Interactions in Epidemiology: Relevance, Identification, and Estimation
The DOI matches its Crossref record; 29 of the 30 references it cites were found in a registry or an index.
Abstract
Sufficient-cause models represent biologic interactions (including synergism) in the form of coparticipation of factors in a sufficient cause1–3 (causal coaction). In this issue of the journal, VanderWeele1 provides new results relating sufficient-cause models to regression models. These results reinforce the point that concepts of biologic interaction do not in general correspond to the concept of statistical interaction, because the latter is only the need for a product term in a statistical model.4,5 They further underscore that, without strict assumptions, sufficient-cause interactions do not correspond to interdependence of effects in potential-outcomes (counterfactual) models. I will examine the practical relevance and limitations of using epidemiologic regressions to study interactions, however the latter are defined. Relevance of Interactions A practical issue concerning interactions is their relevance to public health goals of finding cost-effective intervention strategies for disease reduction. Part of public health involves identifying population subgroups that would benefit most from a given intervention. These are not always high-risk subgroups—it is possible that a lower-risk group would obtain the greatest absolute risk reduction from the proposed intervention. In the absence of bias, departures from risk additivity imply that some subgroups would obtain a greater absolute risk reduction from the intervention than others would. Thus earlier authors noted that if costs were measured in terms of case-load per unit population (average risk), the relevant null model for detecting special groups would be additive on the risk scale.6,7 Risk nonadditivity (or “public health interaction”) implies the existence of such groups.5,6 It also implies existence of interdependent effects in individuals, also known as “interactions” in basic potential-outcome (counterfactual) causal models; that is, risk nonadditivity implies there are individuals such that the presence or direction of effect of one factor depends on the other factor.5,8 Risk nonadditivity is equivalent to heterogeneity or modification of risk differences for one variable across levels of another.5 For factors that can act only causally, excess risk above additivity (superadditivity) among those exposed to both factors signals the presence of individuals who get the disease only when exposed to both factors (“synergism” in the potential-outcomes sense).4,5,8 Identifying such individuals is valuable for preventive medicine and public health, because for them the outcome can be prevented by removing either of the factors. Nonidentifiability of Biologic Interactions Epidemiologic studies observe population distributions of risks, which submerge details of individual experiences. One consequence of this submersion is that many different biologic models will lead to the same population model.9,10 In particular, the implications of biologic models for interaction are unidirectional; for example, risk additivity could hold because no interdependence of effects exists, or instead may hold because of cancellations of the population effects of different individual interactions.5,8 Perfect cancellation of interactions is a higher-order analog of “unfaithfulness” in causal diagrams,11,12 where 2 variables may be unassociated despite having causal connections. For example, a treatment with effects could appear independent of the outcome if it causes and prevents equal numbers of cases. Although perfect cancellation is often implausible, any degree of cancellation will make it more difficult to detect effects or their interdependencies. Even if we could observe each individual's response under each exposure pattern, we still could not determine from those observations which biologic mechanisms or sufficient-cause types were operative8,13–15; hence, individual mechanisms are nonidentifiable even from a perfect factorial or crossover trial.13,15 For example, even with no sufficient-cause synergism, variation in background characteristics could nonetheless produce superadditivity.14 More generally, estimation of degrees of interaction will require further assumptions (eg, monotonicity) that cannot be tested with epidemiologic data, even if the latter are perfect, although valid tests for interactions can be constructed from weaker conditions.1,5,8,13–15 Statistical Modeling Issues Epidemiologic theories of interactions1,4,5,8,9,14–17 derive population causal models from models for individual etiology (eg, risk additivity is derived from the absence of interdependent effects). Study of these models and forms does not require that one use the derived population model as the statistical analysis model. One can estimate risks or rates by using any statistical model, then use these estimates to test or estimate the parameters in the causal model.1,18 For example, one can use risk estimates from a logistic statistical model for inference on parameters in an additiv
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“Regarding studies that employ the product term specification, epidemiologists have called for an end to the practice of assessing interactions based purely on product terms in multiplicative models (Greenland, 2009; Rajaleid, Janszky, & Hallqvist, 2011).”
From Syndemics of psychosocial problems and HIV risk: A systematic review of empirical tests of the disease interaction concept (2015) · cites this paper as Greenland, 2009
In the citing paper
“The concept of disease interaction has had a long and controversial history in the epidemiologic literature (Blot & Day, 1979; Greenland, 2009; Kaufman, 2009; Kupper & Hogan, 1978; Rothman, 1974, 1976a; Rothman, Greenland, & Walker, 1980; Saracci, 1980; Siemiatycki & Thomas, 1981; VanderWeele, 2009a; Walter & Holford, 1978).”
