War and Wages

Costs of Care for Operative and Nonoperative Management of Emergency General Surgery Conditions

OBJECTIVE

Many emergency general surgery (EGS) conditions can be managed operatively or nonoperatively, with outcomes that vary by diagnosis. We hypothesized that operative management would lead to higher in-hospital costs but to cost savings over time.

BACKGROUND

EGS conditions account for $28 billion in health care costs in the United States annually. Compared with scheduled surgery, patients who undergo emergency surgery are at increased risk of complications, readmissions, and death, with accompanying costs of care that are up to 50% higher than elective surgery. Our prior work demonstrated that operative management had variable impacts on clinical outcomes depending on the EGS condition.

METHODS

This was a nationwide, retrospective study using fee-for-service Medicare claims data. We included patients 65.5 years of age or older with a principal diagnosis for an EGS condition 7/1/2015-6/30/2018. EGS conditions were categorized as: colorectal, general abdominal, hepatopancreaticobiliary (HPB), intestinal obstruction, and upper gastrointestinal. We used near-far matching with a preference-based instrumental variable to adjust for confounding and selection bias. Outcomes included Medicare payments for the index hospitalization and at 30, 90, and 180 days.

RESULTS

Of 507,677 patients, 30.6% received an operation. For HPB conditions, costs for operative management were initially higher but became equivalent at 90 and 180 days. For all others, operative management was associated with higher inpatient costs, which persisted, though narrowed, over time. Out-of-pocket costs were nearly equivalent for operative and nonoperative management.

CONCLUSIONS

Compared with nonoperative management, costs were higher or equivalent for operative management of EGS conditions through 180 days, which could impact decision-making for clinicians, patients, and health systems in situations where clinical outcomes are similar.

Sensitivity analysis of G-estimators to invalid instrumental variables

Instrumental variables regression is a tool that is commonly used in the analysis of observational data. The instrumental variables are used to make causal inference about the effect of a certain exposure in the presence of unmeasured confounders. A valid instrumental variable is a variable that is associated with the exposure, affects the outcome only through the exposure (exclusion), and is not confounded with the outcome (exogeneity). Unlike the first assumption, the other two are generally untestable and rely on subject-matter knowledge. Therefore, a sensitivity analysis is desirable to assess the impact of assumptions' violation on the estimated parameters. In this paper, we propose and demonstrate a new method of sensitivity analysis for G-estimators in causal linear and non-linear models. We introduce two novel aspects of sensitivity analysis in instrumental variables studies. The first is a single sensitivity parameter that captures violations of exclusion and exogeneity assumptions. The second is an application of the method to non-linear models. The introduced framework is theoretically justified and is illustrated via a simulation study. Finally, we illustrate the method by application to real-world data and provide guidelines on conducting sensitivity analysis.

The Conditional Effects of Multimorbidity on Operative Versus Nonoperative Management of Emergency General Surgery Conditions: A Retrospective Observational Study Using an Instrumental Variable Analysis

OBJECTIVE

To understand how multimorbidity impacts operative versus nonoperative management of emergency general surgery (EGS) conditions.

BACKGROUND

EGS is a heterogenous field, encompassing operative and nonoperative treatment options. Decision-making is particularly complex for older patients with multimorbidity.

METHODS

Using an instrumental variable approach with near-far matching, this national, retrospective observational cohort study of Medicare beneficiaries examines the conditional effects of multimorbidity, defined using qualifying comorbidity sets, on operative versus nonoperative management of EGS conditions.

RESULTS

Of 507,667 patients with EGS conditions, 155,493 (30.6%) received an operation. Overall, 278,836 (54.9%) were multimorbid. After adjustment, multimorbidity significantly increased the risk of in-hospital mortality associated with operative management for general abdominal patients (+9.8%; P = 0.002) and upper gastrointestinal patients (+19.9%, P < 0.001) and the risk of 30-day mortality (+27.7%, P < 0.001) and nonroutine discharge (+21.8%, P = 0.007) associated with operative management for upper gastrointestinal patients. Regardless of multimorbidity status, operative management was associated with a higher risk of in-hospital mortality among colorectal patients (multimorbid: + 12%, P < 0.001; nonmultimorbid: +4%, P = 0.003), higher risk of nonroutine discharge among colorectal (multimorbid: +42.3%, P < 0.001; nonmultimorbid: +55.1%, P < 0.001) and intestinal obstruction patients (multimorbid: +14.6%, P = 0.001; nonmultimorbid: +14.8%, P = 0.001), and lower risk of nonroutine discharge (multimorbid: -11.5%, P < 0.001; nonmultimorbid: -11.9%, P < 0.001) and 30-day readmissions (multimorbid: -8.2%, P = 0.002; nonmultimorbid: -9.7%, P < 0.001) among hepatobiliary patients.

