabilities and Utilities. Medical Decision Making 1997; 17:136-41. Naimark D, Krahn MD, Naglie G, Redelmeier DA, Detsky AS. Primer on Medical Decision Analysis: Working with Markov Processes. Medical Decision Making 1997; 17:152-9. Doubilet P, Begg CB, Weinstein MC, Braun P, McNeil BJ. Probabilistic sensitivity analysis using Monte Carlo simulations. Medical Decision Making 1985; 5:157-77. Sculpher M, Fenwick E, Claxton K. Assessing quality in decision analytic cost-effectiveness models. A suggested framework and example of application. Pharmacoeconomics 2000; 17(5):461-77. Ryan M and Farrar S ‘Using conjoint analysis to elicit preferences for health care,’ BMJ 2000;320:1530-1533. Claxton K, ‘The irrelevance of inference: a decision-making approach to the stochastic evaluation of health care technologies,’ Journal of Health Economics 1999:18;341-364.

34

APPEAPPEAPPEAPPENDIX 2.NDIX 2.NDIX 2.NDIX 2.

THE COMMENTARY STEPTHE COMMENTARY STEPTHE COMMENTARY STEPTHE COMMENTARY STEP----BYBYBYBY----STEPSTEPSTEPSTEP

Figure 1: Choice of Comparator.

� ��

� �

� ��

� ��

� �

� ��

� �

� � ��

� ��

� �

� �� ! � ��

� �

35

� �� " ��

# ��$ �� % # �� $ �� &

� ��" ��

��

� ��

# ��$ �� '

��

� ��

� �� ! ��

�� (�� (��

�� (�� ) �� (��

�� (�� (��

�� !�� "�� "��

�� (�� (�� *��) ��

� �

� �� ) ��

Figure 2: Effectiveness Source and Benefit.

36

Figure 3: Effectiveness, Type A.

��(��

��(��

��+ ��

��+ ��

�� # $ � � ��

�� % ��% ��

� ��

� ��

��

� ��

� ��

��(�� !�� "�

�� ) ��

��(�� !�� "�

�� ) ��

��

� �

��

��

��

37

, �� - �

� ��

��

��

��

��(�� - ��

��(�� - ��

��

��(��

�� & ��

� ��

��(��

, ��

��

. �� - ��

�� ' ��

� �

��(��

��(��

Figure 4: Effectiveness, Type B.

38

� ��

��

� �

��(��

��

/ ! �� 0 ��

��

�� (��

1 ��

�� ( ��

. �� - ��

��

� ��

��

��/ ��

��/ ��

� ��

� �

, ��

� ��2��

Figure 5: Effectiveness, Type C.

39

� ��- ��3��

�� 4 ��

�� - ��3��

�� )��

� �

��. ��) ��

��. ��) ��

��

$ ��

��

��

� �

��

� ��) ��

��

��

��

$ ��

Figure 6: Cost overview

40

$ �� ) ��- �

� ��) ��

� �� ) ��

� � � ��* �� + �� % ��

�� ) ��

�� + ��

�� + ��

��# �� , �� - ��+ ��

� ��) ��

��

� �� ) ��

��

+ �� 0 ��"�0 ��4 �� 4 ��

� �

Figure 7: Resource Use.

41

� � � ��

��

��

��

��

��

$ �� - �

+ �� 0 ��"�0 �� 2��

� ��

� ��

�� . ��

�� % ��

. �� ! ��

� ��

� ��

� ��

. ��

Figure 8: Prices or unit costs

42

, ��- ��

��

� ��

��(��

��(��

��(��- ��

��(��

, ��

��(��

��(��

, ��2��

��

��(�� !�� % �� "

��

��

, ��

Figure 9: Other issues.

43

APPENDIX 3.APPENDIX 3.APPENDIX 3.APPENDIX 3.

NHS EED AND CODECS EXAMPLE ABSTRACTSNHS EED AND CODECS EXAMPLE ABSTRACTSNHS EED AND CODECS EXAMPLE ABSTRACTSNHS EED AND CODECS EXAMPLE ABSTRACTS 1. Type A abstract 1. Type A abstract 1. Type A abstract 1. Type A abstract –––– where the source of effectiveness d where the source of effectiveness d where the source of effectiveness d where the source of effectiveness data is a single study. ata is a single study. ata is a single study. ata is a single study.

Full record displayFull record displayFull record displayFull record display

Item 1 of 1 --- Document 130559 Cost effectiveness of community leg ulcer clinics: randomised controlled trialCost effectiveness of community leg ulcer clinics: randomised controlled trialCost effectiveness of community leg ulcer clinics: randomised controlled trialCost effectiveness of community leg ulcer clinics: randomised controlled trial Morrell C J, Walters S J, Dixon S, Collins K A, Brereton L M L, Peters J, Brooker C G D. BMJ 1998; 316: 1487-1491. This record was compiled by CRD commissioned reviewers according to a set of guidelines developed in collaboration with a group of leading health economists. Health technologyHealth technologyHealth technologyHealth technology Clinic-based weekly treatment of leg ulcers with four layer bandaging or usual home-based care by district nurses. DiseaseDiseaseDiseaseDisease Skin and connective tissue diseases. Type of interventionType of interventionType of interventionType of intervention Treatment. Hypothesis/study questionHypothesis/study questionHypothesis/study questionHypothesis/study question The objective of the study was to assess the cost-effectiveness of using clinic-based weekly treatment of leg ulcers with four layer bandaging versus using the usual home-based care by district nurses for the treatment of venous leg ulcers. Home-based care was explicitly considered as the comparator since it represented the common practice in dealing with venous leg ulcers. The perspective considered in the economic analysis was explicitly stated as the UK NHS. Economic study typeEconomic study typeEconomic study typeEconomic study type Cost-effectiveness analysis. Study populationStudy populationStudy populationStudy population The study population comprised patients who had had a venous ulcer below the knee to the foot for at least 3 months and who were able to travel to the clinic. SettingSettingSettingSetting Community and primary care. The economic study was carried out in Sheffield, UK. Dates to which data relateDates to which data relateDates to which data relateDates to which data relate The effectiveness data and resources use data were gathered between September 1994 and May 1995. 1995 prices were used. Source of effectiveness dataSource of effectiveness dataSource of effectiveness dataSource of effectiveness data The evidence for the final outcomes was derived from a single study. Links between effectiveness and cost dataLinks between effectiveness and cost dataLinks between effectiveness and cost dataLinks between effectiveness and cost data The costing was undertaken prospectively on the same patient sample as that used in the effectiveness study.

