Cohort profile: the South African National Health Laboratory Service (NHLS) National HIV Cohort.

PURPOSE: South Africa's National Health Laboratory Service (NHLS) National HIV Cohort was established in 2015 to facilitate monitoring, evaluation and research on South Africa's National HIV Treatment Programme. In South Africa, 84.8% of people living with HIV know their HIV status; 70.7% who know their status are on ART; and 87.4% on ART are virologically suppressed. PARTICIPANTS: The NHLS National HIV Cohort includes the laboratory data of nearly all patients receiving HIV care in the public sector since April 2004. Patients are included in the cohort if they have received a CD4 count or HIV RNA viral load (VL) test. Using an anonymised unique patient identifier that we have developed and validated to linked test results, we observe patients prospectively through their laboratory results as they receive HIV care and treatment. Patients in HIV care are seen for laboratory monitoring every 6-12 months. Data collected include age, sex, facility location and test results for CD4 counts, VLs and laboratory tests used to screen for potential treatment complications. FINDINGS TO DATE: From April 2004 to April 2018, 63 million CD4 count and VL tests were conducted at 5483 facilities. 12.6 million unique patients had at least one CD4 count or VL, indicating they had accessed HIV care, and 7.1 million patients had a VL test indicating they had started antiretroviral therapy. The creation of NHLS National HIV Cohort has enabled longitudinal research on all lab-monitored patients in South Africa's national HIV programme, including analyses of (1) patient health at presentation; (2) care outcomes such as 'CD4 recovery', 'retention in care' and 'viral resuppression'; (3) patterns of transfer and re-entry into care; (4) facility-level variation in care outcomes; and (5) impacts of policies and guideline changes. FUTURE PLANS: Continuous updating of the cohort, integration with available clinical data, and expansion to include tuberculosis and other lab-monitored comorbidities.

Iron status and self-reported fatigue in blood donors.

Fatigue is a reported symptom of iron depletion, but studies in blood donors show no conclusive link. We conducted an observational analysis of data from the STRIDE randomized trial to evaluate association of iron status with self-reported fatigue. STUDY DESIGN AND METHODS: Three blood centers randomly assigned 692 frequent donors to education or iron supplementation treatments. Biomarkers for iron status were measured during 20 to 24 months of follow-up. A fatigue score was derived from an 11-item questionnaire at baseline and final visits, and associations between iron status and fatigue were assessed. RESULTS: Final lab and questionnaire data were evaluable from 337 subjects. At baseline, female sex, older age, and anemia were associated with fatigue, but iron status was not. Mean (±SD) fatigue score change was 0.0 (±0.5). Mean (±SD) increase in iron stores was 1.0 (±3.5) mg/kg, but changes in body iron stores were not associated with fatigue score changes (0.01 per mg/kg; 95% CI, -0.01 to 0.02) or with fatigue (RR, 1.01; 95% CI, 0.99 to 1.04). The only factor associated with fatigue score changes was baseline fatigue (0.36; 95% CI, 0.25 to 0.48). CONCLUSION: Among high-frequency donors, neither iron status at baseline nor changes in iron status predicted fatigue during follow-up, with improvements limited to those with higher levels of baseline fatigue. Assessment of the association between iron and fatigue in blood donors benefits from careful consideration of study design and the study population.

Adolescent HIV treatment in South Africa's national HIV programme: a retrospective cohort study.

BACKGROUND: The number of South African adolescents receiving HIV care and treatment in South Africa is growing. By use of routinely collected laboratory data from South Africa's National HIV Programme, we aimed to quantify the numbers of adolescents accessing HIV care and treatment over time, characterise the role of perinatal infection in these trends, and estimate proportions of adolescents seeking HIV care and antiretroviral therapy (ART) in South Africa's public sector. METHODS: We did a retrospective, descriptive cohort study of children and adolescents aged 1-19 years accessing care in South Africa's public sector HIV treatment programme from 2005 to 2016 with a CD4 cell count or viral load recorded in South Africa's National Health Laboratory Service database. We estimated the total number of children and adolescents entering HIV care with a CD4 cell count or viral load test result by calendar period, as well as the proportion in care and receiving ART with at least one viral load test result. We stratified analyses by gender and by whether the patient entered care at younger than 15 years (probably perinatally infected) or at 15-19 years (probably infected in adolescence). FINDINGS: We identified 730 882 patients aged 1-19 years at entry to care between Jan 1, 2005, and Dec 31, 2016. 209 205 (54%) of 388 439 patients entering care younger than 15 years and 301 242 (88%) of 342 443 patients entering care aged 15-19 were female. During the study period, the number of virologically monitored patients aged 15-19 years receiving ART increased from 7949 in 2005-08 to 80 918 in 2013-16. 92 783 (66%) of 140 028 patients aged 15-19 years seeking care started ART by 2016, well below UNAID's target of ART for 90% of those diagnosed. We project that the number of adolescents on ART will continue to rise. INTERPRETATION: The many adolescents aged 15-19 years receiving ART reflect the ageing of children entering care at ages 1-14 years, and increases in care-seeking among horizontally infected adolescents aged 15-19 years. However, many adolescents seeking care do not start ART, suggesting an urgent need for interventions to increase uptake of ART and improve services for this population. FUNDING: US National Institutes of Health, and the President's Emergency Plan for AIDS Relief through the US Agency for International Development.

