Neil A Martinson

University of the Witwatersrand, Johannesburg, Gauteng, South Africa

Are you Neil A Martinson?

Claim your profile

Publications (100)522.18 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: INTRODUCTION Evidence on the efficiency of HIV prevention interventions is limited. ORPHEA project aimed to estimate average costs and analyse their heterogeneity for three HIV prevention interventions: prevention of mother-to-child transmission (PMTCT), voluntary medical male-circumcision (VMMC), and HIV-testing and counseling (HTC), in Kenya, Zambia, Rwanda and South-Africa. METHODS Sample comprises of 60-80 clinics per country. Micro-costing was performed and relevant costs (personnel, supplies, utilities, equipment and property) and intervention-output data were collected retrospectively for 2011/2012. Information on time-allocation, building-space-used and time-motion is also collected . Quality is captured by measuring the facility’s attrition rate at each stage along the service delivery, using provider-vignettes and patient-exit-interviews. We estimated average-cost per each HIV prevention intervention. RESULTS Estimated weighted-average costs were US$18.1 per HTC client, US$50.1 per women tested in PMTCT services, US$93.4 per MC client. Staff costs account for more than 80% of the average costs in each prevention intervention. Variation in average cost for HTC and PMTCT by facility type is limited, except for faith-based hospitals. VMMC costs show important variations among different type of facilities, resulting cheaper to perform VMMC intervention in health centres. CONCLUSIONS Results suggest that there is a large potential to increase efficiency within the current financial and structural constraints of the health system in some countries in Africa. Staff accounts for the larger proportion of average-costs and offer the highest potential to implement actions to increase efficiency and quality. There is also need to explore different types of service delivery-models and their impacts on efficiency.
    142nd APHA Annual Meeting and Exposition 2014; 11/2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: While healthcare services become increasingly available in developing countries, there is heterogeneity in the quality of service delivery across health providers. As part of a cost-efficiency project (ORPHEA) conducted in Kenya, Zambia, South Africa and Rwanda, we assessed the quality of three HIV prevention interventions: Voluntary Medical Male Circumcision (VMMC), Prevention of Mother-to-Child Transmission (PMTCT), and HIV Testing and Counseling (HTC). Description: We used multistage sampling techniques to select 60-80 health facilities per country stratifying by facility location, size and type. We conducted cross-sectional provider vignettes and patient exit interviews (PEI) to assess service quality measures by estimating provider competence and performance scores. Provider competence scores were compared with average scores for provider performance at the facility level to estimate potential performance. Results: Results are currently available for two (Kenya, Zambia) of the four countries*, where we surveyed 799 providers and 1,161 clients. On average, HTC healthcare providers had a vignette score of 59% and a PEI score of 26%; PMTCT providers had a vignette score of 56% and a PEI score of 32%; and VMMC providers had a vignette score of 67% and received a score of 48% from PEI. Conclusions: We found, there is significant opportunity to increase quality of services without additional health resources but rather by reallocating labor resources. There is therefore need to investigate how different service delivery models could be implemented to reach full performance. ------------------------- * Results from the remaining two countries may be available by the time of the conference. Learning Areas: Biostatistics, economics Conduct evaluation related to programs, research, and other areas of practice Provision of health care to the public Public health or related research Social and behavioral sciences Learning Objectives: Assess the quality of three HIV prevention interventions in Kenya, Rwanda, South Africa and Zambia. Keywords: HIV Interventions, Quality of Care Presenting author's disclosure statement: Qualified on the content I am responsible for because: I am a researcher at the Center for Health Systems Research at the Mexico National Institute of Public Health (INSP) where I have been involved in multi-country economic, impact, and efficiency evaluations in in low- and middle-income countries. I hold a master's degree in economics. Any relevant financial relationships? No I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation. Back to: 4330.0: HIV/AIDS in International Settings
    142nd APHA Annual Meeting and Exposition 2014; 11/2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: Introduction: Staff costs comprise the largest share of average facility costs to provide HIV prevention services. Evidence has shown that greater scale of production reduces average costs of these interventions. Little is known about the optimal combination of staff and quality that predict technical efficiency, as to increase value-for-money of HIV spending. Methods: A cost analysis was implemented at 92 facilities offering HTC in Kenya, Zambia, Rwanda and South Africa in 2011-2012. Staff costs were allocated to interventions using time-motion data. Standardized vignettes were administered to score technical proficiency. We estimated average staff cost by dividing total prorated staff costs by the number of HTC clients in each facility. We created quantiles of operation scale and quality, to analyze its tradeoff and explore costs efficiency of HTC delivery. Results: Mean staff costs per client tested were 18.8 USD (IQR: 0.5;18.1), with heterogeneity by facility type. Independent of health providers' competence score, operation scale negatively correlated with average costs: Moving from the first quartile of scale towards the upper quartile yielded a 90% reduction in staff costs. Among the subset of facilities in the uppermost quartile of scale, we found a mean difference in staff costs between the lower and upper quartiles of providers’ competence of 1.2 USD. Conclusions: These results demonstrate the need to understand the mechanisms that curb efficiency in the provision of HIV prevention interventions. We provide evidence of the possibility to achieve higher quality of care without escalating the budget, provided that service utilization is increased.