From Syndemics of psychosocial problems and HIV risk: A systematic review of empirical tests of the disease interaction concept (2015) · cites this paper as Greenland, 2009
In the citing paper
“At the same time, from the public health perspective, additive effects are favourable over multiplicative effects to increase the net benefit by allocating the treatment to the proper subgroup.12,13”
From Understanding of interaction (subgroup) analysis in clinical trials (2019) · cites this paper as [12]
Citations
Cites
The Elements of Statistical Learning: Data Mining, Inference, and Prediction
John Lu, Z. Q. (2010) · Journal of the Royal Statistical Society Series A: Statistics in Society · doi:10.1111/j.1467-985x.2010.00646_6.x · Cited by 12,390
Bootstrap Methods and their Application
Davison, A. C., & Hinkley, D. V. (1997) · doi:10.1017/cbo9780511802843 · Cited by 7,350
Friedman, J., Hastie, T., & Tibshirani, R. (2000) · The Annals of Statistics · doi:10.1214/aos/1016218223 · Cited by 6,884 · cited 1×
“included, this goal leads to fitting methods that can use a very large set of terms, such as shrinkage(27) and machine-learning algorithms such as boosting.(28,29) The same methods can also”
MethodROTHMAN, K. J., GREENLAND, S., & WALKER, A. M. (1980) · American Journal of Epidemiology · doi:10.1093/oxfordjournals.aje.a113015 · Cited by 859 · cited 2×
“Thus earlier authors noted that if costs were measured in terms of case-load per unit population (average risk), the relevant null model for detecting special groups would be additive on the risk scale.(6,7) Risk nonadditivity (or “public health interaction”) implies the existence of such groups.”
BackgroundSelf-citationROTHMAN, K. J. (1976) · American Journal of Epidemiology · doi:10.1093/oxfordjournals.aje.a112335 · Cited by 776
Multiple-bias modelling for analysis of observational data
Greenland, S. (2005) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1111/j.1467-985x.2004.00349.x · Cited by 497
Self-citationTHE EFFECT OF MISCLASSIFICATION IN THE PRESENCE OF COVARIATES
GREENLAND, S. (1980) · American Journal of Epidemiology · doi:10.1093/oxfordjournals.aje.a113025 · Cited by 494 · cited 1×
“ity (such as nonadditivity), and measurement error can cause any of these problems even if it is nondifferential.(21,32)”
BackgroundSelf-citationAn overview of relations among causal modelling methods
Greenland, S., & Brumback, B. (2002) · International Journal of Epidemiology · doi:10.1093/ije/31.5.1030 · Cited by 406 · cited 1×
“Even if we could observe each individual’s response under each exposure pattern, we still could not determine from those observations which biologic mechanisms or sufficient-cause types were operative; hence, individual mechanisms are nonidentifiable even from a perfect factorial or crossover trial.(13,15) For example…”
BackgroundSelf-citationInvited Commentary: Variable Selection versus Shrinkage in the Control of Multiple Confounders
Greenland, S. (2007) · American Journal of Epidemiology · doi:10.1093/aje/kwm355 · Cited by 283
Self-citationTests for interaction in epidemiologic studies: A review and a study of power
Greenland, S. (1983) · Statistics in Medicine · doi:10.1002/sim.4780020219 · Cited by 280
Self-citation
Research works printed but not linked: 1
- Greenland, 2007, Encyclopedia of Epidemiology, Vol. 1., 149
Cited by
Detecting gene–gene interactions that underlie human diseases
Cordell, H. J. (2009) · Nature Reviews Genetics · doi:10.1038/nrg2579 · Cited by 1,350
VanderWeele, T. J., & Knol, M. J. (2014) · Epidemiologic Methods · doi:10.1515/em-2013-0005 · Cited by 1,127
“This will be relevant whenever resources are constrained and thus relevant also for cost-effectiveness (Greenland, 2009).”
BackgroundGene–environment-wide association studies: emerging approaches
Thomas, D. (2010) · Nature Reviews Genetics · doi:10.1038/nrg2764 · Cited by 692
Bally, M., Dendukuri, N., Rich, B., Nadeau, L., Helin-Salmivaara, A., Garbe, E., & Brophy, J. M. (2017) · BMJ · doi:10.1136/bmj.j1909 · Cited by 528
“73 Since methods exist for pooling IPD and aggregate data 51 we considered performing a combined IPD aggregate data meta-analysis.”
MethodHart, C. L., Morrison, D. S., Batty, G. D., Mitchell, R. J., & Davey Smith, G. (2010) · BMJ · doi:10.1136/bmj.c1240 · Cited by 422
“Our findings have important clinical and public health implications. New perspectives on the risk of liver disease may need to be considered for people who are overweight and consume alcohol and lower, BMI specific “safe” limits of alcohol consumption may need to be defined. Liver disease is often advanced when it is…”
Wade, R., Cronholm, P. F., Fein, J. A., Forke, C. M., Davis, M. B., Harkins-Schwarz, M., Pachter, L. M., & Bair-Merritt, M. H. (2016) · Child Abuse & Neglect · doi:10.1016/j.chiabu.2015.11.021 · Cited by 357
The (mis)use of overlap of confidence intervals to assess effect modification
Knol, M. J., Pestman, W. R., & Grobbee, D. E. (2011) · European Journal of Epidemiology · doi:10.1007/s10654-011-9563-8 · Cited by 249
“There are several methods to assess effect modification and the debate on which method is best is still ongoing [2-5].”
