Triangulation in aetiological epidemiology

May combine two claims.

Participants in the RCT were randomized at a median of 29 weeks of gestation, and those in the intervention arm had their glucose monitored from that time; that approach reflects the effect of lowered fasting glucose for the ~last 11 weeks of gestation.

“Participants in the RCT were randomized at a median of 29 weeks of gestation, and those in the intervention arm had their glucose monitored from that time; that approach reflects the effect of lowered fasting glucose for the ∼ last 11 weeks of gestation.”

May combine two claims.

We observe similar effects in the RCT and MR approaches and weaker effects in the multivariable regression analyses.

“We observe none of these, rather we observe similar effects in the RCT and MR approaches and weaker effects in the multivariable regression analyses.”

May combine two claims.

Additional adjustment for maternal socioeconomic position, age, BMI and parity attenuated the associations in White British BiB women towards the minimally adjusted results for Pakistani women, whereas these adjustments in Pakistani women did not notably alter its magnitude.

“Additional adjustment for maternal socioeconomic position, age, body mass index (BMI) and parity attenuated the associations in the White British BiB women towards the minimally adjusted results for the Pakistani women, whereas these adjustments in the Pakistani women did not notably alter its magnitude.”

May combine two claims.

Effect estimates from all three approaches (MR, prospective cohorts, RCT) point in the same direction of lower SBP causing reduced odds of CHD, but magnitudes vary, being greatest for MR, intermediate for prospective cohorts, and least for RCTs.

“Effect estimates from all three approaches point in the same direction of lower SBP causing reduced odds of CHD (Figure 2a), but the magnitudes of this effect vary, being greatest for the MR studies, intermediate for the prospective cohorts and least for the RCTs.53”

• cites [53] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

Observational and RCT studies rarely make the exposure duration explicit as part of the research question.

“However, we believe it is important to highlight this as a separate criterion because observational and RCT studies rarely make the exposure duration explicit as part of the research question.”

May combine two claims.

Positive fasting glucose-birthweight effects were also seen in the IV of glucose as an intermediate in an RCT and IV MR studies.

“Positive fasting glucose-birthweight effects were also seen in the IV of glucose as an intermediate in an RCT57 and IV MR studies.55”

• cites [57] sourceNot checked yet

  Not checked in the free run.

• cites [55] Tyrrell, 2016, Genetic evidence for causal relationships between maternal obesity-related traits and birth weight, JAMA sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The within-sibship analyses in both studies were imprecisely estimated and not reliably different from the results of analyses between unrelated individuals.

“though the within-sibship analyses in both studies were imprecisely estimated and not reliably different from the results of analyses between unrelated individuals.62–64”

• cites [62] Evenhouse, 2005, Improved estimates of the benefits of breastfeeding using sibling comparisons to reduce selection bias, Health Serv Res sourceNot checked yet

  Not checked in the free run.

• cites [63] sourceNot checked yet

  Not checked in the free run.

• cites [64] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

In the LMICs, socioeconomic position was not notably related to breastfeeding or related to it in the opposite direction to that seen in the UK cohort.

“In the LMICs, socioeconomic position was not notably related to breastfeeding or related to it in the opposite direction to that seen in the UK cohort.”

• cites [21] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

When we plot the effect of each approach against estimated duration of exposure, the results provide some support for a cumulative duration effect, with a biased (exaggerated) effect from the IV analyses in the RCT.

“When we plot the effect of each approach against estimated duration of exposure, as done by Ference et al.,53 the results provide some support for a cumulative duration effect, with a biased (exaggerated) effect from the IV analyses in the RCT (Figure 3b).”

• cites [53] sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: For many exposures, greater duration of exposure results in greater risk.

Probably not a substantial claim: may not pass S2 (checkable).

The idea behind triangulation is that when we compare different approaches with assumed unrelated sources of bias, particularly if the expected direction of bias for some of the approaches is different, we would not expect to obtain the same estimates of the causal effect (unless all were unbiased).

“The idea behind triangulation is that when we compare different approaches with assumed unrelated sources of bias, particularly if the expected direction of bias for some of the approaches is different, we would not expect to obtain the same estimates of the causal effect (unless all were unbiased).7,15”

• cites [7] sourceNot checked yet

  Not checked in the free run.

• cites [15] Lipton P. Inference to the best explanation. In: Newton-Smith WH (ed). A Companion to the Philosophy of Science. New York, NY: Blackwell, 2000.Not checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: The key sources of bias of different approaches used in a triangulation are largely unrelated to each other.