44

Single studySingle studySingle studySingle study Study sampleStudy sampleStudy sampleStudy sample Power calculations were used to determine the sample size. A total of 328 subjects attended for assessment, of whom 233 were randomly assigned: 113 to the home group and 120 to the clinic group. Patients (93) from the original 328 with an ankle brachial pressure index of <0.8 were excluded from the clinic group and returned to home care. Two patients refused to participate and were returned to home care. Study desStudy desStudy desStudy designignignign The study was a randomised controlled trial carried out in eight centres (Sheffield, Nottingham, Derby and Sutton-in-Ashfield). The duration of the follow-up was 1 year and the drop out rates were 14.2% for the clinic and 20.3% for the home group. A random assignment schedule and serially numbered, sealed, opaque allocation envelopes were prepared in advance for each of the eight clinic sites. Analysis of effectivenessAnalysis of effectivenessAnalysis of effectivenessAnalysis of effectiveness The analysis of the effectiveness study was based on intention to treat. The primary health outcomes were the initial healing time of the ulcers, the cumulative percentage of ulcer healed at 12 weeks, the ulcer recurrence rate from the initial ulcer healed time to first ulcer recurrence, the average ulcer-free time over the 12 month follow-up period, and health status measured by the 36 item short form health survey (SF-36), the EuroQol (EQ), the McGill short form pain questionnaire (SF-MPQ), and the Frenchay activities index. The groups were shown to be comparable in health status and demographic features. The effects of confounding variables such as age, ulcer area, ulcer duration, and history of deep vein involvement on healing times were adjusted using Cox proportional hazards regression analysis. Effectiveness resultsEffectiveness resultsEffectiveness resultsEffectiveness results The clinic group had a better record in initial healing times (P= 0.03) and had a higher rate of cumulative percentage of ulcer patients healed at 12 weeks (34% versus 24% in the home group, 95% CI for the difference was -2% to 22%). The patients in the clinic group were 1.45 times more likely to be healed from the initial ulcer (95% CI: 1.04 - 2.03). When adjusted for confounding variables the clinic group were 1.65 times (1.15 to 2.35) more likely to be healed. The clinic group experienced a higher rate of ulcer recurrence after the initial healing (35% versus 23% in the home group). The study revealed no difference in the time to first ulcer recurrence between the groups (log rank test 0.78; df=1; P=0.38). The average ulcer-free time over a one year follow-up period was 20.1 weeks for the clinic group and 14.2 weeks for the home group (95% CI for the difference was 1.2 to 10.5). According to the health status measures, no significant differences between the groups were identified. Clinical conclusionsClinical conclusionsClinical conclusionsClinical conclusions The study revealed that the clinic-based treatment using four layer bandaging supplied by specially trained nurses was more effective than the usual home-based care provided by district nurses. Economic analysisEconomic analysisEconomic analysisEconomic analysis Measure of benefits used in the economic analysisMeasure of benefits used in the economic analysisMeasure of benefits used in the economic analysisMeasure of benefits used in the economic analysis Ulcer-free time (weeks) was considered by the authors as the most appropriate measure to be included in the economic analysis. Direct costsDirect costsDirect costsDirect costs Quantities were reported separately from costs for the home group but not for the clinic group (reported in a separate paper). The average annual total costs were divided into leg ulcer treatment cost per year including staff time, materials, transport, and overheads; the cost of GP services per year; and the cost hospital services per year. The average cost per clinic attendance and per home visit were reported. The cost/quantity boundary was that of the NHS. The cost data related to overhead costs were obtained from the service managers. 1995 prices were used. Indirect costsIndirect costsIndirect costsIndirect costs Not included. CurrencyCurrencyCurrencyCurrency UK pounds Sterling (£). Statistical analysis of costsStatistical analysis of costsStatistical analysis of costsStatistical analysis of costs Permutation tests were performed to compare the average costs between the groups.

45

Sensitivity analysisSensitivity analysisSensitivity analysisSensitivity analysis A series of one-way sensitivity analyses was carried out on treatment costs and overhead costs to assess the effects of changes in baseline values on the average total costs. Estimated benefits used in the economic analysisEstimated benefits used in the economic analysisEstimated benefits used in the economic analysisEstimated benefits used in the economic analysis The average ulcer-free time over a one year follow-up period was 20.1 weeks for the clinic group and 14.2 weeks for the home group (95% CI for the difference was 1.2 to 10.5). Cost resultsCost resultsCost resultsCost results The average (SD) total costs for the clinic group were £877.6 (674.3) versus £863.09 (865.32) for the home group. Synthesis of costs and benefitsSynthesis of costs and benefitsSynthesis of costs and benefitsSynthesis of costs and benefits A synthesis of costs and benefits was carried out by calculating a cost-effectiveness ratio (additional cost required per additional ulcer-free week). The value of the ratio was £2.46 (-£31.94 to £99.12) per ulcer free week. None of the parameters in the sensitivity analysis had significant effects on the average costs of the alternative heath technologies. The authors pointed out that, if the grade of the nurses reduced from G or H to E, and if throughput per session increased to more than 10 patients, the clinic-based treatment would dominate the usual home-based care. CoCoCoConclusions, commentary and implicationsnclusions, commentary and implicationsnclusions, commentary and implicationsnclusions, commentary and implications Author's ConclusionsAuthor's ConclusionsAuthor's ConclusionsAuthor's Conclusions Community-based leg ulcer clinics with specially trained nurses, using four layer bandaging, were more effective than traditional home based treatment. This benefit was achieved at a small additional cost and could be delivered at reduced cost if certain service configurations were used. CRD commentaryCRD commentaryCRD commentaryCRD commentary Selection of comparators: The reason for the choice of the comparator is clear. The comparator was chosen because it represented the routine care provided for patients with venous leg ulcers. You should consider whether this is a widely used technology in your own setting. Validity of estimate of measure of benefit: The estimates of the measure of benefit are likely to be valid because of randomisation and accounting for the effects of confounding variables. 93 patients were excluded from the trial as they were contraindicated for four layer compression bandages. Additionally, the care provided in the community was not a standard treatment as in the case of the clinic group. These factors may tend to bias the benefit results in favour of the clinic group. Validity of estimate of costs: The resource quantities were not reported for the clinic-based treatment. As the authors noted, it would have been more appropriate to have included the indirect costs of the alternative heath technologies. The exclusion of indirect costs might have biased the results in favour of the home-based treatment. Other issues: Other issues: Other issues: Other issues: The issue of generalisability to other settings was partially addressed by performing sensitivity analyses on the cost items. The authors made extensive comparisons with other studies and provided discussions concerning observed differences. These included the age distribution of the sample, baseline distributions of leg ulcers, and healing rates with the intervention. The influence of comorbidities on the quality of life estimates was also discussed in that there are problems in isolating the impact of leg ulcers on overall quality of life. Implications of the studyImplications of the studyImplications of the studyImplications of the study The findings of the study suggest that for venous leg ulcer clinics (using four-layer compression bandaging) to be a cost-effective option for the NHS, they should only be implemented if they are staffed by appropriately trained nurses in the relevant grades, with an optimal flow of patients. Subject index termsSubject index termsSubject index termsSubject index terms Subject indexing assigned by NLM:

46

Aged; Ambulatory-Care-Facilities [economics]; Bandages; Cost-Benefit-Analysis; England; Follow-Up-Studies; Home-Care-Services [economics]; Leg-Ulcer [economics]; Patient-Satisfaction; Recurrence; Rural-Health-Services [economics]; Treatment-Outcome; Urban-Health-Services [economics]; Ambulatory-Care [economics]; Community-Health-Services [economics]; Leg-Ulcer [therapy]; Female; Human; Male; Support,-Non-U. S.-Gov't Country codeCountry codeCountry codeCountry code United Kingdom. Review funding bodyReview funding bodyReview funding bodyReview funding body Study commissioned and funded by the former Trent Regional Health Authority (now NHS Executive Trent). Address for correspondenceAddress for correspondenceAddress for correspondenceAddress for correspondence Dr C J Morrell, School of Health and Related Research (ScHARR), University of Sheffield, Sheffield S1 4DA, United Kingdom. Accession number and database entry dateAccession number and database entry dateAccession number and database entry dateAccession number and database entry date 988130 31121998 Web addressWeb addressWeb addressWeb address: http://www.bmj.com/cgi/content/full/316/7143/1487 Copyright:Copyright:Copyright:Copyright: University of York, 1998.

2. Type B abstrac2. Type B abstrac2. Type B abstrac2. Type B abstract t t t –––– the effectiveness evidence is derived from a review of the literature. the effectiveness evidence is derived from a review of the literature. the effectiveness evidence is derived from a review of the literature. the effectiveness evidence is derived from a review of the literature.

Full record display Full record display Full record display Full record display

Item 1 of 1 Item 1 of 1 Item 1 of 1 Item 1 of 1 ------------ Document 130784 Document 130784 Document 130784 Document 130784 The use of CT scanning in dementia: a systematic reviewThe use of CT scanning in dementia: a systematic reviewThe use of CT scanning in dementia: a systematic reviewThe use of CT scanning in dementia: a systematic review Foster G R, Scott D A, Payne S. International Journal of Technology Assessment in Health Care 1999; 15(2): 406-423. This record was compiled by CRD commissioned reviewers according to a set of guidelines developed in collaboration with a group of leading health economists. Health technologyHealth technologyHealth technologyHealth technology The use of CT scanning in dementia. DiseaseDiseaseDiseaseDisease Nervous system diseases. Type of interventionType of interventionType of interventionType of intervention Screening and treatment. Hypothesis/study questionHypothesis/study questionHypothesis/study questionHypothesis/study question To determine the cost-utility of computerised tomography (CT) scanning as a screening test for potentially reversible dementia. The comparator was explicitly stated as "no screening". The aim of the study was to define what proportion of dementias are due to theoretically treatable causes, what proportion would actually be treated if diagnosed by CT scanning and what proportion of those treated would actually benefit. The CT scan offers the possibility of diagnosing structural lesions such as tumours, subdural haematomas, hydrocephalus and strokes at virtually no risk to the patient. The perspective adopted in the economic analysis was not stated but appears to be that of the health service in Scotland. Economic study typeEconomic study typeEconomic study typeEconomic study type

47

Cost-utility analysis. Study populationStudy populationStudy populationStudy population Male and female patients above and below 65 years of age with newly-diagnosed dementia without a theoretically treatable condition. SettingSettingSettingSetting Hospital. The study was carried out in Scotland, UK. Dates to which data relateDates to which data relateDates to which data relateDates to which data relate The effectiveness data were derived from studies previously published up to December, 1997. The dates during which resource use data were collected and the price year were not reported. Source of effectiveness dataSource of effectiveness dataSource of effectiveness dataSource of effectiveness data Effectiveness data were derived from a systematic review of published studies. ModellingModellingModellingModelling A decision tree (spreadsheet) was used to model costs and benefits of the screening strategies. This is available on request from the authors. Review/synthesis of previously published studiesReview/synthesis of previously published studiesReview/synthesis of previously published studiesReview/synthesis of previously published studies Outcomes assessed in the reviewOutcomes assessed in the reviewOutcomes assessed in the reviewOutcomes assessed in the review The following outcomes were assessed in the review: prevalence of dementia, prevalence of reversible dementia, proportion of reversible dementias detectable by CT, the occurrence of normal pressure hydrocephalus (NHH), brain tumours and chronic subdural haematomas (SDH), the proportion of all dementia with lesions detectable by CT, the proportion of CT detectable reversibles, the outcomes of surgery for tumours, SDH or for NPH, the number of life years gained by tumour treatment, and the number of life years gained with and without SDH operation or NPH operation. Study designs and other criteria for inclusion in the reviewStudy designs and other criteria for inclusion in the reviewStudy designs and other criteria for inclusion in the reviewStudy designs and other criteria for inclusion in the review Observational studies were included since few clinical trials were available. Sources searched to identify primary studiesSources searched to identify primary studiesSources searched to identify primary studiesSources searched to identify primary studies The MEDLINE and Embase databases and references from personal contacts were searched. Reference lists of all useful papers obtained were also used. CriteriCriteriCriteriCriteria used to ensure the validity of primary studiesa used to ensure the validity of primary studiesa used to ensure the validity of primary studiesa used to ensure the validity of primary studies Not stated. Methods used to judge relevance and validity, and for extracting dataMethods used to judge relevance and validity, and for extracting dataMethods used to judge relevance and validity, and for extracting dataMethods used to judge relevance and validity, and for extracting data Summary statistics from each study were used. Number of primary studies includedNumber of primary studies includedNumber of primary studies includedNumber of primary studies included Approximately 21 studies were included. MMMMethod of combination of primary studiesethod of combination of primary studiesethod of combination of primary studiesethod of combination of primary studies Most effectiveness estimates were not combined, although on a few occasions weighted averages were used. The data are largely based on Clarfield's meta-analysis (ref. 12), supplemented by other studies (amounting to 4,800 patients). Investigation of differences between studiesInvestigation of differences between studiesInvestigation of differences between studiesInvestigation of differences between studies Not stated. Results of the reviewResults of the reviewResults of the reviewResults of the review The prevalence of dementia was 9.3%. The proportion of reversible dementias and the proportion of reversible dementias detectable by CT were 13.7% and 27.5%, respectively. NPH occurs in less than 2% of dementia cases. Brain tumours occur in 1 to 4% of cases. SDH occurs in 0.4% of demented patients. The proportion of all dementia with lesions detectable by CT was 6.6%. The proportion of CT detectable reversibles due to tumour, SDH or NPH was 53.36, 11.13, and 35.51, respectively. The probabilities of mortality with surgery for tumour, for SDH and for NPH were 0.05 each. The probabilities of improvement with surgery for tumours, SDH, and NPH were 0.18, 0.8, and 0.2, respectively. The number of life years gained by tumour treatment was 6 months. The life years gained after