MethodCorraini, P., Olsen, M., Pedersen, L., Dekkers, O., & Vandenbroucke, J. (2017) · Clinical Epidemiology · doi:10.2147/clep.s129728 · Cited by 236
“Confounding assumptions to assess interaction may be less strict in public health scenarios where only one of the interacting exposures is intended to be modified by intervention.(37,39,40) In this scenario, only unmeasured confounding between the main exposure and the outcome may be a concern.”
Tsai, A. C., & Burns, B. F. O. (2015) · Social Science & Medicine · doi:10.1016/j.socscimed.2015.06.024 · Cited by 227
Formalizing the role of agent-based modeling in causal inference and epidemiology
Marshall, B. D. L., & Galea, S. (2015) · American Journal of Epidemiology · doi:10.1093/aje/kwu274 · Cited by 192
Challenges and opportunities in genome-wide environmental interaction (GWEI) studies
Aschard, H., Lutz, S., Maus, B., Duell, E. J., Fingerlin, T. E., Chatterjee, N., Kraft, P., & Van Steen, K. (2012) · Human Genetics · doi:10.1007/s00439-012-1192-0 · Cited by 146
“A significant test for interaction—whether from a multiplicative odds ratio model or additive absolute risk model for disease traits, or from additivity for log-transformed or untransformed continuous traits—need not imply biological interaction, just as biological interaction need not imply statistical interaction (G…”
InfluentialPreterm birth and mortality in adulthood: a systematic review
Crump, C. (2020) · Journal of Perinatology · doi:10.1038/s41372-019-0563-y · Cited by 119
Understanding of interaction (subgroup) analysis in clinical trials
Brankovic, M., Kardys, I., Steyerberg, E. W., Lemeshow, S., Markovic, M., Rizopoulos, D., & Boersma, E. (2019) · European Journal of Clinical Investigation · doi:10.1111/eci.13145 · Cited by 130
Interactive Effects of Physical Fitness and Body Mass Index on the Risk of Hypertension
Crump, C., Sundquist, J., Winkleby, M. A., & Sundquist, K. (2016) · JAMA Internal Medicine · doi:10.1001/jamainternmed.2015.7444 · Cited by 124
Rod, N. H., Kumari, M., Lange, T., Kivimäki, M., Shipley, M., & Ferrie, J. (2014) · PLoS ONE · doi:10.1371/journal.pone.0091965 · Cited by 113
“There is an ongoing discussion on how to define and measure interaction, but most health researchers would agree that interventions or preventive strategies should be aimed at patients or population sub-groups where most cases could potentially be prevented [14].”
BackgroundCornelis, M. C., Tchetgen, E. J. T., Liang, L., Qi, L., Chatterjee, N., Hu, F. B., & Kraft, P. L. (2011) · American Journal of Epidemiology · doi:10.1093/aje/kwr368 · Cited by 120
Additive Interaction in Survival Analysis: Use of the Additive Hazards Model
Rod, N. H., Lange, T., Andersen, I., Marott, J. L., & Diderichsen, F. (2012) · Epidemiology · doi:10.1097/ede.0b013e31825fa218 · Cited by 131
Attributing Effects to Interactions
VanderWeele, T. J., & Tchetgen Tchetgen, E. J. (2014) · Epidemiology · doi:10.1097/ede.0000000000000096 · Cited by 104
Good Practices for Observational Studies of Maternal Weight and Weight Gain in Pregnancy
Hutcheon, J. A., & Bodnar, L. M. (2018) · Paediatric and Perinatal Epidemiology · doi:10.1111/ppe.12439 · Cited by 100
Galenkamp, H., Braam, A. W., Huisman, M. A., & Deeg, D. J. H. (2012) · European Journal of Public Health · doi:10.1093/eurpub/cks031 · Cited by 96
Genome-Wide Meta-Analysis of Joint Tests for Genetic and Gene-Environment Interaction Effects
Aschard, H., Hancock, D. B., London, S. J., & Kraft, P. (2010) · Human Heredity · doi:10.1159/000323318 · Cited by 90
“In the range of models we tested, the joint test was often more powerful than the marginal test of the gene, with an absolute gain of power over 50p in the best cases. In situations where the marginal test had better power than the joint test, the difference was small (less than 10p absolute power). The standard test…”
ResultUnraveling the complex interplay between genes, environment, and climate in ALS
Vasta, R., Chia, R., Traynor, B. J., & Chiò, A. (2022) · eBioMedicine · doi:10.1016/j.ebiom.2021.103795 · Cited by 82
Interaction between asthma and smoking increases the risk of adult airway obstruction
Aanerud, M., Carsin, A.-E., Sunyer, J., Dratva, J., Gislason, T., Jarvis, D., deMarco, R., Raherison, C., Wjst, M., Dharmage, S. C., et al. (2015) · European Respiratory Journal · doi:10.1183/09031936.00055514 · Cited by 84
A Novel Statistic for Genome-Wide Interaction Analysis
Wu, X., Dong, H., Luo, L., Zhu, Y., Peng, G., Reveille, J. D., & Xiong, M. (2010) · PLoS Genetics · doi:10.1371/journal.pgen.1001131 · Cited by 80
Christensen, H. N., Diderichsen, F., Hvidtfeldt, U. A., Lange, T., Andersen, P. K., Osler, M., Prescott, E., Tjønneland, A. M., Rod, N. H., & Andersen, I. (2017) · Epidemiology · doi:10.1097/ede.0000000000000718 · Cited by 70
Madsen, I. E. H., Lange, T., Borritz, M., & Rugulies, R. (2015) · Journal of Psychiatric Research · doi:10.1016/j.jpsychires.2015.04.004 · Cited by 64
Hutcheon, J. A., Moskosky, S. B., Ananth, C. V., Basso, O., Briss, P. A., Ferré, C. D., Frederiksen, B. N., Harper, S. B., Hernández–Díaz, S., Hirai, A. H., et al. (2018) · Paediatric and Perinatal Epidemiology · doi:10.1111/ppe.12512 · Cited by 69
Pearce, D. C., Dowling, K., & Sim, M. R. (2012) · Journal of Exposure Science & Environmental Epidemiology · doi:10.1038/jes.2012.15 · Cited by 65
“Suggestive of a synergistic effect (Greenland, 2009), cancer incidence was shown to be exacerbated by exposure to arsenic in soil in more socioeconomically disadvantaged SLAs.”
