Pith. sign in

REVIEW 3 major objections 6 minor 25 references

RiM: Record, Improve and Maintain Physical Well-being using Federated Learning

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that a federated MLP fine-tuned with FedAvg predicts student sleep and distance deficits at 60.71% sign accuracy and 0.91 MAE, outperforming FedPer, while keeping raw data on users' phones.

desk verdict Circular evaluation and an unsupported differential-privacy claim sink an otherwise neat undergraduate FL-health thesis. read the letter →

arxiv 2505.06384 v1 pith:P7JW3XUW submitted 2025-05-09 cs.LG cs.CRcs.CY

Aditya Mishra , Haroon Lone This is my paper · ORCID
classification cs.LGcs.CRcs.CY
keywords federatedlearningmobilehealthwell-beingFedAvgPerlifestyledeficitpredictionmultilayerperceptrondifferentialprivacy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

RiM is a mobile app that tracks students' steps, distance, sleep, and meals, then uses a small neural network to predict which lifestyle measures fall short of ideal sleep and distance ranges. The paper tries to establish that a two-stage pipeline—pre-training a multilayer perceptron on simulated data, then fine-tuning it with federated learning on data that stays on each phone—can deliver personalized wellness recommendations without centralizing raw health records. Measured by sign agreement and mean absolute error, the FedAvg variant outperforms the FedPer variant on a ten-client real-world dataset. If correct, this offers a lightweight, privacy-preserving alternative to centralized wellness analytics for a population that rarely shares health data.

What carries the argument

The central object is a five-hidden-layer MLP (32-to-4 neurons) that maps a standardized feature vector $[distance, sleep, BMI, age, breakfast, meal, gender]$ to two outputs: sleep deficit and distance deficit, each defined as the signed distance outside an ideal range. It is pre-trained on simulated data whose features are sampled from stated distributions, then fine-tuned in federated rounds using FedAvg, which averages all client weights, and FedPer, which averages only shared root layers while keeping personal head layers local. A rule-based scoring layer converts predicted deficits into weighted risk scores, aggregates interaction risks, and surfaces the top two recommendation messages. The accuracy metric counts a prediction as correct when the sign of the predicted deficit matches the sign of the true deficit.

What would settle it

Run a membership-inference attack on the shared fine-tuned weights: if an attacker can determine whether a specific day's step count or meal flag was included in a client's training set from the weight updates, the privacy claim is false. A formal DP accountant applied to the unclipped, un-noised updates would also show unbounded $\varepsilon$.

Watch

Extended reading notes

Core claim

The paper's central empirical claim is that the FedAvg-fine-tuned MLP outperforms the FedPer-fine-tuned variant: average accuracy 60.71% versus 46.34%, and MAE 0.91 versus 1.19. It interprets this gap as a personalization trade-off: FedAvg's global weight averaging transfers deficit-detection skill across clients, while FedPer's per-client heads preserve each user's bias toward the healthy zero-deficit pattern but fail to model rare non-ideal events. The paper further claims that the pipeline guarantees differential privacy because only model weights, not raw sensor or demographic data, leave the device.

Load-bearing premise

The load-bearing premise is that only transmitting model weights, with no added noise or formal privacy accounting, yields differential privacy; if shared weights leak information about local data, the privacy guarantee collapses.

Editorial extensions

If this is right

  • A phone with only accelerometer readings can support an MLP that flags sleep and distance deficits, without requiring additional wearable sensors for the core prediction.
  • FedAvg's global averaging works better for students whose lifestyles deviate from ideal, while FedPer's local heads better preserve an all-healthy pattern, so the choice of aggregation scheme changes which users receive useful alerts.
  • The sign-accuracy measure means 60.71% reflects how often the model gets the direction of the deficit right, not the exact magnitude; MAE 0.91 quantifies the average magnitude error in the same units as the deficits.
  • If the privacy assumption holds, the system can be deployed without a central collection of health records, lowering the risk of data breaches and reducing the consent friction associated with sharing personal health data.
  • The top-two risk-scored recommendations give users actionable guidance while avoiding message overload.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the shared weight updates are not clipped or noised, the claimed differential-privacy guarantee does not follow from the algorithm as described; formal differential privacy would require calibrated noise, gradient clipping, and a privacy budget.
  • The accuracy metric is sign agreement, so a model that always predicts 'no deficit' would score highly on healthy clients, which likely explains part of FedPer's advantage on zero-deficit users and should be checked against a no-deficit baseline.
  • A direct extension would repeat the fine-tuning with calibrated noise added to the updates and measure how the 60.71% accuracy and 0.91 MAE degrade, giving an empirical privacy-utility trade-off for this setting.
  • Since pre-training uses synthetic distributions, the pretrained weights may not transfer to populations with different activity or meal patterns; retraining on a broader set of real cohorts is a testable next step.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper presents RiM, an Android application that tracks steps, distance, sleep, and meal information, and uses a multilayer perceptron (MLP) pre-trained on simulated data and fine-tuned with federated learning (FedAvg and FedPer) on data from ten IISER Bhopal students. The model outputs sleep and distance deficits, which are then used by a rule-based recommender to show personalized wellness advice. The paper reports that FedAvg achieves 60.71% sign accuracy and 0.91 MAE, outperforming FedPer, and claims that sharing only model weights rather than raw data provides differential privacy.