48

improvement from SDH operation were 100% (of 16 years), if untreated SDH 25%, or after failure to improve from SDH operation were 25%. The life years gained after improvement from NPH operation, if untreated NDH, or after failure to improve from NDH operation are 75%, of 16 years each. These data formed the principal effectiveness/outcome parameters used in the model. Economic analysisEconomic analysisEconomic analysisEconomic analysis Measure ofMeasure ofMeasure ofMeasure of benefits used in the economic analysis benefits used in the economic analysis benefits used in the economic analysis benefits used in the economic analysis Quality-adjusted life years (QALYs) were used as the measure of benefits. The authors adopted the quality of life values used by Simon and Lubin, which were based on their assessment of quality of life for demented subjects. Direct costsDirect costsDirect costsDirect costs Costs were not discounted given that only screening and initial treatment costs were included. Direct costs included costs of the scan, capital charges, transport and community costs. The quantity/cost boundary adopted was that of the health service. The estimation of quantities was based on actual data. Cost estimates were derived from the Royal Infirmary of Edinburgh NHS Trust and Dundee Teaching Hospitals NHS Trust. The price year was not reported. Indirect costsIndirect costsIndirect costsIndirect costs Not included. CuCuCuCurrencyrrencyrrencyrrency UK pounds sterling (£). Statistical analysis of costsStatistical analysis of costsStatistical analysis of costsStatistical analysis of costs Not stated. Sensitivity analysisSensitivity analysisSensitivity analysisSensitivity analysis A sensitivity analysis was conducted on variables considered to be sensitive. These included the proportion of all cases of dementia that are reversible, the proportion of those reversible cases that can actually be detected by CT scanning, the future life span of the patients presenting for screening, and the cost of each CT scan. Estimated benefits used in the economic analysisEstimated benefits used in the economic analysisEstimated benefits used in the economic analysisEstimated benefits used in the economic analysis A strategy of screening all patients with dementia and only treating those with SDH generated 178 QALYs. A strategy of screening all patients with dementia and only treating those with SDH and NPH generated 174 QALYs. A strategy of screening all patients with dementia and only treating those with SDH and tumours generated 217 QALYs. Cost resultsCost resultsCost resultsCost results The costs associated with screening all patients with dementia and only treating those with SDH amounted to £2,517.329. The costs associated with screening all patients with dementia and only treating those with SDH and NPH amounted to £2,997.639. The costs associated with screening all patients with dementia and only treating those with SDH and tumours amounted to £4,243.857. Synthesis of costs and benefitsSynthesis of costs and benefitsSynthesis of costs and benefitsSynthesis of costs and benefits A strategy of screening all patients with dementia and only treating those with SDH had a cost-utility of £14,171/QALY. A strategy of screening all patients with dementia and only treating those with SDH and NPH had a cost-utility of £17,238/QALY. A strategy of screening all patients with dementia and only treating those with SDH and tumours had a cost-utility of £19,983/QALY. Conclusions, commentary and implicationsConclusions, commentary and implicationsConclusions, commentary and implicationsConclusions, commentary and implications Author's ConclusionsAuthor's ConclusionsAuthor's ConclusionsAuthor's Conclusions CT scanning appears cost-effective in dementia patient aged under 65 years. However, it should be undertaken selectively in more elderly patients. Surgical treatment of NPH may reduce quality-adjusted survival and should only be undertaken within clinical trials. CRD commentaryCRD commentaryCRD commentaryCRD commentary Selection of comparators: The rationale for the choice of the comparator (do nothing) was clear and enabled the effects of the intervention to be assessed. You should consider if this, or other comparators, would be relevant to your own setting.

49

Validity of estimate of measure of benefit: The principal input parameters for the model were derived from a well-conducted and systematic review. Quality of life estimates were specifically derived for demented patients. The results do not seem to have been presented selectively. The authors acknowledged, however, that the model did not take into account the harm done to elderly patients who have to undergo a CT scan as a result of false positive results and any subsequent tests or treatment. Validity of estimate of costs: The authors limited their analysis to direct costs. Opportunity costs of CT scans were not considered. Treatment costs for those patients for whom treatment is unsuccessful were not included. The cost estimates are likely to be specific to the Scottish setting and included only direct costs, although as the authors point out, indirect costs are relevant to the population studied and a societal perspective would therefore be informative in any future studies. Other issues: The robustness of the results were tested using sensitivity analyses. In the absence of CT scanning, 50% of treatable structural lesions are likely to be detected in normal practice, thereby decreasing the cost-utility of CT scanning. The results were not compared with similar studies although comparisons at a data-level were undertaken. The generalisability of the results to other settings or countries was not specifically discussed. The authors did not consider other screening technologies such as magnetic resonance imaging, single photon emission computed tomography, or positron emission tomography. Implications of the studyImplications of the studyImplications of the studyImplications of the study The authors recommend that CT scanning should be limited to those patients for whom it is clearly indicated by the signs and symptoms of the illness. A suggested strategy is CT scan demented patients only when: (a) aged under 65, OR (b) aged 65 and over AND onset < 1 year OR atypical presentation OR rapid unexplained deterioration OR unexplained focal neurological signs or symptoms OR history of recent (before onset) head injury OR urinary incontinence or gait ataxia early in the illness. Subject index termsSubject index termsSubject index termsSubject index terms Subject indexing assigned by NLM: Dementia [radiography]; Mass-Screening [economics]; Tomography,-X-Ray-Computed [economics]; Age-Factors; Aged; Aged,-80-and-over; Cost-Benefit-Analysis; Dementia [epidemiology]; Dementia [etiology]; Dementia [surgery]; Human; Incidence; Mass-Screening [methods]; Middle-Age; Models,-Econometric; Patient-Selection; Prevalence; Quality-Adjusted-Life-Years; Scotland [epidemiology]; State-Medicine [economics]; Survival-Analysis; Tomography,-X-Ray-Computed [standards]; Treatment-Outcome] Address for correspondenceAddress for correspondenceAddress for correspondenceAddress for correspondence R G Foster, Scottish Office Home and Health Department, St Andrew's House, Edinburgh, UK. Accession number and database entry dateAccession number and database entry dateAccession number and database entry dateAccession number and database entry date 19998259 31121999 Copyright: University of YorkCopyright: University of YorkCopyright: University of YorkCopyright: University of York, 1999., 1999., 1999., 1999.

