A tutorial on the use of instrumental variables in pharmacoepidemiology

Intubation Decision Based on Illness Severity and Mortality in COVID-19: An International Study

OBJECTIVES

To evaluate the impact of intubation timing, guided by severity criteria, on mortality in critically ill COVID-19 patients, amidst existing uncertainties regarding optimal intubation practices.

DESIGN

Prospective, multicenter, observational study conducted from February 1, 2020, to November 1, 2022.

SETTING

Ten academic institutions in the United States and Europe.

PATIENTS

Adults (≥ 18 yr old) confirmed with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and hospitalized specifically for COVID-19, requiring intubation postadmission. Exclusion criteria included patients hospitalized for non-COVID-19 reasons despite a positive SARS-CoV-2 test.

INTERVENTIONS

Early invasive mechanical ventilation (EIMV) was defined as intubation in patients with less severe organ dysfunction (Sequential Organ Failure Assessment [SOFA] < 7 or Pa o2 /F io2 ratio > 250), whereas late invasive mechanical ventilation (LIMV) was defined as intubation in patients with SOFA greater than or equal to 7 and Pa o2 /F io2 ratio less than or equal to 250.

MEASUREMENTS AND MAIN RESULTS

The primary outcome was mortality within 30 days of hospital admission. Among 4464 patients, 854 (19.1%) required mechanical ventilation (mean age 60 yr, 61.7% male, 19.3% Black). Of those, 621 (72.7%) were categorized in the EIMV group and 233 (27.3%) in the LIMV group. Death within 30 days after admission occurred in 278 patients (42.2%) in the EIMV and 88 patients (46.6%) in the LIMV group ( p = 0.28). An inverse probability-of-treatment weighting analysis revealed a statistically significant association with mortality, with patients in the EIMV group being 32% less likely to die either within 30 days of admission (adjusted hazard ratio [HR] 0.68; 95% CI, 0.52-0.90; p = 0.008) or within 30 days after intubation irrespective of its timing from admission (adjusted HR 0.70; 95% CI, 0.51-0.90; p = 0.006).

CONCLUSIONS

In severe COVID-19 cases, an early intubation strategy, guided by specific severity criteria, is associated with a reduced risk of death. These findings underscore the importance of timely intervention based on objective severity assessments.

Comparison of instrumental variable methods with continuous exposure and binary outcome: A simulation study

BACKGROUND

Instrumental variable (IV) methods are widely employed to estimate causal effects when concerns regarding unmeasured confounders. Although comparisons among several IV methods for binary outcomes exist, comprehensive evaluations are insufficient. Therefore, in this study, we aimed to conduct a simulation with some settings for a detailed comparison of these methods, focusing on scenarios where IVs are valid and under effect homogeneity with different instrument strengths.

METHODS

We compared six IV methods under 32 simulation scenarios: two-stage least squares (2SLS), two-stage predictor substitutions (2SPS), two-stage residual inclusions (2SRI), limited information maximum likelihood (LIML), inverse-variance weighted methods with a linear outcome model (IVWLI), and inverse-variance weighted methods with a non-linear model (IVWLL). By comparing these methods, we examined three key estimates: the parameter estimates of the exposure variable, the causal risk ratio, and the causal risk differences.

RESULTS

Based on the results, six IV methods could be classified into three groups: 2SLS and IVWLI, 2SRI and 2SPS, and LIML and IVWLL. The first pair showed a clear bias owing to outcome model misspecification. The second pair showed a relatively good performance when strong IVs are available; however, the estimates suffered from a significant bias when only weak IVs are used. The third pair produced relatively conservative results, although they were less affected by weak IV issues.

CONCLUSIONS

The findings indicate that no panacea is available for the bias associated with IV methods. We suggest using multiple IV methods: one for primary analysis and another for sensitivity analysis.

Editor's Choice -- Survival, Limb Salvage, and Management of Patients with Lower Limb Acute Ischaemia: A French National Retrospective Observational Study

OBJECTIVE

The aim was to describe the baseline characteristics of French patients referred with acute limb ischaemia (ALI), and their clinical management and outcome (death, amputation).

METHODS

This retrospective observational cohort study used the National Health Data System. All adults hospitalised for ALI who underwent revascularisation with an endovascular or open surgical approach between 1 January 2015 and 31 December 2020 were included and followed up until death or the end of the study (31 December 2021). A one year look back period was used to capture patients' medical history. The risks of death, and major and minor amputations were described using Kaplan-Meier and Aalen-Johansen estimators. A Cox model was used to report the adjusted association between groups and risk of death and Fine-Gray models for the risk of amputations considering the competing risk of death.

RESULTS

Overall, 51 390 patients (median age 70 years, 69% male) were included and had a median follow up of 2.7 years: 39 411 (76.7%) were treated with an open approach and 11 979 (23.3%) with a percutaneous endovascular approach. The preferred approach for the revascularisation varied between French regions. The one year overall survival was 78.0% and 85.2% in the surgery and endovascular groups, respectively. The surgery group had a higher risk of death (hazard ratio [HR] 1.17, 95% CI 1.12 - 1.21), a higher risk of major amputation (sub-distribution HR 1.20, 95% CI 1.10 - 1.30) and lower risk of minor amputation (sub-distribution HR 0.66, 95% CI 0.60 - 0.71) than the endovascular group. Diabetes and dialysis increased the risk of major amputation by 52% and 78%, respectively. Subsequent ALI was the third most common cause of hospital re-admission within one year.

CONCLUSION

ALI remains a condition at high risk of death and amputation. Individual risk factors and ALI severity need to be considered to choose between approaches. Continued prevention efforts, improved management, and access to the most suitable approach are necessary.

Effect of supplementary private health insurance on out-of-pocket inpatient medical expenditure: evidence from Malaysia

The vital role of healthcare financing in achieving universal health coverage is indisputable. However, most countries, including Malaysia, face challenges in establishing an equitable and sustainable healthcare financing system due to escalating healthcare costs, an ageing population and a growing disease burden. With desirable pre-payment and risk pooling features, private health insurance (PHI) is considered an alternative financing option to reduce out-of-pocket (OOP) medical expenditure. However, ongoing theoretical and empirical debates persist regarding the adequacy of financial risk protection provided by PHI largely because it depends on its role, the benefit design and the regulations in place. Our study aimed to investigate the effect of supplementary PHI on OOP inpatient medical expenditure in Malaysia. Secondary data analysis was conducted using the Malaysian National Health and Morbidity Survey 2019 dataset. A total of 983 respondents with a history of inpatient hospitalization in the past 12 months were included in the study. Instrumental variable analysis using a two-stage residual inclusion was performed to address endogeneity bias, with wealth status and education level as the instrumental variables. Tobit regression model was used in the second stage considering the censored distribution of the outcome variable. Missing data were handled using multiple imputation. About one-fifth of the respondents had PHI. In this study, we found that having PHI significantly increased OOP inpatient medical expenditure in all three marginal effects. Additionally, age, residential location, ethnicity (citizenship), being covered by government guarantee letter, government funding and employer-sponsored health insurance were other significant factors associated with OOP inpatient medical expenditure. Our findings undermine a key justification to advocate PHI uptake among the population, with a need for the Malaysian government to reassess the role of PHI in healthcare financing and reconsider PHI subsidization policy. Regulations should also be strengthened to enhance the financial risk protection provided by PHI.