Significance. The manuscript has clear strengths: it provides a complete mobile-sensing pipeline with two algorithmic descriptions (Algorithms 1 and 2), releases code via GitHub, and compares two federated personalization strategies on a real, albeit very small, dataset. If the evaluation were valid, a lightweight personalized recommender that keeps data local would be useful and topical. However, the headline claims are not supported: the target definitions make the main evaluation circular, and the differential-privacy guarantee is asserted without any mechanism. These issues invalidate the paper's central contributions in their current form.

major comments (3)
  1. [Section 3.2] The target variables d_sleep and d_distance are defined as piecewise-linear functions of the sleep and distance features that are also inputs to the MLP, since the feature vector is x = [distance, sleep, bmi, age, breakfast, meal, gender] and d_j is a function of x_j for j in {sleep, distance}. Therefore the 'prediction' task is to reproduce a known deterministic transformation of the input, not to generalize to an independent outcome. A trivial rule-based system that directly applies the closed-form deficit formula to the test features would achieve near-perfect sign accuracy and near-zero MAE, whereas the reported FedAvg accuracy of 60.71% and MAE of 0.91 are substantially worse than such a baseline. Consequently, Tables 4.1 and 4.2 do not measure predictive efficacy; they measure how well the MLP approximates a known function, and the central claim of the abstract and Chapter 5 does not follow.
  2. [Sections 1.3 and 5] The claim that the approach 'ensures differential privacy by never transmitting raw user data' is unsupported. No clipping, noise injection, or privacy-accounting mechanism is described or implemented; Section 3.3 transmits model weights that can encode information about local data. Differential privacy is a property of the randomized algorithm's output, not of data locality, so the privacy guarantee as stated is false. Establishing a formal guarantee would require modifying the algorithm, e.g., by using DP-FedAvg-style gradient clipping and Gaussian noise with an explicit privacy budget, which is absent from the manuscript.
  3. [Section 4] The experimental evaluation uses only ten participants, with eight days for fine-tuning and seven days for testing, and it lacks a non-trivial baseline such as the closed-form deficit formula or a constant predictor. The client-wise accuracies range from 25.18% to 81.42%, and no confidence intervals, cross-validation, or statistical tests are reported. Given the circular target definitions, the FedAvg-versus-FedPer comparison cannot support the paper's claim of demonstrated efficacy; the observed variation is consistent with noise on tiny samples and with class imbalance in the zero-deficit class.
minor comments (6)
  1. [Section 3.5] The sentence 'We use simulated data to pre-train the model and fine-tune and test in on the collected real-world dara' contains two typos: 'in' should be removed and 'dara' should be 'data'.
  2. [Section 3.2] The threshold symbol τ is overloaded: it is used for step detection in Algorithm 1 and for deficit thresholds in Section 3.2, which can confuse the reader; distinct symbols should be used.
  3. [Section 3.2] Equation (3.3) defines accuracy as sign agreement including the zero class, but the text does not discuss how the prevalence of zero deficits affects the metric; for clients whose true deficits are mostly zero, predicting zero achieves high accuracy without any skill.
  4. [Section 2.2] The sentence 'FedPer algorithm is used for fine-tuning the MLP model' is inconsistent with Chapter 4, where both FedAvg and FedPer are evaluated; the phrasing should be clarified to describe the fine-tuning setup accurately.
  5. [Table 2.1] The age feature is listed as 'Truncated Normal (Empirical Distribution)', but the text and Figure 2.1 refer to 'an empirical distribution'; the distinction between the two descriptions should be resolved.
  6. [Section 3.1] The GitHub link should be cited with an access date or version, following the style used for other web resources in the bibliography.

