A standard reporting format is likely to be helpful both to authors of submissions and to the secretariat of the appraisal group. It would act as a check list to ensure that essential information is not omitted, and it would help in organising the work involved in producing the submission and in appraising its contents. However, such a list cannot be exhaustive and cannot be a substitute for good practice in performing evaluations.
     
This annex sets out, as the basis of consultation, some initial suggestions for the possible contents of industry submissions. A definitive statement of NICE's information requirements will be published following the consultation on this document. It will be updated by NICE as needed (see main paper para 20).
     
3 It is proposed that submissions should present information under the following headings:
 
i
 summary and introductory remarks;
 
ii
 methods and assumptions used to produce the results;
 
iii
 results of cost-effectiveness analysis;
 
iv
 aggregate cost impacts;
 
v
 equity issues;
 
vi
 conclusions
     
Summary and introductory remarks
     
4 The sponsors should provide an executive summary in the form of a structured abstract setting out the "bottom line" results. These should be interpreted in the context of alternative therapies having regard to the limitations of the analysis, and appropriate recommendations for NHS use may be offered.
     
5 This summary should be followed by an introduction to the main submission which should describe the disease, its epidemiology and current practice and its significance in terms of ill health and resource costs imposed on society.
     
6 It should describe the intervention being assessed and the indications for which a recommendation is sought. For pharmaceuticals, for example, the description should include therapeutic classification, brand and generic name, dosage form and route and all indications for which market authorization is being sought or has been obtained. Specific patient subgroups to which the analysis is being applied should be defined.
     
7 Sponsors should disclose their relationship with the author of the study to make it clear how much of the submission the author agrees with and to what extent the authors could be regarded as independent of the sponsor.
     
Methods
     
8 The submission should specify whether the study design was prospective, retrospective, modelled or a mixture. The analytical framework and reasons for adopting it should be set out. Departures from incremental cost-utility analysis should be justified.
   
9 The choice of comparator therapy should be explained and any reasons for not following guideline recommendations. The method adopted for comparing clinical efficacy should be given, eg head-to-head randomised control trial (RCT), meta-analysis or other specified methods.
   
10 The choice of time horizon should also be justified, eg by showing that significant costs or benefits beyond the chosen horizon are unlikely.
   
11 All related studies of the therapy and its comparators should be reviewed and fully referenced, particularly all RCTs. If the results of any of these are excluded from the analysis, the reasons for this should be fully explained.
   
12 The approach to outcome measurement should be documented in some detail. All outcome variables considered should be mentioned. Reasons should be given for rejecting any outcomes considered and for ignoring any outcomes actually measured. For all outcomes included, the method of measurement should be given. The methods, evidence base and assumptions used in deriving clinical efficacy results should be described. Other clinical and socio-demographic data collected in trials should also be listed.
   
13 Where quality of life is a significant factor in the assessment, the methods used to assess quality of life should be described and justified. Thus where an attempt has been made to measure quality of life directly, the focus will be on the performance of the measuring instrument. Where quality of life is modelled indirectly from clinical end points, it will be more important to demonstrate the evidence underpinning the modelling.
   
14 The approach to estimating costs should be specified, and costs should be suitably categorised in terms of who bears them. The sources of information should be given for both resource use and price data, and assumptions should be stated explicitly.
   
15 The major sources of uncertainty should be pointed out and classified according to whether uncertainty arises from sampling error or from the range of plausible assumptions, hence which are handled using confidence intervals and which subjected to sensitivity analysis or other methods.
   
17 Any major assumptions and limitations contained in the analysis should be pointed out.
   
Results - Outcomes
   
18 The analysis should be reported step by step proceeding from the base data, through any systematic review, meta-analysis and modelling, to final estimates of efficacy and effectiveness relative to comparator therapy. The structure for reporting should therefore be:
 
i
 results of individual trials or observational studies;
 
ii
 results of reviews and meta-analysis;
 
iii
 modelled results (showing how derived);
 
iv
 overall conclusions on efficacy and effectiveness, including estimated increase in health gain relative to the comparator.
     
  The level of detail required is the minimum necessary to enable replication of the results from base data. Additional data may, however, be requested.
     
Results - costs related to health treatment
     
19 Again, sufficient information should be reported to enable all cost estimates to be traced to base data. The following items should be separately identified:
 
i
 costs to the NHS;
 
ii
 health-related costs to the government-funded Personal Social Services, where relevant;
 
iii
 costs avoided as a result of the therapy, classified to beneficiary;
 
iv
 capital and overhead costs;
 
v
 costs arising in different years (with and without discounting).
     
  A wider range of costs and benefits may be relevant in particular cases. The secretariat would be able to advise in individual cases how costs and benefits arising outside the NHS should be treated.
     
Results - sensitivity and sub-groups
   
20 Sources of uncertainty not deriving from sampling error should be presented in a sensitivity analysis. A decision tree should be considered where there are several important sources of uncertainty.
 
21 Cost-effectiveness results for each identified sub-group should be set out, identifying any economically significant differences.
 
Aggregate cost impacts
 
22 The submission should estimate the number of patients for whom the therapy is likely to be clinically cost effective, and derive from this an estimate of the total cost to the NHS of adopting it. A time profile of costs should be given. A sensitivity analysis should be reported showing how these estimates would be affected by extending treatment to all patients with indications for which the therapy has, or will have, regulatory approval.
   
Equity issues
   
23 The socio-demographic characteristics of the beneficiaries should be noted where they are a highly skewed sample of the UK population. It is important to reveal instances where a treatment could have a significant impact on health inequalities by gender, ethnic group or socio-economic category or on a group particularly disadvantaged due to the nature of their disease.
   
Other unquantifiable factors
   
24 Any other unquantifiable factors, ie factors which cannot be reflected in quantified assessment of health gain but which in the view of the sponsoring company are relevant to the final assessment of clinical cost-effectiveness, should be briefly set out, together with any relevant detail which would help the appraisal group judge their significance. The appraisal group might in due course issue further guidance on the factors which are likely to be relevant.
   
Conclusions
   
25 The final part of the submission should draw together the main results, assumptions, limitations and uncertainties, in the context of relevant alternative therapies, and present a reasoned conclusion based on the estimated increase in health gain and any other relevant factors.
   

Comments are invited on:

  • the proposed format for submissions;
  • whether the data requested would be sufficient to enable the appraisal group to discharge its remit, or if not what other aspects should be included.