BackgroundHeming, N., Renault, A., Kuperminc, E., Brun‐Buisson, C., Mégarbane, B., Quenot, J., Siami, S., Cariou, A., Forceville, X., Schwebel, C., et al. (2024) · The Lancet Respiratory Medicine · doi:10.1016/s2213-2600(23)00430-7 · Cited by 62
Racial/Ethnic Disparities in Inadequate Gestational Weight Gain Differ by Pre-pregnancy Weight
Headen, I. E., Mujahid, M. S., Cohen, A. K., Rehkopf, D. H., & Abrams, B. F. (2015) · Maternal and Child Health Journal · doi:10.1007/s10995-015-1682-5 · Cited by 63
Nordahl, H., Osler, M., Frederiksen, B., Andersen, I., Prescott, E., Overvad, K., Diderichsen, F., & Rod, N. H. (2014) · Stroke · doi:10.1161/strokeaha.114.005252 · Cited by 70
Is Self-Rated Health Still Sensitive for Changes in Disease and Functioning Among Nonagenarians?
Galenkamp, H., Deeg, D. J. H., Huisman, M. A., Hervonen, A., Braam, A. W., & Jylhä, M. K. (2013) · The Journals of Gerontology Series B · doi:10.1093/geronb/gbt066 · Cited by 63
Interactive Effects of Aerobic Fitness, Strength, and Obesity on Mortality in Men
Crump, C., Sundquist, J., Winkleby, M. A., & Sundquist, K. (2016) · American Journal of Preventive Medicine · doi:10.1016/j.amepre.2016.10.002 · Cited by 59
Stressful social relations and mortality: a prospective cohort study
Lund, R., Christensen, U., Nilsson, C. J., Kriegbaum, M., & Hulvej Rod, N. (2014) · Journal of Epidemiology and Community Health · doi:10.1136/jech-2013-203675 · Cited by 50
The interaction of biopsychological and socio-environmental influences on criminological outcomes
Barnes, J. C., Raine, A., & Farrington, D. P. (2020) · Justice Quarterly · doi:10.1080/07418825.2020.1730425 · Cited by 53
“Fifth, although interactions are often thought of as “… nothing more than a product term” (Greenland, 2009, p. 16), there are different ways to estimate them in a 2For an accessible discussion, see Uri Simonsohn’s blog entry on the paper: http://datacolada.org/57 statistical analysis.”
BackgroundGait Speed and Processing Speed as Clinical Markers for Geriatric Health Outcomes
Sanders, J. B., Bremmer, M. A., Comijs, H. C., van de Ven, P. M., Deeg, D. J. H., & Beekman, A. T. F. (2016) · American Journal of Geriatric Psychiatry · doi:10.1016/j.jagp.2016.12.003 · Cited by 51
Causal theory and the etiology of periodontal diseases
Heaton, B., & Dietrich, T. (2011) · Periodontology 2000 · doi:10.1111/j.1600-0757.2011.00414.x · Cited by 59
Lesko, C. R., Henderson, N. C., & Varadhan, R. (2018) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2018.04.005 · Cited by 51
Bours, M. J. L. (2021) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2021.01.022 · Cited by 54
Galenkamp, H., Huisman, M. A., Braam, A. W., Schellevis, F., & Deeg, D. J. H. (2014) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2014.02.002 · Cited by 50
VanderWeele, T. J., & Vansteelandt, S. (2011) · American Journal of Epidemiology · doi:10.1093/aje/kwr334 · Cited by 46
“Greenland (23) further remarked that parsimonious logistic regression models impose nonadditivity.”
Interactive effects of obesity and physical fitness on risk of ischemic heart disease
Crump, C., Sundquist, J., Winkleby, M. A., & Sundquist, K. (2016) · International Journal of Obesity · doi:10.1038/ijo.2016.209 · Cited by 50
Crump, C., Sundquist, J., Winkleby, M. A., & Sundquist, K. (2016) · International Journal of Stroke · doi:10.1177/1747493016641961 · Cited by 48
“Additive interactions are often unexamined despite being more informative about public health impact.(36,37) Previous studies have reported similar main effects for BMI, aerobic fitness, or self-reported physical activity in relation to stroke, without examining their interactions.”