Circularity Check

1 steps flagged · score 8.0 of 10

Deficit targets are defined in Section 3.2 as deterministic piecewise-linear functions of the same sleep/distance features fed into the MLP, so the reported accuracy/MAE measure fitting a known transform, not predictive efficacy.

  1. self definitional [Section 3.2 (Machine Learning Frameworks), deficit definition; Section 3.4 Eq. (3.3); Table 3.1 ideal ranges]
    "Given a user feature vector x= [distance,sleep,bmi,age,breakfast,meal,gender], we first standardize: ˜x= StandardScaler(x). The MLP model f then predicts deficits d=f( ˜x) = [ d_sleep, d_distance ], where each deficit d_j is defined as ... (ideal_j,min−x_j), x_j < ideal_j,min, −(x_j−ideal_j,max), x_j > ideal_j,max, 0, otherwise."

    The feature vector x includes the raw sleep hours and distance traveled, and the targets d_sleep and d_distance are computed from those same x_j values using the fixed threshold formula, with ideal intervals from Table 3.1 ([7,9] hours sleep, [5,8] km distance). Thus the label is a deterministic function of the input: y_j = g(x_j). A rule-based oracle applying this closed form would obtain 100% sign accuracy and 0 MAE on the same test targets. The reported FedAvg accuracy of 60.71% and MAE of 0.91, being worse than the trivial formula, show only that the MLP imperfectly reproduces a known transformation and do not demonstrate 'efficacy in predicting lifestyle deficits'. Because the evaluation metric (Eq.

full rationale

The core evaluation in Chapter 4 is circular by the paper's own equations: the sleep and distance deficits used as targets are defined in Section 3.2 as piecewise-linear functions of the very same sleep and distance features that enter the model's input vector. Consequently, the accuracy/MAE numbers are not evidence of predictive skill about an external, unknown outcome; they are measures of how well the network learns an identity-like threshold map, and a simple closed-form rule beats the reported numbers. This is a self-definitional circularity that invalidates the headline 'demonstrating its efficacy' claim. I did not score the differential-privacy assertion as circular: the claim that sharing weights without noise guarantees differential privacy is unsupported and incorrect, but that is a correctness/privacy flaw rather than a reduction of a prediction to its inputs. No load-bearing self-citation chain or imported uniqueness theorem appears; citations to FedAvg, FedPer, and Flower are external background. The circularity score is high because the central measured result itself rests on the definitional target construction, not because of author self-citation.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities. The central claim rests on several hand-chosen thresholds and distribution assumptions. Most importantly, the ideal ranges for sleep and distance are used to define the prediction targets themselves, so the targets are not independent of the model inputs. The differential privacy premise is an unvalidated assumption that the paper treats as fact.

free parameters (6)
  • Simulated data distribution parameters = not stated
    Section 2.1 assigns distributions (negative binomial, log-normal, normal, Bernoulli, truncated normal) to features but does not give the parameter values (e.g., means, variances) used to generate the pretraining dataset. These values determine the pretrained model.
  • Ideal sleeping range = [7, 9] hours
    Table 3.1 sets the ideal sleep window used to define sleep deficits; this threshold is chosen by hand and directly defines the target labels.
  • Ideal distance traveled = [5, 8] km
    Table 3.1 sets the ideal distance range used to define distance deficits; chosen by hand and directly defines target labels.
  • Rule weights and thresholds = w_sleep=1, w_distance=0.8, w_bmi=1, w_meal=1, R_high=3, theta=0.5, top_n=2
    Table 3.1 lists weights and thresholds for the rule-based recommendation scoring; these are hand-chosen and affect which recommendations are surfaced.
  • Step and distance tracking parameters = tau=1.8, delta=0.5 m
    Table 3.1 and Algorithm 1: tau is the acceleration magnitude threshold and delta is the distance added per step; both are hand-selected and affect the step and distance inputs.
  • Training hyperparameters = lr=1e-3, batch=32, pretrain epochs=200, fine-tune epochs=10, hidden=[64,32,16,8,4], clients=10
    Table 3.1 provides standard hyperparameters for the MLP and federated learning. They are chosen by hand and influence the reported results.
assumptions (5)
  • domain assumption Simulated data distributions approximate real student lifestyle data
    Section 2.1 generates pretraining data from chosen distributions, assuming these reflect the real IISER Bhopal population; if not, the pretrained representations may not transfer.
  • ad hoc to paper Sharing model weights without noise provides differential privacy
    Sections 1.3 and 5 assert this premise. It is false without formal DP mechanisms, yet the entire privacy-preserving claim rests on it.
  • domain assumption 8 days of fine-tuning data per client is sufficient for personalization
    Section 2.2 splits each client's 15 days into 8 training and 7 test days; no justification is given that this amount supports the federated fine-tuning.
  • domain assumption The sign-based accuracy metric is a meaningful measure of recommendation quality
    Section 3.4 defines accuracy as sign agreement, but with heavily zero-dominated labels this metric rewards trivial constant-zero predictions; no baseline is reported.
  • domain assumption Accelerometer-based sleep and step tracking produce accurate ground truth
    Section 3.1 describes heuristic algorithms; for instance, sleep hours are inferred from inactivity windows, which can misclassify quiet wakefulness as sleep.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RiM: Record, Improve and Maintain Physical Well-being using Federated Learning." pith.science (2026). https://pith.science/paper/P7JW3XUW