Probably not a substantial claim: may not pass S2 (checkable).

In the case where the key source of violation of the exclusion restriction criteria is the same, we would want to avoid comparing these two studies in triangulation.

“In this case we would want to avoid comparing these two studies in triangulation, as the key source of bias is essentially the same.”

• AssumptionThe authors state this as an assumption: The key sources of bias of different approaches used in a triangulation are largely unrelated to each other.

May combine two claims.

Finding similar results from the application of each of these statistical methods to a dataset and concluding that this similar answer is likely to be the correct causal answer would be clearly wrong; the similar answer is likely to be the (residually) confounded answer.

“Finding similar results from the application of each of them to a dataset and concluding that this similar answer is likely to be the correct causal answer would be clearly wrong; the similar answer is likely to be the (residually) confounded answer.”

May combine two claims.

In RCTs, failure to compare outcomes by the original randomized groups or subversion of the randomized process by researchers or participants is a likely key source of confounding.

“In RCTs, failure to compare outcomes by the original randomized groups or subversion of the randomized process by researchers or participants is a likely key source of confounding.”

• cites [17] sourceNot checked yet

  Not checked in the free run.

• cites [18] Furberg CD. How should one analyse and interpret clinical trials in which patients don’t take the treatments assigned to them? James Lind Library Bulletin: Commentaries on the History of Treatment Evaluation . 2009. http://www.jameslindlibrary.org/articles/how-should-one-analyse-and-interpret-clinical-trials-in-which-patients-dont-take-the-treatments-assigned-to-them/ (9 May 2016, date last accessed). sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

If the source of violation of the exclusion restriction criteria is different and unrelated in each of these approaches, it would be appropriate to triangulate them.

“but if the source of violation of this assumption is different, and unrelated, in each of these approaches it would be appropriate to triangulate them”

• AssumptionThe authors state this as an assumption: The key sources of bias of different approaches used in a triangulation are largely unrelated to each other.

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

MR is relatively novel and there has been considerable focus on its limitations and methods for testing and dealing with these, in particular for dealing with violation of the exclusion restriction criteria due to genuine pleiotropy.

“Because MR is relatively novel there has been considerable focus on its limitations and methods for testing and dealing with these, in particular for dealing with violation of the exclusion restriction criteria due to genuine pleiotropy.36,37”

• cites [36] sourceNot checked yet

  Not checked in the free run.

• cites [37] sourceNot checked yet

  Not checked in the free run.

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

They represent refinements of conventional approaches, intended to specifically explore and assess different possible sources of bias, by varying the population(s), exposures and/or outcomes under study.

“They represent refinements of conventional approaches, intended to specifically explore and assess different possible sources of bias, by varying the population(s), exposures and/or outcomes under study.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Two methods (MR-Egger and the weighted median estimator) are likely to become increasingly used in MR sensitivity analyses as two-sample MR using summary GWAS data becomes more common.

“Two of these methods in particular (MR-Egger36 and the weighted median estimator37) are likely to become increasingly used in MR sensitivity analyses, as two-sample MR using summary genome-wide association studies (GWAS) data becomes more common.38”

• cites [36] sourceNot checked yet

  Not checked in the free run.

• cites [37] sourceNot checked yet

  Not checked in the free run.

• cites [38] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Confounding can compromise causal effect estimation in multivariable regression in a prospective cohort, RCT, and MR, but its source is markedly different in each.

“Thus, in each of the three approaches (multivariable regression in a prospective cohort, RCT and MR), confounding can compromise causal effect estimation, but its source is markedly different”

Classified as probably not a substantial claim.

MR-Egger and weighted median estimator could be used as a sensitivity analysis to assess violation of the exclusion restriction criteria when several different treatment IVs are used in RCTs.

“Although these approaches have been developed for use in MR, and in particular in the case of MR-Egger for two-sample MR,36,38 they could clearly be used as a sensitivity analysis to assess violation of the exclusion restriction criteria when several different treatment IVs are used in RCTs.”

• cites [36] sourceNot checked yet

  Not checked in the free run.

• cites [38] sourceNot checked yet

  Not checked in the free run.

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

These additional methods are not inherently different from the 'conventional' approaches and several could be applied to the same dataset.