    142nd APHA Annual Meeting and Exposition 2014; 11/2014
  • AIDS research and human retroviruses. 10/2014; 30 Suppl 1:A64.
  • AIDS research and human retroviruses. 10/2014; 30 Suppl 1:A197.
  • AIDS research and human retroviruses. 10/2014; 30 Suppl 1:A219-20.
  • The International Journal of Tuberculosis and Lung Disease 08/2014; 18(8). · 2.76 Impact Factor
  • The Journal of Infectious Diseases 07/2014; · 5.85 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Introduction: Globally, hepatitis B virus (HBV) infection is the leading cause of liver-related mortality. Newborn vaccination, maternal antiviral therapy and administering hepatitis B immune globulin shortly after birth can greatly reduce the risk of perinatal and infant infection. However, evidence-based policy regarding these interventions in Africa is hampered by gaps in knowledge of HBV epidemiology. We describe maternal chronic hepatitis B (CHB) prevalence and infant infection during the first year of life within a cohort of women living with HIV. Methods: We recruited and prospectively followed pregnant women living with HIV and their infants from prenatal clinics in an urban area of South Africa. Hepatitis B surface antigen, anti-hepatitis B surface antibodies and HBV DNA were assessed in all women. Hepatitis B testing was also performed at 6 and 52 weeks for all infants born to mothers with either positive surface antigen or detectable HBV DNA. Results: We enrolled 189 women with a median age of 29 years and median CD4 count of 348 cells/mm(3). Fourteen had a positive surface antigen (7.4%), of which six were positive for "e" antigen. An additional three had detectable HBV DNA without positive surface antigen. One infant developed CHB and three others had evidence of transmission based on positive HBV DNA assays. HBV vaccinations were delivered at six weeks of life to all infants. Conclusions: Our findings highlight the risk of peripartum HBV transmission in this setting. Approaches to reducing this transmission should be considered.
    Journal of the International AIDS Society 05/2014; 17(1):18871. · 3.94 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The tuberculin skin test (TST) is used to help diagnose tuberculosis (TB) in acutely ill hospitalised children. OBJECTIVE To investigate the potential augmentative effect of topical calcipotriol (a vitamin D analogue) or zinc on TST induration. Three TSTs were performed among 64 hospitalised children; each site was covered with topical aqueous cream (control), calcipotriol or zinc and assessed 24 and 48 h later by investigators blinded to all topical applications. TSTs were reactive in 15 (23.4%) children, of whom 13 (20.3%) were TST-positive. Topical calcipotriol and zinc induced TST positivity in two children with reactive but negative control TSTs. These treatments, however, did not significantly increase TST positivity rates. In children with reactive TSTs, the median 48 h induration diameter was not significantly different between the control, calcipotriol- or zinc-treated groups, which were respectively 12.0 (25%-75% IQR 5.0 - 18.0), 14.0 (25%-75% IQR 10.0 - 15.0) and 12.0 (25%-75% IQR 8.0 - 15.0) mm. Topical treatments did not induce TST reactivity or TST positivity in children with culture-confirmed TB disease (n = 4), human immunodeficiency virus infection (n= 18) or kwashiorkor (n = 9). Topical calcipotriol or zinc does not induce TST reactivity or significantly increase TST positivity rates in acutely ill hospitalised children. However, further studies are required to assess the effects of topical treatments on TST positivity in severely malnourished children.
