CoChat
Sign Up
Cited by 59Statistical Methods in EpidemiologyBirth, Development, and HealthAdvanced Causal Inference Techniques

A structured approach to hypotheses involving continuous exposures over the life course

Open access
Smith, Andrew D.A.C., Hardy, Rebecca, Heron, Jon, Joinson, Carol J., Lawlor, Debbie A., Macdonald-Wallis, Corrie, Tilling, Kate·2016·International Journal of Epidemiology
DOI 10.1093/ije/dyw164PMC PMC5841633
Open this paper in CoChat — free accountOpen PDF

Abstract

BACKGROUND: Epidemiologists are often interested in examining different hypotheses for how exposures measured repeatedly over the life course relate to later-life outcomes. A structured approach for selecting the hypotheses most supported by theory and observed data has been developed for binary exposures. The aim of this paper is to extend this to include continuous exposures and allow for confounding and missing data. METHODS: We studied two examples, the association between: (i) maternal weight during pregnancy and birthweight; and (ii) stressful family events throughout childhood and depression in adolescence. In each example we considered several plausible hypotheses including accumulation, critical periods, sensitive periods, change and effect modification. We used least angle regression to select the hypothesis that explained the most variation in the outcome, demonstrating appropriate methods for adjusting for confounders and dealing with missing data. RESULTS: The structured approach identified a combination of sensitive periods: pre-pregnancy weight, and gestational weight gain 0-20 weeks and 20-40 weeks, as the best explanation for variation in birthweight after adjusting for maternal height. A sensitive period hypothesis best explained variation in adolescent depression, with the association strengthening with the proximity of stressful family events. For each example, these models have theoretical support at least as strong as any competing hypothesis. CONCLUSIONS: We have extended the structured approach to incorporate continuous exposures, confounding and missing data. This approach can be used in either an exploratory or a confirmatory setting. The interpretation, plausibility and consistency with causal assumptions should all be considered when proposing and choosing life course hypotheses.

AI snapshot

Written from the paper’s full text.

TL;DR

This paper extends a structured, LARS-based approach for choosing life-course hypotheses from binary exposures to continuous exposures, while incorporating confounding and missing data. It illustrates the method in two ALSPAC examples: maternal weight in pregnancy and offspring birthweight, and stressful family events in childhood and adolescent depression.

Key findings

  • In 11,499 mother–offspring pairs, the selected model for birthweight included pre-pregnancy weight, total gestational weight gain, and weight gain from 0–20 weeks, with an R-squared of 12% after adjusting for maternal height.
  • Birthweight increased by 10 g per kg pre-pregnancy weight (95% CI 10, 11), 30 g per kg weight gain from 0–20 weeks (95% CI 26, 33), and 10 g per kg weight gain from 20–40 weeks (95% CI 7, 13).
  • In 3,240 female ALSPAC participants, the recency hypothesis best explained variation in depression score at age 14, although all models explained less than 1% of variation; the association increased by 0.014 per additional year of age (95% CI 0.007, 0.022).
  • The LARS procedure selected the recency hypothesis in all 20 imputed datasets, with depression score increases per stressful family event of 0.029 at age 2, 0.059 at age 4, 0.088 at age 6, and 0.117 at age 8.

Methods

  • The authors extended a structured approach for life-course hypotheses using least angle regression (LARS) variable selection, encoding accumulation, critical period, sensitive period, change, effect modification, and threshold hypotheses as variables for continuous exposures.
  • Two ALSPAC examples were analyzed: maternal weight during pregnancy and offspring birthweight in 11,499 mother–offspring pairs, and stressful family events from 6 to 103 months and depressive symptoms at age 14 in 3,240 females.
  • Confounding was addressed using DAGs and adjustment for maternal height; missing data and measurement error were handled with a multilevel model in the birthweight example and multiple imputation by chained equations in the depression example, with elbow plots and covariance tests used to select variables.