Mehta, N. K., & Preston, S. H. (2016) · Social Science & Medicine · doi:10.1016/j.socscimed.2016.02.009 · Cited by 48
“The statistical methods for the assessment of interactions are well developed (Berrington de Gonz alez and Cox, 2005; Greenland, 2009).”
MethodOn the Uses and Abuses of Regression Models: A Call for Reform of Statistical Practice and Teaching
Carlin, J. B., & Moreno‐Betancur, M. (2025) · Statistics in Medicine · doi:10.1002/sim.10244 · Cited by 50
Estimating the Relative Excess Risk Due to Interaction: A Bayesian Approach
Chu, H., Nie, L., & Cole, S. R. (2011) · Epidemiology · doi:10.1097/ede.0b013e318208750e · Cited by 40
Personalized cardiovascular medicine: concepts and methodological considerations
Völzke, H., Schmidt, C. O., Baumeister, S. E., Ittermann, T., Fung, G., Krafczyk-Korth, J., Hoffmann, W. U., Schwab, M. G., zu Schwabedissen, H. E. M., Dörr, M., et al. (2013) · Nature Reviews Cardiology · doi:10.1038/nrcardio.2013.35 · Cited by 44
Timmermans, E. J., Visser, M., Wagtendonk, A., Noordzij, J. M., & Lakerveld, J. (2021) · BMC Public Health · doi:10.1186/s12889-021-11368-6 · Cited by 40
Kramer, M. R., Dunlop, A. L., & HOGUE, C. J. R. (2013) · Maternal and Child Health Journal · doi:10.1007/s10995-013-1244-7 · Cited by 38
“Effect modification of interest to public health practice is ideally described on the additive scale of risk differences, rather than on the multiplicative scale of risk ratios [21, 22].”
BackgroundDeguen, S., Lalloué, B., Bard, D., Havard, S., Arveiler, D., & Zmirou‐Navier, D. (2010) · Epidemiology · doi:10.1097/ede.0b013e3181e09925 · Cited by 35
Adjustment for physical activity in studies of sedentary behaviour
Page, A., Peeters, G. M. E. E., & Merom, D. (2015) · Emerging Themes in Epidemiology · doi:10.1186/s12982-015-0032-9 · Cited by 38
“However, this distinction between statistical and biological interaction is contestable, and is inherently difficult to estimate even with defined biological mechanisms [24, 25].”
BackgroundCook, M. B., Dawsey, S. M., Diaw, L., Blaser, M. J., Perez-Perez, G. I., Abnet, C. C., Taylor, P. R., Albanes, D., Virtamo, J. R. K., & Kamangar, F. (2010) · Cancer Epidemiology Biomarkers & Prevention · doi:10.1158/1055-9965.epi-10-0270 · Cited by 36
Does obesity modify the relationship between physical activity and breast cancer risk?
Neil‐Sztramko, S., Boyle, T., Milosevic, E., Nugent, S. F., Gotay, C. C., & Campbell, K. L. (2017) · Breast Cancer Research and Treatment · doi:10.1007/s10549-017-4449-4 · Cited by 36
“to understand the true shape of the interaction, more precise data, and more sophisticated statistical modeling will be needed in future studies [37].”
BackgroundA nontechnical explanation of the counterfactual definition of confounding
Bours, M. J. L. (2020) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2020.01.021 · Cited by 34
Sanders, J. B., Comijs, H. C., Bremmer, M. A., Deeg, D. J. H., & Beekman, A. T. F. (2014) · International Journal of Geriatric Psychiatry · doi:10.1002/gps.4224 · Cited by 36
“10, because in general, the power of a statistical test is lower for a higher order term than for a first-order term (Aiken and West, 1991; Greenland, 2009).”
BackgroundInfluentialXiao, M., Chu, H., Cole, S. R., Chen, Y., MacLehose, R. F., Richardson, D. B., & Greenland, S. (2021) · Journal of Clinical Epidemiology · doi:10.1016/j.jclinepi.2021.08.002 · Cited by 34
Self-citationRoa‐Díaz, Z. M., Teuscher, J., Gamba, M., Bundo, M., Grisotto, G., Wehrli, F., Gamboa-Delgado, E. M., Rojas, L. Z., Gómez‐Ochoa, S. A., Verhoog, S., et al. (2022) · BMC Cardiovascular Disorders · doi:10.1186/s12872-022-02808-1 · Cited by 31
Interaction on an Additive Scale
de Mutsert, R., de Jager, D. J., Jager, K. J., Zoccali, C., & Dekker, F. W. (2011) · Nephron Clinical Practice · doi:10.1159/000327596 · Cited by 30
Feng, S., Meng, Q., Guo, B., Guo, Y., Chen, G., Pan, Y., Zhou, J. W., Pengcuociren, Xu, J., Zeng, Q., et al. (2022) · Environmental Research · doi:10.1016/j.envres.2022.113699 · Cited by 31
Genetic Interactions in Nonsyndromic Orofacial Clefts in Europe—EUROCRAN Study
Mossey, P. A., Little, J., Steegers‐Theunissen, R. P. M., Molloy, A. M., Peterlin, B., Shaw, W. C., Johnson, C. Y., FitzPatrick, D. R., Franceschelli, P., & Rubini, M. (2016) · The Cleft Palate-Craniofacial Journal · doi:10.1597/16-037 · Cited by 29
Kaufman, J. S. (2009) · Epidemiology · doi:10.1097/ede.0b013e318197c0f5 · Cited by 27
“What can be done to improve this? Greenland has recently recommended that we abandon the word “interaction” altogether.(2) This proposal is not unreasonable (and can be adopted by any sympathetic author).”
