The DOI matches its Crossref record; 37 of the 37 references it cites were found in a registry or an index.
Instrumental Variables: Application and Limitations
Abstract
To correct for confounding, the method of instrumental variables (IV) has been proposed. Its use in medical literature is still rather limited because of unfamiliarity or inapplicability. By introducing the method in a nontechnical way, we show that IV in a linear model is quite easy to understand and easy to apply once an appropriate instrumental variable has been identified. We also point out some limitations of the IV estimator when the instrumental variable is only weakly correlated with the exposure. The IV estimator will be imprecise (large standard error), biased when sample size is small, and biased in large samples when one of the assumptions is only slightly violated. For these reasons, it is advised to use an IV that is strongly correlated with exposure. However, we further show that under the assumptions required for the validity of the method, this correlation between IV and exposure is limited. Its maximum is low when confounding is strong, such as in case of confounding by indication. Finally, we show that in a study in which strong confounding is to be expected and an IV has been used that is moderately or strongly related to exposure, it is likely that the assumptions of IV are violated, resulting in a biased effect estimate. We conclude that instrumental variables can be useful in case of moderate confounding but are less useful when strong confounding exists, because strong instruments cannot be found and assumptions will be easily violated.
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
“IV analyses are an alternative estimation method that can be used to derive an estimate of the causal effect of the treatment (here, CRP) on the outcome (SBP) by accounting for the size of the association between randomization and CRP.3,4,12–15”
From Mendelian randomization (2022) · cites this paper as [13]
Citations
Cites
The central role of the propensity score in observational studies for causal effects
ROSENBAUM, P. R., & RUBIN, D. B. (1983) · Biometrika · doi:10.1093/biomet/70.1.41 · Cited by 31,241
Instrumental Variables Regression with Weak Instruments
Staiger, D., & Stock, J. H. (1997) · Econometrica · doi:10.2307/2171753 · Cited by 7,081
Instrumental Variables Regression with Weak Instruments
Staiger, D., & Stock, J. (1994) · doi:10.3386/t0151 · Cited by 5,302 · cited 4×
“However, it has been argued that the application of this method is limited because of its strong assumptions, making it difficult in practice to find a suitable instrumental variable.(16) The objectives of this article are first to introduce the application of the method of IV in epidemiology in a nontechnical way and…”
BackgroundMethodInfluentialIdentification of Causal Effects Using Instrumental Variables
Angrist, J. D., Imbens, G. W., & Rubin, D. B. (1996) · Journal of the American Statistical Association · doi:10.1080/01621459.1996.10476902 · Cited by 4,230
Bound, J., Jaeger, D. A., & Baker, R. (1995) · Journal of the American Statistical Association · doi:10.1080/01621459.1995.10476536 · Cited by 3,898
(1947) · The American Statistician · doi:10.1080/00031305.1947.10501534 · Cited by 3,745
Randomized, Controlled Trials, Observational Studies, and the Hierarchy of Research Designs
Concato, J., Shah, N., & Horwitz, R. I. (2000) · New England Journal of Medicine · doi:10.1056/nejm200006223422507 · Cited by 3,563
Hannan, E. J., & Theil, H. (1973) · Technometrics · doi:10.2307/1266838 · Cited by 3,238
Cox, D. R. (2008) · Springer eBooks · doi:10.1007/978-0-387-32833-1_5 · Cited by 2,687
Angrist, J. D., & Imbens, G. W. (1995) · Journal of the American Statistical Association · doi:10.1080/01621459.1995.10476535 · Cited by 1,346
Research works printed but not linked: 1
- Bound, 1995, JASA, 90, 443
Cited by
Sorted by citations
Items 401–465 of 465
Akazawa, M. (2019) · Carolina Digital Repository (University of North Carolina at Chapel Hill) · doi:10.17615/qs68-dr16
the HIV-infected population: An observational study
Lohse, N. (2006)
Colonoscopy use by Primary Care Physicians and Colorectal Cancer Incidence and Mortality
Jacob, B. J. (2012) · TSpace (University of Toronto)
Kern, D. M., & Auchincloss, A. H. (2016) · iDEA - Drexel University Institutional Repository · doi:10.17918/etd-7109
Babanezhad, M. (2017) · Iranian Journal Of Health Sciences · doi:10.18869/acadpub.jhs.5.1.13
“risk of LBW (HBW) may not yield a consistent estimate due to the existence of the latter factors associated with LBW (HBW) which were omitted from the analysis (16-19).”
