Improving identification of symptomatic cancer at primary care clinics: A predictive modeling analysis in Botswana.

In resource-limited settings, augmenting primary care provider (PCP)-based referrals with data-derived algorithms could direct scarce resources towards those patients most likely to have a cancer diagnosis and benefit from early treatment. Using data from Botswana, we compared accuracy of predictions of probable cancer using different approaches for identifying symptomatic cancer at primary clinics. We followed cancer suspects until they entered specialized care for cancer treatment (following pathologically confirmed diagnosis), exited from the study following noncancer diagnosis, or died. Routine symptom and demographic data included baseline cancer probability assessed by the primary care provider (low, intermediate, high), age, sex, performance status, baseline cancer probability by study physician, predominant symptom (lump, bleeding, pain or other) and HIV status. Logistic regression with 10-fold cross-validation was used to evaluate classification by different sets of predictors: (1) PCPs, (2) Algorithm-only, (3) External specialist physician review and (4) Primary clinician augmented by algorithm. Classification accuracy was assessed using c-statistics, sensitivity and specificity. Six hundred and twenty-three adult cancer suspects with complete data were retained, of whom 166 (27%) were diagnosed with cancer. Models using PCP augmented by algorithm (c-statistic: 77.2%, 95% CI: 73.4%, 81.0%) and external study physician assessment (77.6%, 95% CI: 73.6%, 81.7%) performed better than algorithm-only (74.9%, 95% CI: 71.0%, 78.9%) and PCP initial assessment (62.8%, 95% CI: 57.9%, 67.7%) in correctly classifying suspected cancer patients. Sensitivity and specificity statistics from models combining PCP classifications and routine data were comparable to physicians, suggesting that incorporating data-driven algorithms into referral systems could improve efficiency.

A Step Toward Timely Referral and Early Diagnosis of Cancer: Implementation and Impact on Knowledge of a Primary Care-Based Training Program in Botswana.

INTRODUCTION: Health system delays in diagnosis of cancer contribute to the glaring disparities in cancer mortality between high-income countries and low- and middle-income countries. In Botswana, approximately 70% of cancers are diagnosed at late stage and median time from first health facility visit for cancer-related symptoms to specialty cancer care was 160 days (IQR 59-653). We describe the implementation and early outcomes of training targeting primary care providers, which is a part of a multi-component implementation study in Kweneng-East district aiming to enhance timely diagnosis of cancers. METHODS: Health-care providers from all public facilities within the district were invited to participate in an 8-h intensive short-course program developed by a multidisciplinary team and adapted to the Botswana health system context. Participants' performance was assessed using a 25-multiple choice question tool, with pre- and post-assessments paired by anonymous identifier. Statistical analysis with Wilcoxon signed-rank test to compare performance at the two time points across eight sub-domains (pathophysiology, epidemiology, social context, symptoms, evaluation, treatment, documentation, follow-up). Linear regression and negative binomial modeling were used to determine change in performance. Participants' satisfaction with the program was measured on a separate survey using a 5-point Likert scale. RESULTS: 176 participants attended the training over 5 days in April 2016. Pooled linear regression controlling for test version showed an overall performance increase of 16.8% after participation (95% CI 15.2-18.4). Statistically significant improvement was observed for seven out of eight subdomains on test A and all eight subdomains on test B. Overall, 71 (40.3%) trainees achieved a score greater than 70% on the pretest, and 161 (91.5%) did so on the posttest. Participants reported a high degree of satisfaction with the training program's content and its relevance to their daily work. CONCLUSION: We describe a successfully implemented primary health care provider-focused training component of an innovative intervention aiming to reduce health systems delays in cancer diagnosis in sub-Saharan Africa. The training achieved district-wide participation, and improvement in the knowledge of primary health-care providers in this setting. CLINICAL TRIAL REGISTRATION: www.ClinicalTrials.gov, identifier NCT02752061.

Similar HIV protection from four weeks of zidovudine versus nevirapine prophylaxis among formula-fed infants in Botswana.