Association with menopausal hormone therapy and asymptomatic gallstones in US women in the third National Health and Nutrition Examination Study

15% of US adults have gallstones, most of which are clinically "silent". Several studies show that menopausal hormone therapy (MHT) increases symptomatic gallstones and cholecystectomy risk. MHT use may be contraindicated in women with gallstones and population studies may be biased by "confounding by contraindication" while the true association between MHT and gallstones remains underestimated. We sought to examine whether MHT use was associated with asymptomatic gallstones using instrumental variable (IV) analysis to account for confounding by contraindication. We used 2018 postmenopausal women from the Third National Health and Nutrition Examination Survey to estimate associations of MHT use with asymptomatic gallstones. A traditional logistic regression analysis was compared to instrumental variable (IV) analysis to account for confounding by contraindication. 12% of women with asymptomatic gallstones and 25% of women without gallstones were current MHT users (P < 0.001). The traditional analysis suggested a decreased odds of asymptomatic gallstones in current versus never users (OR 0.58, 95% CI 0.37, 0.89), but increased odds (OR 1.51, 95% CI 0.44, 5.16) in the IV analysis. The traditional analysis consistently underestimated the odds of asymptomatic gallstones with MHT use compared to the IV analysis. Accounting for confounding by contraindication, we found a suggestive, though imprecise, positive association between MHT use and asymptomatic gallstones among postmenopausal women. Failure to consider contraindication can produce incorrect results.

Addressing unmeasured confounders in cohort studies: Instrumental variable method for a time-fixed exposure on an outcome trajectory

Instrumental variable methods, which handle unmeasured confounding by targeting the part of the exposure explained by an exogenous variable not subject to confounding, have gained much interest in observational studies. We consider the very frequent setting of estimating the unconfounded effect of an exposure measured at baseline on the subsequent trajectory of an outcome repeatedly measured over time. We didactically explain how to apply the instrumental variable method in such setting by adapting the two-stage classical methodology with (1) the prediction of the exposure according to the instrumental variable, (2) its inclusion into a mixed model to quantify the exposure association with the subsequent outcome trajectory, and (3) the computation of the estimated total variance. A simulation study illustrates the consequences of unmeasured confounding in classical analyses and the usefulness of the instrumental variable approach. The methodology is then applied to 6224 participants of the 3C cohort to estimate the association of type-2 diabetes with subsequent cognitive trajectory, using 42 genetic polymorphisms as instrumental variables. This contribution shows how to handle endogeneity when interested in repeated outcomes, along with a R implementation. However, it should still be used with caution as it relies on instrumental variable assumptions hardly testable in practice.

The causal relationship between immune cells and different kidney diseases: A Mendelian randomization study

Studies have suggested that the progress of most kidney diseases from occurrence to course and subsequent related complications are closely related to inflammatory reaction. Increased common leukocytes count in the family (neutrophils, eosinophils, basophils, lymphocytes, etc.) are also involved in the tissue damage of kidney diseases. However, these studies are only traditional observational studies, which cannot prove whether there is a causal relationship between these four kinds of leukocytes count and kidney diseases. We aim to explore the causal relationship between these four kinds of leukocytes count and kidney diseases by Mendelian randomization (MR). Large sample size of the genome-wide association database of four cell traits (neutrophil, basophil, lymphocyte, and eosinophil cell counts) in the leukocyte family were used as exposure variables. The outcome variables were various renal diseases (including chronic renal failure, acute renal failure, hypertensive heart or/and kidney disease, hypertensive renal disease, disorders resulting from impaired renal tubular function, and type 1 diabetes with renal complications). The covariates used in multivariable MR are also four cell traits related to blood cells (neutrophil, basophil, lymphocyte, and eosinophil cell counts). Instrumental variables and single nucleotide polymorphic loci were identified (P < 5 × 10-8. Linkage disequilibrium R2 < 0.001). The causal relationships were studied by inverse variance weighted (IVW), weighted median, and MR-Egger regression. Sensitivity analysis was also performed. In our study, IVW analysis results showed that increased neutrophil cell count was a risk factor for chronic renal failure (OR = 2.0245861, 95% CI = 1.1231207-3.649606, P = 0.01896524), increased basophil cell count was a risk factor for chronic renal failure (OR = 3.975935, 95% CI = 1.4871198-10.62998, P = 0.005942755). Basophil cell count was not a risk factor for acute renal failure (OR = 1.160434, 95% CI = 0.9455132-1.424207, P = 0.15448828). Increased basophil cell count was a protective factor for hypertensive heart and/or renal disease (OR = 0.7716065, 95% CI = 0.6484979-0.9180856, P = 0.003458707). Increased basophil cell count was a risk factor for disorders resulting from impaired renal tubular function (OR = 1.648131, 95% CI = 1.010116-2.689133, P = 0.04546835). Increased lymphocyte cell count was a risk factor for hypertensive renal disease (OR = 1.372961, 95% CI = 1.0189772-1.849915, P = 0.03719874). Increased eosinophil cell count was a risk factor for type 1 diabetes with renal complications (OR = 1.516454, 95% CI = 1.1826453-1.944482, P = 0.001028964). Macrophage inflammatory protein 1b levels was a protective factor for renal failure (OR = 0.9381862, 95% CI = 0.8860402-0.9934013, P = 0.02874872). After multivariable MR was used to correct covariates (neutrophil, basophil, and lymphocyte cell counts), the correlation effect between increased eosinophil cell counts and type 1 diabetes with renal complications was still statistically significant (P = 0.02201152). After adjusting covariates (neutrophil, basophil, and eosinophil cell counts) with multivariable MR, the correlation effect between increased lymphocyte cell counts and hypertensive renal disease was still statistically significant (P = 0.02050226). This study shows that increased basophils can increase the relative risk of chronic renal failure and renal tubular dysfunction, and reduce the risk of hypertensive heart disease and/or hypertensive nephropathy, while increased basophil cell count will not increase the relative risk of acute renal failure, increased neutrophil cell count can increase the risk of chronic renal failure, increased lymphocyte cell count can increase the relative risk of hypertensive nephropathy, and increased eosinophil cell count can increase the relative risk of type 1 diabetes with renal complications. Macrophage inflammatory protein 1b levels was a protective factor for renal failure.

The Association between Direct Oral Anticoagulants Prescribing Behavior and Non-Valvular Atrial Fibrillation Outcomes: An Instrumental Variable Analysis of Real-World Data

Several observational studies have compared apixaban with rivaroxaban in patients with non-valvular atrial fibrillation (NVAF), but these analyses may be confounded by unmeasured characteristics. This study used provider prescribing preference (PPP) as an instrumental variable (IV) to assess the association between prescriber choice of rivaroxaban vs. apixaban and the study outcomes of stroke/systemic embolism (SE), major bleeding, and death in a retrospective cohort of NVAF patients in the US. Initiators of either medication were linked to their prescribers and followed until the first of the study outcome, the end of rivaroxaban/apixaban use, or 365 days after initiation. PPP for each patient was the percent of rivaroxaban initiations issued by the provider for the prior 10 NVAF patients. Cox regression models tested associations between quintiles of PPP and each outcome. A total of 61,155 patients and 1726 providers were included. The IV was a strong predictor of rivaroxaban prescription (OR = 17.9; 95% CI: 16.6, 19.3). There were statistically significant associations between increasing preference for rivaroxaban and rates of major bleeding (ptrend = 0.041) and death (ptrend = 0.031), but not stroke/SE (ptrend = 0.398). This analysis provides evidence of the relative safety of apixaban over rivaroxaban for the risk of major bleeding and death.