@misc{pith2026250506384,
  author       = {Pith},
  title        = {Pith review of: RiM: Record, Improve and Maintain Physical Well-being using Federated Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P7JW3XUW}},
  note         = {Machine review of arXiv:2505.06384}
}
read the original abstract

In academic settings, the demanding environment often forces students to prioritize academic performance over their physical well-being. Moreover, privacy concerns and the inherent risk of data breaches hinder the deployment of traditional machine learning techniques for addressing these health challenges. In this study, we introduce RiM: Record, Improve, and Maintain, a mobile application which incorporates a novel personalized machine learning framework that leverages federated learning to enhance students' physical well-being by analyzing their lifestyle habits. Our approach involves pre-training a multilayer perceptron (MLP) model on a large-scale simulated dataset to generate personalized recommendations. Subsequently, we employ federated learning to fine-tune the model using data from IISER Bhopal students, thereby ensuring its applicability in real-world scenarios. The federated learning approach guarantees differential privacy by exclusively sharing model weights rather than raw data. Experimental results show that the FedAvg-based RiM model achieves an average accuracy of 60.71% and a mean absolute error of 0.91--outperforming the FedPer variant (average accuracy 46.34%, MAE 1.19)--thereby demonstrating its efficacy in predicting lifestyle deficits under privacy-preserving constraints.

Figures

Figures reproduced from arXiv: 2505.06384 by the authors.

Figure 2.1
Figure 2.1. The figure illustrates the distributions of simulated features used for pre￾training the MLP model. Simulated step counts follow a negative binomial distribution, while distance traveled and weight are drawn from log-normal distributions. In addi￾tion, meal consumption and gender are generated using Bernoulli distributions, and age is modeled based on an empirical distribution. The figure also illustrates the normal… view at source ↗
Figure 2.2
Figure 2.2. The figure presents the distribution of meal consumption—breakfast, lunch, and dinner—where a value of 1 indicates the meal was taken and 0 indicates it was skipped. It also shows the gender distribution within the simulated dataset. All variables follow a Bernoulli distribution. 5 [PITH_FULL_IMAGE:figures/full_fig_p014_2_2.png] view at source ↗
Figure 3.1
Figure 3.1. Architecture of the proposed MLP model, which consists of five hidden layers and outputs sleep and distance deficits that feed into the recommendation system to generate personalized recommendations. We employ a two-stage approach that (i) predicts lifestyle deficits via a pretrained MLP regression model and (ii) applies rule-based severity and interaction checks to com￾pute per-parameter risk scores. Only the highe… view at source ↗
Figures from the paper (2 more)
Figure 3.2
Figure 3.2. Figure 3.2: Training loss and validation accuracy curves plotted across pre-training iterations of the MLP model. 3.3 Federated Learning We use the Flower framework to implement federated learning. A central server is set up to load the pre-trained model and distribute its weigh…
Figure 3.3
Figure 3.3. Figure 3.3: The figure illustrates the FedPer weight-update process: orange neurons denote the shared root layers that are sent to the server averaged via FedAvg, while blue neurons denote the personal head layers that remain on the client device to preserve personalization. We …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 20 canonical work pages