CONCLUSIONS

The effects of multimorbidity on operative versus nonoperative management varied by EGS condition category. Physicians and patients should have honest conversations about the expected risks and benefits of treatment options, and future investigations should aim to understand the optimal management of multimorbid EGS patients.

How does a local instrumental variable method perform across settings with instruments of differing strengths? A simulation study and an evaluation of emergency surgery

Local instrumental variable (LIV) approaches use continuous/multi-valued instrumental variables (IV) to generate consistent estimates of average treatment effects (ATEs) and Conditional Average Treatment Effects (CATEs). There is little evidence on how LIV approaches perform according to the strength of the IV or with different sample sizes. Our simulation study examined the performance of an LIV method, and a two-stage least squares (2SLS) approach across different sample sizes and IV strengths. We considered four 'heterogeneity' scenarios: homogeneity, overt heterogeneity (over measured covariates), essential heterogeneity (unmeasured), and overt and essential heterogeneity combined. In all scenarios, LIV reported estimates with low bias even with the smallest sample size, provided that the instrument was strong. Compared to 2SLS, LIV provided estimates for ATE and CATE with lower levels of bias and Root Mean Squared Error. With smaller sample sizes, both approaches required stronger IVs to ensure low bias. We considered both methods in evaluating emergency surgery (ES) for three acute gastrointestinal conditions. Whereas 2SLS found no differences in the effectiveness of ES according to subgroup, LIV reported that frailer patients had worse outcomes following ES. In settings with continuous IVs of moderate strength, LIV approaches are better suited than 2SLS to estimate policy-relevant treatment effect parameters.

Instrumental variable model average with applications in Mendelian randomization

Mendelian randomization is a technique used to examine the causal effect of a modifiable exposure on a trait using an observational study by utilizing genetic variants. The use of many instruments can help to improve the estimation precision but may suffer bias when the instruments are weakly associated with the exposure. To overcome the difficulty of high-dimensionality, we propose a model average estimator which involves using different subsets of instruments (single nucleotide polymorphisms, SNPs) to predict the exposure in the first stage, followed by weighting the submodels' predictions using penalization by common penalty functions such as least absolute shrinkage and selection operator (LASSO), smoothly clipped absolute deviation (SCAD) and minimax concave penalty (MCP). The model averaged predictions are then used as a genetically predicted exposure to obtain the estimation of the causal effect on the response in the second stage. The novelty of our model average estimator also lies in that it allows the number of submodels and the submodels' sizes to grow with the sample size. The practical performance of the estimator is examined in a series of numerical studies. We apply the proposed method on a real genetic dataset investigating the relationship between stature and blood pressure.

Among-Hospital Variation in Intensive Care Unit Admission Practices and Associated Outcomes for Patients with Acute Respiratory Failure

Rationale: We have previously shown that hospital strain is associated with intensive care unit (ICU) admission and that ICU admission, compared with ward admission, may benefit certain patients with acute respiratory failure (ARF). Objectives: To understand how strain-process-outcomes relationships in patients with ARF may vary among hospitals and what hospital practice differences may account for such variation. Methods: We examined high-acuity patients with ARF who did not require mechanical ventilation or vasopressors in the emergency department (ED) and were admitted to 27 U.S. hospitals from 2013 to 2018. Stratifying by hospital, we compared hospital strain-ICU admission relationships and hospital length of stay (LOS) and mortality among patients initially admitted to the ICU versus the ward using hospital strain as a previously validated instrumental variable. We also surveyed hospital practices and, in exploratory analyses, evaluated their associations with the above processes and outcomes. Results: There was significant among-hospital variation in ICU admission rates, in hospital strain-ICU admission relationships, and in the association of ICU admission with hospital LOS and hospital mortality. Overall, ED patients with ARF (n = 45,339) experienced a 0.82-day shorter median hospital LOS if admitted initially to the ICU compared with the ward, but among the 27 hospitals (n = 224-3,324), this effect varied from 5.85 days shorter (95% confidence interval [CI], -8.84 to -2.86; P < 0.001) to 4.38 days longer (95% CI, 1.86-6.90; P = 0.001). Corresponding ranges for in-hospital mortality with ICU compared with ward admission revealed odds ratios from 0.08 (95% CI, 0.01-0.56; P < 0.007) to 8.89 (95% CI, 1.60-79.85; P = 0.016) among patients with ARF (pooled odds ratio, 0.75). In exploratory analyses, only a small number of measured hospital practices-the presence of a sepsis ED disposition guideline and maximum ED patient capacity-were potentially associated with hospital strain-ICU admission relationships. Conclusions: Hospitals vary considerably in ICU admission rates, the sensitivity of those rates to hospital capacity strain, and the benefits of ICU admission for patients with ARF not requiring life support therapies in the ED. Future work is needed to more fully identify hospital-level factors contributing to these relationships.