Simplified clinical algorithm for identifying patients eligible for same-day HIV treatment initiation (SLATE): Results from an individually randomized trial in South Africa and Kenya.

BACKGROUND: The World Health Organization recommends "same-day" initiation of antiretroviral therapy (ART) for HIV patients who are eligible and ready. Identifying efficient, safe, and feasible procedures for determining same-day eligibility and readiness is now a priority. The Simplified Algorithm for Treatment Eligibility (SLATE) study evaluated a clinical algorithm that allows healthcare workers to determine eligibility for same-day treatment and to initiate ART at the patient's first clinic visit. METHODS AND FINDINGS: SLATE was an individually randomized trial at three outpatient clinics in urban settlements in Johannesburg, South Africa and three hospital clinics in western Kenya. Adult, nonpregnant, HIV-positive, ambulatory patients presenting for any HIV care, including HIV testing, but not yet on ART were enrolled and randomized to the SLATE algorithm arm or standard care. The SLATE algorithm used four screening tools-a symptom self-report, medical history questionnaire, physical examination, and readiness assessment-to ascertain eligibility for same-day initiation or refer for further care. Follow-up was by record review, and analysis was conducted by country. We report primary outcomes of 1) ART initiation ≤28 days and 2) initiation ≤28 days and retention in care ≤8 months of enrollment. From March 7, 2017 to April 17, 2018, we enrolled 600 patients (median [IQR] age 34 [29-40] and CD4 count 286 [128-490]; 63% female) in South Africa and 477 patients in Kenya (median [IQR] age 35 [29-43] and CD4 count 283 [117-541]; 58% female). In the intervention arm, 78% of patients initiated ≤28 days in South Africa, compared to 68% in the standard arm (risk difference [RD] [95% confidence interval (CI)] 10% [3%-17%]); in Kenya, 94% of intervention-arm patients initiated ≤28 days compared to 89% in the standard arm (6% [0.5%-11%]). By 8 months in South Africa, 161/298 (54%) intervention-arm patients had initiated and were retained, compared to 146/302 (48%) in the standard arm (6% [(2% to 14%]). By 8 months in Kenya, the corresponding retention outcomes were identical in both arms (137/240 [57%] of intervention-arm patients and 136/237 [57%] of standard-arm patients). Limitations of the trial included limited geographic representativeness, exclusion of patients too ill to participate, missing viral load data, greater study fidelity to the algorithm than might be achieved in standard care, and secular changes in standard care over the course of the study. CONCLUSIONS: In South Africa, the SLATE algorithm increased uptake of ART within 28 days by 10% and showed a numerical increase (6%) in retention at 8 months. In Kenya, the algorithm increased uptake of ART within 28 days by 6% but found no difference in retention at 8 months. Eight-month retention was poor in both arms and both countries. These results suggest that a simple structured algorithm for same-day treatment initiation procedures is feasible and can increase and accelerate ART uptake but that early retention on treatment remains problematic. TRIAL REGISTRATION: Clinicaltrials.gov NCT02891135, registered September 1, 2016. First participant enrolled March 6, 2017 in South Africa and July 13, 2017 in Kenya.

Health facility and skilled birth deliveries among poor women with Jamkesmas health insurance in Indonesia: a mixed-methods study.