  1. [1]

    Effectiveness of digital health interventions in promoting physical activity among college students: Systematic review and meta-analysis

    Siyuan Bi, Junfeng Yuan, Yanling Wang, Wenxin Zhang, Luqin Zhang, Yongjuan Zhang, Rui Zhu, and Lin Luo. Effectiveness of digital health interventions in promoting physical activity among college students: Systematic review and meta-analysis. Journal of medical Internet research , 26:e51714, 2024

    work page 2024

  2. [2]

    Ratings and experiences in using a mobile application to increase physical activity among university students: implications for future design

    Caroline A Figueroa, Laura Gomez-Pathak, Imran Khan, Joseph Jay Williams, Courtney R Lyles, and Adrian Aguilera. Ratings and experiences in using a mobile application to increase physical activity among university students: implications for future design. Universal Access in the Information Society , 23(2):821--830, 2024

    work page 2024

  3. [3]

    The use of mobile apps and fitness trackers to promote healthy behaviors during covid-19: A cross-sectional survey

    Huong Ly Tong, Carol Maher, Kate Parker, Tien Dung Pham, Ana Luisa Neves, Benjamin Riordan, Clara K Chow, Liliana Laranjo, and Juan C Quiroz. The use of mobile apps and fitness trackers to promote healthy behaviors during covid-19: A cross-sectional survey. PLOS Digital Health , 1(8):e0000087, 2022

    work page 2022

  4. [4]

    The effectiveness of a mobile phone--based physical activity program for treating depression, stress, psychological well-being, and quality of life among adults: quantitative study

    Hyungsook Kim, Kikwang Lee, Ye Hoon Lee, Yoonjung Park, Yonghyun Park, Yeonwoo Yu, Jaeyoung Park, and Sihyeon Noh. The effectiveness of a mobile phone--based physical activity program for treating depression, stress, psychological well-being, and quality of life among adults: quantitative study. JMIR mHealth and uHealth , 11:e46286, 2023

    work page 2023

  5. [5]

    Using physical health apps to promote healthy lifestyles in youth mental healthcare: A nationwide perspective-gathering exercise of over 400 service users

    Joseph Firth, Chelsea Sawyer, John Sainsbury, Rachel Morell, Hamish Fibbins, Sandra Bucci, Lamiece Hassan, Josh A Firth, Henry Onyweaka, John Torous, et al. Using physical health apps to promote healthy lifestyles in youth mental healthcare: A nationwide perspective-gathering exercise of over 400 service users. Psychiatry Research , 342:116187, 2024

    work page 2024

  6. [6]

    A survey on multi-task learning

    Yu Zhang and Qiang Yang. A survey on multi-task learning. IEEE transactions on knowledge and data engineering , 34(12):5586--5609, 2021

    work page 2021

  7. [7]

    Multitask learning for automated sleep staging and wearable technology integration

    Hao-Yi Chih, Tanveer Ahmed, Amy P Chiu, Yu-Ting Liu, Hsin-Fu Kuo, Albert C Yang, and Der-Hsien Lien. Multitask learning for automated sleep staging and wearable technology integration. Advanced Intelligent Systems , 6(1):2300270, 2024

    work page 2024

  8. [8]

    Multitask deep learning for cost-effective prediction of patient's length of stay and readmission state using multimodal physical activity sensory data

    Sajid Ali, Shaker El-Sappagh, Farman Ali, Muhammad Imran, and Tamer Abuhmed. Multitask deep learning for cost-effective prediction of patient's length of stay and readmission state using multimodal physical activity sensory data. IEEE Journal of Biomedical and Health Informatics , 26(12):5793--5804, 2022

    work page 2022

Show all 25 references
  1. [9]

    Communication-efficient learning of deep networks from decentralized data

    Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguera y Arcas. Communication-efficient learning of deep networks from decentralized data. In Artificial intelligence and statistics , pages 1273--1282. PMLR, 2017

    2017

  2. [10]

    Differentially private federated learning: A client level perspective

    Robin C Geyer, Tassilo Klein, and Moin Nabi. Differentially private federated learning: A client level perspective. arXiv preprint arXiv:1712.07557 , 2017

    2017 arXiv

  3. [11]

    Federated f-differential privacy

    Qinqing Zheng, Shuxiao Chen, Qi Long, and Weijie Su. Federated f-differential privacy. In International conference on artificial intelligence and statistics , pages 2251--2259. PMLR, 2021