Limitations

  • The paper states that strong correlation between exposure measurements may decrease statistical power; without centering, the interaction variable in the first example would have led to selecting an effect modification hypothesis instead of the additive model.
  • The paper states that it is unclear how much hypothesis selection is affected by measurement error in the exposure, and it assumes the exposure is measured with the same precision; it also did not investigate nonlinear relationships, assumed a linear association between stressful life events and depression, and notes that GWG combines multiple phenotypes and better exposure measurements might change findings.
  • The paper states that the structured approach will fail to identify the underlying association if it is not included in the a priori set of hypotheses, and that the method for different confounders does not allow for combining more than one hypothesized association with different potential confounders, as may occur with effect modification.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1093/ije/dyw164

Cites

38 of 38 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 31 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    In the maternal weight and birthweight example, the structured approach identified a combination of sensitive periods: pre-pregnancy weight, and gestational weight gain 0-20 weeks and 20-40 weeks, as the best explanation for variation in birthweight after adjusting for maternal height.

    “The structured approach identified a combination of sensitive periods: pre-pregnancy weight, and gestational weight gain 0-20 weeks and 20-40 weeks, as the best explanation for variation in birthweight after adjusting for maternal height.”
    • Also stated in: § Abstract · plain
      “In an example investigating the association between maternal weight during pregnancy and offspring birthweight, the structured approach identified a combination of pre-pregnancy weight, weight gain during 0-20 weeks of gestation, and weight gain during 20-40 weeks of gestation, as the best explanation for variation in birthweight.”
    Quote in the paperSubstantial claim
  • Findingown data

    Pre-pregnancy weight is positively associated with birthweight, with a change of 10 g per kg increase (95% CI 10, 11), adjusted for maternal height.

    “| Pre-pregnancy weight (kg) | 60.7 (12.6) | 10 (10, 11) | 0.0001 |”
    Quote in the paperNot a substantial claim
  • Findingown data

    The R-squared value for all models in the depression and stressful family events example is less than 1%.

    “The R-squared value for all models is less than 1%, indicating that very little of the depression score at age 14 can be attributed to stressful family events over the preceding life course.”
    Quote in the paperSubstantial claim
  • Findingown data

    The selected model for the GWG example has an R-squared value of 12%.

    “The clearest elbow is at three additional variables, which are pre-pregnancy weight, total GWG and GWG between 0 and 20 weeks (R-squared value 12%).”
    Quote in the paperNot a substantial claim
  • Findingown data

    In the GWG example, the clearest elbow in the LARS procedure is at three additional variables: pre-pregnancy weight, total GWG, and GWG between 0 and 20 weeks.

    “The clearest elbow is at three additional variables, which are pre-pregnancy weight, total GWG and GWG between 0 and 20 weeks (R-squared value 12%).”
    Quote in the paperSubstantial claim
  • May combine two claims. 8
  • The quote only partly states this claim. 1
  • Classified as probably not a substantial claim. 26

  • May combine two claims.

    Findingown data

    In the stressful family events and depression example, a sensitive period hypothesis best explained variation in adolescent depression, with the association strengthening with the proximity of stressful family events.

    “A sensitive period hypothesis best explained variation in adolescent depression, with the association strengthening with the proximity of stressful family events.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In the proximal model identified by LARS, the change in depression score per stressful family event increased with age at event, from 0.029 (95% CI 0.014, 0.045) at age 2 to 0.117 (0.056, 0.178) at age 8.

    “| Age in years at event(s) | Change in depression score per event (95% confidence interval) | P-value from covariance test |”
    • Also stated in: § Results · plain
      “| 2 | 0.029 (0.014, 0.045) | |”
    • Also stated in: § Results · plain
      “| 8 | 0.117 (0.056, 0.178) | < 0.0001 |”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The association between stressful life events and depression is linear.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The Avon Longitudinal Study of Parents and Children (ALSPAC) is a prospective population-based birth cohort study that recruited 14 541 pregnant women resident in Avon, UK, with expected dates of delivery 1 April 1991 to 31 December 1992.

    “The Avon Longitudinal Study of Parents and Children (ALSPAC) is a prospective population-based birth cohort study that recruited 14 541 pregnant women resident in Avon, UK, with expected dates of delivery 1 April 1991 to 31 December 1992.26,27”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [26] source

      Not checked in the free run.

    • cites [27] source

      Not checked in the free run.

  • May combine two claims.

    Premisenone given

    Individuals with missing data on one or more exposure measurements would not be included in a complete cases analysis, leading to loss of power and potential bias.

    “Individuals with missing data on one or more exposure measurements would not be included in a complete cases analysis, leading to loss of power and potential bias.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Under the developmental overnutrition hypothesis, maternal early-/pre-pregnancy weight and weight gain up to 20 weeks are anticipated to be related to birthweight.