BackgroundMethodPratt, N. L. (2010) · Adelaide Research & Scholarship (AR&S) (University of Adelaide)
The Mortality Cost of Undertriage of Major Trauma in Ontario
Haas, B. (2012) · TSpace (University of Toronto)
Research Methods: Epidemiologic Research in Geriatric Oncology
Bastiaannet, E. (2017) · Springer eBooks · doi:10.1007/978-3-319-44870-1_8-1
Wallis, C. J. D. (2017) · TSpace (University of Toronto)
Ramkumar, N., & Brown, J. R. (2018) · Circulation Cardiovascular Quality and Outcomes · doi:10.1161/circoutcomes.118.005235
Inference about treatment effects using bounds, sensitivity analysis and instrumental variables
Richardson, A. (2019) · Carolina Digital Repository (University of North Carolina at Chapel Hill) · doi:10.17615/04zx-0009
Hasegawa, R. B., & Small, D. S. (2020) · Figshare · doi:10.6084/m9.figshare.13373997
Li, Y., Lee, Y., Port, F. K., & Robinson, B. M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.09780
“It is quite intuitive since µi and the intercept term in Equation (6) completely control for unobserved betweencluster confounders and exhaust the degree of freedom at the cluster level.”
BackgroundInfluentialClay, K. B., Portnykh, M., & Severnini, E. (2019) · SSRN Electronic Journal · doi:10.2139/ssrn.3337156
Instrumental Variables Regression with Measurement Errors and Multicollinearity in Instruments
Sheikhi, A., Rezapoor, M., & Hoseenkhani, H. (2020) · Journal of the Iranian Statistical Society · doi:10.52547/jirss.19.2.15
“Some researchers caution that the challenge of identifying a credible instrument is not trivial (see for example, Martens et al. (2006) and Lousdal (2018)).”
BackgroundCausal Inference in Educational Systems: A Graphical Modeling ApproachPreprint
Tadayon, M., & Pottie, G. J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2108.00654 · Cited by 6
“Other popular quasi-experimental designs are Difference-in-differences (DID) and instrumental variable (IV) that can provide unbiased causal estimates when a randomized experiment is not possible [23],[24], [25],[26], [27], and [28].”
Condiciones de vida y de trabajo en la Marina Mercante decimonónica
Rico, J. M. (2011) · Revista de historia naval
Muñoz, A., & Nieto, F. J. (2021) · doi:10.1093/med/9780198816805.003.0030 · Cited by 1
Kenward, M., & Daniel, R. (2011) · Faculty Opinions – Post-Publication Peer Review of the Biomedical Literature · doi:10.3410/f.13398961.14768062
Muñoz, Á., & Nieto, F. J. (2015) · doi:10.1093/med/9780199661756.003.0108
Befolkningens utdanning og tilgjengelighet til trygdefinansiert tannpleie i Norge
Grytten, J. I. (2021) · Den norske tannlegeforenings Tidende · doi:10.56373/2021-1-5
Saha, P., -, A. A. N., & -, M. N. H. (2022) · International Journal For Multidisciplinary Research · doi:10.36948/ijfmr.2022.v04i06.1195
Cupido, A. J., Petersen, R. S., Schmidt, A. F., Levi, M. M., Cohn, D. M., & Fijen, L. M. (2023) · Journal of Thrombosis and Haemostasis · doi:10.1016/j.jtha.2023.07.008
Zhao, H., Jiang, W., Song, S., Lyu, L., Chang, C., Zhao, Y., Wu, B., Wang, L., & Girgenti, M. J. (2023) · Research Square · doi:10.21203/rs.3.rs-3229244/v1
Zhang, L., Lewsey, J., & McAllister, D. (2024) · Journal of Comparative Effectiveness Research · doi:10.57264/cer-2023-0044
“Rather, it is the sample size that decreases the F-statistics and makes the IV weaker [11] .”
BackgroundSharpening Blunt Instruments: Exploiting Non-Linearities to Enhance IdentificationPreprint
Schwarz, C., Sharif, S., & Oswald, C. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4855678
Advances in the Application of Mendelian Randomization in Neurological Research
吴, 含. (2024) · Advances in Clinical Medicine · doi:10.12677/acm.2024.1472025
Zhu, M., Wei, M., Huang, S., & Pan, X. (2024) · Journal of Thoracic Disease · doi:10.21037/jtd-24-361
“In the current study, we selected the SNPs with P 5×10 −8 for the European population and P 1×10 −5 for the East Asian population, which means that the selected SNPs are associated with exposure factors at a genome-wide level of significance (15,23,24) .”
Bai, Z., Shang, X., & Zhang, H. (2024) · Journal of Public Mental Health · doi:10.1108/jpmh-06-2024-0072
Powerful Test of Heterogeneity in Two‐Sample Summary‐Data Mendelian Randomization
Wang, K., & Alberding, S. Y. (2024) · Statistics in Medicine · doi:10.1002/sim.10279
Pharmacoepidemiology in Cardiovascular Research
Ko, D., Yoo, T. K., & Kim, D. H. (2024) · Elsevier eBooks · doi:10.1016/b978-0-443-15734-9.00031-5
Absence of genetic association between insulin-like growth factors and esophageal cancer
Sun, Z., & Wang, X. (2024) · Medicine · doi:10.1097/md.0000000000040899
“[20] The analysis followed 3 assumptions of MR studies [17] (Fig.”