BACKGROUND: The growing momentum for quality and affordable health care for all has given rise to the recent global universal health coverage (UHC) movement. As part of Indonesia's strategy to achieve the goal of UHC, large investments have been made to increase health access for the poor, resulting in the implementation of various health insurance schemes targeted towards the poor and near-poor, including the Jamkesmas program. In the backdrop of Indonesia's aspiration to reach UHC is the high rate of maternal mortality that disproportionally affects poor women. The objective of this study was to evaluate the association of health facility and skilled birth deliveries among poor women with and without Jamkesmas and explore perceived barriers to health insurance membership and maternal health service utilization. METHODS: We used a mixed-methods design. Utilizing data from the 2012 Indonesian Demographic and Health Survey (n = 45,607), secondary analysis using propensity score matching was performed on key outcomes of interest: health facility delivery (HFD) and skilled birth delivery (SBD). In-depth interviews (n = 51) were conducted in the provinces of Jakarta and Banten among poor women, midwives, and government representatives. Thematic framework analysis was performed on qualitative data to explore perceived barriers. RESULTS: In 2012, 63.0% of women did not have health insurance; 19.1% had Jamkesmas. Poor women with Jamkesmas were 19% (OR = 1.19 [1.03-1.37]) more likely to have HFD and 17% (OR = 1.17 [1.01-1.35]) more likely to have SBD compared to poor women without insurance. Qualitative interviews highlighted key issues, including: lack of proper documentation for health insurance registration; the preference of pregnant women to deliver in their parents' village; the use of traditional birth attendants; distance to health facilities; shortage of qualified health providers; overcrowded health facilities; and lack of health facility accreditation. CONCLUSIONS: Poor women with Jamkesmas membership had a modest increase in HFD and SBD. These findings are consistent with economic theory that health insurance coverage can reduce financial barriers to care and increase service uptake. However, factors such as socio-cultural beliefs, accessibility, and quality of care are important elements that need to be addressed as part of the national UHC agenda to improve maternal health services in Indonesia.

Imputing HIV treatment start dates from routine laboratory data in South Africa: a validation study.

BACKGROUND: Poor clinical record keeping hinders health systems monitoring and patient care in many low resource settings. We develop and validate a novel method to impute dates of antiretroviral treatment (ART) initiation from routine laboratory data in South Africa's public sector HIV program. This method will enable monitoring of the national ART program using real-time laboratory data, avoiding the error potential of chart review. METHODS: We developed an algorithm to impute ART start dates based on the date of a patient's "ART workup", i.e. the laboratory tests used to determine treatment readiness in national guidelines, and the time from ART workup to initiation based on clinical protocols (21 days). To validate the algorithm, we analyzed data from two large clinical HIV cohorts: Hlabisa HIV Treatment and Care Programme in rural KwaZulu-Natal; and Right to Care Cohort in urban Gauteng. Both cohorts contain known ART initiation dates and laboratory results imported directly from the National Health Laboratory Service. We assessed median time from ART workup to ART initiation and calculated sensitivity (SE), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) of our imputed start date vs. the true start date within a 6 month window. Heterogeneity was assessed across individual clinics and over time. RESULTS: We analyzed data from over 80,000 HIV-positive adults. Among patients who had a workup and initiated ART, median time to initiation was 16 days (IQR 7,31) in Hlabisa and 21 (IQR 8,43) in RTC cohort. Among patients with known ART start dates, SE of the imputed start date was 83% in Hlabisa and 88% in RTC, indicating this method accurately predicts ART start dates for about 85% of all ART initiators. In Hlabisa, PPV was 95%, indicating that for patients with a lab workup, true start dates were predicted with high accuracy. SP (100%) and NPV (92%) were also very high. CONCLUSIONS: Routine laboratory data can be used to infer ART initiation dates in South Africa's public sector. Where care is provided based on protocols, laboratory data can be used to monitor health systems performance and improve accuracy and completeness of clinical records.

Tenofovir stock shortages have limited impact on clinic- and patient-level HIV treatment outcomes in public sector clinics in South Africa.

OBJECTIVE: Using data from four public sector clinics in South Africa, we sought to investigate provider- and patient-level outcomes, to understand how the 2012 tenofovir stock shortage affected the HIV care and monitoring of ART patients. METHODS: Prospective cohort analysis of ART-naïve, non-pregnant, HIV-infected patients >18 years initiating first-line ART between 1 July 2011-31 March 2013. Linear regression was used for all outcomes (number of ART initiates, days between pharmacy visits, transfers, single-drug substitutions, treatment interruptions, missed pharmacy visits, loss to follow-up and elevated viral load). We fit splines to smooth curves with knots at the beginning (1 February 2012) and end (31 August 2012) of the stock shortage and displayed results graphically by clinic. Difference-in-difference models were used to evaluate the effect of the stock shortage on outcomes. RESULTS: Results suggest a potential shift in the management of patients during the shortage, mainly fewer average days between visits during the shortage vs. before or after at all four clinics, and a significant difference in the proportion of patients missing visits during vs. before (RD: 1.2%; 95% CI: 0.5%, 2.0%). No significant difference was seen in other outcomes. CONCLUSION: While South Africa has made great strides to extend access to ART and increase the quality of the health services provided, patient care can be affected when stock shortages/outs occur. While our results show little effect on treatment outcomes, this most likely reflects the clinics' ability to mitigate the crisis by continuing to keep patient care and treatment as consistent as possible.

