Browsing by Author "Mehta, Ushma"

Now showing 1 - 11 of 11
  • Thumbnail Image
    Item
    Open Access
    Cohort Profile: The Western Cape Pregnancy Exposure Registry (WCPER)
    (2022-06) Kalk, Emma; Heekes, Alexa; Slogrove, Amy; Phelanyane, Florence; Davies, Mary-Ann; Myer, Landon; Euvrard, Jonathan; Kroon, Max; Petro, Greg; Fieggen, Karen; Stewart, Chantal; Rhoda, Natasha; Gebhardt, Stefan; Osman, Ayesha; Anderson, Kim; Boulle, Andrew; Mehta, Ushma
    Purpose: The Western Cape Pregnancy Exposure Registry (PER) was established at two public sector healthcare sentinel sites in the Western Cape province, South Africa, to provide ongoing surveillance of drug exposures in pregnancy and associations with pregnancy outcomes. Participants: Established in 2016, all women attending their first antenatal visit at primary care obstetric facilities were enrolled and followed to pregnancy outcome regardless of the site (ie, primary, secondary, tertiary facility). Routine operational obstetric and medical data are digitised from the clinical stationery at the healthcare facilities. Data collection has been integrated into existing services and information platforms and supports routine operations. The PER is situated within the Provincial Health Data Centre, an information exchange that harmonises and consolidates all health-related electronic data in the province. Data are contributed via linkage across a unique identifier. This relationship limits the missing data in the PER, allows validation and avoids misclassification in the population-level data set. Findings to date: Approximately 5000 and 3500 pregnant women enter the data set annually at the urban and rural sites, respectively. As of August 2021, >30 000 pregnancies have been recorded and outcomes have been determined for 93%. Analysis of key obstetric and neonatal health indicators derived from the PER are consistent with the aggregate data in the District Health Information System. Future plans: This represents significant infrastructure, able to address clinical and epidemiological concerns in a low/middle-income setting.
  • Thumbnail Image
    Item
    Open Access
    Effect of artemether-lumefantrine policy and improved vector control on malaria burden in KwaZulu-Natal, South Africa
    (Public Library of Science, 2005) Barnes, Karen I; Durrheim, David N; Little, Francesca; Jackson, Amanda; Mehta, Ushma; Allen, Elizabeth; Dlamini, Sicelo S; Tsoka, Joyce; Bredenkamp, Barry; Mthembu, D Jotham
    In KwaZulu-Natal strengthening of vector control and a change in antimalarial treatment policy to use of artemether-lumefantrine has been associated with a decrease in malaria cases, admissions, and deaths.
  • Thumbnail Image
    Item
    Open Access
    Effect of artemether-lumefantrine policy and improved vector control on malaria burden in KwaZulu-Natal, South Africa
    (2005) Barnes, Karen I; Durrheim, David N; Little, Francesca; Jackson, Amanda; Mehta, Ushma; Allen, Elizabeth; Dlamini, Sicelo S; Tsoka, Joyce; Bredekamp, Barry; Mthembu, D Jotham; White, Nicholas J; Sharp, Brian L
    Between 1995 and 2000, KwaZulu–Natal province, South Africa, experienced a marked increase in Plasmodium falciparum malaria, fuelled by pyrethroid and sulfadoxine-pyrimethamine resistance. In response, vector control was strengthened and artemether-lumefantrine (AL) was deployed in the first Ministry of Health artemisinin-based combination treatment policy in Africa. In South Africa, effective vector and parasite control had historically ensured low-intensity malaria transmission. Malaria is diagnosed definitively and treatment is provided free of charge in reasonably accessible public-sector health-care facilities.