50

CODECS example abstractCODECS example abstractCODECS example abstractCODECS example abstract AuthorsAuthorsAuthorsAuthors DETOURNAY B, PLANE A, VOCHELLE N and FAGNANI F. TitleTitleTitleTitle Cost effectiveness of a low-molecular-weight-heparin in prolonged prophylaxis against deep vein thrombosis after total hip replacement. PeriodicalPeriodicalPeriodicalPeriodical Pharmacoeconomics Publication date, Publication date, Publication date, Publication date, Volume, number, pagesVolume, number, pagesVolume, number, pagesVolume, number, pages 1998 ; 13 (1 part 1), p. 81-89 Publication languagePublication languagePublication languagePublication language English SUBJECT OF STUDYSUBJECT OF STUDYSUBJECT OF STUDYSUBJECT OF STUDY Health technologyHealth technologyHealth technologyHealth technology Prolonged used of Low-Molecular-Weight-Heparin (LMWH), or enoxaparin, following hip replacement surgery for the prophylaxis of deep vein thrombosis (DVT) and pulmonary embolism (PE). DiDiDiDiseaseseaseseasesease Cardiovascular diseases. Disease (Complement of information)Disease (Complement of information)Disease (Complement of information)Disease (Complement of information) Deep vein thrombosis and pulmonary embolism Target populationTarget populationTarget populationTarget population Patients having undergone hip replacement surgery. Type of intervention Type of intervention Type of intervention Type of intervention Secondary prevention. Hypothesis / Study questionHypothesis / Study questionHypothesis / Study questionHypothesis / Study question The purpose of the study is to compare the economic implications of prolonged prophylaxis with LMWH or enoxaparin, in patients having undergone total hip replacement surgery after discharge. The study examined medication prescribed during hospitalisation (placebo group) versus those subject to a three-week prolongation of treatment after discharge (enoxaparin group). This medical approach to the study, using prolonged prophylaxis by subcutaneous injection, is justified by the authors firstly by the fact that two clinical studies have demonstrated the effectiveness of prolonged rather than short-term anti-coagulant therapy. Secondly, the effectiveness of long-term, oral anti-coagulant therapy following total hip replacement surgery has never been evaluated in clinical trials using a controlled placebo. This study was conducted from a societal point of view taking into account only direct medical costs. KEY ELEMENTS OF THE STUDYKEY ELEMENTS OF THE STUDYKEY ELEMENTS OF THE STUDYKEY ELEMENTS OF THE STUDY Economic study type Economic study type Economic study type Economic study type Cost-effectiveness analysis. Study populationStudy populationStudy populationStudy population Patients having undergone total hip replacement. Setting Setting Setting Setting Ambulatory care. The study was carried out in France.

51

Dates to which data relate Dates to which data relate Dates to which data relate Dates to which data relate Efficiency and resource use estimates were obtained from literature reviews published in 1987, 1994, 1995 and 1996 and were validated by expert opinions. Except for the cost of a hospital day, which was for 1996, published prices indicated are 1997 values and were obtained from the French National Diagnostic-Related Group Cost Study. Sources of effectiveness and tolerance data Sources of effectiveness and tolerance data Sources of effectiveness and tolerance data Sources of effectiveness and tolerance data A review of the literature. Modelling Modelling Modelling Modelling The authors developed a decision tree with the purpose of combining available clinical resource and cost data in France along with the natural history of the disease. The incidence rate of a DVT/PE event in the LMWH and short-term prophylaxis arm of the decision tree are derived from two clinical trials. The other probabilities used in the tree correspond to the average values of the data available from the four studies. In view of the uncertainties associated with most of the data the authors used a Monte Carlo simulation. Link between effectiveness and cost data Link between effectiveness and cost data Link between effectiveness and cost data Link between effectiveness and cost data Not stated. CHARACTERISTICS OF THE EVALUATION OF EFFECTIVENESS CHARACTERISTICS OF THE EVALUATION OF EFFECTIVENESS CHARACTERISTICS OF THE EVALUATION OF EFFECTIVENESS CHARACTERISTICS OF THE EVALUATION OF EFFECTIVENESS AND TOLERANCE RESULTSAND TOLERANCE RESULTSAND TOLERANCE RESULTSAND TOLERANCE RESULTS Literature reviewLiterature reviewLiterature reviewLiterature review Judgement criteria used to interrogate the databasesJudgement criteria used to interrogate the databasesJudgement criteria used to interrogate the databasesJudgement criteria used to interrogate the databases The effectiveness data used in the two clinical trials are based on the incidence of DVT and/or PE events of both groups (enoxaparin and placebo) in the post operative period. Study designs and other criteria for inclusion in the review Study designs and other criteria for inclusion in the review Study designs and other criteria for inclusion in the review Study designs and other criteria for inclusion in the review Not stated. Sources searched to identify primary studies Sources searched to identify primary studies Sources searched to identify primary studies Sources searched to identify primary studies Not stated. Criteria used to ensure the validity of primary studies Criteria used to ensure the validity of primary studies Criteria used to ensure the validity of primary studies Criteria used to ensure the validity of primary studies Not stated. Methods used to judge relevance andMethods used to judge relevance andMethods used to judge relevance andMethods used to judge relevance and the validity, and for extracting data the validity, and for extracting data the validity, and for extracting data the validity, and for extracting data Not stated. Number and type of primary studies included Number and type of primary studies included Number and type of primary studies included Number and type of primary studies included Two double blind, randomised clinical studies were included. Four additional studies were used to better estimate the mean probability associated with the different outcomes of the decision tree. The characteristics of these studies were not described. Method of combination of primary studies Method of combination of primary studies Method of combination of primary studies Method of combination of primary studies Decision analysis method. Investigation of differences between primary studies Investigation of differences between primary studies Investigation of differences between primary studies Investigation of differences between primary studies The two clinical studies differed in: the duration of LMWH treatment prior to randomisation (14 days v. 7 days); the evidence showing there was no DVT/PE at the moment of discharge (venography v. clinical signs or symptoms); and on the duration of follow-up (35 days v. 30 days). ResultResultResultResults of the literature review s of the literature review s of the literature review s of the literature review Single-centre, double-blind, placebo-controlled, randomised trials were included. In both trials prophylaxis with subcutaneous enoxaparin 40mg/day was administered during hospitalisation. Upon discharge, eligible patients were randomly assigned to receive enoxaparin or a matching placebo during three weeks. The study also required that patients have no previous history (either symptomatic or with venography) of a DVT/PE event. In one study, 35-days after surgery, the enoxaparin group experienced a 12.2% reduction [95% confidence interval (CI) 2.4 to 22%; p= 0.018] in the incidence of thromboembolitic related events. The other study, 30-days after surgery, indicated an incidence for thromboembolism of 18% and 39% for the enoxaparin and placebo group