BackgroundInfluentialSewell, K., Paul, S., Polt, K. D., Sugg, M. M., Leeper, R. D., Rao, Y. D., & Runkle, J. D. (2024) · Discover Mental Health · doi:10.1007/s44192-023-00055-0 · Cited by 30
“We estimated the relative excess risk due to interaction (RERI) to approximate additive interaction using risk ratios in lieu of risks [70].”
MethodReligious and Spiritual Factors in Depression
Vasegh, S., Rosmarin, D. H., Koenig, H. G., Dew, R. E., & Bonelli, R. M. (2012) · Depression Research and Treatment · doi:10.1155/2012/298056 · Cited by 30
“[28] J.”
BackgroundA Kernel Regression Approach to Gene‐Gene Interaction Detection for Case‐Control Studies
Larson, N. B., & Schaid, D. J. (2013) · Genetic Epidemiology · doi:10.1002/gepi.21749 · Cited by 30
“It is important to note that this definition of interaction does necessarily coincide with the biological interpretation of interaction, and that one does not necessarily imply the other [Greenland, 2009].”
BackgroundStoicescu, C., Ameilia, R., Irwanto, I., Praptoraharjo, I., & Mahanani, M. M. (2019) · Journal of Urban Health · doi:10.1007/s11524-019-00352-6 · Cited by 28
“To this end, epidemiologists have argued for an expansion of the assessment of interactions based purely on product terms in multiplicative models of cross-sectional data to also include additive interaction [24, 27, 28].”
BackgroundFerraro, L., Cascia, C. L., Quattrone, D., Sideli, L., Matranga, D., Capuccio, V., Tripoli, G., Gayer-Anderson, C. E., Morgan, C., Sami, M. B., et al. (2019) · Schizophrenia Bulletin · doi:10.1093/schbul/sbz077 · Cited by 27
“People with psychotic disorders might be more likely to recall risk factors 95 and, for example, could recall greater disadvantages in their early life in PAS or to have used drugs in CEQ interview.”
Ziyab, A. H., Karmaus, W. J. J., Zhang, H., Holloway, J. W., Steck, S. E., Ewart, S., & Arshad, S. H. (2014) · International Archives of Allergy and Immunology · doi:10.1159/000365990 · Cited by 25
Prior, J. A., Jordan, K. P., & Kadam, U. T. (2014) · Rheumatology · doi:10.1093/rheumatology/keu175 · Cited by 28
Malla, G., Long, D. L., Judd, S. E., Irvin, M. R., Kissela, B. M., Lackland, D. T., Safford, M. M., Levine, D. A., Howard, V. J., Howard, G., et al. (2019) · Diabetes Care · doi:10.2337/dc19-0442 · Cited by 25
“We used a higher a to assess effect-measure modification given the lower power to detect statistical interaction in epidemiological studies (27,28); however, the tradeoff with this approach is an increase in type I error (26).”
BackgroundMethodeQTL Epistasis – Challenges and Computational Approaches
Huang, Y., Wuchty, S., & Przytycka, T. M. (2013) · Frontiers in Genetics · doi:10.3389/fgene.2013.00051 · Cited by 25
“In the context of epidemiological studies of disease risk it has been argued that making inferences about biologic interactions from statistical interactions is not straightforward and even sometimes inappropriate (e.g., Thompson, 1991; Greenland, 2009).”
BackgroundAbrahamowicz, M., Beauchamp, M., Fournier, P. F., & Dumont, A. (2013) · Pharmacoepidemiology and Drug Safety · doi:10.1002/pds.3485 · Cited by 24
Banack, H. R., & Kaufman, J. S. (2015) · American Journal of Epidemiology · doi:10.1093/aje/kwv168 · Cited by 25
Anthropometric and DXA-derived measures of body composition in relation to pre-diabetes among adults
Mohammad, A., Ziyab, A. H., & Mohammad, T. (2023) · BMJ Open Diabetes Research & Care · doi:10.1136/bmjdrc-2023-003412 · Cited by 25
Dennis, J. K., Krewski, D. R., Côté, F.-S., Fafard, E., Little, J., & Ghadirian, P. (2011) · The Breast Journal · doi:10.1111/j.1524-4741.2011.01133.x · Cited by 22
“Although the concept of interaction in epidemiology is often debated (44), departures from multiplicative models are appropriate for many”
BackgroundKenneth J. Rothman and multicausality in epidemiology
Bizouarn, P. (2012) · Revue d Épidémiologie et de Santé Publique · doi:10.1016/j.respe.2011.08.067 · Cited by 21
Rosa, R. J., de Araújo, J. S. T., Berra, T. Z., Ramos, A. C. V., Moura, H. S. D., do Nascimento, M. C., Tártaro, A. F., Silva, R. V. D. S., Delpino, F. M., Fiorati, R. C., et al. (2023) · PLOS Global Public Health · doi:10.1371/journal.pgph.0001636 · Cited by 23
“This is a cross-sectional, descriptive and analytical study [12], with data collection via web survey.”