    “Under this hypothesis we might anticipate maternal early-/pre-pregnancy weight (reflecting her level of adiposity) and weight gain up to 20 weeks (when maternal fat deposition contributes more to weight gain than other constituents including fetal weight, placenta and amniotic fluid) to be related to birthweight.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    Methods for handling missing data can be incorporated into the structured approach.

    “Methods for handling confounding and missing data can be incorporated into the structured approach.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given

    The structured approach can be used in either an exploratory or a confirmatory setting.

    “This approach can be used in either an exploratory or a confirmatory setting.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given

    Methods for handling confounding can be incorporated into the structured approach.

    “Methods for handling confounding and missing data can be incorporated into the structured approach.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    One might anticipate pre-pregnancy weight to interact positively with GWG.

    “Indeed, one might anticipate pre-pregnancy weight to interact positively with GWG, since the developing fetus of women with greater adiposity will gain more weight throughout pregnancy (under the developmental overnutrition hypothesis) and fetal weight gain will directly contribute to their greater birthweight.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The developmental overnutrition hypothesis suggests that greater maternal adiposity, and associated greater maternal circulating glucose and other nutrients, overfeed the developing fetus resulting in greater birthweight and adiposity throughout life.

    “The ‘developmental overnutrition hypothesis’ suggests that greater maternal adiposity, and associated greater maternal circulating glucose and other nutrients, overfeed the developing fetus resulting in greater birthweight and adiposity throughout life.23,24”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [23] source

      Not checked in the free run.

    • cites [24] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Since GWG includes birthweight, later GWG where the fetus contributes more may also be important.

    “However, since GWG includes birthweight (the final attained weight of the fetus), later GWG where the fetus contributes more may also be important.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The Institute of Medicine publishes recommended upper and lower thresholds for GWG according to pre-pregnancy BMI, with the aim of reducing the implications of GWG including child obesity.

    “The Institute of Medicine (IOM) publishes recommended upper and lower thresholds for GWG according to pre-pregnancy body mass index (BMI), with the aim of reducing the implications of GWG including child obesity.25”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [25] Institute of Medicine and National Research Council. Weight Gain During Pregnancy: Reexamining the Guidelines. Washington, DC: National Academies Press, 2009.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    For the alternative explanations that genetic variants influence adiposity or that mother's weight reflects pelvic size, only pre-pregnancy weight is anticipated to be associated with birthweight.

    “For both of these we might anticipate that only pre-pregnancy weight will be associated with birthweight.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Unlike stepwise variable selection procedures, LARS does not over-inflate effect size estimates during variable selection nor bias hypothesis tests of those effect sizes.

    “Unlike stepwise variable selection procedures, LARS does not over-inflate effect size estimates during variable selection,14 nor bias hypothesis tests of those effect sizes.19”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [14] source

      Not checked in the free run.

    • cites [19] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    The use of a DAG to specify causal assumptions can aid the identification of potential confounders, as variables blocking back-door paths from exposure to outcome.

    “The use of a DAG to specify causal assumptions can aid the identification of potential confounders, as variables blocking back-door paths from exposure to outcome.20”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [20] Pearl, 2014, Confounding equivalence in causal inference, J Causal Infer source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Family stability in the first 5 years of life may play a role in subsequent development of depression.

    “Family stability in the first 5 years of life may play a role in subsequent development of depression.33”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [33] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    C and Δ3 were centred to reduce their correlation with their product.

    “LARS can suffer from low power if some variables are highly correlated,13 therefore we centred C and Δ3 to reduce their correlation with their product.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

    Premisecitation

    The recency hypothesis holds that adverse exposures have time-limited depressogenic effects, meaning proximal exposures are predicted to have stronger associations with depression compared with distal exposures.

    “Another potential mechanism is the ‘recency hypothesis’,34 in which adverse exposures have time-limited depressogenic effects, meaning proximal exposures are predicted to have stronger associations with depression compared with distal exposures.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [34] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The analysis adjusted for potential confounding by maternal height.

    “We adjusted for potential confounding by maternal height using the method discussed above.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S1 (asserted).

    Premisecitation

    Exposure to multiple stressors has a greater adverse effect on mental health than exposure to single stressors.

    “It has been suggested that exposure to multiple stressors has a greater adverse effect on mental health than exposure to single stressors,31 and therefore it is likely that mental health outcomes such as depression will increase in severity with increased number of stressful family events.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [31] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Height may have a non-linear (power) relationship with weight.