“These additional methods are not inherently different from the ‘conventional’ approaches and several could be applied to the same dataset.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In some triangulation examples, new sources of bias in some approaches might only become apparent once analysis begins.

“In some triangulation examples, new sources of bias in some of the approaches might only become apparent once analysis begins.”

May combine two claims.

The key sources of bias in the prospective cohort studies meta-analysis (possible residual confounding) and RCT (ignoring DBP in addition to SBP) would likely exaggerate a positive effect, whereas additional analyses suggest MR results likely have very little or no bias.

“However, the key sources of bias in the prospective cohort studies meta-analysis (possible residual confounding) and RCT (ignoring the impact of diastolic blood pressure (DBP) in addition to SBP on CHD) would likely exaggerate a positive effect, whereas our additional analyses suggest that the MR results likely have very little or no bias (Table 3; Figures S1–3, available at IJE online).”

May combine two claims.

MR generally examines exposure across a large part of the life course and might be unable to distinguish critical or sensitive-period exposure effects.

“For example, MR generally examines exposure across a large part of the life course and might be unable to distinguish critical or sensitive-period exposure effects.”

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

Timing of exposure might be included as part of a research question, for example many developmental origins questions concern exposure during a particular developmental time period.

“Although timing of exposure might be included as part of a research question–for example, many developmental origins questions are concerned with exposure during a particular (developmental) time period–this will not always be the case.”

Probably not a substantial claim: may not pass S2 (checkable).

Including one or all of cross-context comparison, use of different controls in case-control studies, and negative control studies in a set of additional approaches using triangulation to address a causal question is advantageous.

“Thus, including one or all of these three approaches (cross-context comparison, use of different controls in case control studies, and negative control studies) in a set of (additional) approaches using triangulation to address a causal question is advantageous.”

May combine two claims.

The MR study will likely reflect mean differences in maternal glucose levels across the whole of pregnancy, whereas the multivariable regression analyses in the pregnancy cohorts are based on one baseline measure of fasting glucose taken between 26 and 28 weeks of gestation and thus reflect lower glucose during the last 12-14 weeks of gestation.

“The MR study will likely reflect mean differences in maternal glucose levels across the whole of pregnancy, whereas the multivariable regression analyses in the pregnancy cohorts are based on one (baseline) measure of fasting glucose taken between 26 and 28 weeks of gestation and thus reflect lower glucose during the last 12–14 weeks of gestation (assuming, as in Ference et al.,53 that exposure”

• cites [53] sourceNot checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: In prospective cohort studies, exposure duration is from baseline assessment to end of follow-up.

Probably not a substantial claim: may not pass S2 (checkable).

In a triangulation framework, assumptions about how different approaches assess exposure duration or timing need to be explored explicitly.

“In a triangulation framework this needs to be explored explicitly, as we do in the examples below.”

May combine two claims.

Prospective cohort studies often study exposures over longer time periods than RCTs, and attenuation by errors (regression dilution bias) will influence prospective cohort studies more than trials with reasonable estimates of sustained differences in exposure generated by the intervention.

“For example, whereas prospective cohort studies often will be studying exposures over longer time periods than RCTs, attenuation by errors (i.e. regression dilution bias) will influence prospective cohort studies more than trials in which you have reasonable estimates of the sustained differences in exposure generated by the intervention.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

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”

• cites [49] sourceNot checked yet

  Not checked in the free run.

• cites [50] Smith ADAC Hardy R Heron J . Model selection of life course hypotheses involving continuous exposures. Int J Epidemiolernational Journal of Epidemiology 2016, July 1. pii: dyw164. [Epub ahead of print.]. sourceNot checked yet

  Not checked in the free run.

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

With other cancer controls, selective recall bias is less likely, but it is plausible that the risk factor of interest has a general carcinogenic effect or a specific effect on some other cancers in the same direction as the effect in bowel cancer cases; if true, bias would be towards the null.

“With other (not bowel) cancer controls, selective recall bias is less likely, but it is plausible that the risk factor of interest (unknowingly) has a general carcinogenic effect or a specific effect on some of the ‘other cancers’ in the same direction as the effect in the bowel cancer cases. If that were true in analyses with other cancer controls, the bias would be towards the null.”

May combine two claims.

In many negative control studies, the negative control may show a weaker association than the real study, suggesting some bias in the real study but also some causal effect under the assumptions.