Antiretroviral therapy, labor productivity, and sex: a longitudinal cohort study of tea pluckers in Kenya.

OBJECTIVE: To estimate the impact of antiretroviral therapy (ART) on labor productivity and income using detailed employment data from two large tea plantations in western Kenya for HIV-infected tea pluckers who initiated ART. DESIGN: Longitudinal study using primary data on key employment outcomes for a group of HIV-infected workers receiving antiretroviral therapy (ART) and workers in the general workforce. METHODS: We used nearest-neighbor matching methods to estimate the impacts of HIV/AIDS and ART among 237 HIV-positive pluckers on ART (index group) over a 4-year period (2 years pre-ART and post-ART) on 4 monthly employment outcomes - days plucking tea, total kilograms (kgs) harvested, total days working, and total labor income. Outcomes for the index group were compared with those for a matched reference group from the general workforce. RESULTS: We observed a rapid deterioration in all four outcomes for HIV-infected individuals in the period before ART initiation and then a rapid improvement after treatment initiation. By 18-24 months after treatment initiation, the index group harvested 8% (men) and 19% (women) less tea than reference individuals. The index group earned 6% (men) and 9% (women) less income from labor than reference individuals. Women's income would have dropped further if they had not been able to offset their decline in tea plucking by spending more time on nonplucking assignments. CONCLUSION: HIV-infected workers experienced long-term income reductions before and after initiating ART. The implications of such long-term impacts in low-income countries have not been adequately addressed.

Effect of training traditional birth attendants on neonatal mortality (Lufwanyama Neonatal Survival Project): randomised controlled study.

OBJECTIVE: To determine whether training traditional birth attendants to manage several common perinatal conditions could reduce neonatal mortality in the setting of a resource poor country with limited access to healthcare. DESIGN: Prospective, cluster randomised and controlled effectiveness study. SETTING: Lufwanyama, an agrarian, poorly developed district located in the Copperbelt province, Zambia. All births carried out by study birth attendants occurred at mothers' homes, in rural village settings. PARTICIPANTS: 127 traditional birth attendants and mothers and their newborns (3559 infants delivered regardless of vital status) from Lufwanyama district. INTERVENTIONS: Using an unblinded design, birth attendants were cluster randomised to intervention or control groups. The intervention had two components: training in a modified version of the neonatal resuscitation protocol, and single dose amoxicillin coupled with facilitated referral of infants to a health centre. Control birth attendants continued their existing standard of care (basic obstetric skills and use of clean delivery kits). MAIN OUTCOME MEASURES: The primary outcome was the proportion of liveborn infants who died by day 28 after birth, with rate ratios statistically adjusted for clustering. Secondary outcomes were mortality at different time points; and comparison of causes of death based on verbal autopsy data. RESULTS: Among 3497 deliveries with reliable information, mortality at day 28 after birth was 45% lower among liveborn infants delivered by intervention birth attendants than control birth attendants (rate ratio 0.55, 95% confidence interval 0.33 to 0.90). The greatest reductions in mortality were in the first 24 hours after birth: 7.8 deaths per 1000 live births for infants delivered by intervention birth attendants compared with 19.9 per 1000 for infants delivered by control birth attendants (0.40, 0.19 to 0.83). Deaths due to birth asphyxia were reduced by 63% among infants delivered by intervention birth attendants (0.37, 0.17 to 0.81) and by 81% within the first two days after birth (0.19, 0.07 to 0.52). Stillbirths and deaths from serious infection occurred at similar rates in both groups. CONCLUSIONS: Training traditional birth attendants to manage common perinatal conditions significantly reduced neonatal mortality in a rural African setting. This approach has high potential to be applied to similar settings with dispersed rural populations. Trial registration Clinicaltrials.gov NCT00518856.

Improvements in physical wellbeing over the first two years on antiretroviral therapy in western Kenya.