Ni, M. Y., Jiang, C., Cheng, K. K., Zhang, W., Gilman, S. E., Lam, T. H., Leung, G. M., & Schooling, C. M. (2015) · International Journal of Geriatric Psychiatry · doi:10.1002/gps.4370 · Cited by 21
“…consensus in the epidemiological community that interactions of public health importance are additive interactions (Saracci, 1980; Rothman et al., 2008; Knol and VanderWeele, 2012), which provide evidence for targeting subpopulations where interventions would be most effective (Greenland, 2009).”
BackgroundBödeker, W., & Backhaus, T. (2010) · European Journal of Epidemiology · doi:10.1007/s10654-010-9464-2 · Cited by 21
“Given the scale dependency of interaction terms as well as residual confounding it is debated whether epidemiological approaches in general are sufficient for a reliable assessment of combined effects [ 15 ].”
BackgroundDiabetes, depressive symptoms, and mortality risk in old age: The role of inflammation
Costa, E., Diniz, B. S., Firmo, J. O. A., Peixoto, S. V., Filho, A. I. D. L., Lima‐Costa, M. F., & Blay, S. L. (2019) · Depression and Anxiety · doi:10.1002/da.22908 · Cited by 22
“Assessing the additive scale is particularly relevant for assessing the significance of effect modification (Greenland, 2009).”
BackgroundArsenic, tobacco use, and lung cancer: An occupational cohort with 27 follow-up years
Su, Z., Wei, M., Jia, X., Fan, Y., Zhao, F., Zhou, Q., Taylor, P. R., & Qiao, Y. (2021) · Environmental Research · doi:10.1016/j.envres.2021.112611 · Cited by 23
Social Factors as Modifiers of Hurricane Irene Evacuation Behavior in Beaufort County, NC
Ricchetti-Masterson, K., & Horney, J. A. (2013) · PLoS Currents · doi:10.1371/currents.dis.620b6c2ec4408c217788bb1c091ef919 · Cited by 20
Ancestry, ethnicity, and race: explaining inequalities in cardiometabolic disease
Eastwood, S. V., Hemani, G., Watkins, S., Scally, A., Smith, G. D., & Chaturvedi, N. (2024) · Trends in Molecular Medicine · doi:10.1016/j.molmed.2024.04.002 · Cited by 22
Biological Interaction: Time to Drop the Term?
Lawlor, D. A. (2011) · Epidemiology · doi:10.1097/ede.0b013e3182093298 · Cited by 20
Headen, I., Laraia, B., Coleman‐Phox, K., Vieten, C., Adler, N., & Epel, E. (2019) · Obesity · doi:10.1002/oby.22356 · Cited by 18
Smoking, HIV, and risk of pregnancy loss
Westreich, D. J., Cates, J., Cohen, M. H., Weber, K. M., Seidman, D. L., Cropsey, K. L., Wright, R. L., Milam, J. E., Young, M., Mehta, C. C., et al. (2016) · AIDS · doi:10.1097/qad.0000000000001342 · Cited by 18
Identifying Heterogeneous Effect Using Latent Supervised Clustering With Adaptive Fusion
Chen, J., Tran-Dinh, Q., Kosorok, M. R., & Liu, Y. (2020) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2020.1763808 · Cited by 18
“However, such a method requires strong parametric assumptions requiring that the underlying heterogeneity can be determined by those interactions (Greenland 2009).”
MethodNkwata, A. K., Song, X., Zhang, M., & Ezeamama, A. E. (2020) · BMC Public Health · doi:10.1186/s12889-020-08842-y · Cited by 19
Nkwata, A. K., Zhang, M., Song, X., Giordani, B., & Ezeamama, A. E. (2021) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph18031358 · Cited by 18
“10 because the power of statistical tests for higher order terms is generally lower than for first-order terms [44,45].”
BackgroundReverse attenuation in interaction terms due to covariate measurement error
Muff, S., & Keller, L. F. (2015) · Biometrical Journal · doi:10.1002/bimj.201400157 · Cited by 17
“In other contexts the interest may not be in the interactions themselves, but it is vital to detect them, for instance in the evaluation of intervention strategies in epidemiology (Greenland, 2009).”
