Missingness and algorithmic bias: an example from the United States National Outbreak Reporting System, 2009-2019.

Growing debates about algorithmic bias in public health surveillance lack specific examples. We tested a common assumption that exposure and illness periods coincide and demonstrated how algorithmic bias can arise due to missingness of critical information related to illness and exposure durations. We examined 9407 outbreaks recorded by the United States National Outbreak Reporting System (NORS) from January 1, 2009 through December 31, 2019 and detected algorithmic bias, a systematic over- or under-estimation of foodborne disease outbreak (FBDO) durations due to missing start and end dates. For 7037 (75%) FBDOs with complete date-time information, ~ 60% reported that the exposure period ended before the illness period started. For 2079 (87.7%) FBDOs with missing exposure dates, average illness durations were ~ 5.3 times longer (p < 0.001) than those with complete information, prompting the potential for algorithmic bias. Modern surveillance systems must be equipped with investigative capacities to examine and assess structural data missingness that can lead to bias.

How well global dietary intake estimates agree: a case of sodium consumption.

Global dietary data repositories are key components of nutrition surveillance. The two most comprehensive databases, the Global Dietary Database (GDD) and the Global Burden Disease (GBD), provide national dietary intake estimates but use different data sources and models to generate estimates. To explore the agreement between GDD and GBD estimates, we compared country-specific average daily sodium intakes in 169 countries over a 28-year period using descriptive statistics, the Bland-Altman method, and prevalence exceeding the intake reference level of 2.3 g/day. We detected a staggering 36% difference between GDD and GBD estimates of global mean intakes (2.68 ± 0.74 vs. 3.88 ± 1.15 g/day, respectively; p < 0.0001). As 104 (61.5%) countries reported to have over-consumed sodium by both databases, the development of standardized approaches for national dietary intake estimation is critical for monitoring global sodium intake in a systematic and comprehensive way and for implementing global strategies to reduce sodium intake.

Weight Loss Trajectories and Short-Term Prediction in an Online Weight Management Program.

The extent to which early weight loss in behavioral weight control interventions predicts long-term success remains unclear. In this study, we developed an algorithm aimed at classifying weight change trajectories and examined its ability to predict long-term weight loss based on weight early change. We utilized data from 667 de-identified individuals who participated in a commercial weight loss program (Instinct Health Science), comprising 69,363 weight records. Sequential polynomial regression models were employed to classify participants into distinct weight trajectory patterns based on key model parameters. Next, we applied multinomial logistic models to evaluate if early weight loss in the first 14 days and prolonged duration of participation were significantly associated with long-term weight loss patterns. The mean percentage of weight loss was 7.9 ± 5.1% over 133 ± 69 days. Our analysis revealed four main weight loss trajectory patterns: a steady decrease over time (30.6%), a decrease to a plateau with subsequent decline (15.8%), a decrease to a plateau with subsequent increase (46.9%), and no substantial decrease (6.7%). Early weight change rate and total participating duration emerged as significant factors in differentiating long-term weight loss patterns. These findings contribute to support the provision of tailored advice in the early phase of behavioral interventions for weight loss.

Peak timing of slowest growth velocity among young children coincides with highest ambient temperatures in Burkina Faso: a longitudinal study.

BACKGROUND: Seasonal cycles in climatic factors affect drivers of child growth and contribute to seasonal fluctuations in undernutrition. Current growth seasonality models are limited by categorical definitions of seasons that rely on assumptions about their timing and fail to consider their magnitude. OBJECTIVE: We disentangle the relationship between climatic factors and growth indicators, using harmonic regression to determine how child growth is related to peaks in temperature, precipitation, and vegetation. METHODS: Longitudinal anthropometric data collected between August 2014 and December 2016 from 5039 Burkinabè children measured monthly from age 6 to 28 mo (108,580 observations) were linked with remotely sensed daily precipitation, vegetation, and maximum air temperature. Our models parsimoniously extract a cyclic signal with multiple potential peaks, to compare the magnitude and timing of seasonal peaks in climatic factors and morbidity with that of nadirs in growth velocity (cm/mo, kg/mo). RESULTS: Length and weight velocity were slowest twice a year, coinciding both times with the highest temperatures, and peak fever incidence. Length velocity is slowest 13 d after the first temperature peak in April, and 5 d after the second. Similarly, weight velocity is slowest 13 d before the first temperature peak, and 11 d before the second. The statistical relationship between temperature and anthropometry shows that when the current temperature is higher, weight velocity is lower (ß = -0.0048; 95% CI: -0.0059, -0.0038), and length velocity is higher (ß = 0.0088; 95% CI: 0.0070, 0.0105). CONCLUSIONS: Results suggest that child health and development are more affected by high temperatures than by other aspects of climatic seasonality such as rainfall. Emerging shifts in climatic conditions will pose challenges to optimal growth, highlighting the importance of changes that optimize the timing of nutrition interventions and address environmental growth-limiting conditions. CLINICAL TRIAL REGISTRY: Clinicaltrials.gov: NCT02071563.

Reducing the risks of nuclear war - The role of health professionals.

No abstract available.