  • Thumbnail Image
    Item
    Open Access
    Eliciting harms data from trial participants: how perceptions of illness and treatment mediate recognition of relevant information to report
    (BioMed Central Ltd, 2011) Allen, Elizabeth; Barnes, Karen; Mushi, Adiel; Massawe, Isolide; Staedke, Sarah; Mehta, Ushma; Vestergaard, Lasse; Lemnge, Martha; Chandler, Clare
    Background: There is no consensus on the ideal methodology for eliciting participant-reported harms, but question methods influence the extent and nature of data detected. This gives potential for measurement error and undermines meta-analyses of adverse effects. We undertook to identify barriers to accurate and complete reporting of harms data, by qualitatively exploring participants’ experiences of illness and treatment, and reporting behaviours; and compared the number and nature of data detected by three enquiry methods. Methods: Participants within antiretroviral/antimalarial interaction trials in South Africa and Tanzania were asked about medical history, treatments and/or adverse events by general enquiries followed by checklists. Those reporting differently between these two question methods were invited to an in-depth interview and focus group discussion. Health narratives were analysed to investigate accuracy and completeness of case record form data and to understand reasons for differential reporting between question methods. Outcomes were the number and nature of data by question method, themes from qualitative analyses and a theoretical interpretation of participants’ experiences. Results: We observed a cumulative increase in sensitivity of detection of all types of reports while progressing from general enquiry, through checklist, to in-depth interview. Questioning detail and terminology influenced participants’ recognition of health issues and treatments. Reporting patterns and vocabulary suggest influence from the relative importance that illnesses and treatments have for participants. Perceptions were often dichotomised (e.g. ‘street’ versus clinic treatments, symptoms experienced versus tests and examinations performed, chronic versus acute illness, persistent versus intermittent symptoms, activity- versus malaria-related symptoms) and this differentiation extended to ideas of relevance to report. South African participants displayed a ‘trial citizenship’, taking responsibility for the impact of their reporting on trial results, and even reaching reporting decisions by consensus. In contrast, Tanzanians perceived their role more as patients than participants; the locus of responsibility for knowing information relevant to the trial fell with trial staff as doctors rather than with themselves. Conclusions: Our observations of how reporting relates to participant perceptions inside and outside trials could help optimise how harms data are elicited. Questions reflecting the different ways that biomedically defined illness and treatment data are perceived by participants may help them understand relevance for reporting. We will theorise how these two disparate trial environments may have influenced how participants understood their role, as this could help researchers achieve empowered participation in similar trials.
  • Thumbnail Image
    Item
    Open Access
    Evaluating harm associated with anti-malarial drugs: a survey of methods used by clinical researchers to elicit, assess and record participant-reported adverse events and related data
    (BioMed Central Ltd, 2013) Allen, Elizabeth; Chandler, Clare; Mandimika, Nyaradzo; Pace, Cheryl; Mehta, Ushma; Barnes, Karen
    BACKGROUND:Participant reports of medical histories, adverse events (AE) and non-study drugs are integral to evaluating harm in clinical research. However, interpreting or synthesizing results is complicated if studies use different methods for ascertaining and assessing these data. To explore how these data are obtained in malaria drug studies, a descriptive online survey of clinical researchers was conducted during 2012 and 2013. METHODS: The survey was advertised through e-mails, collaborators and at conferences. Questions aimed to capture the detail, rationale and application of methods used to obtain relevant data within various study designs and populations. Closed responses were analysed using proportions, open responses through identifying repeating ideas and underlying concepts. RESULTS: Of fifty-two respondents from 25 counties, 87% worked at an investigational site and 75% reported about an interventional study. Studies employed a range of methods to elicit, assess and record participant-reported AEs and related data. Questioning about AEs in 31% of interventional studies was a combination of general (open questions about health) and structured (reference to specific health-related items), 26% used structured only and 18% general only. No observational studies used general questioning alone. A minority incorporated pictorial tools. Rationales for the questioning approach included: standardization of assessment or data capture, specificity or comprehensiveness of data sought, avoidance of suggestion, feasibility, and understanding participants' perceptions. Most respondents considered the approach they reported was optimal, though several reconsidered this. Four AE grading, and three causality assessment approaches were reported. Combining general and structured questions about non-study drug use were considered useful for revealing and identifying specific medicines, while pictures could enhance reports, particularly in areas of low literacy. CONCLUSIONS: It is critical to evaluate the safety of anti-malarial drugs being deployed in large, diverse populations. Many studies would be suitable for contributing to a larger body of evidence for answering questions on harm. However this survey showed that various methods are used to obtain relevant data, which could influence study results. As the best practices for obtaining such data are unclear, anti-malarial clinical researchers should work towards consensus about the selection and/or design of optimal methods.