    “Height may have a non-linear (power) relationship with weight,29,30 so we used LARS to identify a linear representation of height as the most appropriate power to include.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [29] source

      Not checked in the free run.

    • cites [30] source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductcitation

    A multilevel model was used to obtain estimated maternal weight at 0, 20 and 40 weeks of gestation in 11,499 mother-offspring pairs, and these estimates overcame the issues of irregular sampling and missing data in the exposure measurements.

    “we used a multilevel model to obtain estimated maternal weight at 0 (i.e. pre-pregnancy), 20 and 40 weeks of gestation in 11 499 mother-offspring pairs.28 Such estimates overcame the issues of irregular sampling and missing data in the exposure measurements.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [28] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted).

    Interpretationcitation

    Mental health outcomes such as depression will increase in severity with increased number of stressful family events.

    “It has been suggested that exposure to multiple stressors has a greater adverse effect on mental health than exposure to single stressors,31 and therefore it is likely that mental health outcomes such as depression will increase in severity with increased number of stressful family events.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Exposure to adversities in early childhood is more strongly associated with subsequent mental health problems than exposure to these events in later childhood and adolescence.

    “Alternatively, there is growing evidence that exposure to adversities in early childhood are more strongly associated with subsequent mental health problems than exposure to these events in later childhood and adolescence, as specific regions of the developing brain during early childhood might be vulnerable to adverse exposures that could increase risk for depression.32”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [32] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductown data

    LARS was used to identify a linear representation of height as the most appropriate power to include.

    “Height may have a non-linear (power) relationship with weight,29,30 so we used LARS to identify a linear representation of height as the most appropriate power to include.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Specific regions of the developing brain during early childhood might be vulnerable to adverse exposures that could increase risk for depression.

    “Alternatively, there is growing evidence that exposure to adversities in early childhood are more strongly associated with subsequent mental health problems than exposure to these events in later childhood and adolescence, as specific regions of the developing brain during early childhood might be vulnerable to adverse exposures that could increase risk for depression.32”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [32] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Missing data in the exposure were imputed using chained equations.

    “We imputed missing data in the exposure using chained equations.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    In the stressful family events example, imputation using chained equations may be used in conjunction with the LARS structured approach and was robust to random changes between imputed datasets.

    “Finally, in our stressful family events example, we showed how the alternative method for handling missing data, imputation using chained equations, may be used in conjunction with the LARS structured approach and was robust to random changes between imputed datasets.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Women with greater pre-pregnancy weight might on average have greater fat mass, which across pregnancy results in overnutrition of their infant.

    “Thus the observed data provide some support for the developmental overnutrition hypothesis, as women with greater pre-pregnancy weight might on average have greater fat mass, which across pregnancy results in overnutrition of their infant.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    According to the hopelessness theory of depression, distal contributors (e.g. attributional style) influence causal attributions related to proximal life events.

    “According to the hopelessness theory of depression, distal contributors (e.g. attributional style) influence causal attributions related to proximal life events.36”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [36] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The structured approach may be used to assess the extent to which hypotheses concerning the association between continuous exposure measurements and an outcome are supported by observed data.

    “The structured approach may be used to assess the extent to which hypotheses concerning the association between continuous exposure measurements and an outcome are supported by observed data.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    This is likely to occur when the set of potential hypotheses is general enough to include, in combination, all possible associations permitted by the causal assumptions.

    “This is likely to occur when the set of potential hypotheses is general enough to include, in combination, all possible associations permitted by the causal assumptions.”
    Quote in the paperNot a substantial claim
32 of 32

  • The paper relies on this without support.

    The impact of confounding, measurement error in the exposure, and missing data in the exposure and outcome had not yet been explored within the structured approach.

    “Furthermore, the impact of confounding, measurement error in the exposure and missing data in the exposure and outcome have not yet been explored within the structured approach.”

  • The authors state this as an assumption.

    In the examples, all hypotheses considered have the same potential confounders.

    “In our examples we have assumed that all of the hypotheses considered have the same potential confounders”

  • The paper relies on this without support.

    The causal relationships depicted in Figure 1 are correct.

    “Figure 1 shows a DAG depicting the potential causal relationships between the exposure and outcome measurements referred to in hypotheses (i)-(v), and also maternal height, which is a potential confounder.”