52

respectively (p <0.001). The incidence of toxicities and other major adverse events were similar in the placebo and enoxaparin-treated groups. ECONOMIC ANALYSIS: METHODOLOGYECONOMIC ANALYSIS: METHODOLOGYECONOMIC ANALYSIS: METHODOLOGYECONOMIC ANALYSIS: METHODOLOGY Judgement criteria for the economic analysis Judgement criteria for the economic analysis Judgement criteria for the economic analysis Judgement criteria for the economic analysis The judgement criteria for this economic analysis are the number of DVT/PE events that would be detected by routine venography, DVT/PE events that are diagnosed and treated in current practice and preventable deaths. A decision tree was used to estimate the number of DVT/PE diagnosed in current practice in France and life years gained. The number of life years gained was obtained from the transformation of the number of deaths avoided, which itself, is obtained by the extrapolation of the results of clinical effectiveness with a natural history model of post operative DVT/PE. In the absence of long term follow-up data for patients having undergone surgery for hip replacement, the authors estimate each death in the model will imply a loss of 12.5 life-years (optimistic hypothesis) or as a conservative more pessimistic alternative, a loss of 6.25 life-years. National data were taken from hospitals’ diagnosis related group (DRG) and French life expectancy tables. Direct costs Direct costs Direct costs Direct costs Not all costs and quantities were reported separately. Direct costs identified are: treatment cost for DVT/PE, rehospitalisation costs after surgery in the event of DVT/PE, and prophylaxis treatment with enoxaparin. Quantities consumed were determined according to four experts opinion and from the characteristics of the product: duplex scans in 100% of the patients and contrast venography in 25%, at least three hospitalisation days in a medical ward to diagnose a pulmonary embolism. A routine 40mg injection of enoxaparin administered at the patient’s home by a visiting nurse and one platelet count per week. Unit costs were issued by the French National Sickness Fund 1997 (for the diagnosis of DVT), a study at Assisstance Publique de Paris 1995 (for the diagnosis of PE), the costs issued from the French National Diagnostic-Related Group Cost Study 1996 (for the cost of treating DVT/PE), the French health authorities and the VIDAL index for drug costs 1997 (for the price of enoxaparin). The costs of laboratory tests and injections administered were based on current rates, as determined by the French National Sickness Fund (1997). Discounting was unnecessary due to the short duration of therapy. Median and differential costs were reported. The costs of adverse effects were excluded from the analysis because the two main trials reported a low incidence and mild effects. In view of the uncertainties associated with estimating cost data, a Monte Carlo simulation technique was used to create a distribution. Indirect costs Indirect costs Indirect costs Indirect costs Not taken into account. CurrencyCurrencyCurrencyCurrency French francs (FF). ECONOMIC ANALYSIS: RESULTSECONOMIC ANALYSIS: RESULTSECONOMIC ANALYSIS: RESULTSECONOMIC ANALYSIS: RESULTS Judgement criteria measurementJudgement criteria measurementJudgement criteria measurementJudgement criteria measurement Based on a hypothetical cohort of 100 000 patients, the mean distribution (between the 5th and 95th percentile) of the number of deaths avoided due to prolonging the prophylaxis treatment with LMWH three weeks after hospital discharge was 601 to 783. Similarly, treatment could prevent between 16 012 to 21 122 thromboembolic events. Of the latter, however, it is argued that only 3623 to 4812 would have been diagnosed and treated in practice. The incremental benefits are also presented using the value for the 5th and 95th percentile. Cost results Cost results Cost results Cost results The cost differential of prolonging prophylaxis with LMWH was 1118 FF1996 to 1300 FF1996 per patient (net median cost). The values are also presented at the 5th and 95th percentiles. Synthesis of costs and benefits Synthesis of costs and benefits Synthesis of costs and benefits Synthesis of costs and benefits The median cost-effectiveness ratio of each potentially avoidable death varies between 139 480 FF1996 and 216 188 FF1996; between 11 158 FF1996 and 34 590 FF1996 per year of life saved; between 5 286 FF1996 and 8 031 FF1996 per DVT/PE event potentially avoided (in the clinical setting) detectable through routine contrasted venography and between 23 532 FF1996 and 35 268 FF1996 per DVT/PE event avoided in standard practice. The average annual cost of life years gained varies between 11 158 FF1996 and 17 295 FF1996 (under the hypothesis that a death corresponds to a loss of 12.5 life years) and between 22 317 FF1996 and 34 590 FF1996 (for a death associated with a loss of 6.25 life years). If the 5th percentile (which corresponds to the less favourable hypothesis) and the 95th percentile (the most