“Rather, the negative control study may show a weaker (than the real study) association, suggesting some bias in the real study but, under the assumptions of a negative control study, there is also some causal effect.”

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

We anticipated a priori that the RCT effects would be exaggerated due to assumptions about change in glucose levels in the control arm, and possibly due to violation of the exclusion restriction criteria, but thought that the MR results might be weakly biased towards the null because of adjustment for offspring genetic variants introducing a path from mother's genetic IV to father's genotype and到

“We anticipated a priori that the RCT effects would be exaggerated due to the assumptions we had to make about change in glucose levels in the control arm, and also possibly due to violation of the exclusion restriction criteria (Table 3; and Supplementary text, available at IJE online), but thought that the MR results might be weakly biased towards the null because of adjustment for offspring”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Some approaches in a specific example might have several important sources of bias with different directions rather than one key source.

“Some approaches in a specific example might have several important sources of bias that have different directions (rather than one key source).”

Probably not a substantial claim: may not pass S4 (based).

Assumptions made about how different approaches assess exposure duration or timing can be biased.

“However, it is important to acknowledge that assumptions made about how different approaches assess exposure duration or timing can be biased.”

May combine two claims.

Evidence from the cross-context comparison suggested that there was residual confounding in the European multivariable approach, but that may not have fully explained a positive effect.

“Evidence from our cross-context comparison suggested that there was residual confounding in the European multivariable approach, but that may not have fully explained a positive effect.”

May combine two claims.

In case-control studies using population controls, the key source of bias is likely selective recall bias of exposure, which in most cases would tend to bias results away from the null and towards the risk factor increasing cancer risk.

“The key source of bias with population controls is likely to be selective recall bias of the exposure (i.e. a tendency to over-report exposures that are thought to be risk factors in the population being studied), which in most cases would tend to bias results away from the null and towards the risk factor increasing cancer risk.”

Probably not a substantial claim: may not pass S2 (checkable).

Once approaches and data are decided on, it is important that key sources of bias and the likely overall direction of these for each approach are articulated and taken into account when comparing results in a triangulation framework.

“As we demonstrate in the illustrative examples, once approaches and data are decided on, it is important that key sources of bias and the likely overall direction of these for each approach are articulated and taken into account when comparing results in a triangulation framework.”

May combine two claims.

RCTs will, in general, test exposure over a shorter duration than conventional approaches applied within prospective cohort studies, and MR studies will often test exposure duration across a much longer period of the life course, potentially from conception.

“In relation to the different approaches that might be compared in a triangulation framework, RCTs will, in general, test exposure over a shorter duration than conventional approaches applied within prospective cohort studies, and MR studies will often test exposure duration across a much longer period of the life course (potentially from conception).”

Classified as probably not a substantial claim.

In cross-context comparisons, results are compared between two or more different populations, or the same population in two or more different contexts, under the a priori assumption that confounding structures differ between the contexts/populations.

“In cross-context comparisons, results are compared between two or more different populations, or the same population in two or more different contexts (e.g. over different time periods), under the a priori assumption that confounding structures differ between the different contexts/populations.21”

• cites [21] sourceNot checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: Confounding structures differ between the different contexts/populations in cross-context comparisons.

May combine two claims.

Using different controls in case-control studies has similar underlying principles to cross-context studies, but differences in exposure misclassification rather than confounding structures is the most common use.

“Using different controls in case control studies22 has similar underlying principles to those described above for cross-context studies, but here differences in exposure misclassification (rather than confounding structures) is the most common use of this approach.22–24”

• cites [22] Smith, 1988, Cancer case-control studies with other cancers as controls, Int J Epidemiol sourceNot checked yet

  Not checked in the free run.

• cites [23] sourceNot checked yet

  Not checked in the free run.

• cites [24] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

By emphasizing the importance of taking account of duration and timing of exposure in the criteria, triangulation might highlight the possibility of time-specific effects, which can be further tested in a triangulation framework focused on the specific time-exposure question.

“By emphasizing the importance of taking account of duration and timing of exposure in our criteria, triangulation might highlight the possibility of time-specific effects (i.e. generate a hypothesis), which can be further tested in a triangulation framework focused on the specific time-exposure question.”

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

Triangulation works best if within the set of approaches being compared there are some with differences in their expected directions of bias.

“Triangulation works best if within the set of approaches being compared there are some with differences in their expected directions of bias.”