Instrumental Variable Analysis of Racial Discrimination and Blood Pressure in a Sample of Young Adults

Racial inequities in blood pressure levels have been extensively documented. Experiences of racial discrimination could explain some of this disparity, although findings from previous studies have been inconsistent. To address limitations of prior literature, including measurement error, we implemented instrumental variable analysis to assess the relationship between racial discrimination in institutional settings and blood pressure. Using data from 3,876 Black and White adults with an average age of 32 years from examination 4 (1992-1993) of the Coronary Artery Risk Development in Young Adults Study, our primary analysis examined the relationship between self-reported experiences of racial discrimination in institutional settings and blood pressure using reflectance meter measurement of skin color as an instrument. Findings suggested that an increase in experiences of racial discrimination was associated with higher systolic and diastolic blood pressure (β = 2.23 mm Hg (95% confidence interval: 1.85, 2.61) and β = 1.31 (95% confidence interval: 1.00, 1.62), respectively). Our instrumental variable estimates suggest that experiences of racial discrimination within institutional settings contribute to racial inequities in elevated blood pressure and cardiovascular disease outcomes in a relatively young cohort of adults and may yield clinically relevant differences in cardiovascular health over the life course.

Can cognitive inflexibility reduce symptoms of anxiety and depression? Promoting the structural nested mean model in psychotherapy research

OBJECTIVE

To estimate the causal effect of executive functioning on the remission of depression and anxiety symptoms in an observational dataset from a vocational rehabilitation program. It is also an aim to promote a method from the causal inference literature and to illustrate its value in this setting.

METHOD

With longitudinal (four-time points over 13 months) data from four independent sites, we compiled a dataset with 390 participants. At each time point, participants were tested on executive function and self-reported symptoms of anxiety and depression. We used g-estimation to evaluate whether objectively tested cognitive flexibility affected depressive/anxious symptoms and tested for moderation. Multiple imputations were used to handle missing data.

RESULTS

The g-estimation showed a strong causal effect of cognitive inflexibility reducing depression and anxiety and modified by education level. In a counterfactual framework, a hypothetical intervention that could lower cognitive flexibility seemed to cause improvement in mental distress at the subsequent time-point (negative sign) for low education. The less flexibility, the larger improvement. For high education, the same but weaker effect was found, with a change in sign, negative during the intervention and positive during follow-up.

DISCUSSION

An unexpected and strong effect was found from cognitive inflexibility on symptom improvement. This study demonstrates how to estimate causal psychological effects with standard software in an observational dataset with substantial missing and shows the value of such methods.

Relationship between the lower limit of systolic blood pressure target and kidney function decline in advanced chronic kidney disease: an instrumental variable analysis from the REACH-J CKD cohort study

Intensive antihypertensive treatment decreases cardiovascular disease and mortality risks in chronic kidney disease (CKD), whereas extremely low systolic blood pressure (SBP) is associated with worsening kidney function and poor prognosis. Although the SBP variation is particularly large in patients with CKD, the optimal lower limit of SBP target is unclear. In a nationwide, multicenter cohort study of patients with an estimated glomerular filtration rate (eGFR) < 45 mL/min/1.73 m2, we evaluated the association between the eGFR slopes and the lower limit of SBP target at ≥110 mmHg using a linear mixed-effects model and an instrumental variable method. The instrumental variable was calculated as the facility-level percentage of nephrologists who answered in the survey that their lower limit of SBP target was 110 mmHg or higher. A total of 1320 patients (mean age 70 years; 66% men) were included. The mean eGFR slope ± standard deviation over the four years to baseline was -2.48 ± 2.15 mL/min/1.73 m2/year. The instrumental variable for the lower limit of SBP target at ≥110 mmHg (vs. ≤100 mmHg) was associated with less eGFR decline (coefficient: +1.05 mL/min/1.73 m2/year; 95% confidence interval: 0.33-1.77), while unassociated with a history of cardiovascular disease. The renoprotective effect was particularly larger in the subgroups of the elderly and those with a history of cardiovascular disease. In conclusion, the lower limit of SBP target at 110 mmHg or higher was associated with improved eGFR slope, suggesting the importance of aiming at avoiding excessively low SBP in patients with advanced CKD.

Bicarbonate, calcium, and magnesium for in-hospital cardiac arrest - An instrumental variable analysis

INTRODUCTION

Bicarbonate, calcium, and magnesium are commonly used during in-hospital cardiac arrest. Whether these drugs are associated with survival in cardiac arrest patients is uncertain.

METHODS

This was an observational study using data from the Get With The Guidelines registry. Adult patients with an in-hospital cardiac arrest between January 2008 and December 2021 were included. An instrumental variable approach was used based on hospital preferences for bicarbonate, calcium, and magnesium, respectively. The primary outcome was survival to hospital discharge.

RESULTS

A total of 319,230 patients were included. The median age was 66 years, 59% patients were male, and 85% patients presented with a non-shockable rhythm. Bicarbonate was administered in 58% patients, calcium in 33% patients, and magnesium in 10% patients. When considering drug use in the previous cardiac arrest patient at a given hospital as an instrument, the absolute difference in survival to hospital discharge was estimated at -14.2% (95% CI, -19.9 to -8.6) for bicarbonate, -3.0% (95% CI, -8.6 to 2.6) for calcium, and 10.7% (95% CI, -0.8 to 22.2) for magnesium as compared to no drug. When considering the proportion of drug use within the past year at a given hospital as an instrument, the confidence intervals were very wide, making the results difficult to interpret.

CONCLUSIONS

In this analysis, the results for bicarbonate, calcium, and magnesium were inconclusive due to wide confidence intervals and inconsistencies in estimates across instrumental variables. Randomized trials are needed to investigate the effect of these drugs on patient outcomes.

Breastfeeding is associated with the intelligence of school-age children in Mexico

Breastfeeding has been consistently associated with higher intelligence since childhood. However, this relation could be confounded due to maternal selection bias. We estimated the association between predominant breastfeeding and intelligence in school-age children considering potential selection bias and we simulated the intelligence gap reduction between low versus higher socioeconomic status children by increasing breastfeeding. We analysed predominant breastfeeding practices (breastmilk and water-based liquids) of children 0-3 years included in the Mexican Family Life Survey (MxFLS-1). Intelligence was estimated as the z-score of the abbreviated Raven score, measured at 6-12 years in the MxFLS-2 or MxFLS-3. We predicted breastfeeding duration among children with censored data with a Poisson model. We used the Heckman selection model to assess the association between breastfeeding and intelligence, correcting for selection bias and stratified by socioeconomic status. Results show after controlling for selection bias, a 1-month increase in predominant breastfeeding duration was associated with a 0.02 SD increase in the Raven z-score (p < 0.05). The children who were predominantly breastfed for 4-6 months versus <1 month had 0.16 SD higher Raven z-score (p < 0.05). No associations were found using multiple linear regression models. Among low socioeconomic status children, increasing predominantly breastfeeding duration to 6 months would increase their mean Raven z-score from -0.14 to -0.07 SD and reduce by 12.5% the intelligence gap with high socioeconomic status children. In conclusion, predominant breastfeeding duration was significantly associated with childhood intelligence after controlling for maternal selection bias. Increased breastfeeding duration may reduce poverty-driven intelligence inequities.

Group sequential testing under instrumented difference-in-differences approach

Unmeasured confounding is a major obstacle to reliable causal inference based on observational studies. Instrumented difference-in-differences (iDiD), a novel idea connecting instrumental variable and standard DiD, ameliorates the above issue by explicitly leveraging exogenous randomness in an exposure trend. In this article, we utilize the above idea of iDiD, and propose a novel group sequential testing method that provides valid inference even in the presence of unmeasured confounders. At each time point, we estimate the average or conditional average treatment effect under iDiD setting using the data accumulated up to that time point, and test the significance of the treatment effect. We derive the joint distribution of the test statistics under the null using the asymptotic properties of M-estimation, and the group sequential boundaries are obtained using theα $$ \alpha $$ -spending functions. The performance of our proposed approach is evaluated on both synthetic data and Clinformatics Data Mart Database (OptumInsight, Eden Prairie, MN) to examine the association between rofecoxib and acute myocardial infarction, and our method detects significant adverse effect of rofecoxib much earlier than the time when it was finally withdrawn from the market.