BackgroundSexual Behaviors Linked to Drug and Alcohol Use Among Men Who Have Sex With Men in China
Zhang, C., Qian, H.-Z., Yin, L., Liu, Y., Strauss, S. M., Ruan, Y., Shao, Y., Lu, H., & Vermund, S. H. (2016) · Substance Use & Misuse · doi:10.1080/10826084.2016.1197264 · Cited by 16
“The higher p-value can be more conservative to capture all possible interaction effects (Greenland, 2009; Kaufman, 2009; Knol, Egger, Scott, Geerlings, & Vandenbroucke, 2009).”
BackgroundEvaluating Additive Interaction Using Survival Percentiles
Bellavia, A., Bottai, M., & Orsini, N. (2016) · Epidemiology · doi:10.1097/ede.0000000000000449 · Cited by 16
Graff, M., Richardson, A. S., Young, K. L., Mazul, A. L., Highland, H. M., North, K. E., Mohlke, K. L., Lange, L. A., Lange, E. M., Harris, K. M., et al. (2016) · Health & Place · doi:10.1016/j.healthplace.2016.09.003 · Cited by 16
“Second, interaction estimation has greater random error than main effect estimations due to smaller sample size (Greenland 2009) and this is especially a limitation in non-European ancestry populations.”
BackgroundMethods in causal inference. Part 2: Interaction, mediation, and time-varying treatments
Bulbulia, J. A. (2024) · Evolutionary Human Sciences · doi:10.1017/ehs.2024.32 · Cited by 16
“…in the distribution of features that modify a treatment effect, and no correction is applied, effect estimates will be biased for the target population in at least one measure of effect (Bulbulia, 2024b; Greenland, 2009; Lash et al., 2020) We present two candidate effect modifiers in Table 5 G 3 .”
Galenkamp, H., Deeg, D. J. H., Braam, A. W., & Huisman, M. A. (2012) · Geriatrics and gerontology international/Geriatrics & gerontology international · doi:10.1111/j.1447-0594.2012.00963.x · Cited by 15
Factors Associated with Hurricane Evacuation: A Statistical Meta-Analysis of Studies, 1999-2018
Karaye, I. M., Taylor, N., Perez-Patron, M. J., Thompson, C. M., & Horney, J. A. (2021) · Disaster Medicine and Public Health Preparedness · doi:10.1017/dmp.2021.24 · Cited by 15
Smith, B. T., Warren, C. M., Andreacchi, A. T., Schwartz, N., & Hobin, E. P. (2023) · Epidemiology · doi:10.1097/ede.0000000000001674 · Cited by 14
“15 In joint effect models, we dichotomized education (low: less than high school or high school diploma/some postsecondary vs. high: trades/certificates below bachelor’s, bachelor’s or above) to increase the precision of our estimates and present individual effects of both exposures and their joint…”
BackgroundMethodCausation and causal inference for genetic effects
Vansteelandt, S., & Lange, C. (2012) · Human Genetics · doi:10.1007/s00439-012-1208-9 · Cited by 15
“…logistic) models for risks, then they require repeated tests at each level of the covariate S; moreover, the model’s nonlinear structure implies an a priori predisposition toward non-additivity, thereby possibly inducing an inflation of Type I errors (Vansteelandt et al. 2012a; Greenland 2009).”
BackgroundAschard, H., Zaitlen, N., Lindström, S., & Kraft, P. L. (2014) · Epidemiology · doi:10.1097/ede.0000000000000195 · Cited by 14
Morfeld, P., & Spallek, M. (2016) · Zentralblatt für Arbeitsmedizin Arbeitsschutz und Ergonomie · doi:10.1007/s40664-016-0118-4 · Cited by 13
Methods in causal inference. Part 3: measurement error and external validity threats
Bulbulia, J. A. (2024) · Evolutionary Human Sciences · doi:10.1017/ehs.2024.33 · Cited by 13
“If no correction is applied, unbiased effect estimates for the analytic sample will bias causal effect estimates for the target population in at least one measure of effect (Greenland, 2009; Lash et al., 2020; VanderWeele, 2012).”
Metadata
- Authors
- Greenland, S.
- Affiliations
- University of California, Los Angeles (US)
- Corresponding author
- Greenland, Sander
- ORCID
- Greenland, Sander
- Venue
- Epidemiology
- Published
- 2009-01
- Volume
- 20
- Issue
- 1
- Pages
- 14-17
- ISSN
- 1044-3983 (ISSN-L)
- OpenAlex
- W1989872151
- Cited by
- 204
- References
- 30
- Citations by year
- 2026: 9 · 2025: 15 · 2024: 15 · 2023: 10 · 2022: 11 · 2021: 13 · 2020: 9 · 2019: 10 · 2018: 6 · 2017: 6 · 2016: 16 · 2015: 14 · 2014: 17 · 2013: 14 · 2012: 15
- FWCI
- 7.57
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Epidemiology (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Social Sciences › Arts and Humanities › History and Philosophy of Science › Philosophy of Biology and Scientific Explanation
- Keywords
- statistical interaction, synergistic effect, superadditivity, non-identifiability, intervention targeting
- MeSH terms
- Models, Theoretical, Terminology as Topic, Population Surveillance, Regression Analysis, Reproducibility of Results, Causality, Epidemiologic Studies, Risk Assessment
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.
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.