53

favourable hypothesis) are considered, the results give a value for the cost per year of life saved in the interval [2 421 Frs1996; 141 984 Frs1996] (under the hypothesis that a death corresponds to a loss of 12.5 or 6.25 life years). Verification of the robustness of the resultsVerification of the robustness of the resultsVerification of the robustness of the resultsVerification of the robustness of the results In addition to a multi-parameter sensitivity analysis on all the variables of the decision tree by the Monte Carlo technique, the authors performed a single-criterion sensitivity analysis so as to identify the parameters having the most important impact on the cost-effectiveness ratio. Cost per year of life saved was sensitive to the cost of enoxaparin, the incidence of PE for those patients waiting a non-symptomatic DVT, mortality rate of non-symptomatic PE among those who survived more than one hour after the start of the PE, and the rate of actual DVTs (potentially detectable by a mandatory venography) in each arm of the tree. AUTHORS ' CONCLUSIONSAUTHORS ' CONCLUSIONSAUTHORS ' CONCLUSIONSAUTHORS ' CONCLUSIONS Despite many uncertainties over cost and effectiveness data, the authors conclude that prolonged prophylaxis with LMWH (enoxaparin) during hospitalisation and for three weeks after discharge in patients having undergone hip replacement surgery is not only more effective, but is also cost-effective when compared to the short term LMWH (enoxaparin) prophylaxis treatment given only during hospitalisation. Further clinical and economic studies are needed to demonstrate the potential interest of other oral anticoagulant therapies in this indication. METHODOLOGICAL DISCUSSIONMETHODOLOGICAL DISCUSSIONMETHODOLOGICAL DISCUSSIONMETHODOLOGICAL DISCUSSION According to the CODECS expertAccording to the CODECS expertAccording to the CODECS expertAccording to the CODECS expert CCCChoice of comparator: In the absence of a valid protocol, the choice of a placebo seems justifiable. Validity of result estimates: No systematic procedure of interrogation of the databases seems to have been undertaken. Nosological characteristics of the diagnostic tests are not stated; the absence of the false positives should have been justified. The intermediate judgement criteria of the clinical trials (number of venous thromboembolism in each arm) were transformed into final effectiveness indicators (number of deaths avoided; number of life-years gained; number of diagnosed and treated venous thromboembolic events). The choice of a Gaussian distribution function of the results in the sensitivity analysis is not justified by the authors. Validity of cost estimates: In the absence of an observational study, the quantities of resources consumed were estimated by experts, costs per unit seem to be somewhat homogeneous despite their coming from different years of data entry. The choice of a different valuation unit for hospitalisation in case of a pulmonary embolism is not justified. All hospital costs are complete costs; the analysis of the variable costs of GHM is missing. Cost evaluation is valid only for France. Other issues: the medical and economic study is well structured. A one-way sensitivity analysis and the Monte Carlo simulation technique confirm the robustness of the results. Other publications of related interestOther publications of related interestOther publications of related interestOther publications of related interest Anderson D, O'Brien B, Levine M, Roberts R, Wells P, Hirsh J. Efficacy and cost of low-molecular-weight compared with standard heparin for the prevention of deep vein thrombosis after total hip arthroplasty. Ann Int Med 1993;199:1105-12. CODECS EXPERT COMMENTARY (Implications of the study)CODECS EXPERT COMMENTARY (Implications of the study)CODECS EXPERT COMMENTARY (Implications of the study)CODECS EXPERT COMMENTARY (Implications of the study) A randomised, controlled trial is necessary to confirm these results. Further comparisons with other treatments must be considered. Sources of Financial Support Sources of Financial Support Sources of Financial Support Sources of Financial Support Rhône Poulenc Rorer Address for correspondenceAddress for correspondenceAddress for correspondenceAddress for correspondence Dr Bruno Detournay CEMKA 43 Bld Maréchal Joffre, 92 340 Bourg la Reine, France.

54

BIBLIOGRAPHYBIBLIOGRAPHYBIBLIOGRAPHYBIBLIOGRAPHY The following readings constituted useful sources for the preparation of the Guidance for writing critical abstracts for the NHS Economic Evaluation Database: 1. NHS Centre for Reviews and Dissemination. Improving Access to Cost-effectiveness Information for Health

care Decision-Making: The NHS Economic Evaluation Database. CRD Report 6 (2nd Edition). University of York: York Publishing Services Ltd, 2001.

2. Methodological guide used for the elaboration of the structured abstracts. CODECS Project. Collège des Économistes de la Santé, 2002.

3. Methodological guide on the economic evaluation of health care interventions. Collège des Économistes de la Santé, Paris, July 2003.

4. Office of Technology Assessment. Assessing the efficacy and safety of medical technologies. Publication No. OTA-H-75. Washington: Government Printing Office, 1978

5. Manual for Selecting reviews and Writing Abstracts for the Database of Abstracts of Reviews of Effectiveness (DARE): NHS CRD, October 2000.

6. Assessment CCOfHT. Guidelines for economic evaluation of pharmaceuticals: Canada. Ottawa: Canadian Coordinating Office for Health Technology Assessment, 1994. Available from:

7. Birch S, Donaldson C. Applications of cost-benefit analysis to health care: departures from welfare economic theory. Journal of Health Economics 1987;6:211-25.

8. Birch S, Gafni A. Cost effectiveness/utility analyses. Do current decision rules lead us to where we want to be? Journal of Health Economics 1992;11:279-96.

9. Birch S, Gafni A. Changing the problem to fit the solution: Johannesson and Weinstein's (mis)application of economics to real world problems. Journal of Health Economics 1993;12:469-76.

10. Birch S, Gafni A. Cost-effectiveness ratios: in a league of their own. Health Policy 1994;28:133-41. 11. Briggs A, Sculpher M, Buxton M. Uncertainty in the economic evaluation of health care technologies:

the role of sensitivity analysis. Health Economics 1994;3:95-104. 12. Chalmers I, Haynes B. Reporting, updating, and correcting systematic reviews of the effects of health

care. British Medical Journal 1994;309:862-5. 13. Chalmers TCCPSHSSH. Bias in treatment assignment in controlled clinical trials. The New England

Journal of Medicine 1983:1358-61. 14. Department of Clinical Epidemiology and Biostatistics, Centre MUHS. How to read clinical journals:

VII. To understand an economic evaluation (part A). Canadian Medical Association Journal 1984;130:1428-33.

15. Drummond M. Economic analysis alongside controlled trials: an introduction for clinical researchers.: Department of Health, 1994.

16. Drummond M, Torrance GW, Mason J. Cost-effectiveness league tables: more harm than good? Soc. Sci. Med. 1993;37:33-40.

17. Drummond M, Brandt A, Luce B, et al. Standardizing methodologies for economic evaluation in health care. International Journal of Technology Assessment in Health care 1993;9:26-36.

18. Drummond M, O'Brien B. Clinical importance, statistical significance and the assessment of economic and quality-of-life outcomes. Health Economics 1993;2:205-12.