Pre-processing data to reduce biases: full matching incorporating an instrumental variable in population-based studies

BACKGROUND

Epidemiologists are often concerned with unobserved biases that produce confounding in population-based studies. We introduce a new design approach-'full matching incorporating an instrumental variable (IV)' or 'Full-IV Matching'-and illustrate its utility in reducing observed and unobserved biases to increase inference accuracy. Our motivating example is tailored to a central question in humanitarian emergencies-the difference in sexual violence risk by displacement setting.

METHODS

We conducted a series of 1000 Monte Carlo simulations generated from a population-based survey after the 2010 Haitian earthquake and included earthquake damage severity as an IV and the unmeasured variable of 'social capital'. We compared standardized mean differences (SMDs) for covariates after different designs to understand potential biases. Mean risk differences (RDs) were used to assess each design's accuracy in estimating the oracle of the simulated data set.

RESULTS

Naive analysis and pair matching equivalently performed. Full matching reduced imbalances between exposed and comparison groups across covariates, except for the unobserved covariate of 'social capital'. Pair and full matching overstated differences in sexual violence risk when displaced to a camp vs a community [pair: RD = 0.13, 95% simulation interval (SI) 0.09-0.16; full: RD = 0.11, 95% SI 0.08-0.14). Full-IV Matching reduced imbalances across observed covariates and importantly 'social capital'. The estimated risk difference (RD = 0.07, 95% SI 0.03-0.11) was closest to the oracle (RD = 0.06, 95% SI 0.4-0.8).

CONCLUSION

Full-IV Matching is a novel approach that is promising for increasing inference accuracy when unmeasured sources of bias likely exist.

Bridging preference-based instrumental variable studies and cluster-randomized encouragement experiments: Study design, noncompliance, and average cluster effect ratio

Instrumental variable (IV) methods are widely used in medical research to draw causal conclusions when the treatment and outcome are confounded by unmeasured confounding variables. One important feature of such studies is that the IV is often applied at the cluster level, for example, hospitals' or physicians' preference for a certain treatment where each hospital or physician naturally defines a cluster. This paper proposes to embed such observational IV data into a cluster-randomized encouragement experiment using nonbipartite matching. Potential outcomes and causal assumptions underpinning the design are formalized and examined. Testing procedures for two commonly used estimands, Fisher's sharp null hypothesis and the pooled effect ratio (PER), are extended to the current setting. We then introduce a novel cluster-heterogeneous proportional treatment effect model and the relevant estimand: the average cluster effect ratio. This new estimand is advantageous over the structural parameter in a constant proportional treatment effect model in that it allows treatment heterogeneity, and is advantageous over the PER estimand in that it does not suffer from Simpson's paradox. We develop an asymptotically valid randomization-based testing procedure for this new estimand based on solving a mixed-integer quadratically constrained optimization problem. The proposed design and inferential methods are applied to a study of the effect of using transesophageal echocardiography during coronary artery bypass graft surgery on patients' 30-day mortality rate. R package ivdesign implements the proposed method.

