Knowledge, attitudes and perceptions of stroke: a cross-sectional survey in rural and urban Uganda

Study design and population

This study was part of a larger U.S. National Institute of Heath-funded Medical Education
Partnership Initiative (MEPI) neurological disorder survey, where we assessed community
knowledge and attitudes on stroke and stroke risk factors among some of the study
participants.

This was a cross-sectional study conducted within an ongoing population survey of
3000 participants on prevalence and incidence of neurological diseases in Mukono district.
Face to face interviews were conducted between August and November 2014. Multistage
stratified sampling technique was used as described below 11]. At the subcounty level, urban Mukono Town Council (TC) was randomly selected and
Nakisunga sub–county as rural were randomly selected out of 13 sub-counties. Mapping
of the selected urban and rural areas was based on the Uganda Population and Housing
census where 11,373 and 9570 households were identified in Mukono TC and Nakisunga
respectively 12]. The sampling frame was all households in these areas. Systematic sampling was used
to select households in each village to total 2000 in the urban area and 2200 in the
rural area that would participate in the large population survey. Out of the 1500
participants in the urban area and 1500 in the rural area, systematic sampling technique
was used to select every tenth household for this interview. If the selected household
objected participation, then the next household would be considered. A total of 377
participants participated in this study with 177 and 200 from the urban and rural
areas, respectively. They were interviewed on selected aspects of stroke knowledge,
attitudes and perception. The Inclusion criteria included; usual resident who is present
in the sampled household on the night before the survey, aged 18 years or older (adult)
and willingness to provide informed consent. We excluded those who were physically
unable to undergo interview. One adult, randomly selected from each household was
approached and consented, participated in the study. Only one participant was randomly
selected from each household using simple random method. We employed randomly selected
cards by the potential participants written on “Yes” and “No” for study participation.
If similar cards were drawn then the process would be repeated until one participant
was selected. The selected households were visited by the research team. The randomly
selected participant was informed about the research and the intended use of the information
obtained. A request for a written informed consent was then sought. To address potential
sources of bias, a standardised questionnaire was used with a wide range of responses
for the study participants which were read to the participants. The study interviewers
received trained on the study protocols for data collection in order to minimize inter-observer
variability during data gathering and entering data.

Sample size determination

The sample size calculation for stroke knowledge and attitudes was based on the prevalence
of hypertension which us an important risk factor for stroke. The sample size was
calculated using formula: { } where p = prevalence of hypertension, q = complement of the prevalence, margin
of error is error = d, ? = significance level. Setting significance at 0.05 and error
margin at 5 %, adjusted sample requirement for an assumed 10 % level of non-response
(nr = 10 %) = N*. Based on a previous study in Mukono 13] where hypertension prevalence was 27 % and N* = 336, we recruited 370 participants.

Questionnaire and measurements

We used a modified standardized questionnaire that assessed knowledge and attitudes
towards stroke already used in the sub-Sahara African settings 5], 10], 14] (see Additional file 1). Participant’s knowledge of stroke warning signs was categorized based on the numbers
of stroke warning signs 15], 16]. Individuals with good knowledge could identify 5–10 stroke warning signs, fair knowledge
2–4 signs, and poor knowledge one stroke warning sign. A similar categorization was
used for participant’s knowledge of stroke risk factors.

Ethical considerations

Ethical approval for the study was obtained from Makerere University College of Health
Sciences’ School of Medicine review board and ethics committee Ref number 2013-145
and UNCST Ref Number. HS 1551. Written informed consent was obtained before enrolling
the participants into the study.

Data analysis

Descriptive statistics of mean, frequency, and percentages were used to summarise
data on socio-demographic variables and stroke knowledge and perceptions. Chi square
or Fisher’s exact tests were used as appropriate to assess associations between stroke
knowledge and perception and demographic variables and self-reported risk factors.
Logistic regression was used to determine predictors of knowing the organ affected
by stroke, good level of knowledge of stroke warning signs and risk factors. All tests
of hypothesis were two tailed with a level of significance at 0.05. All statistical
analysis was performed using STATA software version 12 (Stata Corporation, College
Station, TX, USA).