19. Drummond MF. Principles of economic appraisal in health care. Oxford: Oxford University Press, 1980.

20. Drummond MF, Davies L. Economic analysis alongside clinical trials. International. Journal of Technology Assessment in Health Care 1991;7:561-73.

21. Drummond MF, Jefferson TD. Guidelines for authors and peer reviewers of economic submissions to the BMJ. British Medical Journal 1996;313:275-83.

22. Drummond MF, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press, 1988.

55

23. Egger M, Davey Smith G, Song F, et al. Making sense of meta-analysis. Pharmacoepidemiology and Drug Safety 1993;2:65-72.

24. Elixhauser A, Luce BR, Taylor WR, et al. Health care CBA/CEA: an update on the growth and composition of the literature. Medical Care 1993;31:JS1-JS11.

25. Eysenk HJ. Meta-analysis and its problems. British Medical Journal 1994;309:789-92. 26. Gafni A, Birch S. Searching for a common currency: critical appraisal of the scientific basis underlying

European harmonization of the measurement of Health Related Quality of Life (EuroQol). Health Policy 1993;23:219-28.

27. Glick H, Kinosian B, Schulman K. Decision analytic modelling: some uses in the evaluation of new pharmaceuticals. Drug Information Journal 1994;28:691-707.

28. Goodman C. Literature searching and evidence interpretation for assessing health care practices. Stockholm: SBU - The Swedish Council on Technology Assessment in Health Care, 1993.

29. Gudex C, Kind P, van Dalen H, et al. Comparing scaling methods for health state valuations - Rosser revisited. York: Centre for Health Economics (Discussion Paper 107), 1993.

30. Guyatt GH, Sackett DL, Cook DJ. Users' guide to the medical literature: how to use an article about therapy and prevention. A. Are the results of the study valid? Journal of the American Medical Association 1994;270:2598-601.

31. Guyatt GH, Sackett DL, Cook DJ. Users' guide to the medical literature: how to use an article about therapy and prevention. What were the results and will they help me in caring for my patients? Journal of the American Medical Association 1994;271:59-63.

32. Haines A, Jones R. Implementing findings of research. British Medical Journal 1994;308:1488-92. 33. Health Do. DH Register of cost-effectiveness studies. London: Department of Health, Economics and

Operational Research Division, 1994. 34. Kaplan RM, Bush JW, Berry CC. Health status: types of validity and the index of well-being. Health

Services Research 1976:478-507. 35. L'Abbe KA, Detsky AS, O'Rourke K. Meta-analysis in clinical research. Annals of Internal Medicine

1987;107:224-33. 36. Laupacis A, Feeny D, Detsky AS, et al. How attractive does a new technology have to be to warrant

adoption and utilisation? Tentative guidelines for using clinical and economic evaluations. J Can. Med. Assoc. 1992;146:473-81.

37. Mason J, Drummond M. The DH register of cost-effectiveness studies: a review of study content and quality. York: Centre for Health Economics (Discussion Paper 128) 1995.

38. Mason J, Drummond M, Torrance J. Some guidelines on the use of cost effectiveness league tables. British Medical Journal 1996;306:570-2.

39. McGuire A, Henderson J, Mooney G. The economics of health care. London: Routledge and Kegan Paul, 1988.

40. Mooney GH. Economics, medicine and health care. Brighton: Wheatsheaf Books, 1984. 41. Mulrow CD. Rational for systematic reviews. British Medical Journal 1994;309:597-9. 42. O'Brien B, Drummond MF, Labelle RJ, et al. In search of power and significance: issues in the design

and analysis of stochastic cost-effectiveness studies in health care. Medical Care 1994;32:150-63. 43. O'Brien BJ, Drummond MF. Statistical versus quantitative significance in the socioeconomic

evaluation of medicines. PharmacoEconomics 1994;5:389-98. 44. Oxman AD. Checklist for review articles. British Medical Journal 1994;309:648-51. 45. Oxman AD, Guyatt GH. Guidelines for reading literature reviews. Canadian Medical Association

Journal 1988;138:697-703. 46. Oxman AD, Guyatt GH. Validation of an index of the quality of review articles. J. Clin. Epidemiology

1991;44:1271-8. 47. Patrick DL, Erickson P. Health status and health policy: allocating resources to health care. New York:

Oxford University Press, 1993. 48. Petitti DB. Meta-analysis, decision analysis and cost-effectiveness analysis: methods and quantitative

synthesis in medicine. New York: Oxford University Press, 1994. 49. Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia: Lippincott-Raven Publishers, 1998. 50. Rutten F, Drummond M. Making decisions about health technologies: a cost-effectiveness perspective.

York: Centre for Health Economics (Occasional Paper Series), 1994.

56

51. Sackett D.L, Haynes RB, Guyatt GH, et al. Clinical epidemiology. Boston: Little, Brown and Company, 1991.

52. Sackett DL, Torrance GW. The utility of different health states as perceived by the general public. J. Chronic Dis. 1978;31:697-704.

53. Sheldon TA. Discounting in health care decision-making: time for a change? Journal of Public Health Medicine 1992;14:250-6.

54. Sheldon TA. Problems of using modelling in the economic evaluation of health care. Health Economics 1996;5:1-11.

55. Sheldon TA, Song F, Davey Smith G. Critical appraisal of the medical literature: how to assess whether health-care interventions do more good than harm. In: Drummond MF, Maynard A, editors. Purchasing and providing cost-effective health care. Edinburgh: Churchill Livingstone, 1993.

56. Sonnenberg FA, Roberts MS, Tsevat J, et al. Toward a peer review process for medical decision analysis models. Medical Care 1994;32:JS52-JS64.

57. Thompson SG, Pocock SJ. Can meta-analyses be trusted? The Lancet 1991;338:1127-30. 58. Torrance GW, Boyle MH, Horwood SP. Application of multi-attribute utility theory to measure

social preferences for health states. Operations Research 1982;30:1043-69. 59. Udvarhelyi IS, Colditz GA, Rai A, et al. Cost-effectiveness and cost-benefit analyses in the medical

literature: Are the methods being used correctly? Annals of Internal Medicine 1992;116:234-44. 60. Van Hout BA, Al MJ, Gordon GS, et al. Costs, effects and C/E-ratio's alongside a clinical trial. Health

Economics 1994;3:309-19.

EURONHEED Methodological Guide V3 - unirioja.es · 2 ACKNOWLEDGEMENTS This is the third draft of the European Network of Health Economic Evaluation Databases (EURONHEED) methodological

Documents