Improvements in physical wellbeing during the first six months on antiretroviral therapy (ART) are well known, but little is known regarding long-term follow-up. We conducted a prospective cohort study among 222 HIV-positive adult tea plantation workers in western Kenya to assess wellbeing over their first two years on ART. Study subjects completed a standardized questionnaire during repeat ART clinic visits. A 30-day recall period was used to elicit the number of days when subjects experienced poor health and the number of days that pain made it difficult to complete usual activities at home and work. A seven-day recall period was used to assess the severity of bodily pain, nausea, fatigue, and rash. Prevalence of most symptoms declined over time. A median of seven days poor health during the first month on ART declined to three days in the 24th month (p=0.043). For pain making usual activities difficult, a median of seven days during the first month on ART fell to zero by 12 months (p< or =0.0001) but increased to three days by two years. Any bodily pain (range 59-83%) and fatigue (range 51-84%) over the past seven days were common through two years. However, pain and fatigue often over the past seven days declined over two years (from 24-10% (p=0.067) and 41-15% (p=0.002)). Skin rash was rare at all times, though higher at two years (8.6%) than any other time. Initial improvements in physical wellbeing were sustained over two years, however, increased pain and skin rash at year two may indicate problems as treatment programs mature. These improvements in physical wellbeing will be important in sustaining the long-term success of HIV treatment programs.

Ambulatory short-course high-dose oral amoxicillin for treatment of severe pneumonia in children: a randomised equivalency trial.

BACKGROUND: WHO case management guidelines for severe pneumonia involve referral to hospital for treatment with parenteral antibiotics. If equally as effective as parenteral treatment, home-based oral antibiotic treatment could reduce referral, admission, and treatment costs. Our aim was to determine whether home treatment with high-dose oral amoxicillin and inpatient treatment with parenteral ampicillin were equivalent for the treatment of severe pneumonia in children. METHODS: This randomised, open-label equivalency trial was done at seven study sites in Pakistan. 2037 children aged 3-59 months with severe pneumonia were randomly allocated to either initial hospitalisation and parenteral ampicillin (100 mg/kg per day in four doses) for 48 h, followed by 3 days of oral amoxicillin (80-90 mg/kg per day; n=1012) or to home-based treatment for 5 days with oral amoxicillin (80-90 mg/kg per day in two doses; n=1025). Follow-up assessments were done at 1, 3, 6, and 14 days after enrollment. The primary outcome was treatment failure (clinical deterioration) by day 6. Analyses were done per protocol and by intention to treat. This trial is registered, ISRCTN95821329. FINDINGS: In the per-protocol population, 36 individuals were excluded from the hospitalised group and 37 from the ambulatory group, mainly because of protocol violations or loss to follow-up. There were 87 (8.6%) treatment failures in the hospitalised group and 77 (7.5%) in the ambulatory group (risk difference 1.1%; 95% CI -1.3 to 3.5) by day 6. Five (0.2%) children died within 14 days of enrollment, one in the ambulatory group and four in the hospitalised group. In each case, treatment failure was declared before death and the antibiotic had been changed. None of the deaths were considered to be associated with treatment allocation; there were no serious adverse events reported in the trial. INTERPRETATION: Home treatment with high-dose oral amoxicillin is equivalent to currently recommended hospitalisation and parenteral ampicillin for treatment of severe pneumonia without underlying complications, suggesting that WHO recommendations for treatment of severe pneumonia need to be revised.

Early effects of antiretroviral therapy on work performance: preliminary results from a cohort study of Kenyan agricultural workers.

OBJECTIVE: This paper estimates the impact of antiretroviral therapy (ART) on days harvesting tea per month for tea-estate workers in Kenya. Such information is needed to assess the potential economic benefits of providing treatment to working adults. METHODS: Data for this analysis come from company payroll records for 59 HIV-infected workers and a comparison group of all workers assigned to the same work teams (reference group, n = 1992) for a period covering 2 years before and 1 year after initiating ART. Mean difference tests were used to obtain overall trends in days harvesting tea by month. A difference in difference approach was used to estimate the impact of HIV/AIDS on days working in the pre-ART period. Information on likely trends in the absence of the therapy was used to estimate the positive impacts on days harvesting tea over the initial 12 months on ART. RESULTS: No significant difference existed in days plucking tea each month until the ninth month before initiating ART, when workers worked -2.79 fewer days than references (15% less). This difference grew to 5.09 fewer days (27% less) in the final month before initiating ART. After 12 months on ART, we conservatively estimate that workers worked at least twice as many days in the month than they would have in the absence of ART. CONCLUSIONS: Treatment had a large, positive impact on the ability of workers to undertake their primary work activity, harvesting tea, in the first year on ART.