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

In cross-context comparisons, additional populations are sought in which confounders of concern are unrelated to exposure or outcome, related in the opposite direction, or do not vary; thus in additional populations there would be no confounding or confounding in the opposite direction.

“The idea with this approach is that where there are concerns about the association being driven by residual confounding in one context/population, additional populations are sought in which the confounders of concern are unrelated to exposure or outcome (or related in the opposite direction) or do not vary in the additional populations. Thus, in the additional populations there would either be no”

• AssumptionThe authors state this as an assumption: Confounding structures differ between the different contexts/populations in cross-context comparisons.

Classified as probably not a substantial claim.

There would be violation of the exclusion restriction criteria if the intervention, in addition to influencing breastfeeding, also affected the mothers, such that they had a tendency to a healthier lifestyle more generally.

“There would be violation of the exclusion restriction criteria if the intervention, in addition to influencing breastfeeding, also affected the mothers, such that they had a tendency to a healthier lifestyle more generally”

Classified as probably not a substantial claim.

Three studies were identified which examined associations of being breastfed with BMI.

“Three studies were identified which examined associations of being breastfed with BMI.”

• cites [62] Evenhouse, 2005, Improved estimates of the benefits of breastfeeding using sibling comparisons to reduce selection bias, Health Serv Res sourceNot checked yet

  Not checked in the free run.

• cites [63] sourceNot checked yet

  Not checked in the free run.

• cites [64] sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

An intention-to-treat analysis was used to assess the impact of these differences on BMI at age 6.5 (n = 13,879) and on BMI and fat mass index (FMI) at age 11.5 (n = 13,866).

“An intention- to-treat analysis was used to assess the impact of these differences on BMI at age 6.5 (n = 13 879) and on BMI and fat mass index (FMI) at age 11.5 (n = 13 866).”

• cites [65] Kramer, 2007, Long-term effects of prolonged and exclusive breastfeeding on child height, weight, adiposity, and blood pressure: new evidence from a large randomized trial, Am J Clin Nutr sourceNot checked yet

  Not checked in the free run.

• cites [66] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

We explored associations of observed confounders with these outcomes and these verified that these were appropriate negative controls.

“We explored associations of observed confounders with these outcomes and these verified that these were appropriate negative controls (Table S3)”

• cites Table S3Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Timing of exposure will be similar in all approaches.

“timing will be similar in all approaches”

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

As the length of exposure to breastfeeding varied across the studies or was not measured in some, we have not commented on duration of exposure.

“As the length of exposure (to breastfeeding) varied across the studies or was not measured in some, we have not commented on duration of exposure”

Probably not a substantial claim: may not pass S2 (checkable).

For this example, our triangulation effort highlights the need for additional evidence.

“However, for this example, we would suggest our triangulation effort highlights the need for additional evidence.”

May combine two claims.

The intervention resulted in marked differences in ever breastfeeding, duration of breastfeeding and whether breastfeeding was exclusive.

“The intervention resulted in marked differences in ever breastfeeding, duration of breastfeeding and whether breastfeeding was exclusive.”

• cites [65] Kramer, 2007, Long-term effects of prolonged and exclusive breastfeeding on child height, weight, adiposity, and blood pressure: new evidence from a large randomized trial, Am J Clin Nutr sourceNot checked yet

  Not checked in the free run.

• cites [66] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

In the cross-context comparison, SEP did not relate to breastfeeding in LMICs or was in the opposite direction to that seen in the UK; if an association in UK participants was due to SEP confounding, a null or opposite direction association would be expected in those from LMICs.

“It was demonstrated that SEP did not relate to breastfeeding in the LMICs or was in the opposite direction (more affluent and educated women being less likely to breastfeed) to that seen in the UK (breastfeeding more common in the more affluent and educated women). If an association in UK participants was due to SEP confounding, we would expect a null or opposite direction association in those”

• cites [21] sourceNot checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: Confounding structures differ between the different contexts/populations in cross-context comparisons.

May combine two claims.

Two of the identified studies used the same data from the US National Longitudinal Study of Adolescent Health (Add Health) and had similar results.

“Two of these used the same data from the US National Longitudinal Study of Adolescent Health (Add Health) and had similar results;62,63”

• cites [62] Evenhouse, 2005, Improved estimates of the benefits of breastfeeding using sibling comparisons to reduce selection bias, Health Serv Res sourceNot checked yet

  Not checked in the free run.

• cites [63] sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The third study was US-based and included 488 sib-pairs whose BMI was assessed at ages 9-19 years.