Average Causal Effect Estimation Via Instrumental Variables: the No Simultaneous Heterogeneity Assumption

BACKGROUND

Instrumental variables (IVs) can be used to provide evidence as to whether a treatment has a causal effect on an outcome . Even if the instrument satisfies the three core IV assumptions of relevance, independence, and exclusion restriction, further assumptions are required to identify the average causal effect (ACE) of on . Sufficient assumptions for this include homogeneity in the causal effect of on ; homogeneity in the association of with ; and no effect modification.

METHODS

We describe the no simultaneous heterogeneity assumption, which requires the heterogeneity in the - causal effect to be mean independent of (i.e., uncorrelated with) both and heterogeneity in the - association. This happens, for example, if there are no common modifiers of the - effect and the - association, and the - effect is additive linear. We illustrate the assumption of no simultaneous heterogeneity using simulations and by re-examining selected published studies.

RESULTS

Under no simultaneous heterogeneity, the Wald estimand equals the ACE even if both homogeneity assumptions and no effect modification (which we demonstrate to be special cases of-and therefore stronger than-no simultaneous heterogeneity) are violated.

CONCLUSIONS

The assumption of no simultaneous heterogeneity is sufficient for identifying the ACE using IVs. Since this assumption is weaker than existing assumptions for ACE identification, doing so may be more plausible than previously anticipated.

How strong is the evidence to support statins repurposing for the treatment of cirrhosis?

In this editorial, we briefly mention the studies that support the use of statins to change the natural history of liver cirrhosis, alongside potential biases and flaws that need to be considered when analyzing data. The key message that we want to communicate is that even if current evidence is somehow compelling, it is limited, mostly from observational studies, and in general not enough to formally recommend the prescription of statins in patients with cirrhosis as disease-modifying agents. Finally, we also mention some important facts about the safety of statins in the context of patients with underlying liver disease.

Generating Practice-Based Evidence in the Use of Guideline-Recommended Combination Therapy for Secondary Prevention of Acute Myocardial Infarction

Background: Clinical guidelines recommend beta-blockers, angiotensin-converting enzyme inhibitors/angiotensin-receptor blockers, and statins for the secondary prevention of acute myocardial infarction (AMI). It is not clear whether variation in real-world practice reflects poor quality-of-care or a balance of outcome tradeoffs across patients. Methods: The study cohort included Medicare fee-for-service beneficiaries hospitalized 2007-2008 for AMI. Treatment within 30-days post-discharge was grouped into one of eight possible combinations for the three drug classes. Outcomes included one-year overall survival, one-year cardiovascular-event-free survival, and 90-day adverse events. Treatment effects were estimated using an Instrumental Variables (IV) approach with instruments based on measures of local-area practice style. Pre-specified data elements were abstracted from hospital medical records for a stratified, random sample to create "unmeasured confounders" (per claims data) and assess model assumptions. Results: Each drug combination was observed in the final sample (N = 124,695), with 35.7% having all three, and 13.5% having none. Higher rates of guideline-recommended treatment were associated with both better survival and more adverse events. Unmeasured confounders were not associated with instrumental variable values. Conclusions: The results from this study suggest that providers consider both treatment benefits and harms in patients with AMIs. The investigation of estimator assumptions support the validity of the estimates.

Unmeasured confounding in nonrandomized studies: quantitative bias analysis in health technology assessment

Evidence generated from nonrandomized studies (NRS) is increasingly submitted to health technology assessment (HTA) agencies. Unmeasured confounding is a primary concern with this type of evidence, as it may result in biased treatment effect estimates, which has led to much criticism of NRS by HTA agencies. Quantitative bias analyses are a group of methods that have been developed in the epidemiological literature to quantify the impact of unmeasured confounding and adjust effect estimates from NRS. Key considerations for application in HTA proposed in this article reflect the need to balance methodological complexity with ease of application and interpretation, and the need to ensure the methods fit within the existing frameworks used to assess nonrandomized evidence by HTA bodies.

Propensity Score and Instrumental Variable Techniques in Observational Transplantation Studies: An Overview and Worked Example Relating to Pre-Transplant Cardiac Screening

Inferring causality from observational studies is difficult due to inherent differences in patient characteristics between treated and untreated groups. The randomised controlled trial is the gold standard study design as the random allocation of individuals to treatment and control arms should result in an equal distribution of known and unknown prognostic factors at baseline. However, it is not always ethically or practically possible to perform such a study in the field of transplantation. Propensity score and instrumental variable techniques have theoretical advantages over conventional multivariable regression methods and are increasingly being used within observational studies to reduce the risk of confounding bias. An understanding of these techniques is required to critically appraise the literature. We provide an overview of propensity score and instrumental variable techniques for transplant clinicians, describing their principles, assumptions, strengths, and weaknesses. We discuss the different patient populations included in analyses and how to interpret results. We illustrate these points using data from the Access to Transplant and Transplant Outcome Measures study examining the association between pre-transplant cardiac screening in kidney transplant recipients and post-transplant cardiac events.

A National Study to Compare Effective Management of Constipation in Children Receiving Concurrent Versus Standard Hospice Care

Constipation is a distressing and uncomfortable symptom children experience at end of life. There is a gap in knowledge about how different approaches to hospice care delivery might improve pediatric symptom management of constipation. The purpose of this study was to evaluate the effectiveness of pediatric concurrent hospice versus standard hospice care to manage constipation. Medicaid data (2011-2013) were analyzed. Children who were younger than 21 years enrolled in hospice care and had a hospice enrollment between January 1, 2011, and December 31, 2013, were included. Instrumental variable analysis was used to test the effectiveness of concurrent versus standard hospice care. Among the 18 152 children, approximately 14% of participants were diagnosed or treated for constipation from a nonhospice provider during hospice enrollment. A higher proportion of children received nonhospice care for constipation in concurrent hospice care, compared with standard hospice (19.5% vs 13.2%), although this was not significant (β = .22, P < .05) after adjusting for covariates. The findings demonstrated that concurrent care was no more effective than standard hospice care in managing pediatric constipation. Hospice and nonhospice providers may be doing a sufficient job ordering bowel regimens before constipation becomes a serious problem for children at end of life.

Risk of Serious Infection With Low-dose Glucocorticoids in Patients With Rheumatoid Arthritis: An Instrumental Variable Analysis

BACKGROUND

Low-dose glucocorticoids are commonly used in the treatment of rheumatoid arthritis (RA). Observational studies have found an increased risk of serious infection associated with low-dose glucocorticoids, but concerns about residual confounding remain.

METHODS

We identified adults with RA on stable immunomodulatory therapy for >6 months receiving no glucocorticoids or ≤5 mg/day using Medicare data from 2006 to 2015. We used provider preference for glucocorticoids as an instrumental variable (IV) to assess associations between low-dose glucocorticoid use and the risk of infection requiring hospitalization using a cause-specific proportional hazards model.

RESULTS

We identified 163,603 qualifying treatment episodes among 120,656 patients. Glucocorticoids ≤5 mg/day were used by 25,373/81,802 (31.0%) of patients seen by a rheumatologist with low provider preference for glucocorticoids and by 36,087/81,801 (44.1%) of patients seen by a rheumatologist with high provider preference for glucocorticoids (adjusted odds ratio 1.81, 95% confidence interval 1.77, 1.84 for association between provider preference and glucocorticoids). Chronic obstructive pulmonary disease, opioids, antibiotics, previous emergency department visits, hospitalizations, and infections requiring hospitalization infections were unbalanced with regard to exposure but not to the IV. The incidence of infection requiring hospitalization was 8.0/100 person-years among patients unexposed to glucocorticoids versus 11.7/100 person-years among those exposed. The association between glucocorticoids and infection requiring hospitalization from IV analysis (hazard ratio 1.26 [1.02-1.56]) was similar to results from a standard multivariable model (hazard ratio 1.24 [1.21-1.28]).

