Background When we were unable to identify an electronic data capture (EDC) package that supported our requirements for clinical research in resource-limited regions, we set out to build our own reusable EDC framework. We needed to capture data when offline, synchronize data on demand, and enforce strict eligibility requirements and complex longitudinal protocols. Based on previous experience, the geographical areas in which we conduct our research often have unreliable, slow internet access that would make web-based EDC platforms impractical. We were unwilling to fall back on paper-based data capture as we wanted other benefits of EDC. Therefore, we decided to build our own reusable software platform. In this paper, we describe our customizable EDC framework and highlight how we have used it in our ongoing surveillance programs, clinic-based cross-sectional studies, and randomized controlled trials (RCTs) in various settings in India and Ecuador. Objective This paper describes the creation of a mobile framework to support complex clinical research protocols in a variety of settings including clinical, surveillance, and RCTs. Methods We developed ConnEDCt, a mobile EDC framework for iOS devices and personal computers, using Claris FileMaker software for electronic data capture and data storage. Results ConnEDCt was tested in the field in our clinical, surveillance, and clinical trial research contexts in India and Ecuador and continuously refined for ease of use and optimization, including specific user roles; simultaneous synchronization across multiple locations; complex randomization schemes and informed consent processes; and collecting diverse types of data (laboratory, growth measurements, sociodemographic, health history, dietary recall and feeding practices, environmental exposures, and biological specimen collection). Conclusions ConnEDCt is customizable, with regulatory-compliant security, data synchronization, and other useful features for data collection in a variety of settings and study designs. Furthermore, ConnEDCt is user friendly and lowers the risks for errors in data entry because of real time error checking and protocol enforcement.

Pregnancy and lactation are critical life stages with unique nutritional requirements, including for vitamin A (VA). Current Dietary Reference Intakes (DRIs) for VA were published in 2001. The objective of this review was to identify and categorize evidence related to VA requirements in pregnancy and lactation since these DRIs were formulated. We searched MEDLINE and included articles according to an analytic framework of maternal VA exposure on status and health outcomes in the mother-child dyad. Intermediate and indirect evidence supports that maternal VA intakes can impact the mother's VA status, breastmilk, and health outcomes, as well as the child's VA status and select health outcomes. Food-based approaches may lead to more sustained, sufficient VA status in mothers and children. Research needs include further study linking maternal VA intakes on maternal and child VA status, and further associations with outcomes to determine intake requirements to optimize health.

BACKGROUND When we were unable to identify an electronic data capture (EDC) package that supported our requirements for clinical research in resource-limited regions, we set out to build our own reusable EDC framework. We needed to capture data when offline, synchronize data on demand, and enforce strict eligibility requirements and complex longitudinal protocols. Based on previous experience, the geographical areas in which we conduct our research often have unreliable, slow internet access that would make web-based EDC platforms impractical. We were unwilling to fall back on paper-based data capture as we wanted other benefits of EDC. Therefore, we decided to build our own reusable software platform. In this paper, we describe our customizable EDC framework and highlight how we have used it in our ongoing surveillance programs, clinic-based cross-sectional studies, and randomized controlled trials (RCTs) in various settings in India and Ecuador. OBJECTIVE This paper describes the creation of a mobile framework to support complex clinical research protocols in a variety of settings including clinical, surveillance, and RCTs. METHODS We developed ConnEDCt, a mobile EDC framework for iOS devices and personal computers, using Claris FileMaker software for electronic data capture and data storage. RESULTS ConnEDCt was tested in the field in our clinical, surveillance, and clinical trial research contexts in India and Ecuador and continuously refined for ease of use and optimization, including specific user roles; simultaneous synchronization across multiple locations; complex randomization schemes and informed consent processes; and collecting diverse types of data (laboratory, growth measurements, sociodemographic, health history, dietary recall and feeding practices, environmental exposures, and biological specimen collection). CONCLUSIONS ConnEDCt is customizable, with regulatory-compliant security, data synchronization, and other useful features for data collection in a variety of settings and study designs. Furthermore, ConnEDCt is user friendly and lowers the risks for errors in data entry because of real time error checking and protocol enforcement.

Objectives To characterize anemia and anthropometry in children in rural South India:1.Report the prevalence of undernutrition and anemia2.Examine associations with demographic, socio-economic, and clinical characteristics among 6–24-month-old children screened for participation in a randomized controlled feeding trial. Methods This cross-sectional assessment included 339 children screened for participation in a randomized controlled feeding trial from March – September 2019. Complete blood counts were performed immediately after blood and anthropometric measurements were collected. Sociodemographic data were collected from mothers. We assessed the outcomes hemoglobin (Hb, g/dL), anemia (Hb < 11 g/dL), stunting (LAZ < −2), wasting (WLZ < −2) and underweight (WAZ < −2) with correlates at the individual (child age, sex, birthweight), maternal (maternal age, maternal education) and household (family income) level. Covariates were included based on previous literature and biological plausibility. Linear regression (ß, SE) was used to determine covariates for continuous outcomes (such as Hb) and binomial and Poisson regressions (RR, 95% CI) were used in the case of categorical outcomes (Hb < 11 g/dL, LAZ < −2, WLZ < −2, WAZ < −2). All multivariate models included a priori determined covariates age and sex and covariates with univariate p-values ≤ 0.20. Variables with a p-value of ≤ 0.05 were retained in the final multivariate models (SAS 9.4). Results Almost half (46.9%) of this population was ≤ 12 months of age at the time of assessment. Over half (52.7%) of participants were female. Anemia was present in 72.6% of children. 15.1% of children were stunted, 14.2% underweight and 8.2% were wasted based on the corresponding WHO Z-scores. A higher birth weight of 1 kg was associated with a lower risk of later underweight [0.47 (0.29, 0.79); P = 0.01] and wasting [0.42 (0.21, 0.85); P = 0.02]. Children with families whose monthly income was ≤ 5000 INR (∼USD 70) per month were more likely to be stunted [2.32 (1.17, 4.59); P = 0.01] and underweight [2.07 (1.05, 4.06); P = 0.03]. Conclusions In these 6–24-month-old children, anemia and poor growth are widely prevalent. The results from this analysis provide evidence of a potential to benefit from nutritional interventions targeted at children in this population. Funding Sources Cornell Division of Nutritional Sciences Harvest Plus.