“The third study was also US-based and included 488 sib-pairs whose BMI was assessed at ages 9–19 years64”

• cites [64] sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

We tried to identify outcomes in the ALSPAC cohort that would be affected by confounders important in the relation of breastfeeding to later offspring outcomes but for which a biological/causal effect of breastfeeding was unlikely.

“We tried to identify outcomes in the ALSPAC cohort67,68 that would be affected by confounders that are important in the relation of breastfeeding to later offspring outcomes but for which a biological/causal effect of breastfeeding was unlikely”

• cites [67] sourceNot checked yet

  Not checked in the free run.

• cites [68] sourceNot checked yet

  Not checked in the free run.

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

The study that focused on mean BMI included 2734 sib-pairs aged 12-18.

“we present results from the study that focused on mean BMI (2734 sib-pairs; ages 12–18).62”

• cites [62] Evenhouse, 2005, Improved estimates of the benefits of breastfeeding using sibling comparisons to reduce selection bias, Health Serv Res sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Some of the approaches that we describe could be considered a form of triangulation within themselves.

“Some of the approaches that we describe could be considered a form of triangulation within themselves.”

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

The process for conducting IARC monographs also has some similarities to triangulation, in that the aim is to identify and review all relevant papers on cancer in humans and experimental animals.

“The process for conducting International Agency for Research on Cancer (IARC) monographs also has some similarities to triangulation, in that the aim is to identify and review ‘all relevant papers on cancer in humans and experimental animals …’74”

• cites [74] Pearce, 2015, IARC Monographs: 40 years of evaluating carcinogenic hazards to humans, Environ Health Perspect sourceNot checked yet

  Not checked in the free run.

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

As more observational studies have genetic data, it will be increasingly possible to compare conventional multivariable regression approaches, MR and possibly negative control (exposure or outcome) studies, within the same datasets.

“As more observational studies have genetic data, it will be increasingly possible to compare conventional multivariable regression approaches, MR and possibly negative control (exposure or outcome) studies, within the same datasets.”

Classified as probably not a substantial claim.

IARC monographs are not explicit about using a triangulation approach.

“However, these monographs are not explicit about using a triangulation approach.”

Probably not a substantial claim: may not pass S1 (asserted). May combine two claims.

We anticipated that if the 'real' inverse association of being breastfed with obesity was due to confounding that was mimicked in the negative controls, we would see a positive association between home invasion of mice and obesity and an inverse association between invasion of pigeons and obesity.

“we anticipated that if the ‘real’ inverse association of being breastfed with obesity was due to confounding that was mimicked in the negative controls, we would see a positive association between home invasion of mice and obesity and an inverse association between invasion of pigeons and obesity (Supplementary text).”

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

Pointing to the same conclusion does not mean that the results are statistically consistent and could be pooled; currently triangulation will mostly provide a qualitative assessment of the strength of evidence regarding causality.

“Pointing to the same conclusion does not mean that the results are statistically consistent and could be pooled; currently triangulation will mostly provide a qualitative assessment of the strength of evidence regarding causality.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Attempts to use triangulation are likely to yield further developments including analytical methods that could improve how to quantitatively combine data to obtain bounds of likely magnitudes of causal effect.

“As with other areas of data integration and development of novel methods for improving causal inference, attempts to use this approach are likely to yield further developments including analytical methods that could improve how to quantitatively combine data to obtain bounds of likely magnitudes of causal effect.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

The application of IV analyses to test causal effects of intermediates in RCTs, or the use of RCTs to test effects of the primary intervention on a range of outcomes, would be enhanced by greater sharing of data.

“The application of IV analyses to test causal effects of intermediates in RCTs, or the use of RCTs to test effects of the primary intervention on a range of outcomes, would be enhanced by greater sharing of data.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Pooling of data (as in meta-analyses) assumes that all studies give unbiased estimates of the same underlying causal effect, and can be pooled to give one estimate.

“Pooling of data (as in meta-analyses) assumes that all studies give unbiased estimates of the same underlying causal effect, and can be pooled to give one estimate.”

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

Greater data access is increasingly promoted and there are some examples of good practice.

“Greater data access is increasingly promoted and there are some examples of good practice.”

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

The aim of IJE cohort profiles is to increase the ease of data sharing globally, with all profiles having to include a section on 'How can I get hold of the data?'.

