The ScHARR Guide to
Writing Consultancy Proposals
Objective
This document aims to help staff to write and cost consultancy proposals as efficiently and comprehensively as possible.
Before writing a proposal always check that the proposal has met the consultancy criteria outlined in n:/general/consultancy/IKT Form
Section A: Consultancy Proposal Template (p 2-15)
This basic template will give you a framework for writing a consultancy proposal. Within each section of the template there are notes in green text providing advice and highlighting issues to consider when completing the section. For more detailed discussion of key issues summarised in the template you can refer to section B (using the links provided)
The template includes advice on costing to ensure that all relevant costs are included in the proposal.
There is also a separate template for estimating number of days for consultancy systematic review work, located in n:/general/consultancy/systematic review template to assist the systematic reviewers.
This template is only a guide and is based on a short proposal for development of a cost effectiveness model developed in Excel. The amount of detail required in a proposal will vary, depending on the client, the topic area and scope of the project. If previous proposals have been done for the client in question, these should also be used as a guide.
Be explicit wherever possible – for instance, do not say you will perform a systematic literature search for costs and QoL data if you will be using berry-picking or other techniques.
It is important to state exactly what work will be undertaken as part of this proposal and, where ever possible, make it clear what work is beyond the scope of the project.
Section B: Detailed Advice on Writing Consultancy Proposals (p 16 -27)
A more detailed guide on specific issues relating to writing and costing proposals (p15 onwards of this document)
Steps in writing a consultancy proposal
TEMPLATE
School ofHealth
And
Related
Research.
Title
Date
Name of ScHARR personnel
Name of ScHARR personnel
Name of ScHARR personnel
1. INTRODUCTION
This proposal has been produced following telephone conversations between (name) at the University of Sheffield and (name) at (company name) on (date)
The proposal includes:
o An outline of the objectives and proposed modelling techniques
o The proposed work programme and deliverables
o Potential timescales and costs
The proposal is a draft and is subject to refinement if required.
Objectives
The primary objectives of the project are :
· to explore the cost effectiveness of x compared with y for patients with xxxxxxx
· to provide estimates of the budget impact of introducing x as a first line therapy in England and Wales
· to provide a project report outlining detailed methodologies and results.
The project will not include xxxxxxxxxxxxxxxxxxxxxxxxxx .(eg the writing of regulatory submission documentation)
HINTS
The scope needs to be as focussed as possible and, where possible, to specify which issues are to be excluded. If you are unfamiliar with the subject area some preliminary scoping work might be worthwhile, as might including contingency time for unexpected issues that may arise. For further information, see B.1.1 Defining the objectives.
The intended therapeutic positioning of an intervention and the strength of data supporting it can make a big difference to the type and volume of work needed. For further information, see B.1.2 Drivers of the type and volume of work required.
2. BACKGROUND
Prevalence
Symptoms
Prognosis
Current Treatment Options
New drug
3. EVIDENCE
Review of effectiveness and safety of new drug and comparators
Trial evidence
In some cases cost effectiveness models will be based on the results of one key trial. This is relatively straight forward. See typical text below
Evidence for the effectiveness of x in patients with xxxxxxxxxxxx will be based on results of the phase III study. The study will provide data on time to tumour progression, overall survival, progression-free survival and objective response. It is anticipated that the on-going phase III study will also provide evidence on quality of life based on EQ-5D.
Data from the trial will be analysed and used together with published evidence to populate the model. Clinical expert opinion may be required to verify both the clinical pathway and the magnitude of parameter values used to populate the model.
Literature reviews and evidence synthesis
Additional text is required where there is a wider evidence base and a full clinical review and/or Bayesian synthesis are required. Example of additional text for Mixed Treatment Comparisons is given below.
Not all RCTs included all treatments of interest so there may be no direct evidence comparing all treatment against each other. A synthesis of the available evidence will be conducted using a Bayesian hierarchical model using the statistical software package WinBUGS, and treatment effects will be estimated using the resulting mixed treatment comparisons. Mixed treatment comparison (MTC) is an extension of standard meta-analysis to allow multiple treatments to be compared simultaneously. It involves constructing a network of evidence and using both direct and indirect information to estimate treatment effects. Bayesian inference is often used in MTC, as it allows parameter uncertainty to be modelled explicitly. The benefits of Bayesian MTC are that it provides a coherent theoretical framework for assessing the evidence base as a whole, it allows for more precise estimates of treatment through ‘borrowing strength’ across studies, and it illustrates the degree to which direct and indirect evidence is consistent. (Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 2004 Oct 30;23(20):3105-24).
Estimates of resource use and costs are based on the likelihood of there being a maximum of X RCTs eligible for inclusion in the review. This is based on the number of hits in the scoping searches. In the event that the number of RCTs eligible for inclusion exceeds this number, additional resources will need to be assigned to the project and costed accordingly. This will be discussed and agreed with the client, if required.
NB It is anticipated that clinical input will be required, to confirm clinical data and to assist in identifying exact clinical pathway. External advisors to be identified and paid by client.
HINTS
Payment of external advisors - new financial requirements within the University make it more complicated for ScHARR to pay external advisors – a contract is required and relevant forms need to be completed by the external advisor. It is therefore preferable to request that the client handles payments to advisors directly. See Vanessa Wright for further details.
If the efficacy and side effect data isn’t available from a head-to-head study, then Mixed Treatment Comparisons or Indirect Comparisons may be needed. This add considerably to the volume of work. For further information, see B.2.5 Bayesian Evidence Synthesis .
Review of Quality of Life (QoL)
A systematic search of the literature will be conducted to identify QOL data and utility data for use in the economic model.
Review of Economic Evidence
A systematic search will be undertaken in order to identify published economic evaluations or cost of illness studies. Additional searches will be undertaken to identify possible sources of cost parameters for the economic model including hospitalisation and costs of adverse events.