Objectives We aimed to determine the prevalence of baseline vitamin A insufficiency (VAI; serum retinol (SR) < 1.05µmol/L) and to identify potential correlates of SR and VAI in a population of 12–18-month-old children participating in a randomized controlled trial in urban slums in Mumbai, India. Methods In this cross-sectional analysis, we determined SR concentrations in archived serum samples (N = 118) that were collected from 12–18-month-old children in urban slum communities of Western Mumbai in March–October 2017. We adjusted SR measurements using the BRINDA adjustment method for C-reactive protein.1 Child, maternal and household characteristics were assessed as potential correlates. Linear [β(SE)] and binomial [RR (95% CI)] regressions were used to identify correlates of SR and VAI, respectively. Age and sex were retained in all models. Results The children in this population had a median (IQR) age of 14.5 (12.4, 16.7) months, and 46.6% were girls. Almost a third (27.1%) were underweight (weight-for-age Z-score <-2) and 9.3% were wasted (weight-for-length Z-score <-2). One third (30.5%) were anemic (hemoglobin < 11g/dL) and 19.5% of children were zinc-deficient (Zn < 70µg/dL). Median (IQR) unadjusted SR was 1.1 (0.9, 1.4) µmol/L and VAI was present among 50 (42.4%) of the population. After adjusting for inflammation, SR was 1.2 (1.0, 1.5) µmol/L and VAI was present in 38 (32.2%) of children. In multivariate regressions, each nmol/L increase in vitamin D [25(OH)D] was associated with a 0.01 µmol/L increase in SR [β (SE) 0.01 (0.004), p = 0.004], and each additional child under 5 years living in the household was associated with lower SR [−0.13 (0.04), p = 0.003]. Conclusions In these 12–18-month-old children, VAI was prevalent in nearly two out of every five children. Funding Sources Division of Nutritional Sciences, Cornell University Harvest Plus.

Objectives This secondary analysis was conducted to examine the association of iron status with changes in insulin sensitivity (SI, µU*10–4*min1*ml−1) and visceral adipose tissue (VAT, L) as part of a randomized, double blind, placebo-controlled trial. This analysis was conducted to examine if serum ferritin (µg/L) and hemoglobin (Hb, g/dL)·independently predict changes in SI or VAT·are affected by treatment with rosiglitazone or recombinant human growth hormone (rhGH)·modify the effect of changes in SI or VAT in response to treatment among adults living with HIV with abdominal obesity and insulin-resistance participating in a randomized trial. Methods In a 2 × 2 double-blinded factorial design, participants were provided 4 mg rosiglitazone twice daily, 3 mg rhGH daily, a combination of rhGH + rosiglitazone, or double-placebo (control) for 12 weeks. Serum ferritin was adjusted for inflammation (C-reactive protein, mg/L; α-1-acid glycoprotein, g/L) using the Thurnham method. Generalized linear models were used to assess the association of baseline Hb and ferritin with changes in SI and VAT. Models were adjusted for the a priori determined covariates, age, treatment group, CD4 T-cell count and HIV-1 RNA viral load. Effect modification was assessed by including an interaction term between biomarker and intervention group. Generalized estimating equations were used to assess the effects of treatment on Hb and serum ferritin concentrations. Results At baseline, there was no difference in serum ferritin (n = 47) or Hb (n = 45) across treatments; overall median (Q1, Q3) ferritin 95.60 (48.13, 151.00) µg/L, P = 0.13, Hb 14.40 (13.10,15.10) g/dL, P = 0.48. Neither marker of iron status was an independent predictor of changes in SI or in VAT. Compared to placebo at 12 weeks, Hb was significantly reduced by rosiglitazone (ß = −1.46, SE = 0.35, P < 0.0001), and by rhGH/rosiglitazone (ß = −0.93, SE = 0.30, P = 0.002) but not by rhGH alone. There was no significant effect of treatment on serum ferritin concentrations. Neither ferritin nor Hb modified the effect of intervention on changes in SI or VAT. Conclusions In this study among PLHIV, iron status was not an independent predictor or effect modifier of changes in SI or VAT. Treatments with rosiglitazone significantly reduced Hb concentration, which is consistent with adverse effects reported in previous clinical trials. Funding Sources Division of Nutritional Sciences, Cornell University. Weill Cornell Medicine, Cornell University