  • Thumbnail Image
    Item
    Open Access
    How experiences become data: the process of eliciting adverse event, medical history and concomitant medication reports in antimalarial and antiretroviral interaction trials
    (BioMed Central Ltd, 2013) Allen, Elizabeth; Mushi, Adiel; Massawe, Isolide; Vestergaard, Lasse; Lemnge, Martha; Staedke, Sarah; Mehta, Ushma; Barnes, Karen; Chandler, Clare
    BACKGROUND:Accurately characterizing a drug's safety profile is essential. Trial harm and tolerability assessments rely, in part, on participants' reports of medical histories, adverse events (AEs), and concomitant medications. Optimal methods for questioning participants are unclear, but different methods giving different results can undermine meta-analyses. This study compared methods for eliciting such data and explored reasons for dissimilar participant responses. METHODS: Participants from open-label antimalarial and antiretroviral interaction trials in two distinct sites (South Africa, n=18 [all HIV positive]; Tanzania, n=80 [86% HIV positive]) were asked about ill health and treatment use by sequential use of (1) general enquiries without reference to particular conditions, body systems or treatments, (2) checklists of potential health issues and treatments, (3) in-depth interviews. Participants' experiences of illness and treatment and their reporting behaviour were explored qualitatively, as were trial clinicians' experiences with obtaining participant reports. Outcomes were the number and nature of data by questioning method, themes from qualitative analyses and a theoretical interpretation of participants' experiences. RESULTS: There was an overall cumulative increase in the number of reports from general enquiry through checklists to in-depth interview; in South Africa, an additional 12 medical histories, 21 AEs and 27 medications; in Tanzania an additional 260 medical histories, 1 AE and 11 medications. Checklists and interviews facilitated recognition of health issues and treatments, and consideration of what to report. Information was sometimes not reported because participants forgot, it was considered irrelevant or insignificant, or they feared reporting. Some medicine names were not known and answers to questions were considered inferior to blood tests for detecting ill health. South African inpatient volunteers exhibited a "trial citizenship", working to achieve researchers' goals, while Tanzanian outpatients sometimes deferred responsibility for identifying items to report to trial clinicians. CONCLUSIONS: Questioning methods and trial contexts influence the detection of adverse events, medical histories and concomitant medications. There should be further methodological work to investigate these influences and find appropriate questioning methods.
  • Thumbnail Image
    Item
    Open Access
    Influences on participant reporting in the World Health Organisation drugs exposure pregnancy registry; a qualitative study
    (BioMed Central, 2014-10-31) Allen, Elizabeth N; Gomes, Melba; Yevoo, Lucy; Egesah, Omar; Clerk, Christine; Byamugisha, Josaphat; Mbonye, Anthony; Were, Edwin; Mehta, Ushma; Atuyambe, Lynn M
    Background: The World Health Organisation has designed a pregnancy registry to investigate the effect of maternal drug use on pregnancy outcomes in resource-limited settings. In this sentinel surveillance system, detailed health and drug use data are prospectively collected from the first antenatal clinic visit until delivery. Over and above other clinical records, the registry relies on accurate participant reports about the drugs they use. Qualitative methods were incorporated into a pilot registry study during 2010 and 2011 to examine barriers to women reporting these drugs and other exposures at antenatal clinics, and how they might be overcome. Methods: Twenty-seven focus group discussions were conducted in Ghana, Kenya and Uganda with a total of 208 women either enrolled in the registry or from its source communities. A question guide was designed to uncover the types of exposure data under- or inaccurately reported at antenatal clinics, the underlying reasons, and how women prefer to be asked questions. Transcripts were analysed thematically. Results: Women said it was important for them to report everything they had used during pregnancy. However, they expressed reservations about revealing their consumption of traditional, over-the-counter medicines and alcohol to antenatal staff because of anticipated negative reactions. Some enrolled participants' improved relationship with registry staff facilitated information sharing and the registry tools helped overcome problems with recall and naming of medicines. Decisions about where women sought care, which influenced medicines used and antenatal clinic attendance, were influenced by pressure within and outside of the formal healthcare system to conform to conflicting behaviours. Conversations also reflected women's responsibilities for producing a healthy baby. Conclusions: Women in this study commonly take traditional medicines in pregnancy, and to a lesser extent over-the-counter medicines and alcohol. The World Health Organisation pregnancy registry shows potential to enhance their reporting of these substances at the antenatal clinic. However, more work is needed to find optimal techniques for eliciting accurate reports, especially where the detail of constituents may never be known. It will also be important to find ways of sustaining such drug exposure surveillance systems in busy antenatal clinics.