CONCLUSIONS

Among patients with RA on stable immunomodulatory therapy, IV analysis based on provider preference demonstrated an increased risk of infection requiring hospitalization associated with low-dose glucocorticoids, similar to a traditional analysis.

Effectiveness of Pediatric Concurrent Hospice Care to Improve Continuity of Care

BACKGROUND

The 2010 Patient Protection and Affordable Care Act (ACA) mandated landmark hospice care legislation for children at end of life. Little is known about the impact of pediatric concurrent hospice care.

OBJECTIVE

The purpose of this study was to examine the effect of pediatric concurrent vs standard hospice care on end-of-life care continuity among Medicaid beneficiaries.

METHODS

Using national Medicaid data, we conducted a quasi-experimental designed study to estimate the effect of concurrent vs standard hospice care to improve end-of-life care continuity for children. Care continuity (i.e., hospice length of stay, hospice disenrollment, emergency room transition, and inpatient transition) was measured via claims data. Exposures were concurrent hospice vs standard hospice care. Using instrumental variable analysis, the effectiveness of exposures on care continuity was compared.

RESULTS

Concurrent hospice care affected care continuity. It resulted in longer lengths of stays in hospice (β = 2.76, P < .001) and reduced hospice live discharges (β = -2.80, P < .05), compared to standard hospice care. Concurrent care was not effective at reducing emergency room (β = 2.09, P < .001) or inpatient care (β = .007, P < .05) transitions during hospice enrollment.

CONCLUSION

Our study provides critical insight into the quality of care delivered for children at end of life. These findings have policy implications.

Inferring causality from observational studies: the role of instrumental variable analysis

Inferring causality from observational studies can be challenging because of the perennial threat of biases from selection, measurement, and confounding. The gold standard study design in clinical research is the randomized controlled trial, because random allocation to treatment ensures that, on average, comparison groups are balanced with respect to both known and unknown prognostic factors. However, most clinically relevant exposure-outcome relationships are not amendable (logistically or ethically) to randomization. Thus, there has been an emergence of analytical approaches over the last several years to improve the validity of inferences made from observational studies. This review presents such an approach, instrumental variable analysis, a technique that has been used by economists for many years but has only recently seen increasing use in the health care literature. This review provides a description of the method, the assumptions underlying it, and recent applications in nephrology outcomes research. A more detailed review of the underlying mathematics, properties of an instrumental variable, and suggested elements for reporting an instrumental variable analysis are provided in the Supplementary Appendix.

Pharmacoepidemiological Evaluation in Prostate Cancer-Common Pitfalls and How to Avoid Them

Simple Summary

Pharmacoepidemiologic research provides opportunities to evaluate how commonly used drug groups, such as cholesterol-lowering drugs, may affect the prostate cancer risk or mortality. However, such studies need to be carefully designed in order to avoid biases caused by systematic differences between medication users and non-users. Similarly, data must be carefully analyzed and interpreted while acknowledging possible biases that can lead to erroneous conclusions. Here, we review common pitfalls in such studies and describe ways to avoid them in an effort to aid future research.

Abstract

Pharmacoepidemiologic research provides opportunities to evaluate how commonly used drug groups, such as cholesterol-lowering or antidiabetic drugs, may affect the prostate cancer risk or mortality. This type of research is valuable in estimating real-life drug effects. Nonetheless, pharmacoepidemiological studies are prone to multiple sources of bias that mainly arise from systematic differences between medication users and non-users. If these are not appreciated and properly controlled for, there is a risk of obtaining biased results and reaching erroneous conclusions. Therefore, in order to improve the quality of future research, we describe common biases in pharmacoepidemiological studies, particularly in the context of prostate cancer research. We also list common ways to mitigate these biases and to estimate causality between medication use and cancer outcomes.

Does Performing the Chinese Eye Exercises Help Protect Children's Vision? - New Evidence from Primary Schools in Rural Northwestern China

Background

The high prevalence of myopia among school-age children in China has raised serious concerns about protecting Chinese students’ vision. While the regular performance of the Chinese eye exercises has been adopted as a preventive approach in China since the mid-1960s, these exercises’ effectiveness at protecting students’ vision has remained largely unknown. This study attempts to provide new evidence of the impact of regularly performing the exercises on Chinese students’ visual outcomes, based on a large-scale dataset.

Methods

A school-based survey was conducted among 9842 randomly selected students (fourth graders) from 252 primary schools in rural Northwestern China in 2012. To address potential estimation bias, we adopted both an instrumental variable (IV) approach and a bivariate-probit model to estimate the impacts on students’ visual acuity and the incidences of visual impairment and myopia.

Results

Both IV and bivariate-probit estimates reveal a detrimental impact of regularly performing the Chinese eye exercises on students’ vision. Compared with students who did not regularly perform the exercises, those who did were 6.2 percentage points more likely to have impaired vision and 7.6 percentage points more likely to be myopic. The estimates are robust to different estimation strategies, various specifications, and the majority of subsamples.

Conclusion

Under the assumption that the correct performance of the Chinese eye exercises would not undermine students’ vision, our findings suggest that the commonly-observed incorrect performance of these exercises among Chinese students imposes non-trivial threats to their vision health.

Real-world evidence: the devil is in the detail

Much has been written about real-world evidence (RWE), a concept that offers an understanding of the effects of healthcare interventions using routine clinical data. The reflection of diverse real-world practices is a double-edged sword that makes RWE attractive but also opens doors to several biases that need to be minimised both in the design and analytical phases of non-experimental studies. Additionally, it is critical to ensure that researchers who conduct these studies possess adequate methodological expertise and ability to accurately implement these methods. Critical design elements to be considered should include a clearly defined research question using a causal inference framework, choice of a fit-for-purpose data source, inclusion of new users of a treatment with comparators that are as similar as possible to that group, accurately classifying person-time and deciding censoring approaches. Having taken measures to minimise bias 'by design', the next step is to implement appropriate analytical techniques (for example propensity scores) to minimise the remnant potential biases. A clear protocol should be provided at the beginning of the study and a report of the results after, including caveats to consider. We also point the readers to readings on some novel analytical methods as well as newer areas of application of RWE. While there is no one-size-fits-all solution to evaluating RWE studies, we have focused our discussion on key methods and issues commonly encountered in comparative observational cohort studies with the hope that readers are better equipped to evaluate non-experimental studies that they encounter in the future. Graphical abstract.

Evidence-based statistical analysis and methods in biomedical research (SAMBR) checklists according to design features

BACKGROUND

Statistical analysis according to design features and objectives is essential to ensure the validity and reliability of the study findings and conclusions in biomedical research. Heterogeneity in reporting study design elements and conducting statistical analyses is often observed for the same study design and study objective in medical literatures. Sometimes, researchers face a lot of predicaments using appropriate statistical approaches highlighted by methodologists for a specific study design either due to lack of accessibility or understanding of statistical methods or unavailability of checklists related to design and analysis in a concise format. The purpose of this review is to provide the checklist of statistical analysis and methods in biomedical research (SAMBR) to applied researchers.

RECENT FINDINGS

We initially identified the important steps of reporting design features that may influence the choice of statistical analysis in biomedical research and essential steps of data analysis of common studies. We subsequently searched for statistical approaches employed for each study design/study objective available in publications and other resources. Compilation of these steps produced SAMBR guidance document, which includes three parts. Applied researchers can use part (A) and part (B) of SAMBR to describe or evaluate research design features and quality of statistical analysis, respectively, in reviewing studies or designing protocols. Part (C) of SAMBR can be used to perform essential and preferred evidence-based data analysis specific to study design and objective.