  • The authors state this as an assumption.

    The exposure is always measured with the same level of precision.

    “We have assumed that the exposure is always measured with the same level of precision, as variable selection may be affected by the variance of the exposures.”

  • The authors state this as an assumption.

    Different exposure variables have the same level of measurement error.

    “Although not shown here, LARS can also be used to test different hypothesized life course associations in different exposure variables. These too would have to be assumed to have the same level of measurement error.”

  • The authors state this as an assumption.

    The association between stressful life events and depression is linear.

    “Our analysis assumed a linear association between stressful life events and depression, which would need to be investigated further, as would the definition of a stressful life event and whether some could be more influential than others.”
6 of 6

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Whether using a triangulation approach or not, if the timing of exposure is central to the research question then that needs to be specifically included in the hypotheses tested and the comparisons made.

    “Whether using a triangulation approach or not, if the timing of exposure is central to the research question then that needs to be specifically included in the hypotheses tested and the comparisons made.49,50”

    From Triangulation in aetiological epidemiology (2017) · cites this paper as [50]

  • Timing of exposure, such as whether an exposure has an effect only during a critical period, or has different magnitudes of effect at different times in the life course (sensitive periods), are important to consider.

    “Timing of exposure, such as whether or not an exposure has an effect only during a critical period, or has different magnitudes of effect at different times in the life course (sensitive periods) are also important to consider.49,50”

    From Triangulation in aetiological epidemiology (2017) · cites this paper as [50]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Key message names the wrong outcome for stressful family events

    Minor

    The key message says the stressful-family-events example concerns offspring birthweight, but the abstract identifies its outcome as depression in adolescence.

1 of 1

Cites

Flagged first, then by citations

1 of the 39 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Related

Authors
Smith, A. D. A. C., Hardy, R., Heron, J., Joinson, C. J., Lawlor, D. A., Macdonald-Wallis, C., & Tilling, K.
Affiliations
  • University of Bristol (GB)
  • MRC Unit for Lifelong Health and Ageing (GB)
Corresponding author
Smith, Andrew D.A.C.
ORCID
Venue
International Journal of Epidemiology
Publisher
Oxford University Press (OUP)
Published
2016-07-01
Volume
45
Issue
4
Pages
dyw164
ISSN
  • 0300-5771 (ISSN-L)
  • 1464-3685
DOI
10.1093/ije/dyw164
PMC
PMC5841633
OpenAlex
W2344236281
Cited by
59
References
39
Citations by year
2026: 4 · 2025: 4 · 2024: 10 · 2023: 10 · 2022: 9 · 2021: 8 · 2020: 3 · 2019: 4 · 2018: 4 · 2017: 1 · 2016: 2
FWCI
3.71
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • International Journal of Epidemiology (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • UWE Research Repository (UWE Bristol) (submittedVersion, cc-by)
  • UCL Discovery (University College London) (submittedVersion)
  • HighWire Press Open Archive (submittedVersion)
  • Explore Bristol Research (submittedVersion, cc-by)
Funders
  • Wellcome Trust (NF-SI-0611-10196, 10196, MC_UU_12013/5, 102215, MC_UU_12013/9)
  • National Institute for Health and Care Research (NF-SI-0611-10196, SI-0611-10196)
  • University of Bristol (MC_UU_12013/9, NF-SI-0611-10196, MC_UU_12013, MC_UU_12013/5)
  • Medical Research Council (102215, G1000726, MC_UU_12013_5, MC_UU_12013, MR/J011932/1, MC_PC_15018, NF-SI-0611-10196, MC_UU_12019/2, MC_UU_12013/9)
  • Economic and Social Research Council (NF-SI-0611-10196, MC_UU_12019/2, RES-000-)
Topics
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
  • Health Sciences › Medicine › Pediatrics, Perinatology and Child Health › Birth, Development, and Health
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
Keywords
life course exposure, continuous exposure, hypothesis selection, sensitive period, critical period, accumulation, effect modification, confounding, missing data, least angle regression, birth weight, gestational weight gain, adolescent depression
MeSH terms
Adolescent, Adult, Birth Weight, Body Weight, Data Interpretation, Statistical, Depression, Family, Female, Humans, Life Change Events, Pregnancy, Regression Analysis, Socioeconomic Factors, Stress, Psychological, Weight Gain, Causality, Confounding Factors, Epidemiologic, Adolescent Development

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review