CARDIOVASCULAR JOURNAL OF AFRICA • Vol 21, No 2, March/April 2010
AFRICA
87
prescribed treatment and the overall purpose of CHFmanagement.
This represents a key substudy of the previously described
Heart of Soweto study,
3,13
which is currently mapping the emer-
gence and spectrum of heart disease in a black urban population.
This study benefits from the unique setting at the Chris Hani-
Baragwanath Hospital, representing the only tertiary-care centre
for the population of Soweto and surrounding communities.
Soweto, a township situated in the south west of Johannesburg,
South Africa, has a population of 1.1 million, which has under-
gone economic development, enabling a large proportion of its
population to achieve a more affluent lifestyle. The township
comprises people from different ethnic backgrounds, however
the black African population is predominant.
As Chris Hani-Baragwanath Hospital is the only point of
specialist cardiac care for the population of Soweto, the cardiol-
ogy clinic probably attends to nearly every patient presenting
with symptoms evolving from a cardiac condition. The popula-
tion of the cardiology out-patient department consists of patients
with a suspected cardiac disorder, seen and referred by 12 local
Soweto primary-care clinics. Other patients are initially seen at
the general medicine out-patient facilities, specialist medical
registrar clinic, diabetes clinic or are in-patients admitted to any
other ward of the Chris Hani-Baragwanath hospital and need a
cardiac consultation.
During November 2006 and April 2007, we recruited 200
consecutive patients with a confirmed diagnosis of CHF present-
ing at the cardiology clinic. For the purpose of this study, patients
were included if they had a confirmed diagnosis of CHF by
an attending cardiologist, based on typical clinical symptoms
(shortness of breath, oedema, fatigue) and a documented left
ventricular ejection fraction (LVEF) of
≤
45% using echocar-
diography. Non-English-speaking patients were excluded only
if a translator was unavailable. Patients were approached in the
out-patient cardiology clinic and invited to participate. Data were
collected after an informed consent form was signed.
Prior to study commencement, the relevant local Ethics of
Human Research Committee approved the study. The study
conforms to the standard statements on ethics outlined by the
Declaration of Helsinki.
During a pilot study consisting of 20 participants, a ‘medica-
tion adherence and knowledge on heart failure’ survey specific
to this predominantly black African community was developed
in October 2006. The questionnaire was then applied to the
above study population on a one-to-one basis in an interview of
approximately 20 minutes. All interviews were executed by the
same investigator (VR) and if needed, a translator assisted the
applicant.
A pill count of prescribed CHF treatment was conducted in
those 82 patients (41%) who returned for a scheduled one-month
post-interview appointment. If appropriate, a telephone reminder
was made two days prior to this appointment; however, contact
details were available for few patients.
The medication adherence and knowledge on heart failure
survey addresses the following sections: demographic and clini-
cal data, medication adherence, self-care behaviour (adherence
to follow-up appointments, weighing behaviour, dietary restric-
tion, regular physical activity, smoking abstinence and alcohol
intake), knowledge concerning CHF medication and overall
CHF management. Questionnaires in other studies included the
same sections,
7,9,14
however the answer possibilities in our ques-
tionnaire were mostly dichotomous (yes or no) instead of offer-
ing a range of answers (no, a little, some, a lot).
In accordance with previous studies of this type,
7,14
we defined
treatment adherence as
≥
75% of the prescribed pills taken.
Similarly, we defined appointment adherence as being present
at
≥
75% of the assigned appointments consisting of quarterly
check-ups and monthly medication refills at the hospital phar-
macy. In accordance with the European Society of Cardiology
guidelines,
6
we defined behavioural adherence as daily weight
monitoring, daily intake of five servings of fruit and vegetables,
drinking less than two litres of fluids per day, being physically
active, with compensated CHF two to three times per week,
refraining from smoking and keeping a moderate alcohol intake
(one beer, one to two glasses of wine per day).
Statistics
Descriptive statistics and measures of frequency were conducted
in Microsoft EXCEL
®
and were used to describe the study popu-
lation and various adherences. Data are presented as means
±
standard deviation or percentages. To compare groups we used
χ
2
analyses for discrete variables and the Student’s
t
-test for
continuous variables. Binary logistic regression models were
performed in SPSS 11.5 to determine variables predicting adher-
ence. Determinants for medication adherence were presented
by odds ratio (OR) and 95% confidence intervals (CI), where
an OR
=
1 indicated no influence on medication adherence.
Significance was accepted at the two-sided level of 0.05.
Results
The overall demographic and clinical profile of the study cohort
is presented in Table 1. Overall, black Africans predominated [
n
=
157 (79%)] and there were more men [
n
=
109 (55%)] than
women [
n
=
91 (45%)] with no difference in age profile (mean
age 56
±
13 vs 56
±
15 years). Apart from black African patients
there were Asian Indians (
n
=
10), coloureds (
n
=
8) and white
Africans (
n
=
25), which we combined as ‘other races’. Almost
half of the patients were retired and nearly all lived in a shared
household. Black Africans were significantly more likely to have
no or standard education than the other races combined [128
(82%) vs 24 (56%),
p
=
0.001].
When questioned about their self-perceived level of social
support, black Africans were less likely to report having ‘a lot
of’ practical support than other races combined [91 (58%) vs 33
(78%),
p
=
0.038], however there was no major difference found
in respect of reported emotional support.
Clinical profile
Overall, 90% of our study patients were classified as New York
Heart Association functional class (NYHA) II and III at the point
of being diagnosed with CHF. Overall, the mean left ventricular
ejection fraction was 32
±
8%. Black Africans were less likely
to live longer than five years with CHF than the other races
combined [61 (39%) vs 23 (53%),
p
=
0.085] and additionally,
they were more likely to have been admitted to hospital before
the point of investigation due to their CHF [135 (86%) vs 34
(79%),
p
=
0.188]. However, that did not reach statistical signifi-
cance.
As represented in Fig. 1, the three most common underlying