CONCLUSIONS

We believe that the statistical methods checklists may improve reporting of research design, standardize methodological practices, and promote consistent application of statistical approaches, thus improving the quality of research studies. The checklists do not enforce the use of suggested statistical methods but rather highlight and encourage to conduct the best statistical practices. There is a need to develop an interactive web-based application of the checklists for users for its wide applications.

Cluster randomised trials of prescribing policy: an ethical approach to generating drug safety evidence? A discussion of the ethical application of a new research method

For most chronic medical conditions, multiple medications are available and prescribers often have limited evidence about which therapy is likely to be the most effective and safe for an individual patient. As many patients are exposed every day to medicines that may be less effective than available alternatives, this is of public health importance. Cluster randomised trials of prescribing policy offer an opportunity to rapidly obtain evidence of comparative effectiveness and safety. These trials can pose a low risk to patients and cause minimal disruption to usual care. Despite the potential scientific value of this approach, there remain valid concerns about consent, medication switching and the use of routinely collected data in research. We discuss these concerns with reference to an ongoing pilot study (Evaluating Diuretics in Normal Care (EVIDENCE) - a cluster randomised evaluation of hypertension prescribing policy, ISRCTN 46635087, registered 11 August 2017).

Hypomethylating agent (HMA) therapy use and survival in older adults with Refractory Anemia with Excess Blasts (RAEB) in the United States (USA): a large propensity score-matched population-based study†

Hypomethylating agents (HMA) showed overall survival (OS) benefits in patients with higher-risk myelodysplastic syndromes (HR-MDS) in clinical trials. We conducted a retrospective cohort study of Surveillance, Epidemiology, and End Results (SEER)-Medicare data of patients ≥66 years diagnosed with refractory anemia with excess blasts (RAEB), a proxy for HR-MDS, in 01/2001-04/2004 (pre-period) or 01/2006-12/2011 (post-period). Association between post-period diagnosis and OS was examined using propensity scores (PS)-matched samples. Among 1876 RAEB patients, median OS was 9 months and 30.8% received HMAs (3.6% in pre-period; 43.0% in post-period) with no association between post-period diagnosis and OS. In the top PS quartile, post-period diagnosis was associated with a 74% lower risk of death (Hazard ratio [HR] = 0.26, 95%-CI: 0.10-0.69, p = 0.007), while outcomes were worse in the lowest PS quartile (HR = 2.80, 95%-CI: 1.06-7.36, p = 0.037). HMA lead to a 3-month OS benefit for patients most likely to receive HMA but not for unselected RAEB cohort.

Association Between Uncemented vs Cemented Hemiarthroplasty and Revision Surgery Among Patients With Hip Fracture

IMPORTANCE

Consensus guidelines and systematic reviews have suggested that cemented fixation is more effective than uncemented fixation in hemiarthroplasty for displaced femoral neck fractures. Given that these recommendations are based on research performed outside the United States, it is uncertain whether these findings also reflect the US experience.

OBJECTIVE

To compare the outcomes associated with cemented vs uncemented hemiarthroplasty in a large US integrated health care system.

DESIGN, SETTING, AND PARTICIPANTS

Retrospective cohort study of 12 491 patients aged 60 years and older who underwent hemiarthroplasty treatment of a hip fracture between 2009 and 2017 at 1 of the 36 hospitals owned by Kaiser Permanente, a large US health maintenance organization. Patients were followed up until membership termination, death, or the study end date of December 31, 2017.

EXPOSURES

Hemiarthroplasty (prosthetic replacement of the femoral head) fixation via bony growth into a porous-coated implant (uncemented) or with cement.

MAIN OUTCOMES AND MEASURES

The primary outcome measure was aseptic revision, defined as any reoperation performed after the index procedure involving exchange of the existing implant for reasons other than infection. Secondary outcomes were mortality (in-hospital, postdischarge, and overall), 90-day medical complications, 90-day emergency department visits, and 90-day unplanned readmissions.

RESULTS

Among 12 491 patients in the study cohort who underwent hemiarthroplasty for hip fracture (median age, 83 years; 8660 women [69.3%]), 6042 (48.4%) had undergone uncemented fixation and 6449 (51.6%) had undergone cemented fixation, and the median length of follow-up was 3.8 years. In the multivariable regression analysis controlling for confounders, uncemented fixation was associated with a significantly higher risk of aseptic revision (cumulative incidence at 1 year after operation, 3.0% vs 1.3%; absolute difference, 1.7% [95% CI, 1.1%-2.2%]; hazard ratio [HR], 1.77 [95% CI, 1.43-2.19]; P < .001). Of the 6 prespecified secondary end points, none showed a statistically significant difference between groups, including in-hospital mortality (1.7% for uncemented fixation vs 2.0% for cemented fixation; HR, 0.94 [95% CI, 0.73-1.21]; P = .61) and overall mortality (cumulative incidence at 1 year after operation: 20.0% for uncemented fixation vs 22.8% for cemented fixation; HR, 0.95 [95% CI, 0.90-1.01]; P = .08).

CONCLUSIONS AND RELEVANCE

Among patients with hip fracture treated with hemiarthroplasty in a large US integrated health care system, uncemented fixation, compared with cemented fixation, was associated with a statistically significantly higher risk of aseptic revision. These findings suggest that US surgeons should consider cemented fixation in the hemiarthroplasty treatment of displaced femoral neck fractures in the absence of contraindications.

Real-World Evidence to Assess Medication Safety or Effectiveness in Children: Systematic Review

Background

The promise of real-world evidence (RWE) is especially relevant to pediatrics, where medicines prescribed for children are often used without evidence derived from randomized clinical trials.

Objectives

The aim of this systematic review was to describe the state of RWE in pediatrics by identifying observational studies published during 2016 that used RWE to assess medication safety or effectiveness in children.

Methods

An electronic search of PubMed was combined with an extended search of references within systematic reviews and expert suggestions. Studies were included if they reported on an infant or child under 18 years with exposure to medications; assessed safety or effectiveness; specified a comparison or control group, and were published in English in 2016. Data extraction was conducted by one team member using a standardized form and reviewed by a second team member. Study quality was assessed using the GRACE checklist for rating the quality of observational studies.

Results

After removing duplicates, 915 citations were screened and 29 studies met the eligibility criteria. Most of the eligible studies relied on primary data collection or chart review at a single institution and did not use the growing number of administrative or electronic health record databases available. One-quarter of the studies did not use well-established statistical methods to control for confounders. No single disease group or medication predominated, and age groups ranged from infants to adolescents.

Conclusions

A small body of observational studies published in 2016 were categorized by the study team as using real-world data to assess medication safety or effectiveness in children. Studies varied in age groups, diseases or conditions, and methods, and may not have fully met the FDA definition of RWE. Our review indicates that the use of RWE is not fully developed in pediatrics, and suggests an opportunity to further develop capabilities and more fully leverage administrative and electronic health record databases to study medication safety and effectiveness in children. Our systematic review appears generalizable to pediatrics broadly, and documents that the high level of activity in RWE in general has had less of an impact on pediatrics.

Electronic supplementary material

The online version of this article (10.1007/s40801-020-00182-y) contains supplementary material, which is available to authorized users.