    The International Journal of Tuberculosis and Lung Disease 04/2014; 18(4):388-93. · 2.76 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To report the viral load and CD4 count in HIV-infected, antiretroviral naïve, first -time HIV-testers, not immediately eligible for treatment initiation by current South Africa treatment guidelines. This was a cross-sectional study in a high-volume, free-of-charge HIV testing centre in Soweto, South Africa. We enrolled first time HIV testers and collected demographic and risk-behaviour data and measured CD4 count and viral load. Between March and October 2011, a total of 4793 adults attended VCT and 1062 (22%) tested positive. Of the 1062, 799 (75%) were ART naïve and 348/799 (44%) were first-time HIV testers. Of this group of 348, 225 (65%) were female. Overall their median age, CD4 count and viral load was 34 years (IQR: 28-41), 364 (IQR: 238-542) cells/mm3 and 13,000 (IQR: 2050-98171) copies/ml, respectively. Female first time HIV testers had higher CD4 counts (419 IQR: 262-582 vs. 303 IQR: 199-418 cells/mm3) and lower viral loads (9,100 vs. 34,000 copies/ml) compared to males. Of 183 participants with CD4 count >350 cells/mm3, 62 (34%) had viral loads > 10,000 copies/ml. A large proportion of HIV infected adults not qualifying for immediate ART at the CD4 count threshold of 350 cells/mm3 have high viral loads. HIV-infected men at their first HIV diagnosis are more likely to have lower CD4 counts and higher viral loads than women.
    PLoS ONE 01/2014; 9(3):e90754. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: CD4 count is a proxy for the extent of immune deficiency and declines in CD4 count are a measure of disease progression. Decline in CD4 count is an important component: for estimating benefits of ARV treatment; for individual level counselling on the rapidity of untreated disease progression and prognosis; and can be used in planning demand for health services. Our objective is to report CD4 decline and changes in viral load (VL) in a group of HIV-infected adults enrolled in a randomized trial of preventive treatment for TB in South Africa where clade C infection predominates. HIV-infected, tuberculin skin test positive adults who were not eligible for antiretroviral (ARV) treatment were randomized to a trial of preventive treatment from 2003-2005. VL and CD4 count were assessed at enrollment and CD4 counts repeated at least annually. During follow-up, individuals whose CD4 counts decreased to <200 cells/mm3 were referred for antiretroviral therapy (ART) and were analytically censored. 1106 ARV naïve adults were enrolled. Their median age was 30 years and male to female ratio was 1∶5. Median baseline CD4 count was 490 cells/mm3 (IQR 351-675). The overall mean decline in CD4 count was 61 cells/mm3 per annum. Adjusting for age, gender, baseline hemoglobin, smoking and alcohol use had little impact on the estimate of CD4 decline. However, VL at baseline had a major impact on CD4 decline. The percent decline in CD4 count was 13.3% (95% CI 12.0%, 14.7%), 10.6% (95% CI 8.8%, 12.4%), and 13.8% (95% CI 12.1%, 15.5%) per annum for baseline VLs of <10,000 (N = 314), 10,001-100,000 (N = 338), >100,000 (N = 122) copies/ml. Our data suggests that six and a half years will elapse for an individual's CD4 count to decline from 750 to 350 cells/mm3 in the absence of ART.