Lü, N., Wen, T.-J., & Huang, K.-C. (2025) · OMEGA - Journal of Death and Dying · doi:10.1177/00302228241312744
“…which can be measured in the available dataset, and is highly correlated with the probability of the exposure variable but is unrelated to the outcome variable, except through its association with the exposure variable (Aso & Yasunaga, 2020; Brookhart et al., 2010; Martens et al., 2006).”
Guzmán-Monteza, Y., Fernández-Luna, J. M., & Ribadas, F. J. (2025) · Electronics · doi:10.3390/electronics14081676 · Cited by 1
“Martens et al. [22] defined three requirements for the validity of an IV: relevance, exclusion, and independence of error.”
Humphreys, J., Nicolay, N., Braeye, T., Evercooren, I. V., Hansen, C. H., Moustsen‐Helms, I. R., Sacco, C., Fabiani, M., Castilla, J., Martínez‐Baz, I., et al. (2025) · medRxiv · doi:10.1101/2025.04.30.25326722
“This may involve identifying more-appropriate negative control outcomes, employing techniques like negative control exposures or instrumental variables (45-48), and expanding data linkage to capture additional determinants (e.g., health-seeking behaviours, socio-economic factors).”
Zhang, X. (2025) · Diabetology & Metabolic Syndrome · doi:10.1186/s13098-025-01858-7
“Our MR analysis was conducted based on three IV assumptions [22]: (a) it does not affect the outcome; (b) it is not associated with the outcomes due to confounding pathways; (c) it is associated with the exposure.”
Moe, K., Skarpsno, E. S., Nilsen, T. I. L., Kaspersen, S. L., Ose, S. O., Carslake, D., Mork, P. J., & Aasdahl, L. (2025) · European Journal of Epidemiology · doi:10.1007/s10654-025-01299-6 · Cited by 1
“The causal effect of BMI on sick leave was estimated using the instrumental variable ratio method [ 9, 24], with offspring BMI as an instrument.”
Tao, H., Qin, X. S., Liang, Z., Tan, M., Jiao, J., Wang, M., Tan, H., Huang, C., Tang, S., & Ding, J. (2025) · Journal of Affective Disorders · doi:10.1016/j.jad.2025.120296
Holmes, T. H., Couse, M., Rosenberg‐Hasson, Y., Fernandez, R., Fernandez, A., Mathi, K., Guthridge, J. M., James, J. A., Yabu, J. M., Haddad, F., et al. (2025) · Journal of Probability and Statistics · doi:10.1155/jpas/5014500
“We abandoned that model when it became evident that the prior batch mean was not meeting all criteria for a strong instrument [63].”
Ivabradine and Atrial Fibrillation Incidence
Kim, K., Keur, J., Anderson, H., Radwański, P., & Munger, M. A. (2025) · JACC Advances · doi:10.1016/j.jacadv.2025.102282
Wang, Z., Mao, S., Zhou, D., & Zhang, W. (2025) · Medicine · doi:10.1097/md.0000000000044963
“As illustrated in Figure 1, our 2-sample MR study was structured around 3 principal assumptions [12] : relevance assumption: The IVs demonstrated a strong association with the exposure.”
Roessler, M., Korthauer, L., Petrautzki, I., Bobeth, C., Schulte, C., Repschläger, U., Straub, C., Wende, D., Hertle, D., Deckert, S., et al. (2025) · Value in Health · doi:10.1016/j.jval.2025.11.014
Additional file 1 of Modelling hospital outcome: problems with endogeneity
Moran, J. L., Santamaria, J. D., & Duke, G. J. (2021) · Figshare · doi:10.6084/m9.figshare.14821162
Islam, S. N., Ahammed, T., Anjum, A., Albalawi, O., & Uddin, M. J. (2022) · Figshare · doi:10.6084/m9.figshare.18517702
Chan, C. H., & Wong, A. (2022) · Figshare · doi:10.6084/m9.figshare.20103919
Chan, C. H., & Wong, A. (2022) · Figshare · doi:10.6084/m9.figshare.20103922
Chan, C. H., & Wong, A. (2022) · Figshare · doi:10.6084/m9.figshare.20103925
Li, B., Yiwei, W., Huang, C., & Shi, Y. (2026) · Medicine · doi:10.1097/md.0000000000047334
“[12] A genetic variant can be considered as an instrumental variable (IV) for a given exposure if it satisfies the IV assumptions [13] :it does not affect the outcome; it is not associated with the outcomes due to confounding pathways; and it is associated with the exposure.”
