Cardiovascular Journal of Africa: Vol 26 No 5 (September/October 2015)

38 / 56

CARDIOVASCULAR JOURNAL OF AFRICA • Vol 26, No 5, October/November 2015 36 AFRICA
Conclusion: This study has shown that obesity and overweight
was common, being higher in the urban than the rural popula-
tion. High BMI was more prevalent among females and middle-
aged persons in both settings as well as in urban dwellers with
higher levels of education.
LIMITED SUPPLIES OF ANTIHYPERTENSIVE MEDI-
CATIONS IN PRIMARY CARE SETTINGS IN THREE
AFRICAN COUNTRIES
Watkins David
University of Cape Town, Groote Schuur Hospital, Cape Town,
South Africa;
Background and objectives: Treatment guidelines and targets
for reducing cardiovascular mortality in Africa assume that
antihypertensive medications are routinely available in primary
care settings. I sought to explore the availability of common
antihypertensive drugs across three countries.
Methods: I obtained data from nationally representative health
facility surveys used in the Access, Bottlenecks, Costs, and
Equity (ABCE) study. I included all facilities that provide hyper-
tension care in Ghana, Kenya and Uganda. The antihyperten-
sive drugs assessed in these surveys were captopril, lisinopril,
atenolol, propranolol, nifedipine and hydralazine; both stocks
and stock-out rates were recorded. I first conducted descriptive
analyses of drug availability by country during 2011 (the last
survey year); this included the presence of (1) individual drugs,
(2) drug classes, and (3) at least one or two drugs (from different
classes). I then used the longitudinal datasets from Kenya and
Uganda (2007–2011) to estimate associations between facility
characteristics and drug availability, employing logistic regres-
sion models with facility-level random intercepts.
Results: While individual drug availability varied across countries,
stock-out rates were generally < 10% across countries. However
only 27% of Kenyan facilities reported carrying any ACE inhibitor,
and only 37% of Ghanaian facilities carried any beta-blocker. In
Ghana, 100% of facilities carried at least one drug, but only 51%
of facilities had two or more drug options. Estimates for at least
one versus at least two drugs were 67 vs 53% for Kenyan facilities
and 82 vs 58% for Ugandan facilities, respectively. In both Kenya
and Uganda, public facilities were significantly less likely and
higher-level facilities significantly more likely to carry more than
one antihypertensive drug; however the marginal effects were much
larger in Uganda. In neither country was there an independent
association across facility locale (rural vs semi-urban vs urban).
Conclusions: Primary healthcare providers in three African
countries appear to lack drug options for managing hyperten-
sion. About half of the surveyed facilities carried only one
drug, and public and first-level clinics in Kenya and Uganda
were more likely to have no antihypertensive drugs at all. Hence,
important supply-side barriers remain to implementing hyper-
tension ‘roadmaps’ and guidelines in Africa.
STRATEGIES TO REDUCE THE BURDEN OF RHEU-
MATIC FEVER AND RHEUMATIC HEART DISEASE IN
AFRICA: A COST-EFFECTIVENESS ANALYSIS TOOL
FOR LOCAL DECISION-MAKING
Watkins David*, Lubinga Solomon J1, Babigumira Joseph B1
*University of Cape Town, Cape Town, South Africa; University
of Washington, Seattle, USA;
1University of Washington, Seattle, USA
Background and objectives: Mortality rates from acute rheu-
matic fever (ARF) and rheumatic heart disease (RHD) remain
high in many parts of Africa. However, it is not clear whether to
focus on preventing RHD or treating existing cases with surgery.
We developed a model to assess the cost-effectiveness of scaling
up primary prevention (PP), secondary prevention (SP), and
valve surgery (VS), given local resource constraints.
Methods: We created a Markov model of the natural history of
ARF and RHD, taking transition probabilities, healthcare and
programme costs, current intervention coverage levels, and inter-
vention effectiveness data from previously published studies and
expert opinion. We took a healthcare system perspective on costs,
and measured outcomes as disability-adjusted life-years (DALYs),
discounting both at 3%. We calculated incremental cost-effective-
ness ratios (ICERs) for increasing coverage of PP, SP and VS to
70, 92 and 95%, respectively compared to the status quo level of
coverage. We assessed three scenarios: (1) upper-middle-income
country with existing surgical platform, (2) lower-middle-income
country with no existing surgical platform, and (3) low-income
country with no existing surgical platform. We used per capita
gross domestic product (GDPpc) levels to assess cost effective-
ness. We also estimated the impact of PP, SP and VS on per capita
government health expenditures (GHEpc) in each scenario.
Results: Across all scenarios, scaling PP was cost saving (1 900–
2 300 DALYs averted per 100 000). Scaling SP was cost effec-
tive (ICERs $116–268/DALY) and consumed less than 1% of
GHEpc. Scaling VS was not cost effective in scenario 1 (ICER
$25 800/DALY; 17.5 times GDPpc) or scenario 2 (ICER $49 700/
DALY; 7.4 times GDPpc), but it was cost effective in scenario
3 (ICER $33 200/DALY; 2.6 times GDPpc). The latter result
was driven by the lower cost of using existing surgical capacity;
however programme expenditures consumed 8.8% of GHEpc.
Conclusions: This preliminary analysis suggests that PP and
SP are cost effective in Africa. The cost effectiveness of build-
ing new surgical platforms to scale VS depends highly on
programme costs. Future work will assess the cost effectiveness
of high-volume surgical platforms serving multiple countries.
This model could be used locally as a tool to assist in planning
ARF and RHD programmes.
GLOBAL DEMAND FOR CIGARETTES: A CROSS-COUN-
TRY ANALYSIS OF CONSUMPTION, INCOME AND
PRICE ELASTICITY; 2008–2012
Watkins David*, Conner Ruben 1, Dieleman Joseph,2 Ng Marie2
University of Cape Town, Cape Town, South Africa; University
of Washington, Seattle, USA;