“The aim of IJE cohort profiles is to increase the ease of data sharing globally, with all profiles having to include a section on ‘How can I get hold of the data?’.”

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

By necessity the IARC review process will involve consideration of the different sources of bias in each approach used in the identified papers.

“By necessity this review process will involve consideration of the different sources of bias in each approach used in the identified papers.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Cross-context comparisons with many different populations each with different confounding structures, RCTs of intermediates that use multiple different treatments, and MR with multiple genetic instruments that relate to the exposure of interest through different mechanisms, might be considered a form of triangulation within themselves.

“Cross-context comparisons with many different populations each with different confounding structures, RCTs of intermediates that use multiple different treatments (for example different classes of antihypertensives to test the effect of blood pressure on CHD31) and MR with multiple genetic instruments that relate to the exposure of interest through different mechanisms, might be considered in this”

• cites [31] sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S1 (asserted), S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Methods could be developed to apply to triangulation to give plausible bounds of causal effect, including taking account of duration and timing of effect, and possibly exploring whether there are different bounds of causal effects in different subgroups.

“Methods could be developed to apply to triangulation to give plausible bounds of causal effect, including taking account of duration and timing of effect, and possibly exploring whether there are different bounds of causal effects in different subgroups (suggesting effect modification)”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Cross-context comparisons would be enhanced by greater sharing of research data from cohorts across the globe.

“Cross-context comparisons would be enhanced by greater sharing of research data from cohorts across the globe.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Considerable advances could be made within the triangulation framework, at modest cost but with greater access to research data.

“Considerable advances could be made within the triangulation framework, at modest cost but with greater access to research data.”

May combine two claims.

IARC monographs aim to identify all published evidence but do not specifically seek to identify approaches with different key sources of bias or make inferences on the basis of explicitly comparing studies with different biases.

“They aim to identify all published evidence but do not specifically seek to identify approaches with different key sources of bias or make inferences on the basis of explicitly comparing studies with different biases.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Our definition and criteria for triangulation are similar to some other approaches to causality, including Hill's concept of consistency and work by Susser, Morris and more recently Shipley.

“Our definition and criteria for triangulation are similar to some other approaches to causality, including Hill’s concept of consistency16 and work by Susser,69,70 Morris71 and more recently Shipley,72”

• cites [16] Hill, 1965, The environment and disease: association or causation, Proc R Soc MedNot checked yet

  Not checked in the free run.

• cites [69] Susser M. Procedures for establishing causal association. In: Stewart GT (ed). Trends in Epidemiology:Application to Health Service Research and Training. Springfield, IL: C.C. Thomas, 1972.Not checked yet

  Not checked in the free run.

• cites [70] Susser M. Causal Thinking in the Health Sciences. Concepts and Strategies in Epidemiology. New York, NY: Oxford University Press, 1973.Not checked yet

  Not checked in the free run.

• cites [71] Morris JN. Uses of Epidemiology. 2nd edn. Edinburgh, UK: Livingstone, 1964.Not checked yet

  Not checked in the free run.

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

Increased use of triangulation in aetiological epidemiology is likely to require multidisciplinary collaboration, more extensive data sharing and additional resources in comparison with applying just one (conventional) approach.

“Increased use of triangulation in aetiological epidemiology is likely to require multidisciplinary collaboration, more extensive data sharing and additional resources in comparison with applying just one (conventional) approach.”

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

Where results point to different causal answers, understanding the key sources of bias can help direct researchers to what further research is needed to answer the causal question.

“Where results point to different causal answers, understanding the key sources of bias can help direct researchers to what further research is needed to answer the causal question.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

In conclusion, triangulation has considerable potential to improve causal inference in aetiological epidemiology, which will be enhanced with increased data access and multidisciplinary collaborative work.

“In conclusion, we believe that triangulation has considerable potential to improve causal inference in aetiological epidemiology, which will be enhanced with increased data access and multidisciplinary collaborative work.”

Probably not a substantial claim: may not pass S2 (checkable).

Comparisons across multiple populations or multiple IVs are really a way of testing the extent to which the key sources of bias within those approaches are minimized, rather than triangulation, in which different approaches with different (and unrelated) key sources of bias are compared.

“However, comparisons across multiple populations or multiple IVs are really a way of testing the extent to which the key sources of bias within those approaches are minimized, rather than triangulation, in which different approaches with different (and unrelated) key sources of bias are compared.”