    PLoS ONE 01/2014; 9(5):e96369. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To report the incidence rates of TB and HIV in household contacts of index patients diagnosed with TB. A prospective cohort study in the Matlosana sub-district of North West Province, South Africa. Contacts of index TB patients received TB and HIV testing after counseling at their first household visit and were then followed up a year later, in 2010. TB or HIV diagnoses that occurred during the period were determined. For 2,377 household contacts, the overall observed TB incidence rate was 1.3 per 100 person years (95% CI 0.9-1.9/100py) and TB incidence for individuals who were HIV-infected and HIV seronegative at baseline was 5.4/100py (95% CI 2.9-9.0/100py) and 0.7/100py (95% CI 0.3-1.4/100py), respectively. The overall HIV incidence rate was 2.2/100py (95% CI 1.3-8.4/100py). In the year following a household case finding visit when household contacts were tested for TB and HIV, the incidence rate of both active TB and HIV infection was found to be extremely high. Clearly, implementing proven strategies to prevent HIV acquisition and preventing TB transmission and progression to disease remains a priority in settings such as South Africa.
    PLoS ONE 01/2014; 9(4):e95372. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background: Factors associated with mortality in HIV-infected people in sub-Saharan Africa are widely reported. However rural-urban disparities and their association with all-cause mortality remain unclear. Furthermore, commonly used classical Cox regression ignores unmeasured variables and frailty. Objective: To incorporate frailty in assessing factors associated with mortality in HIV-infected people in rural and urban South Africa. Design: Using data from a prospective cohort following 6,690 HIV-infected participants from Soweto (urban) and Mpumalanga (rural) enrolled from 2003 to 2010; covariates of mortality were assessed by the integrated nested Laplace approximation method. Results: We enrolled 2,221 (33%) rural and 4,469 (67%) urban participants of whom 1,555 (70%) and 3,480 (78%) were females respectively. Median age (IQR) was 36.4 (31.0-44.1) in rural and 32.7 (28.2-38.1) in the urban participants. The mortality rate per 100 person-years was 11 (9.7-12.5) and 4 (3.6-4.5) in the rural and urban participants, respectively. Compared to those not on HAART, rural participants had a reduced risk of mortality if on HAART for 6-12 (HR: 0.20, 95% CI: 0.10-0.39) and >12 months (HR: 0.10, 95% CI: 0.05-0.18). Relative to those not on HAART, urban participants had a lower risk if on HAART >12 months (HR: 0.35, 95% CI: 0.27-0.46). The frailty variance was significant and >1 in rural participants indicating more heterogeneity. Similarly it was significant but <1 in the urban participants indicating less heterogeneity. Conclusion: The frailty model findings suggest an elevated risk of mortality in rural participants relative to the urban participants potentially due to unmeasured variables that could be biological, socio-economic, or healthcare related. Use of robust methods that optimise data and account for unmeasured variables could be helpful in assessing the effect of unknown risk factors thus improving patient management and care in South Africa and elsewhere.
    Global Health Action 01/2014; 7:25488. · 2.06 Impact Factor
  • Source
    JAIDS Journal of Acquired Immune Deficiency Syndromes 01/2014; 65(1):e29-32. · 4.65 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The prevalence of illicitly traded cigarettes in South Africa has been reported to be 40-50%. However, these estimates do not account for the more nuanced characteristics of the illicit cigarette trade. With the goal of better understanding contraband cigarettes in South Africa, this study piloted three methods for assessing the price, brands, pack features and smoker's views about illicit cigarettes in five cities/towns. Data were collected in June and July 2012. A convenience sample of three South African cities (Johannesburg, Durban and Nelspruit) and two smaller towns (Musina and Ficksburg) were chosen for this study. Three cross-sectional approaches were used to assess the characteristics of contraband cigarettes: (1) a dummy purchase of cigarettes from informal retailers, (2) the collection of discarded cigarette packs and (3) a survey of tobacco smokers. For the purposes of the survey, 40 self-reported smokers were recruited at taxi ranks in each downtown site. Adults who were over the age of 18 were asked to verbally consent to participate in the study and answer a questionnaire administered by a researcher. The leading reason for labelling a pack as illicit in each city/town was the absence of an excise stamp (28.6% overall), and the least common reason was an illegal tar or nicotine level (11.1% overall). The overall proportion of informal vendors who sold illicit cigarettes was 41%. Singles and packs of 20 were consistently cheaper at informal vendors. Survey participants' responses reflected varied perspectives on illicit cigarettes and purchasing preferences. Each approach generated an interesting insight into physical aspects of illicit cigarettes. While this pilot study cannot be used to generate generalisable statistics on illicit cigarettes, more systematic surveys of this nature could inform researchers' and practitioners' initiatives to combat illicit and legal cigarette sales and usage.