Unintended educational consequences of the Bolsa Família program in Brazil
Maciel, F. A. (2026) · International Journal of Social Economics · doi:10.1108/ijse-05-2025-0495
Yu, B., & Finkel, S. E. (2026) · Methodology · doi:10.5964/meth.16875
Sharpening Blunt Instruments: Exploiting Non-Linearities for Causal Identification
Schwarz, C., Sharif, S., & Oswald, C. (2026) · International Studies Quarterly · doi:10.1093/isq/sqag020
Peng, W., Zhulan, H., & Zhang, M. (2026) · International Journal of Endocrinology · doi:10.1155/ije/7371037
“The selected genetic IVs were required to satisfy the three key MR assumptions [25–27]: relevance, independence, and exclusion restriction.”
Access to Mental Health Care Services and Suicide Rates in the United States
Shakya, S., & Choudhury, A. R. (2026) · Journal of Policy Analysis and Management · doi:10.1002/pam.70110
Vachuska, K. F., & Adamson, C. (2026) · Social Networks · doi:10.1016/j.socnet.2026.04.018
Post-injury health and economic recovery among agricultural workers
Kim, N., Lee, K., & Suh, T. (2026) · Applied Economics Letters · doi:10.1080/13504851.2026.2685227
Deconfounding via Profiled Transfer Learning
Chen, Z., Jiang, Y., Liu, J., & Yao, F. (2026) · Journal of the American Statistical Association · doi:10.1080/01621459.2026.2686441
“Nevertheless, these approaches often face practical challenges in choosing suitable auxiliaries or rely on stringent assumptions, such as the exclusion restriction and the parallel trends assumption (Martens et al., 2006; Zhao and Cui, 2025).”
Wang, Z. G., Ma, M., Gu, J., Lu, W., Deng, T., & Shi, L. (2026) · The Journal of Maternal-Fetal & Neonatal Medicine · doi:10.1080/14767058.2026.2662131
LLM-Generated Summaries for Teachers: A Randomized Field Experiment in a Digital Learning Platform
Lim, W.-C. I., Worden, E., Sales, A., & Heffernan, N. T. (2026) · ACM Conference on Learning @ Scale (L@S) · doi:10.1145/3774398.3811606
“To estimate the causal effect of using the summary (rather than just access), we use two-stage least squares (2SLS) regression [27], treating the random assignment ( 𝑍 ) as an instrument for actual summary usage.”
Xiang, J., Yan, J., Shen, J., Mao, M., Zheng, Z., & Deng, S. (2026) · International Journal of Endocrinology · doi:10.1155/ije/6688740
Wang, J., Wang, K., & Wang, Z. (2026) · Cancer Biomarkers · doi:10.1177/18758592261463308
Sandeep, B., Cheng, H., Liu, X., Huang, X., Yan, Y., Jiang, F., Xiong, D., Wu, Q., Gao, K., & Xiao, Z. (2026) · Current Problems in Cardiology · doi:10.1016/j.cpcardiol.2026.103398
Data Equity in Quantitative Health Research
Anderson, A., Wang, J. M., Nag, A., & Bothara, R. (2026) · doi:10.1002/9781394311446.ch07
Valid Inference for Two Sample Summary Statistics Mendelian Randomization
Wang, K., & Wang, G. Z. (2026) · Statistics in Medicine · doi:10.1002/sim.70784
Metadata
- Authors
- Martens, E. P., Pestman, W. R., de Boer, A., Belitser, S. V., & Klungel, O. H.
- Affiliations
- Utrecht University (NL)
- Corresponding author
- Martens, Edwin P.
- Belitser, Svetlana V.
- Klungel, Olaf H.
- Venue
- Epidemiology
- Published
- 2006-05
- Volume
- 17
- Issue
- 3
- Pages
- 260-267
- ISSN
- 1044-3983 (ISSN-L)
- OpenAlex
- W2127065386
- Cited by
- 464
- References
- 37
- Citations by year
- 2026: 19 · 2025: 35 · 2024: 31 · 2023: 29 · 2022: 35 · 2021: 29 · 2020: 26 · 2019: 26 · 2018: 21 · 2017: 22 · 2016: 18 · 2015: 28 · 2014: 24 · 2013: 22 · 2012: 27
- FWCI
- 10.31
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Epidemiology (publishedVersion)
- PubMed (publishedVersion)
- Utrecht University Repository (Utrecht University) (submittedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Keywords
- instrumental variables, confounding, confounding by indication, weak instruments, linear regression, exposure measurement error
- MeSH terms
- United States, Models, Statistical, Bias, Confounding Factors, Epidemiologic, Epidemiologic Studies
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