Under utilization of surveillance mammography among older breast cancer survivors.

BACKGROUND: Annual surveillance mammography is recommended for follow-up of women with a history of breast cancer. We examined surveillance mammography among breast cancer survivors who were enrolled in integrated healthcare systems. METHODS: Women in this study were 65 or older when diagnosed with early stage invasive breast cancer (N = 1,762). We assessed mammography use during 4 years of follow-up, using generalized estimating equations to account for repeated measurements. RESULTS: Eighty-two percent had mammograms during the first year after treatment; the percentage declined to 68.5% in the fourth year of follow-up. Controlling for age and comorbidity, women who were at higher risk of recurrence by being diagnosed at stage II or receiving breast-conserving surgery (BCS) without radiation therapy were less likely to have yearly mammograms (compared to stage I, odds ratio [OR] for stage IIA 0.72, confidence interval [CI] 0.59, 0.87, OR for stage IIB 0.75, CI 0.57, 1.0; compared to BCS with radiation, OR 0.58, CI 0.43, 0.77). Women with visits to a breast cancer surgeon or oncologist were more likely to receive mammograms (OR for breast cancer surgeon 6.0, CI 4.9, 7.4, OR for oncologist 7.4, CI 6.1, 9.0). CONCLUSIONS: Breast cancer survivors who are at greater risk of recurrence are less likely to receive surveillance mammograms. Women without a visit to an oncologist or breast cancer surgeon during a year have particularly low rates of mammography. Improvements to surveillance care for breast cancer survivors may require active participation by primary care physicians and improvements in cancer survivorship programs by healthcare systems.

Mammography surveillance and mortality in older breast cancer survivors.

PURPOSE: There are more than 2,000,000 breast cancer survivors in the United States today. While surveillance for asymptomatic recurrence and second primary is included in consensus recommendations, the effectiveness of this surveillance has not been well characterized. Our purpose is to estimate the effectiveness of surveillance mammography in a cohort of breast cancer survivors with complete ascertainment of surveillance mammograms and negligible losses to follow-up. PATIENTS AND METHODS: We enrolled 1,846 stage I and II breast cancer patients who were at least 65 years old at six integrated health care delivery systems. We used medical record review and existing databases to ascertain patient, tumor, and therapy characteristics, as well as receipt of surveillance mammograms. We linked personal identifiers to the National Death Index to ascertain date and cause of death. We matched four controls to each breast cancer decedent to estimate the association between receipt of surveillance mammogram and breast cancer mortality. RESULTS: One hundred seventy-eight women died of breast cancer during 5 years of follow-up. Each additional surveillance mammogram was associated with a 0.69-fold decrease in the odds of breast cancer mortality (95% CI, 0.52 to 0.92). The protective association was strongest among women with stage I disease, those who received mastectomy, and those in the oldest age group. CONCLUSION: Given existing recommendations for post-therapy surveillance, trials to compare surveillance with no surveillance are unlikely. This large observational study provides support for the recommendations, suggesting that receipt of surveillance mammograms reduces the rate of breast cancer mortality in older patients diagnosed with early-stage disease.

A sensitivity analysis of a randomized controlled trial of zinc in treatment of falciparum malaria in children.

BACKGROUND: The randomized trial has long been recognized as a means to assess the efficacy of new interventions, because the investigator can reduce or eliminate many sources of error. As such, clinical trials often do not include quantitative assessments of the extent that systematic error could affect their results. We examined the impact of different sources of bias on a randomized controlled trial of the efficacy of zinc as an adjuvant to malaria therapy in reducing time to total parasite clearance. METHODS: Using data from a previously published study, we identified two sources of bias and used the sensitivity analysis technique developed by Lash and Fink to assess the impact of each source of bias on the outcome. RESULTS: After correcting for each source of bias and reincorporating random error into our results, the point estimate of effect comparing those who received placebo to those who received zinc changed slightly (SMR changed from 0.92 to 0.90) but the 95% interval increased 22% (changing from 0.73-1.16 in the conventional analysis to 0.65-1.26 in the sensitivity analysis). CONCLUSIONS: The findings of this sensitivity analysis serve as a reminder that the frequentist confidence interval underestimates the total error, even in a randomized controlled trial. Authors of randomized controlled trial investigations ought to conduct a complete assessment of the impact of potential sources of bias in their studies. CONSORT guidelines for reporting trial results should be updated to encourage authors to assess the impact of non-random errors on their studies.