    BMJ Open 01/2014; 4(5):e004562. · 2.06 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Many randomized and cohort studies have reported a survival benefit with cotrimoxazole prophylaxis without detecting a difference in tuberculosis (TB) incidence by cotrimoxazole status. However, several in vitro studies have reported that cotrimoxazole possesses anti-TB activity. We sought to compare TB incidence and TB diagnostic yield by cotrimoxazole use among participants in a well characterized cohort of HIV-infected adults living in a high TB prevalence region. We analyzed prospective data from a long-term longitudinal cohort of adults receiving HIV care and TB investigations in Soweto, South Africa. Using longitudinal analysis, we compared total and laboratory confirmed TB incidence by cotrimoxazole status as well as all-cause mortality. In addition, we compared TB culture results by cotrimoxazole status. In a multivariable analysis, adjusted for sex, body mass index, WHO clinical stage, time-updated CD4 count, and antiretroviral therapy status, we observed an association between cotrimoxazole and an increase in TB incidence (hazard ratio 1.7, 95% CI: 1.2, 2.2). However, when restricted to laboratory-confirmed TB, there was no association between cotrimoxazole and TB incidence (hazard ratio: 0.97, 95% CI: 0.39, 2.4). In TB cases, we found no difference in the proportion of positive sputum cultures or days to culture positivity by cotrimoxazole status. Cotrimoxazole was associated with a reduction in mortality. In this cohort with a mortality benefit from cotrimoxazole, we found an increased risk of all TB among individuals using cotrimoxazole, likely a result of residual confounding, but no association between use of cotrimoxazole and laboratory-confirmed TB. Cotrimoxazole did not compromise TB diagnosis.
    PLoS ONE 01/2014; 9(1):e83750. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Research in the predictors of all-cause mortality in HIV-infected people has widely been reported in literature. Making an informed decision requires understanding the methods used. We present a review on study designs, statistical methods and their appropriateness in original articles reporting on predictors of all-cause mortality in HIV-infected people between January 2002 and December 2011. Statistical methods were compared between 2002-2006 and 2007-2011. Time-to-event analysis techniques were considered appropriate. Pubmed/Medline. Original English-language articles were abstracted. Letters to the editor, editorials, reviews, systematic reviews, meta-analysis, case reports and any other ineligible articles were excluded. A total of 189 studies were identified (n = 91 in 2002-2006 and n = 98 in 2007-2011) out of which 130 (69%) were prospective and 56 (30%) were retrospective. One hundred and eighty-two (96%) studies described their sample using descriptive statistics while 32 (17%) made comparisons using t-tests. Kaplan-Meier methods for time-to-event analysis were commonly used in the earlier period (n = 69, 76% vs. n = 53, 54%, p = 0.002). Predictors of mortality in the two periods were commonly determined using Cox regression analysis (n = 67, 75% vs. n = 63, 64%, p = 0.12). Only 7 (4%) used advanced survival analysis methods of Cox regression analysis with frailty in which 6 (3%) were used in the later period. Thirty-two (17%) used logistic regression while 8 (4%) used other methods. There were significantly more articles from the first period using appropriate methods compared to the second (n = 80, 88% vs. n = 69, 70%, p-value = 0.003). Descriptive statistics and survival analysis techniques remain the most common methods of analysis in publications on predictors of all-cause mortality in HIV-infected cohorts while prospective research designs are favoured. Sophisticated techniques of time-dependent Cox regression and Cox regression with frailty are scarce. This motivates for more training in the use of advanced time-to-event methods.