BACKGROUND: The World Health Organization HIV guidelines recommend either infant zidovudine (ZDV) or nevirapine (NVP) prophylaxis for the prevention of intrapartum mother-to-child HIV transmission (MTCT) among formula-fed infants. No study has evaluated the comparative efficacy of infant prophylaxis with twice daily ZDV versus once daily NVP in exclusively formula-fed HIV-exposed infants. METHODS: Using data from the Mpepu Study, a Botswana-based clinical trial investigating whether prophylactic co-trimoxazole could improve infant survival, retrospective analyses of MTCT events and Division of AIDS (DAIDS) Grade 3 or Grade 4 occurrences of anaemia or neutropenia were performed among infants born full-term (≥ 37 weeks gestation), with a birth weight ≥ 2500 g and who were formula-fed from birth. ZDV infant prophylaxis was used from Mpepu Study inception. A protocol modification mid-way through the study led to the subsequent use of NVP infant prophylaxis. RESULTS: Among infants qualifying for this secondary retrospective analysis, a total of 695 (52%) infants received ZDV, while 646 (48%) received NVP from birth for at least 25 days but no more than 35 days. Confirmed intrapartum HIV infection occurred in two (0.29%) ZDV recipients and three (0.46%) NVP recipients (p = 0.68). Anaemia occurred in 19 (2.7%) ZDV versus 12 (1.9%) NVP (p = 0.36) recipients. Neutropenia occurred in 28 (4.0%) ZDV versus 21 (3.3%) NVP recipients (p = 0.47). CONCLUSIONS: Both ZDV and NVP resulted in low intrapartum transmission rates and no significant differences in severe infant haematologic toxicity (DAIDS Grade 3 or Grade 4) among formula-fed full-term infants with a birthweight ≥ 2500 g.

Detecting congenital malformations - Lessons learned from the Mpepu study, Botswana.

INTRODUCTION: A large and increasing number of HIV-infected women are conceiving on antiretroviral treatment (ART). While most antiretrovirals are considered safe in pregnancy, monitoring for rare pregnancy and infant adverse outcomes is warranted. METHODS: We conducted a retrospective secondary analysis nested within a clinical trial of infant cotrimoxazole vs. placebo prophylaxis in Botswana (the Mpepu Study). Infants were examined at birth, and at least every 3 months through 18 months of age. Abnormal physical findings and diagnostic testing revealing malformations were documented. Post hoc, a geneticist classified all reported malformations based on available documentation. Structural malformations with surgical, medical or cosmetic importance were classified as major malformations. We present a descriptive analysis of identified malformations. RESULTS: Between 2011 and 2014, 2,933 HIV-infected women who enrolled in the Mpepu study delivered 2,971 live-born infants. Study staff conducted 2,944 (99%) newborn exams. One thousand eighty-eight (38%) women were taking ART at conception; 1,147 (40%) started ART during pregnancy; 442 (15%) received zidovudine monotherapy; and 223 (7%) received no antiretroviral during pregnancy. Of 33 reported anomalies, 25 (76%) met congenital malformations criteria, 10 (30%) were classified as major malformations, 4 (40%) of which were identified after the birth exam. DISCUSSION: Our results highlight the importance of staff training on identification of congenital malformations, programmatic monitoring beyond the birth examination and the value of geneticist involvement in the malformations classification process in resource-limited settings. These elements will be important to fully define antiretroviral drug safety in pregnancy. SIGNIFICANCE: Surveillance systems for monitoring the safety of antiretroviral use during pregnancy among HIV-infected women in resource-limited setting are lacking. The World Health Organization's published programmatic recommendations for such surveillance systems represents the gold standard. We employed data from a clinical trial in Botswana, a country with a generalized HIV epidemic and high antiretroviral uptake by HIV-infected women, to highlight practical opportunities to strengthen congenital malformation surveillance systems in these settings where over 1 million HIV infected pregnant women reside. TRIAL REGISTRATION: Clinical Trials.gov NCT01229761.