    2021

  4. [12]

    Flower: A friendly federated learning research framework

    Daniel J Beutel, Taner Topal, Akhil Mathur, Xinchi Qiu, Javier Fernandez-Marques, Yan Gao, Lorenzo Sani, Kwing Hei Li, Titouan Parcollet, Pedro Porto Buarque de Gusm \ a o, et al. Flower: A friendly federated learning research framework. arXiv preprint arXiv:2007.14390 , 2020

    2007 arXiv

  5. [13]

    Brendan McMahan, Eider Moore, Daniel Ramage, and Blaise Ag \" u era y Arcas

    H. Brendan McMahan, Eider Moore, Daniel Ramage, and Blaise Ag \" u era y Arcas. Federated learning of deep networks using model averaging. CoRR , abs/1602.05629, 2016

    2016 arXiv

  6. [14]

    Federated learning with personalization layers

    Manoj Ghuhan Arivazhagan, Vinay Aggarwal, Aaditya Kumar Singh, and Sunav Choudhary. Federated learning with personalization layers. arXiv preprint arXiv:1912.00818 , 2019

    1912 arXiv

  7. [15]

    How many steps/day are enough? preliminary pedometer indices for public health

    Catrine Tudor-Locke and David R Bassett. How many steps/day are enough? preliminary pedometer indices for public health. Sports medicine , 34:1--8, 2004

    2004

  8. [16]

    The scaling laws of human travel

    Dirk Brockmann, Lars Hufnagel, and Theo Geisel. The scaling laws of human travel. Nature , 439(7075):462--465, 2006

    2006

  9. [17]

    Self-reported and measured sleep duration: how similar are they? Epidemiology , 19(6):838--845, 2008

    Diane S Lauderdale, Kristen L Knutson, Lijing L Yan, Kiang Liu, and Paul J Rathouz. Self-reported and measured sleep duration: how similar are they? Epidemiology , 19(6):838--845, 2008

    2008

  10. [18]

    Adolescents’ eating patterns influence their nutrient intakes

    JOHANNA T DWYER, MARGUERITE EVANS, ELAINE J STONE, HENRY A FELDMAN, LESLIE LYTLE, DEANNA HOELSCHER, CAROLYN JOHNSON, MICHELLE ZIVE, and MINHUA YANG. Adolescents’ eating patterns influence their nutrient intakes. Journal of the American Dietetic Association , 101(7):798--802, 2001

    2001

  11. [19]

    Determinants of variation in adult body height

    Karri Silventoinen. Determinants of variation in adult body height. Journal of biosocial science , 35(2):263--285, 2003

    2003

  12. [20]

    Prevalence of obesity and trends in the distribution of body mass index among us adults, 1999-2010

    Katherine M Flegal, Margaret D Carroll, Brian K Kit, and Cynthia L Ogden. Prevalence of obesity and trends in the distribution of body mass index among us adults, 1999-2010. Jama , 307(5):491--497, 2012

    1999

  13. [21]

    World population prospects 2019: Highlights

    United Nations, Department of Economic and Social Affairs, Population Division . World population prospects 2019: Highlights. https://population.un.org/wpp/Download/Standard/, 2019. Accessed: 2025-04-03

    2019

  14. [22]

    Global health observatory data repository: Sex ratio at birth

    World Health Organization . Global health observatory data repository: Sex ratio at birth. https://www.who.int/data/gho/data/indicators/indicator-details/GHO/sex-ratio-at-birth, 2018. Accessed: 2025-04-03

    2018

  15. [23]

    Obesity: preventing and managing the global epidemic

    World Health Organization . Obesity: preventing and managing the global epidemic. Technical Report 894, World Health Organization, Geneva, Switzerland, 2000

    2000

  16. [24]

    Breakfast habits, nutritional status, body weight, and academic performance in children and adolescents

    Gail C Rampersaud, Mark A Pereira, Beverly L Girard, Judi Adams, and Jordan D Metzl. Breakfast habits, nutritional status, body weight, and academic performance in children and adolescents. Journal of the american dietetic association , 105(5):743--760, 2005

    2005

  17. [25]

    A. K. Kant and B. I. Graubard. A little change goes a long way: the impact of meal definition on nutrient intake estimates. Journal of Nutrition , 140(2):353--358, 2010

    2010

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.