Abstracts from the research protocol for J-PEM 2000

1) RP1. List of Test and Control drugs
J-PEM2000 Studies on antidiabetics
Test drugs: Pioglitazone, Nateglinide, Glimepiride
Control drugs: All other oral antidiabetics currently available

J-PEM2000 Studies on antihyperlipidaemias
 Test drugs: Cerivastatin, Atorvastatin, Colestimide, Fenofibrate
Control drugs: All other oral antihyperlipidaemias currently available


2) RP2. Patient Registration
› In J-PEM, patients are registered from the record of prescriptions available in individual pharmacies, hospitals and medical offices.
› Possible study subjects are identified from original prescriptions or computerized records.
› A possible subject is given a leaflet before registration (see Progress in J-PEM 2000 "2. Ethics Review Board")
› When registering subjects, patient's name is not given in the registration form but only gender, age, drug and other information which do not allow identification of the subject, and a PEM-code number specifically created for J-PEM2000.
› The subject's name can be identified from the PEM-code when the list of PEM-code vs. patient's name is used. This list is kept by those who have registered the subject but not by those in the office of J-PEM (DSRU Japan). The office of J-PEM (DSRU Japan) will ask those who have registered subjects to discard the list when the studies, including the follow-up studies, have been finished.
› Control patients are those who started the control drug for the first time after the day when the test drug was marketed. (Detail for finding the control patients is given in Pharmacoepidemiology and Drug Safety 1999; 8: 447-456. )


RP2.1. PEM Registration Assistance Tool (PEM-RAT)
› PEM Registration Assistance Tool (PEM-RAT) is a computer software working on the WindowsNT, Windows95, Windows98 and WindowsMe
› PEM-RAT is distributed for free from the DSRU Japan.
› PEM-RAT is installed in a personal computer in individual pharmacies, medical offices and hospitals where possible study subjects are registered.
› PEM-RAT has a flexibility to convert computer records of prescriptions in various formats into records in a standard format.
› Once the records on prescriptions are stored inside PEM-RAT, PEM-RAT picks up possible study subjects who took test and control drugs.
› PEM-RAT also picks up the information of co-administered drugs when the study subjects is specified.
RP2.2. Informed Consent in J-PEM2000
According to the Ethics Review Board (ERB) , a consent is obtained from possible study subjects. The detail is given in Progress in J-PEM 2000 "2. Ethics Review Board")


3) RP3. Questionnaires (English translation is sent on request)
RP3.1. Questionnaires for doctors
3-1-1FJ-PEM2000 Studies on antidiabetics
English translation is sent on request
Envelope Questionnaire for doctors(Blue) Back side

3-1-2FJ-PEM2000 Studies on antihyperlipidaemias

English translation is sent on request
Envelope Questionnaire for doctors(Blue) Back side
RP3.2. Questionnaires for pharmacists
3-2-1FJ-PEM2000 Studies on antidiabetics
English translation is sent on request
Questionnaire for pharmacists in community pharmacies(Green) Back side
Questionnare for pharmacists in hospital(Pink) Back side

3-2-2FJ-PEM2000 Studies on antihyperlipidaemias

English translation is sent on request
Questionnaire for pharmacists in community pharmacies(Green) Back side
Questionnaire for pharmacists in hospital(Pink) Back side
› For one patient, a pair of questionnaires are sent, one for the pharmacist and the other for the doctor.
› As in the UK PEM, questionnaires are sent when more than 6 months have elapsed after the individual subject is prescribed a study drug.
› Questionnaires for pharmacists are unique to J-PEM as questionnaires are sent only to doctors in the UK PEM.
› Pharmacists are asked to report events and detailed information for drugs including those given concomitantly.
› Doctors are asked to report events and concomitant disease, some laboratory test results and other information. Amounts of questions are larger than those in a "green form" in the UK PEM.
› A small amount of money (\500 per questionnaire) is paid per each questionnaire returned with useful information.


4) RP4. Coding events with MedDRA/J
Events are coded using lowest level terms (LLTs) of MedDRA/J but the analysis is done
mainly by using preferred terms (PTs).

[A] Provisional diagnoses and ambiguous/vague information are not coded to avoid non-differential misclassification.
[B] Pre-existing medical conditions are not classified as an event unless a definite change is reported. If an aggravation of a condition is reported, a term without any additional term for the condition is selected, and a box for 'aggravation' is checked.
[C]: Death, hospitalization, or operation is coded as an event.
[D] Choice of term: a lowest level term (LLT) is chosen when regarded proper in terms of the placement in the MedDRA structure and the concept meant by the equivalent English term rather than the extent of matching between the reported term and Japanese term in MedDRA/J.
[E] If two or more events are simultaneously reported for an individual patient, each event is coded and whether or not to be counted is determined for each event.
[F] If an event was repeatedly described, the first event was distinctively entered as one to be counted.
[G] When LLTs belonging to different PTs are found to be synonymous as Japanese terms, one of the LLTs is selected with priority in coding.
[H] Regarding liver function analyses or renal function analyses, medical judgment is often used to add the term "hepatic dysfunction NOS" or "renal impairment NOS" to the laboratory results.

Although the above considerations are not in full accord with the 'MedDRATM Term Selection: Points to Consider (Release 2.0)', the need to conform data to functional event listing and statistical analysis for generating signals necessitate the proper solutions in data entry.


5) RP5. Method of Signal Generation
The best method for signal generation is currently being searched.
However, from the small amount of experiences in J-PEM pilot studies, the followings are suggested.
EComparison of crude event rates (number/patient/year) between a test drug and control drug(s) is effective to pick up adverse drug reactions (ADRs) as an event with the event rate for the test drug significantly larger than that for the control drug(s).
EThe regression method may help refine the information.
EThe follow-up studies also help refine the information.


6) RP6. Long-term survey
₯Long-term survey is designed to find the long-term safety and the long-term effectiveness of the study drug where subjects who gave a consent to participate in the long-term survey are followed by questionnaires sent directly to the patient.
₯The registration form for the long-term survey is handed from the doctor to the patient when the doctor considers this is appropriate.
₯We have not had enough amount of experience on this type of study and more information for the long-term survey will be shown in the future.