    PLoS ONE 01/2014; 9(2):e87356. · 3.53 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: BACKGROUND Despite limited resources for HIV prevention interventions, few empirical studies investigate production efficiency of these services. We describe the methods of an ongoing project, which aims to assess average unit costs and determinants of efficiency for four HIV prevention interventions: prevention of mother-to-child transmission (PMTCT), voluntary medical male circumcision (VMMC), HIV testing and counseling (HTC), and HIV prevention for sex workers (SW) in Kenya, Rwanda, South Africa and Zambia. METHODS Our research methods were developed following: i) comprehensive review of methodologies on efficiency, ii) a peer review process, iii) consultations with governments and stakeholders and, iv) comments from implementing partners. RESULTS In each country, the study sample consists of 60-80 health facilities (40 sites per clinical-based intervention and 15 sites for SW). All relevant input costs and intervention output data are being collected for 2011 or 2012; data sources include registers, reported data, and time-motion methods. Process quality is captured from exit interviews, vignettes, and by measuring the facility's attrition rate at each stage along the service delivery cascade for each intervention. Instruments and tools were adapted to the local contexts, while still allowing cross-country comparisons. CONCLUSIONS Reliable data on costs and efficiency are critical to best inform budget allocation and financial decisions; this requires using rigorous estimation methods. This is one of the first studies to employ a complex package of survey methods to assess technical efficiency, especially in the context of multi-country HIV studies.
    141st APHA Annual Meeting and Exposition 2013; 11/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: In South Africa, the majority of tuberculosis (TB) patients are co-infected with the human immunodeficiency virus (HIV), and delays in diagnosis and treatment likely exacerbate morbidity and mortality. To determine predictors of delays in the diagnosis and treatment of hospitalised suspected pulmonary TB patients co-infected with HIV. Post-analysis of data collected in a three-centre prospective cohort of in-patients clinically diagnosed with active TB in three hospitals in South Africa between 2006 and 2009 during the first 24 h of admission. Delay was assessed by asking time of first symptoms and any prior health-seeking behaviour for this episode of illness. Data from a total of 891 participants with a median age of 36 years and a CD4 count of 67 cells/mm(3) were analysed. Median patient, system and total delays were respectively 28, 1 and 28 days. Unemployment, treatment at Tshepong Hospital, alcohol consumption, crowding index, seeking prior treatment, cotrimoxazole treatment and WHO Stage 4 disease predicted prolonged total delay. Patient delay in seeking care for TB in this high HIV prevalence setting is substantial. Factors identified with delay could be used to develop interventions to improve care seeking and earlier diagnosis of TB.
    The International Journal of Tuberculosis and Lung Disease 09/2013; 17(9):1199-205. · 2.76 Impact Factor

Publication Stats

1k Citations
522.18 Total Impact Points

Institutions

  • 2006–2014
    • University of the Witwatersrand
      • • Perinatal HIV Research Unit
      • • Division of Anatomical Pathology
      • • School of Public Health
      • • School of Electrical and Information Engineering
      Johannesburg, Gauteng, South Africa
  • 2007–2013
    • Johns Hopkins University
      • • Center for Tuberculosis Research
      • • Department of Medicine
      Baltimore, Maryland, United States
  • 2012
    • University of KwaZulu-Natal
      • Centre for the AIDS Programme of Research in South Africa (CAPRISA)
      Durban, KwaZulu-Natal, South Africa
  • 2010–2012
    • Johns Hopkins Bloomberg School of Public Health
      • Department of Health Policy and Management
      Baltimore, MD, United States
  • 2011
    • National University of Singapore
      • Department of Statistics and Applied Probability
      Singapore, Singapore
    • Johns Hopkins Medicine
      Baltimore, Maryland, United States
    • University of Rochester
      • Department of Community and Preventive Medicine
      Rochester, NY, United States
  • 2010–2011
    • Brown University
      • Alpert Medical School
      Providence, RI, United States
  • 2008
    • University of California, San Francisco
      San Francisco, California, United States
    • Alpert Medical School - Brown University
      Providence, Rhode Island, United States
    • Chris Hani Baragwanath Hospital
      Johannesburg, Gauteng, South Africa