Transcript Oral Presentation 1

Methodological Issues in
Measuring Adherence to
Antimalaria Drugs
Irene A. Agyepong, Evelyn K. Ansah, Margaret
Gyapong, David Evans, Guy Barnish
Challenging Methodological
Issues
 Defining Adherence
 Collecting data to enable measurement of
adherence
 Analyzing the data
 Relating the data to other variables
(identification, understanding and
measurement of other variables)
 In the rest of the presentation each issue
above is discussed a little more as well as:
 How it was addressed in our study
 The strengths and weaknesses of how it was
addressed
Defining “adherence”
 Adherence – Comparing how the patient
took the drug with how they were supposed
to take it
 Doing exactly what the prescriber or
dispenser wrote or said – Ethical and
Technical Issues
 In measuring adherence is it relevant or
irrelevant whether the prescription is technically
“correct” or not?
 What do you do when the prescription is
technically “incorrect”
 What do you do when the patient happens to
know or be told the “correct” dose by another
person and adheres to that?
Defining “adherence”
 Several possibilities for defining and
therefore measuring adherence
 Example of Categories
 Strict Full Adherence – Patient takes the exact
dose prescribed everyday for the full duration of
treatment. A higher or a lower dose taken on
any one day leads to the patient being classified
as non adherent
 Minimum Daily Adherence – A sub-optimal dose
on any one day is regarded as non adherence.
However if the client takes a little more than the
recommended dose e.g. 5 tablets instead of 4,
they are still registered as adherent
 Other category possibilities e.g. MTA
Collecting the Data
 Ideally to be sure you have 100% accuracy
in recording how the patient took the drug
you need to observe the process
 However being there to observe on a daily
basis is expensive & introduces bias related
to ‘instrumentation’
 The presence of the observer each time a
dose is due to be taken becomes an
intervention that could in itself increase
adherence
 Realism vs. precision i.e. How strict to be
on detail (e.g. 8 hrly vs 3 times daily)
Collecting the data
 We settled for data collection using one visit
to the home within 24 hours after the last
dose was supposed to have been taken
 This was feasible since only a 3 day recall
was involved
 The data collector took the clients
statement of how they took the drug as a
fact, but also crosschecked by examining
the bottle or package to see if anything was
left and compared it to what was supposed
to have been taken
Collecting the data
 The problems and the expense involved in
tracking clients to their homes in the developing
country setting of no house numbers, street
names, telephone numbers etc
 The problem of syrups
 A tablet is pre-measured so to speak
 But what do you do about the definition of a teaspoon
in the home in a situation where child caretakers are
not given measures at the clinic with the syrups
 Is a mother who gave “one teaspoon” daily non
adherent because her understanding of a teaspoon is
a dessert spoon or a coffee spoon or the cover of the
bottle?
Analyzing the data
 Need to simplify and quantify
definitions for statistical analysis of
data
 Manual comparison of prescription
given and the way the drug was taken
to code adherence
 Coding and entry of codes
 Strict Full Adherence (SFA yes=1, no=0)
 Minimum Daily Adherence (MDA yes=1,
no=0)
Relating the data to other
variables
 Written prescription vs verbal instructions vs
labeling on bottle – do they tally
 Prescriber and dispenser instructions – do
they tally?
 In our initial study they tallied over 95% and
we just took dispenser instructions
 Issues of measurement coming up in the
current qualitative & exploratory study of
factors affecting adherence
 e.g. quality of communication – what the
dispenser said vs. what the client heard,
understood & remembered and its influence on
adherence
Improving adherence
measurements in the future
 Would daily charting of dose taken and
frequency by patients themselves be a
workable alternative to recall from memory
and what kind of bias might it introduce
since it is in itself perhaps an intervention
 Especially important to ask in measuring
adherence where dosages are more
complex than once daily for 3 days as in the
case of chloroquine since memory recall
will worsen with longer duration and more
complex regimens
Improving adherence
measurements in the future
 Possible to agree on certain standard
definitions for measurement to make
different studies comparable on a core
set of variables?
 Better understanding, measurement
and analysis of factors that affect
adherence in the developing country
context and development of means of
quantifying complex variables like
“quality of communication”