Pharmaceutical Benefits Advisory Committee
1995 Guidelines for the Pharmaceutical Industry on Preparation of Submissions to PBAC: Appendix A-C
Guidelines for the Pharmaceutical Industry on Preparation of Submissions to the Pharamaceutical Benefits Advisory Committee
>> Part 1
>> Part 2
>> Part 3
>>> Section 1
>>> Section 2
>>> Section 3
>>> Apprendix A-C
>>> Appendix D-F
>>> Appendix G-I
>>> Appendix J-L
>>> Appendix M-O
>>> Appendix N
>> Part 4
Appendices A - C
Description of the search of published literature
REFER: Section 2.1
The methodology used to search the literature is pivotal to assessing the completeness of the search. Specify:
(a) the medium (eg dial-up, CD-ROM etc) and service provider(s) (eg Dialog, Silver Platter) used to conduct the search;
(b) the specific databases searched (including at least MEDLINE, EMBASE and possibly SCISEARCH, but may include databases internal to the company);
(c) the date the search was conducted;
(d) the date span of the search (which should be up to date to the most recent database update);
(e) the complete search strategies used, including the search terms (key or MeSH words) and the relationship (sets and boolean) between the search terms; and
(f) any supplementary searches, especially manual checking of references in the retrieved papers from the database searches.
Measures taken by investigators to minimise bias in each trial listed in response to Section 2.2
REFER: Section 2.3; Appendix D (b) (iv)
For each of the following methodological topics, choose the description that best fits each trial and answer the supplementary question for each trial. If there is more than one trial, tabulate the responses. Randomisation: it is important that clinical staff are unable to predict which treatment a patient will receive prior to a final decision being made regarding entry to the trial. Which of the following best describes the randomisation technique used?
1. No details of randomisation were reported, or the method used was inadequate (eg randomisation according to the day of the week, even/odd medical record numbers).
2. An insecure randomisation method was used, where clinical staff could possibly learn of the treatment assignment (eg randomisation sequence kept in the clinical area and open/unblinded trial; treatment assignment kept in consecutive "sealed" envelopes and open/unblinded trial).
3. A secure randomisation method was used, where the randomisation sequence was kept away from the clinical area and administered by staff not directly involved in patient care (eg randomisation performed at a separate site available through a toll-free telephone number or by the pharmacy department after the decision has been made to enter the subject in the trial).
Adequacy of follow-up: it is important that an attempt is made to summarise the trial outcomes for all subjects who were included in the trial. A full "intention-to-treat" analysis is the preferred basis for an economic evaluation that attempts to model the likely impact of the drug in the community. Which of the following best describes the adequacy of follow-up?
1. There were significant numbers of drop-outs with no assessment of trial outcome(s) in the subjects who dropped-out, and drop-out rates differed between treated and control groups.
2. There were some drop-outs with no assessment of trial outcome(s) in the subjects who dropped-out, and drop-out rates were (approximately) equivalent in treated and control groups.
3. Trial outcome(s) were assessed in all treated and control subjects who did not withdraw from the trial.
SUPPLEMENTARY QUESTION: summarise for each comparison group the number randomised to treatment, the number of drop-outs and the number of subjects who were lost to follow-up.
NOTES: a drop-out stops the trial medication for a medical reason or a protocol violation but can and, particularly for an economic evaluation, should still be followed-up, whereas a subject who unilaterally elects to withdraw from the trial is deemed to be lost to follow-up.
Blinding of outcomes assessment: it is important that where the comparator is not indistinguishable by visual inspection or taste, or where there is a high chance of "unblinding" (eg oestrogen or beta-blocker treatment), that the observer responsible for measuring the trial outcome remains unaware of the treatment assignment. Which of the following best describes the blinding of the outcomes assessment?
1. There was an inadequate attempt (or no attempt) to blind observer(s), and the measurement technique was subject to observer bias (eg blood pressure measurement with standard sphygmomanometer, measurement of vertebral height on an X-ray, quality of life instrument).
2. The observer(s) were kept fully blinded to treatment assignment, or the measurement technique was not subject to observer bias (eg measurement of bone mineral density or survival).
NOTES: the observer may be a trial investigator and/or a subject. To maintain "full blinding", it is usually necessary to blind all people directly involved in the care of the trial subjects and the trial subjects themselves (ie double-blinding) to prevent "unblinding" of the observer.
Purpose of these assessments
The intention of these assessments is to provide the sponsor and the PBAC with a clear idea of which trials are of the highest scientific rigour and which are therefore likely to give the most accurate estimate of how well the proposed drug works. There is no minimum standard, but the PBAC is most likely to be persuaded by the data from the trials of the highest scientific rigour.
Characteristics of each trial listed in response to Section 2.2
REFER: Sections 2.4 and 2.5; Appendices J (c) and M
Answer each of the following questions for each trial. If there is more than one trial, tabulate the responses.
(b) Was the trial conducted in Australia (or were one or more centres of the multi-national trial located in Australia)?
(c) How do the subjects included in the trial compare with patients who are likely to receive the drug on the PBS? Consider factors known to affect outcomes in the main indication such as demographics, epidemiology, disease severity, setting.
(d) What dosage regimens were used in the trial - are they within those recommended in the current TGA-approved product information?
(e) What was the median (and range) duration of follow-up of the trial?
FOR (a) If the submission includes one or more cross-over trials, indicate for each such trial whether a carry-over effect is likely.
FOR (b) This may be particularly useful in assessing the extent to which there is a change in the patterns of resource provision. For several reasons (such as different incentives), patterns of resource provision seem to differ between health care systems more than patient responses to a drug across different health care systems.
FOR (c) This forms the basis of the consideration of the following three points.
Firstly, how do the trial subjects compare with typical Australian patients suffering from the relevant condition(s), for example in terms of age and sex distribution or of the natural history of the condition(s)? Are any differences likely to matter?
Secondly, how do the trial subjects compare with Australian patients in terms of disease severity? This can be important. A new drug may be cost-effective when use is confined to patients with severe disease but not when it is used to treat patients with milder disease who may respond to less effective and less expensive therapies. It may be possible to estimate the likely impact of this by performing sensitivity analyses in a modelled evaluation (see Section 3.7).
Thirdly, is the trial setting relevant to that of the PBS? For example, most PBS drug use is in the community rather than in a hospital, so a trial in subjects with severe disease requiring hospitalisation may only be relevant in particular circumstances (such as a Highly Specialised Drug or a drug for use in private hospitals).
FOR (d) The trial should use the correct doses of the proposed drug and the main comparator (and a suitable duration of therapy where this is relevant). Doses and duration should be those recommended in the product information as optimal for the relevant indication. These may differ from those shown by market research to be actually used in the community. However prescribing of higher than recommended doses (at higher cost) of a comparator drug is unlikely to be accepted as an argument for a higher price for the proposed drug.
FOR (e) The duration of follow-up for a trial subject is the length of time between randomisation and the end of blinded follow-up of that subject. The duration of non-blinded follow-up of drop-outs should be excluded from the calculations.