HINTS
Find out in detail what data, if any, will be provided by the sponsor, and the format that it is to be provided in. This will help prevent delays and ensure a realistic assessment is made of any work needed by ScHARR to obtain usable data. For further information, see B.2.1 Data - Needs from the Sponsor.
The scale of clinical review will vary considerably from project to project. The exact research question needs to identified as this will dictate the potential scale of the project
In addition, it is important to understand the evidence requirements of any regulatory bodies, such as NICE, where relevant, as this will influence the type of review undertaken e.g. full systematic reviewing. For further information, see B.2.4 Literature Review and Critical Appraisal.
If a clinical review is required suggested wording can be obtained from the Systematic reviewers. Systematic reviewers need to be involved in the proposal writing and are required to take ownership of the number of days estimated based on the research question A separate template for estimating number of days for consultancy systematic review work, located in N:\HEDS\General\Consultancy Proposals Cheats Guide\ to assist the systematic reviewers in allocating the correct number of days
Discuss time requirements for evidence reviews and synthesis. . Decide which parameters will require a comprehensive search and how much synthesis of the evidence is appropriate. For further information, see B.2.4 Literature Review and Critical Appraisal.
Scoping searches are recommended as the volume of papers likely to be identified by searches varies greatly, depending on the scope of the project and disease area so. For further information, see B.2.4 Literature Review and Critical Appraisal.
4. MODEL DEVELOPMENT
A cost effectiveness model will be developed to provide estimates of costs and quality adjusted life years (QALYS) of patients with xxxxxx taking x compared to patients on y. Results will be presented in terms of an incremental cost per QALY, a figure used by policy makers in a number of countries including the UK.
In addition the budget impact of the use of x in first line treatment of xxxxxx will be analysed and presented. The budget impact model will be developed initially to assess the impact for England and Wales, but with the flexibility to allow the budget impact in other regions/countries to be explored.
Model Structure
A mathematical model will be constructed using Microsoft Excel. At this stage it is anticipated that a Markov type cohort model may be used to simulate the clinical pathway of patients with xxxxxxxxxxx treated with xxxxxxxxxxx. A Markov model assumes that a patient is always in one of a finite number of health states. Probabilities are used to determine the movement between these states, and only designated transitions are allowed. Simple assumptions may be required to allow forecasting of the model beyond the end of the trial. Each state can be assigned a quality of life utility value and a cost. The model will compute the average length of time spent in each state and the accrued utility values and costs over the period of analysis.
The potential use of surrogate markers in predicting longer term outcomes in the model will be explored, if required.
Whichever modelling technique is selected, the model will be developed in an EXCEL spreadsheet format to allow transparency and aid future adaptations and modifications.
Time horizon
In economic modelling it is common practice to extrapolate beyond trial data to estimate the costs and benefits accrued over patients’ lifetimes. However, modelling beyond the time horizon of clinical trials is often contentious, particularly when it is not possible to validate the results generated by the model against published evidence. Because of the short duration of the clinical trials used as evidence for effectiveness, and the lack of supporting evidence from long term usage in general clinical practice, it is anticipated that the results will be presented for several time horizons.
Sensitivity Analysis
To forecast costs and QALYs for both x and the comparator it will be necessary to synthesise both primary and secondary data from a number of sources. A probabilistic model will be developed with each variable described as a distribution (i.e. normal, beta etc). By using Monte Carlo simulation, a full uncertainty analysis on the Incremental Cost Effectiveness Ratio (ICER) can be performed. The results of this will be presented on a cost effectiveness acceptability curve (CEAC) or cost effectiveness plane.
HINTS
Time required for model development will depend on existing level of experience within the disease area, whether the project requires development of a new model or revision of an existing model and the potential complexity of the model structure. For further information, see B.3 MODEL STRUCTURE.
The appropriate amount of validation will depend on the mechanics of the model, standards within a particular disease area, and available time amongst other things. Internal testing and debugging should be seen as a minimum. For further information, see B.3 MODEL STRUCTURE.
The amount of time to complete model simulations may be increased by the need for additional scenarios, possibly subgroup analyses, probabilistic sensitivity analysis, and where applicable, more advanced analyses, such as EVI. Always state explicitly the univariate sensitivity analyses that you are costing for e.g. range for cost, QoL, any structural for disease progression/survival assumptions, discount. Include a statement that additional SA to be charged at cost (this is particularly important when doing country adaptations). For further information, see B.4 MODEL SIMULATION AND SENSITIVITY ANALYSIS.
5. DELIVERABLES & TIMELINES
Deliverables
A report providing detailed description on methodology, together with probabilistic and stochastic results in the form of tables and charts.
The format of the report will be a project report intended for internal use by the client.
HINTS
The format and level of detail required of reports may vary according to the client’s use of reports, e.g. for submission to a Regulatory Authority, internal research report. For further information, see B.5.1 Deliverables.
For projects performed to support submission to reimbursement authorities. Always allow x days for additional support after final report has been sent. (This can be costed as an additional optional extra). For further information, see B.5.1 Deliverables.
The costs of writing a manuscript should be added as an additional extra to the proposal as standard. Where appropriate conference costs should also be added as an optional extra. For further information, see B.5.1 Deliverables.
Timelines
Once the project is agreed the work should be completed within x weeks. These timelines are dependent on xxxxxx (e.g delivery of data from the client).
For further information, see B.5.2. Timescales.
6. COSTS
It is vital to estimate the number of days as accurately as possible as well as naming members of staff for SWAP purposes. Input should be obtained from relevant staff, e.g a systematic reviewer for literature reviewing. A range of “typical” values is shown in the person days column – a specific figure will need to be estimated for the project in question