  • Thumbnail Image
    Item
    Open Access
    Noma in northwest Nigeria: a neglected disease in neglected populations
    (2020) Farley, Elise Sarah; Mehta, Ushma; Lenglet, Annick; Davies, Mary-Ann
    Background Noma, also known as cancrum oris, is a gangrenous infection of the oral cavity, which causes widespread orofacial destruction. If untreated, noma has a reported 90% mortality rate within weeks after the onset of first symptoms. Noma progresses through distinct stages defined by the World Health Organisation (WHO); Stage 0: simple gingivitis; Stage 1: acute necrotizing gingivitis; Stage 2: oedema; Stage 3: gangrene; Stage 4: scarring. Stage 5: sequelae. It is unclear how many patients with the early stages of noma will progress to the later stages of disease. Treatment in the early reversible stages with antibiotics, wound debridement and nutritional support greatly reduces morbidity and mortality. Acute noma is most often reported in children aged between two and five years. Many patients who survive the acute stages of the disease suffer into adulthood with disfigurement and disability of varying degrees. Noma is thought to be most prevalent in developing countries in Africa and Asia. Estimates for noma prevalence and incidence vary. In 1998, the WHO estimated an annual incidence of 140,000 cases of acute noma and 770,000 noma survivors living with sequelae. Two Nigerian studies estimated the burden of disease ranged from seven cases per 1,000 children aged between one and 16 years (2003) to 6.4 per 1,000 children (2003). A study from 2019 estimated the period prevalence of noma from 2010 to 2018 was 1.6 per 100,000 population at risk in Nigeria. These estimates are based on expert opinion, number of hospital admissions and retrospectively collected hospital-based data and it is unclear which stages of noma were included. Risk factors for the disease include poor oral hygiene, malnutrition, comorbidities and low socioeconomic status. Despite its ancient history (reported by Hippocrates (460 - 370 BC)), noma-related literature remains mainly confined to case reports and case series. By employing both qualitative and quantitative methods, we sought to examine the biopsychosocial features of noma, its epidemiology and treatment in northwest Nigeria in order to inform advocacy and prevention efforts. The three overarching objectives to fulfil this aim were to assess the distribution of noma among children in northwest Nigeria; identify factors associated with noma (including factors influencing health-seeking behaviour and risk factors for the development of noma) and gain an understanding of the biomedical and non-biomedical care provided to noma patients in this setting. The knowledge gained through this thesis will support the assessment of the need for advocacy around noma, effective resource allocation and the planning of intervention strategies. Methods We conducted a scoping literature review, three quantitative studies (risk factors, outcomes, prevalence) and two qualitative studies (language and beliefs and traditional healing practices) in northwest Nigeria. Data were collected from patient caretakers at the Noma Children's Hospital, hospital staff, children and traditional healers in villages within Sokoto and Kebbi States. Data collection methods included quantitative surveys, oral screenings, anthropometric measurements, quality of life questionnaires, qualitative in-depth interviews and focus group discussions. Consenting adult respondents answered questionnaires and participated in interviews, and where applicable, data was collected from assenting children. Quantitative analyses included descriptive statistics as well as univariable and multivariable risk factor analyses. Qualitative data was manually coded and analysed thematically. Findings We included 74 cases (noma patients presenting at the hospital in the year preceding data collection) and 222 controls (both median age of five years (inter-quartile range 3, 15 years)) in the risk factor study. Vaccination coverage for polio and measles was below 7% in both cases and controls. The multivariable analysis identified the child being fed pap every day (adjusted odds ratio (aOR) 9.8; 95% confidence interval (CI 1.5, 62.7) as a risk factor. The mother being the primary caretaker (aOR 0.08; CI 0.01, 0.5) and the caretaker being married (aOR 0.006; CI 0.0006, 0.5) were protective factors. Of the 37 patients with noma sequelae included in the outcomes study, 21 (56.8%) were male and 22 (62.9%) were aged six years or older. Fifteen patients (40.5%) had two to three surgeries. The most frequently used surgical procedure was a deltopectoral flap (n=16 patients; 43.2%). Trismus was released in 12 patients (32.4%), of these; none had a normal mouth opening at the follow-up visit. Despite this finding, all respondents reported that the surgery had improved their quality of life. In the cross-sectional study assessing the prevalence of all stages of noma, we included 3,499 households and 7,122 children aged <15 years; 4,239 (59.8%) were aged 0 to 5 years. Simple gingivitis was identified in 3.1% (n=181; CI 2.6-3.8), acute necrotizing gingivitis in 0.1% (n=10; CI 0.1-0.3), and oedema in 0.05% (n=3; CI 0.02-0.2). No cases of late-stage noma were detected. Naming of