Instrumental Variables II: Instrumental Variable Application – In 25 Variations, the Physician Prescribing Preference Generally Was Strong and Reduced Covariate Imbalance

Brigham & Women's Hospital, Boston, MA 02120, USA.
Journal of clinical epidemiology (Impact Factor: 3.42). 05/2009; 62(12):1233-41. DOI: 10.1016/j.jclinepi.2008.12.006
Source: PubMed


An instrumental variable (IV) is an unconfounded proxy for a study exposure that can be used to estimate a causal effect in the presence of unmeasured confounding. To provide reliably consistent estimates of effect, IVs should be both valid and reasonably strong. Physician prescribing preference (PPP) is an IV that uses variation in doctors' prescribing to predict drug treatment. As reduction in covariate imbalance may suggest increased IV validity, we sought to examine the covariate balance and instrument strength in 25 formulations of the PPP IV in two cohort studies.
We applied the PPP IV to assess antipsychotic medication (APM) use and subsequent death among two cohorts of elderly patients. We varied the measurement of PPP, plus performed cohort restriction and stratification. We modeled risk differences with two-stage least square regression. First-stage partial r(2) values characterized the strength of the instrument. The Mahalanobis distance summarized balance across multiple covariates.
Partial r(2) ranged from 0.028 to 0.099. PPP generally alleviated imbalances in nonpsychiatry-related patient characteristics, and the overall imbalance was reduced by an average of 36% (+/-40%) over the two cohorts.
In our study setting, most of the 25 formulations of the PPP IV were strong IVs and resulted in a strong reduction of imbalance in many variations. The association between strength and imbalance was mixed.

21 Reads
  • Source
    • "To verify that such an approach to IV selection was also appropriate for binary variables, we also examined the ability of each potential IV to reduce the imbalance in the major covariates. To do so, we compared the mean standardized difference as stratified by the actual treatment with the mean standardized difference as stratified by the IV, as proposed by Rassen et al. [27]. According to these criteria, the best instrument was the variable associated with the highest F-statistics and partial r2 and with the greater reduction in the mean standardized differences. "
    [Show abstract] [Hide abstract]
    ABSTRACT: The assessment of the causal effect of Intensive Care Unit (ICU) admission generally involves usual observational designs and thus requires controlling for confounding variables. Instrumental variable analysis is an econometric technique that allows causal inferences of the effectiveness of some treatments during situations to be made when a randomized trial has not been or cannot be conducted. This technique relies on the existence of one variable or "instrument" that is supposed to achieve similar observations with a different treatment for "arbitrary" reasons, thus inducing substantial variation in the treatment decision with no direct effect on the outcome. The objective of the study was to assess the benefit in terms of hospital mortality of ICU admission in a cohort of patients proposed for ICU admission (ELDICUS cohort). Using this cohort of 8,201 patients triaged for ICU (including 6,752 (82.3%) patients admitted), the benefit of ICU admission was evaluated using 3 different approaches: instrumental variables, standard regression and propensity score matched analyses. We further evaluated the results obtained using different instrumental variable methods that have been proposed for dichotomous outcomes. The physician's main specialization was found to be the best instrument. All instrumental variable models adequately reduced baseline imbalances, but failed to show a significant effect of ICU admission on hospital mortality, with confidence intervals far higher than those obtained in standard or propensity-based analyses. Instrumental variable methods offer an appealing alternative to handle the selection bias related to nonrandomized designs, especially when the presence of significant unmeasured confounding is suspected. Applied to the ELDICUS database, this analysis failed to show any significant beneficial effect of ICU admission on hospital mortality. This result could be due to the lack of statistical power of these methods.
    BMC Medical Research Methodology 09/2011; 11(1):132. DOI:10.1186/1471-2288-11-132 · 2.27 Impact Factor
  • Source
    • "The Shea partial correlation coefficient is the square of the partial correlation between the instrument and the treatment, conditional on other covariates in the model. The partial F statistic greater than 10 and a reasonable value of r2 indicate that the instrument is not weak and contributes substantially to the prediction of treatment [15,43]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Intention-to-treat (ITT) is the standard data analysis method which includes all patients regardless of receiving treatment. Although the aim of ITT analysis is to prevent bias due to prognostic dissimilarity, it is also a counter-intuitive type of analysis as it counts patients who did not receive treatment, and may lead to "bias toward the null." As treated (AT) method analyzes patients according to the treatment actually received rather than intended, but is affected by the selection bias. Both ITT and AT analyses can produce biased estimates of treatment effect, so instrumental variable (IV) analysis has been proposed as a technique to control for bias when using AT data. Our objective is to correct for bias in non-experimental data from previously published individual patient data meta-analysis by applying IV methods Center prescribing preference was used as an IV to assess the effects of methotrexate (MTX) in preventing debilitating complications of chronic graft-versus-host-disease (cGVHD) in patients who received peripheral blood stem cell (PBSCT) or bone marrow transplant (BMT) in nine randomized controlled trials (1107 patients). IV methods are applied using 2-stage logistic, 2-stage probit and generalized method of moments models. ITT analysis showed a statistically significant detrimental effect with the use of day 11 MTX, resulting in cGVHD odds ratio (OR) of 1.34 (95% CI 1.02-1.76). AT results showed no difference in the odds of cGVHD with the use of MTX [OR 1.31 (95%CI 0.99-1.73)]. IV analysis further corrected the results toward no difference in the odds of cGVHD between PBSCT vs. BMT, allowing for a possibility of beneficial effects of MTX in preventing cGVHD in PBSCT recipients (OR 1.14; 95%CI 0.83-1.56). All instrumental variable models produce similar results. IV estimates correct for bias and do not exclude the possibility that MTX may be beneficial, contradicting the ITT analysis.
    BMC Medical Research Methodology 04/2011; 11(1):55. DOI:10.1186/1471-2288-11-55 · 2.27 Impact Factor
  • Source
    • "For randomized trials with partial compliance, where the effect of the actual treatment taken is of interest, the natural IV is the randomization to treatment [29]; but, of course, this is not an option when considering exposures that cannot be randomized as mentioned earlier. Examples in epidemiological contexts are the physician's prescription preference as an IV to assess drug effects [8] [55], cigarette price to assess the effects of smoking [41] or genetic variants that are associated with exposures of interest [16] [34] [39]. The latter has become known as Mendelian randomization and, due to the fact that it is currently generating a lot of interest in the epidemiological literature, will serve as illustration throughout (see Section 2). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Instrumental variable (IV) methods are becoming increasingly popular as they seem to offer the only viable way to overcome the problem of unobserved confounding in observational studies. However, some attention has to be paid to the details, as not all such methods target the same causal parameters and some rely on more restrictive parametric assumptions than others. We therefore discuss and contrast the most common IV approaches with relevance to typical applications in observational epidemiology. Further, we illustrate and compare the asymptotic bias of these IV estimators when underlying assumptions are violated in a numerical study. One of our conclusions is that all IV methods encounter problems in the presence of effect modification by unobserved confounders. Since this can never be ruled out for sure, we recommend that practical applications of IV estimators be accompanied routinely by a sensitivity analysis. Comment: Published in at the Statistical Science ( by the Institute of Mathematical Statistics (
    Statistical Science 11/2010; 25(1). DOI:10.1214/09-STS316 · 2.74 Impact Factor
Show more

Similar Publications


21 Reads
Available from