How Confident Are We about Observational Findings in Healthcare: A Benchmark Study

Healthcare professionals increasingly rely on observational healthcare data, such as administrative claims and electronic health records, to estimate the causal effects of interventions. However, limited prior studies raise concerns about the real-world performance of the statistical and epidemiological methods that are used. We present the "OHDSI Methods Benchmark" that aims to evaluate the performance of effect estimation methods on real data. The benchmark comprises a gold standard, a set of metrics, and a set of open source software tools. The gold standard is a collection of real negative controls (drug-outcome pairs where no causal effect appears to exist) and synthetic positive controls (drug-outcome pairs that augment negative controls with simulated causal effects). We apply the benchmark using four large healthcare databases to evaluate methods commonly used in practice: the new-user cohort, self-controlled cohort, case-control, case-crossover, and self-controlled case series designs. The results confirm the concerns about these methods, showing that for most methods the operating characteristics deviate considerably from nominal levels. For example, in most contexts, only half of the 95% confidence intervals we calculated contain the corresponding true effect size. We previously developed an "empirical calibration" procedure to restore these characteristics and we also evaluate this procedure. While no one method dominates, self-controlled methods such as the empirically calibrated self-controlled case series perform well across a wide range of scenarios.

Quality of Care in Chronic Kidney Disease and Incidence of End-stage Renal Disease in Older Patients: A Cohort Study

Supplemental Digital Content is available in the text.

The quality of care received by a growing number of older patients with chronic kidney disease (CKD) has not been adequately examined.

Nonparametric Bayesian Instrumental Variable Analysis: Evaluating Heterogeneous Effects of Coronary Arterial Access Site Strategies

Percutaneous coronary interventions (PCIs) are nonsurgical procedures to open blocked blood vessels to the heart, frequently using a catheter to place a stent. The catheter can be inserted into the blood vessels using an artery in the groin or an artery in the wrist. Because clinical trials have indicated that access via the wrist may result in fewer post procedure complications, shortening the length of stay, and ultimately cost less than groin access, adoption of access via the wrist has been encouraged. However, patients treated in usual care are likely to differ from those participating in clinical trials, and there is reason to believe that the effectiveness of wrist access may differ between males and females. Moreover, the choice of artery access strategy is likely to be influenced by patient or physician unmeasured factors. To study the effectiveness of the two artery access site strategies on hospitalization charges, we use data from a state-mandated clinical registry including 7,963 patients undergoing PCI. A hierarchical Bayesian likelihood-based instrumental variable analysis under a latent index modeling framework is introduced to jointly model outcomes and treatment status. Our approach accounts for unobserved heterogeneity via a latent factor structure, and permits nonparametric error distributions with Dirichlet process mixture models. Our results demonstrate that artery access in the wrist reduces hospitalization charges compared to access in the groin, with a higher mean reduction for male patients.

Long-term use of hydrocodone vs. oxycodone in primary care

BACKGROUND

Hydrocodone and oxycodone are the Schedule II opioids most often prescribed in primary care. Notwithstanding the dangers of prescription opioid use, the likelihood of long-term use with either drug is presently unknown.

METHODS

Using a retrospective cohort design and data from a commerical healthcare claims repository, we compared the likelihood of long-term use of hydrocodone and oxycodone in primary care patients presenting with acute back pain. Treatment was categorized as long-term if the prescription dates spanned ≥90 days from initial prescription to the run-out date of the last prescription, and included ≥120 days' supply or ≥10 fills. Instrumental variable methods and probit regression were used to model the effect of drug choice on long-term use, estimate the average treatment effect, and correct for confounding by indication.

RESULTS

A total of 3,983 patients who were prescribed only hydrocodone or only oxycodone were followed for 270 days in 2016. Long-term opioid use was observed in 320 patients (8%). Controlling for potential confounders including morphine milligram equivalents and dosage, an estimated 12% (95 CI, 10%-14%) treated with hydrocodone transitioned to long-term use vs. 2% (95 CI, 1%-3%) on oxycodone. Among patients who received more than one prescription (n = 1,866), an estimated 23% (95 CI, 19%-26%) treated with hydrocodone transitioned to long-term use vs. 5% (95 CI, 3%-7%) on oxycodone. The difference between drugs was supported in sensitivity and subgroup analyses. Sample selection bias was not detected.

CONCLUSIONS

Long-term use was substantially greater for patients treated with hydrocodone than oxycodone, despite equianalgesia.

Selection Bias When Estimating Average Treatment Effects Using One-sample Instrumental Variable Analysis

Supplemental Digital Content is available in the text.

Participants in epidemiologic and genetic studies are rarely true random samples of the populations they are intended to represent, and both known and unknown factors can influence participation in a study (known as selection into a study). The circumstances in which selection causes bias in an instrumental variable (IV) analysis are not widely understood by practitioners of IV analyses. We use directed acyclic graphs (DAGs) to depict assumptions about the selection mechanism (factors affecting selection) and show how DAGs can be used to determine when a two-stage least squares IV analysis is biased by different selection mechanisms. Through simulations, we show that selection can result in a biased IV estimate with substantial confidence interval (CI) undercoverage, and the level of bias can differ between instrument strengths, a linear and nonlinear exposure–instrument association, and a causal and noncausal exposure effect. We present an application from the UK Biobank study, which is known to be a selected sample of the general population. Of interest was the causal effect of staying in school at least 1 extra year on the decision to smoke. Based on 22,138 participants, the two-stage least squares exposure estimates were very different between the IV analysis ignoring selection and the IV analysis which adjusted for selection (e.g., risk differences, 1.8% [95% CI, −1.5%, 5.0%] and −4.5% [95% CI, −6.6%, −2.4%], respectively). We conclude that selection bias can have a major effect on an IV analysis, and further research is needed on how to conduct sensitivity analyses when selection depends on unmeasured data.

Impact of TNF inhibitor therapy on joint replacement rates in rheumatoid arthritis: a matched cohort analysis of BSRBR-RA UK registry data

OBJECTIVES

Previous ecological data suggest a decline in the need for joint replacements in RA patients following the introduction of TNF inhibitor (TNFi) therapy, although patient-level data are lacking. Our primary aim was to estimate the association between TNFi use and subsequent incidence of total hip replacement (THR) and total knee replacement.

METHODS

A propensity score matched cohort was analysed using the British Society for Rheumatology Biologics Registry (2001-2016) for RA data. Propensity score estimates were used to match TNFi users to similar conventional synthetic DMARD users (with replacement) using a 1:1 ratio. Weighted multivariable Cox regression was used to estimate the impact of TNFi on study outcomes. Effect modification by baseline age and disease severity were investigated. Joint replacement at other sites was also analysed. An instrumental variable sensitivity analysis was also performed.

RESULTS

The matched analysis contained a total of 19 116 patient records. Overall, there was no significant association between TNFi use vs conventional synthetic DMARD on rates of THR (hazard ratios = 0.86 [95% CI: 0.60, 1.22]) although there was significant effect modification by age (P < 0.001). TNFi was associated with a reduction in THR among those >60 years old (hazard ratio = 0.60 [CI: 0.41, 0.87]) but not in younger patients. No significant associations were found for total knee replacement or other joint replacement.

CONCLUSION

Overall, no association was found between the use of TNFi and subsequent incidence of joint replacement. However, TNFi was associated with a 40% relative reduction in THR rates among older patients.

Measurement Error and Misclassification in Electronic Medical Records: Methods to Mitigate Bias

PURPOSE OF REVIEW

We sought to: 1) examine common sources of measurement error in research using data from electronic medical records (EMR), 2) discuss methods to assess the extent and type of measurement error, and 3) describe recent developments in methods to address this source of bias.

RECENT FINDINGS

We identified eight sources of measurement error frequently encountered in EMR studies, the most prominent being that EMR data usually reflect only the health services and medications delivered within the specific health facility/system contributing to the EMR data. Methods for assessing measurement error in EMR data usually require gold standard or validation data, which may be possible using data linkage. Recent methodological developments to address the impact of measurement error in EMR analyses were particularly rich in the multiple imputation literature.

