methodologic issues 2h

Noel R. Juban, MD, MScUP College of Medicine

BlindingMatchingRandom SamplingTreatment AllocationRandomizationContaminationCo-InterventionWithdrawalsCompliance/Adherence

Refers to a lack of knowledge of the identity of the study treatment.

1. Open Label = no blind is used 2. Single Blind= the patient is unaware of which treatment is

being received, but the investigator has this information. 3. Double Blind = neither the patient nor the investigator is

aware of which treatment the patient is receiving. 4. Triple Blind

Refers to the pairing of one or more controls to each case on the basis of their similarity with respect to selected variables.

1. Controls for confounding or confusing variables.

2. Rule out some particular mechanism in a postulated causal pathway between exposure and disease.

3. Removal of bias; reduction of variance

1. Stratified sampling formation of subgroups by strata according to a specified variable(s) and sampling predetermined number of cases and a predetermined number of controls within each stratum prior to start of the study.

2. Frequency Matching= selection of cases at random, with controls being taken from corresponding subgroups in proportion to the number of cases prior start of the study.

Ex: If 30% of the cases were males of Northern European extraction aged 60-64 years, then 30% of the controls would be taken to have similar

characteristics. 3. Post-stratification

4. Regression analysis

Technique whereby each sampling unit has the same probability of being selected.

Basic Procedure: 1. Prepare a sampling frame or a list showing all the units from

which the sample is to be selected, arranged in any order (i.e, obtain an up-to-date list of elderly people admitted in your hospital)

2. Decide on the size of the sample. 3. Select the required number of units at random, by drawing lots

or (more conveniently) using a table of random numbers.

– the process of deciding who (patient, subject, unit) should get which treatment.

The process of allocating patients (subjects, units) to treatment or exposure of interest according to an auditable random process, wherein each unit has an equal probability of being included. This avoids bias, allows for blinding and provides basis for standard methods of statistical analysis.

1. Begin at any haphazardly chosen point. 2. Proceed in any predetermined direction and follow

this direction faithfully until your sample size is not completed

3. Select sampling units by reading numbers off the table until required numbers of units have been selected.

4. Number not appearing in the list and numbers which reappear are ignored.

Contamination = when the control subject takes the test drug inadvertently (or vice versa)

Co-intervention = when the study subject takes another active drug inadvertently in addition to the test or control drug.

The process of removing a subject (patient, etc.) from the remaining portion of the treatment allocation.

1. Ineligibility 2. Non-compliance – dropouts; still included in the

analysis a. Patient moved from area b. Withdrawal of patient’s informed consent c. Failure of patient to continue to meet criteria for

continuation in protocol d. Decreased patient cooperation

3. Losses to follow-up – failure of patient to return for an appointment or for a specified number of appointments.

4. Competing events – includes contamination and co-intervention

5. Outliers – extreme value significantly different from remaining values; analyze with and without

1. Inflate sample size

Example: drop-outs = 10%

n = n/(1-0.1) = n/0.9

e.g., n = 42, n = 46.6 = 47

2. Measure them anyway

3. Keep them in the group of original allocation

Adherence to dose and dosing schedule, follow-up schedule and restriction from partaking/participating in competing events

1. Simplify the demands of the protocol on patients.

2. Minimize the number and duration of unpleasant or painful tests.

3. Maintain relatively frequent contact with patients, especially at emotionally or physically difficult periods for the patient; contact the patient prior to scheduled visits.

4. Allow for flexible dosing regimens to deal with adverse reactions, toxicity, and unanticipated situations.

5. Provide the patient with appropriate information on the study.

6. Avoid making the patient feel guilty about poor compliance and provide positive physician-patient relationship.

7. Establish a therapeutic goal in conjunction with the patient and assess the patient’s progress towards the goal.

8. Plan patient visits at a mutually convenient time and insure that the patient has a minimal delay in waiting to see the physician or staff.

9. Allow for and encourage patients to participate in their own care (e.g., with self-monitoring of their disease and treatment).

10. Involve the patient’s spouse, family, or support group in the clinical study.

BlindingMatchingRandom SamplingTreatment AllocationRandomizationContaminationCo-InterventionWithdrawalsCompliance/Adherence