the disease differed between caretakers and healthcare workers in the language and beliefs study. Beliefs about the causes of noma were varied (spirits, animals, insects, previous infections). Noma patient caretakers spoke of the mental health strain due to stigmatization as a key issue. Difficulty in accessing care was evident. A lack of trust in the health system was mentioned as a barrier to care. Traditional healers offered specialised forms of care for specific conditions and referral guidance. They viewed the stages of noma as different conditions with individualised remedies and were willing to refer noma patients. Caretakers trusted traditional healers. Conclusion Social conditions and childhood feeding practices are associated with the occurrence of noma in northwest Nigeria. This thesis has shown that following their last surgical intervention, noma patients do experience some improvements in their quality of life, but continue to face functional challenges that inhibit their daily life. We found many, widely distributed, early-stage noma cases in northwest Nigeria indicating a large population at risk of progressing to the later stages of disease. Caretaker and practitioner perspectives may enlighten efforts to improve case finding, and to understand barriers to accessing health care. Differences in disease naming illustrated the difference in beliefs about the disease. Traditional healers could play a crucial role in the early detection of noma and the health-seeking decision-making process of patients. Intervention programmes should include traditional healers through training and referral partnerships. In conclusion, this thesis provides a unique view of the biopsychosocial features, epidemiology and treatment options for noma in northwest Nigeria. Noma is a disease, which is indicative of a weak health system and socio-economic environments of extreme deprivation. Intervention programmes should include widespread health system improvements that could address a host of risk factors for noma, and simultaneously other childhood diseases. These include increasing access to quality health care (including vaccinations), ensuring effective referral mechanisms, predominantly in rural areas, and the creation of a robust surveillance network. Health financing initiatives would need to be paired with these improvements. Nutritional programs aimed at caretakers of young children and community-based oral health initiatives could be effective mechanisms to curb the number of noma cases. Awareness-building initiatives targeting healthcare workers and community members are necessary to improve the detection and timely management of noma in endemic settings. The combined findings of this thesis highlight the neglected nature of noma and make a strong case for placing noma on the WHO neglected tropical diseases list. This initiative could foster awareness among policy-makers and governments and direct much needed funding to facilitate further research, surveillance and targeted health interventions that would contribute to the eradication of noma.
  • Thumbnail Image
    Item
    Open Access
    Protocol for a drugs exposure pregnancy registry for implementation in resource-limited settings
    (BioMed Central Ltd, 2012) Mehta, Ushma; Clerk, Christine; Allen, Elizabeth; Yore, Mackensie; Sevene, Esperanca; Singlovic, Jan; Petzold, Max; Mangiaterra, Viviana; Elefant, Elizabeth; Sullivan, Frank; Holmes, Lewis; Gomes, Melba
    BACKGROUND: The absence of robust evidence of safety of medicines in pregnancy, particularly those for major diseases provided by public health programmes in developing countries, has resulted in cautious recommendations on their use. We describe a protocol for a Pregnancy Registry adapted to resource-limited settings aimed at providing evidence on the safety of medicines in pregnancy.METHODS/DESIGN:Sentinel health facilities are chosen where women come for prenatal care and are likely to come for delivery. Staff capacity is improved to provide better care during the pregnancy, to identify visible birth defects at delivery and refer infants with major anomalies for surgical or clinical evaluation and treatment. Consenting women are enrolled at their first antenatal visit and careful medical, obstetric and drug-exposure histories taken; medical record linkage is encouraged. Enrolled women are followed up prospectively and their histories are updated at each subsequent visit. The enrolled woman is encouraged to deliver at the facility, where she and her baby can be assessed.DISCUSSION:In addition to data pooling into a common WHO database, the WHO Pregnancy Registry has three important features: First is the inclusion of pregnant women coming for antenatal care, enabling comparison of birth outcomes of women who have been exposed to a medicine with those who have not. Second is its applicability to resource-poor settings regardless of drug or disease. Third is improvement of reproductive health care during pregnancies and at delivery. Facility delivery enables better health outcomes, timely evaluation and management of the newborn, and the collection of reliable clinical data. The Registry aims to improve maternal and neonatal care and also provide much needed information on the safety of medicines in pregnancy.