SUMMARY

Presently, sources of measurement error impacting EMR studies are still being elucidated, as are methods for assessing and addressing them. Given the magnitude of measurement error that has been reported, investigators are urged to carefully evaluate and rigorously address this potential source of bias in studies based in EMR data.

Comparative Effectiveness of Generic Latanoprost Versus Branded Prostaglandin Analogs for Primary Open Angle Glaucoma

PURPOSE

The purpose of the study is to determine the comparative effectiveness of generic latanoprost (GL) to its branded (BL) counterpart and other brand-name prostaglandin analogs (bPGAs) in preventing the need for additional therapy in the treatment of primary open angle glaucoma (POAG).

METHODS

Retrospective cohort study using US commercial medical claims data. All new POAG patients from 2000 to 2015 who initiated treatment with either GL or BL were included. Exclusion occurred for having less than 2 years of time in the plan prior to diagnosis, previous use of glaucoma medications, or any diagnosis of glaucoma other than POAG at any time. Analyses compared GL to BL and also to those who received any branded PGA after 2011 (when the generic became available). Cox proportional hazard regression was used to assess the hazards of filling a prescription for a second IOP-lowering medication or having surgical intervention, individually and either outcome combined.

RESULTS

A total of 6317 and 4150 POAG patients were treated with GL and BL, respectively. After 2010, 6317 GL and 3703 brand-name prostaglandin analog (bPGA) POAG patients met criteria for inclusion. After adjustment, compared to BL, the GL conferred a reduced hazard of having a glaucoma procedure performed (hazard ratio [HR] = 0.72, 95% CI: 0.62-0.84, p < 0.001) and the combined outcome (HR = 0.90, 95% CI: 0.83-0.97, p = 0.006) but was not associated with having a second IOP medication (HR = 0.95, 95% CI: 0.87-1.03, p = 0.18). Compared to the bPGAs, however, GL conferred a reduced hazard in each comparison (second IOP medication HR = 0.87, 95% CI: 0.81-0.94, p < 0.001; surgery HR = 0.70, 95% CI: 0.61-0.81, p < 0.001; combined HR = 0.85, 95% CI: 0.79-0.92, p < 0.001). Sensitivity analyses confirmed these findings.

CONCLUSIONS

GL was no less and possibly more effective in preventing the need for additional intervention than branded PGAs.

Calendar time as an instrumental variable in assessing the risk of heart failure with antihyperglycemic drugs

OBJECTIVE

In recent years, second-line diabetes treatment with dipeptidyl peptidase-4 inhibitors (DPP-4i) increased with a corresponding decrease in thiazolidinediones (TZDs). Using hospitalization for heart failure (HF) as a positive control outcome, we explored the use of calendar time as an instrumental variable (IV) and compared this approach to an active comparator new-user study.

METHODS

We identified DPP-4i or TZD initiators after a 6-month washout using Medicare claims 2006-2013. The IV was defined as a binary variable comparing initiators during October 2010 to December 2013 (postperiod) versus January 2008 to May 2010 (preperiod). We examined IV strength and estimated risk differences (RDs) for HF using Kaplan-Meier curves, which were compared with propensity score (PS)-weighted RD for DPP-4i versus TZD.

RESULTS

The IV compared 22 696 initiators (78% DPP-4i) in the postperiod versus 20 283 initiators (38% DPP-4i) in the preperiod, resulting in 40% compliance. The active-comparator (PS-weighted) approach compared 26 198 DPP-4i and 18 842 TZD initiators. Covariate balance across IV levels was slightly better than across treatments (standardized difference, 3% vs 4.5%). The 1- and 2-year local average treatment effects of RD of HF per 100 patients in the "compliers" (95% confidence intervals) were -0.62 (-0.99 to -0.25) and -0.88 (-1.46 to -0.25). Corresponding PS-weighted results were -0.20 (-0.33 to -0.05) and -0.18 (-0.30 to 0.03).

CONCLUSION

Both approaches indicated lesser risk of HF hospitalizations among DPP-4i vs TZD initiators. The magnitude of the estimated effects may differ due to differences in the target populations and assumptions. Calendar time can be leveraged as an IV when market dynamics lead to profound changes in treatments.

[Propensity score: Interests, use and limitations. A practical guide for clinicians]

Propensity scores have been proposed in the early 1980s, and are increasingly used in epidemiology since the 2000s. They are is used to minimize the selection bias in observational studies, leading to a comparability between the exposure groups close to that observed in randomized trials. However, they have important limitations. Besides, new statistical techniques to improve the propensity score performances are more and more complex, while the build and the use of propensity score require a strict methodology to avoid bias, imprecision and non-reproducibility. This overview, designed for clinicians, is aimed at describing the advantages, techniques of use and limitations of propensity scores. A reading grid is provided in order to help interpreting studies using propensity scores.

Instrumental variable methods for a binary outcome were used to informatively address noncompliance in a randomized trial in surgery

Objectives

Randomization can be used as an instrumental variable (IV) to account for unmeasured confounding when seeking to assess the impact of noncompliance with treatment allocation in a randomized trial. We present and compare different methods to calculate the treatment effect on a binary outcome as a rate ratio in a randomized surgical trial.

Study Design and Setting

The effectiveness of peeling versus not peeling the internal limiting membrane of the retina as part of the surgery for a full thickness macular hole. We compared the IV-based estimates (nonparametric causal bound and two-stage residual inclusion approach [2SRI]) with standard treatment effect measures (intention to treat, per protocol and treatment received [TR]). Compliance was defined in two ways (initial and up to the time point of interest). Poisson regression was used for the model-based approaches with robust standard errors to calculate the risk ratio (RR) with 95% confidence intervals.

Results

Results were similar for 1-month macular hole status across methods. For 3- and 6-month macular hole status, nonparametric causal bounds provided a narrower range of uncertainty than other methods, though still had substantial imprecision. For 3-month macular hole status, the TR estimate was substantially different from the other point estimates.

Conclusion

Nonparametric causal bound approaches are a useful addition to an IV estimation approach, which tend to have large levels of uncertainty. Methods which allow RRs to be calculated when addressing noncompliance in randomized trials exist and may be superior to standard estimates. Further research is needed to explore the properties of different IV methods in a broad range of randomized controlled trial scenarios.

Instrumental Variable Analyses in Pharmacoepidemiology: What Target Trials Do We Emulate?

Purpose of review

When leveraging observational data to estimate treatment effects, it is useful to explicitly specify the “target trial” the investigators aspire to emulate. One concern is whether a proposed analysis plan can address the realities of the differences between the available non-randomized observational study and the target trial. When large or unknown sources of unmeasured confounding are suspected, investigators might consider turning to instrumental variable (IV) methods. Of course, the interpretation and appropriateness of IV analyses need to be considered carefully. The purpose of this review is to summarize recent methodologic advancements in how epidemiologists weigh the validity of an IV analysis and to place these methodologic advancements in the context of the feasible target trial’s protocol components.

Recent findings

There have been increased development and application of tools for sensitivity analyses, falsification strategies, and the identification of previously overlooked problems with IV analyses as applied in pharmacoepidemiology. Many of these recent insights can be seen as articulating restrictions on or tradeoffs between the types of target trials that can be validly emulated when using a classical IV analysis.

Summary

Putting classical IV methods in the context of target trials underscores the importance of recent methodologic developments and, more generally, when and how an IV analysis would be appropriate. We see that some tradeoffs in defining the target trials are unavoidable, that some tradeoffs may be offset or explored via sensitivity analyses, and that this serves as a framework for scientific discourse regarding IV and non-IV results emulating potentially different trials with different tradeoffs.