  • Thumbnail Image
    Item
    Open Access
    Self-Reported Antenatal Medicines Use Among Women Living with and Without HIV in Western Cape, South Africa: A Sub-Analysis of the B Positive Cohort Study
    (2022) Elzouki, Zaineb; Mehta, Ushma; Blockman, Marc
    Globally, innumerable women take medication while they are pregnant, and this trend is growing. The pipeline of medicines targeting maternal comorbidities is expanding. However, for most medicines, there is insufficient data on their safety in pregnancy. In addition, women may be taking medication for chronic or acute conditions before they recognize that they are pregnant. This study compared the self-reported pattern of medicine use during the course of pregnancy in a cohort of pregnant women either living with or without HIV; seeking care at Gugulethu primary health care obstetric clinic in Western Cape, South Africa. Data on medicine use was collected over 3 antenatal visits. Medications reported were manually classified and coded by a clinical pharmacist and medical doctor. Structured interviews using a detailed questionnaire on medication use were administered to n=989 pregnant women. Women who had an ectopic pregnancy or an elective termination of pregnancy (TOP) were excluded from the analysis. 982 of these women were included in our analysis (n=507 HIV-negative and n=475 HIV positive). Of these, 39 (4.0%) did not report taking any medicine during pregnancy. Most 907 (92.3%) pregnant women reported using at least one over-the-counter medicine (OTC) and the majority, 601 (61.2%), at least one prescription medicine. A total of 36 (3.7%) reported using at least one herbal or traditional medicine over the course of the pregnancy. Pregnant women living with HIV were significantly less likely to report use of OTC medicine (56.2% vs 77.7%, p=<0.001). Pregnant women living with HIV also reported less herbal medicine use (2.9% vs 4.7%, p=0.07) compared to pregnant women living without HIV, though the effect was non-significant within this sample. Excluding antiretroviral medicines, prescription medicine use was essentially the same among pregnant women living with and without HIV (30.5% vs 30.2%, p=0.96). Exposure to medicines known to be potentially teratogenic or unsafe in pregnancy was reported in 300 (30.65%) pregnant women, with aspirin 238 (24.2%) and nonsteroidal anti-inflammatory medicines 46 (4.7%) medicines being the most reported. This study provides valuable information on self-reported medication use among pregnant women living with and without HIV in a South African primary healthcare setting. Medicine use was widespread in the study cohort, particularly OTC, with high prevalence of potentially unsafe medicines used during pregnancy. Our finding highlights the urgent need to build awareness around rational and safe medicine use among antenatal staff; pharmacists; and women of child-bearing age in South Africa, encouraging the taking of a thorough history of medicine exposure throughout the antenatal period.
  • Thumbnail Image
    Item
    Open Access
    Serious adverse drug reactions at two children’s hospitals in South Africa
    (2020-01-04) Mouton, Johannes P; Fortuin-de Smidt, Melony C; Jobanputra, Nicole; Mehta, Ushma; Stewart, Annemie; de Waal, Reneé; Technau, Karl-Günter; Argent, Andrew; Kroon, Max; Scott, Christiaan; Cohen, Karen
    Abstract Background The high HIV prevalence in South Africa may potentially be shaping the local adverse drug reaction (ADR) burden. We aimed to describe the prevalence and characteristics of serious ADRs at admission, and during admission, to two South African children’s hospitals. Methods We reviewed the folders of children admitted over sequential 30-day periods in 2015 to the medical wards and intensive care units of each hospital. We identified potential ADRs using a trigger tool developed for this study. A multidisciplinary team assessed ADR causality, type, seriousness, and preventability through consensus discussion. We used multivariate logistic regression to explore associations with serious ADRs. Results Among 1050 patients (median age 11 months, 56% male, 2.8% HIV-infected) with 1106 admissions we found 40 serious ADRs (3.8 per 100 drug-exposed admissions), including 9/40 (23%) preventable serious ADRs, and 8/40 (20%) fatal or near-fatal serious ADRs. Antibacterials, corticosteroids, psycholeptics, immunosuppressants, and antivirals were the most commonly implicated drug classes. Preterm neonates and children in middle childhood (6 to 11 years) were at increased risk of serious ADRs compared to infants (under 1 year) and term neonates: adjusted odds ratio (aOR) 5.97 (95% confidence interval 1.30 to 27.3) and aOR 3.63 (1.24 to 10.6) respectively. Other risk factors for serious ADRs were HIV infection (aOR 3.87 (1.14 to 13.2) versus HIV-negative) and increasing drug count (aOR 1.08 (1.04 to 1.12) per additional drug). Conclusions Serious ADR prevalence in our survey was similar to the prevalence found elsewhere. In our setting, serious ADRs were associated with HIV-infection and the antiviral drug class was one of the most commonly implicated. Similar to other sub-Saharan African studies, a large proportion of serious ADRs were fatal or near-fatal. Many serious ADRs were preventable.