Weak-instrument robust tests in two-sample summary-data Mendelian randomization

Mendelian randomization (MR) has been a popular method in genetic epidemiology to estimate the effect of an exposure on an outcome using genetic variants as instrumental variables (IV), with two-sample summary-data MR being the most popular. Unfortunately, instruments in MR studies are often weakly associated with the exposure, which can bias effect estimates and inflate Type I errors. In this work, we propose test statistics that are robust under weak-instrument asymptotics by extending the Anderson-Rubin, Kleibergen, and the conditional likelihood ratio test in econometrics to two-sample summary-data MR. We also use the proposed Anderson-Rubin test to develop a point estimator and to detect invalid instruments. We conclude with a simulation and an empirical study and show that the proposed tests control size and have better power than existing methods with weak instruments.

Improving the design stage of air pollution studies based on wind patterns

A growing literature in economics and epidemiology has exploited changes in wind patterns as a source of exogenous variation to better measure the acute health effects of air pollution. Since the distribution of wind components is not randomly distributed over time and related to other weather parameters, multivariate regression models are used to adjust for these confounding factors. However, this type of analysis relies on its ability to correctly adjust for all confounding factors and extrapolate to units without empirical counterfactuals. As an alternative to current practices and to gauge the extent of these issues, we propose to implement a causal inference pipeline to embed this type of observational study within an hypothetical randomized experiment. We illustrate this approach using daily data from Paris, France, over the 2008-2018 period. Using the Neyman-Rubin potential outcomes framework, we first define the treatment of interest as the effect of North-East winds on particulate matter concentrations compared to the effects of other wind directions. We then implement a matching algorithm to approximate a pairwise randomized experiment. It adjusts nonparametrically for observed confounders while avoiding model extrapolation by discarding treated days without similar control days. We find that the effective sample size for which treated and control units are comparable is surprisingly small. It is however reassuring that results on the matched sample are consistent with a standard regression analysis of the initial data. We finally carry out a quantitative bias analysis to check whether our results could be altered by an unmeasured confounder: estimated effects seem robust to a relatively large hidden bias. Our causal inference pipeline is a principled approach to improve the design of air pollution studies based on wind patterns.

Analysis approaches to address treatment nonadherence in pragmatic trials with point-treatment settings: a simulation study

Background

Two-stage least square [2SLS] and two-stage residual inclusion [2SRI] are popularly used instrumental variable (IV) methods to address medication nonadherence in pragmatic trials with point treatment settings. These methods require assumptions, e.g., exclusion restriction, although they are known to handle unmeasured confounding. The newer IV-method, nonparametric causal bound [NPCB], showed promise in reducing uncertainty compared to usual IV-methods. The inverse probability-weighted per-protocol [IP-weighted PP] method is useful in the same setting but requires different assumptions, e.g., no unmeasured confounding. Although all of these methods are aimed to address the same nonadherence problem, comprehensive simulations to compare performances of them are absent in the literature.

Methods

We performed extensive simulations to compare the performances of the above methods in addressing nonadherence when: (1) exclusion restriction satisfied and no unmeasured confounding, (2) exclusion restriction is met but unmeasured confounding present, and (3) exclusion restriction is violated. Our simulations varied parameters such as, levels of adherence rates, unmeasured confounding, and exclusion restriction violations. Risk differences were estimated, and we compared performances in terms of bias, standard error (SE), mean squared error (MSE), and 95% confidence interval coverage probability.

Results

For setting (1), 2SLS and 2SRI have small bias and nominal coverage. IP-weighted PP outperforms these IV-methods in terms of smaller MSE but produces high MSE when nonadherence is very high. For setting (2), IP-weighted-PP generally performs poorly compared to 2SLS and 2SRI in term of bias, and both-stages adjusted IV-methods improve precision than naive IV-methods. For setting (3), IV-methods perform worst in all scenarios, and IP-weighted-PP produces unbiased estimates and small MSE when confounders are adjusted. NPCB produces larger uncertainty bound width in almost all scenarios. We also analyze a two-arm trial to estimate vitamin-A supplementation effect on childhood mortality after addressing nonadherence.

Conclusions

Understanding finite sample characteristics of these methods will guide future researchers in determining suitable analysis strategies. Since assumptions are different and often untestable for IP-weighted PP and IV methods, we suggest analyzing data using both IP-weighted PP and IV approaches in search of a robust conclusion.

Reporting methodological issues of the mendelian randomization studies in health and medical research: a systematic review

Background

Mendelian randomization (MR) studies using Genetic risk scores (GRS) as an instrumental variable (IV) have increasingly been used to control for unmeasured confounding in observational healthcare databases. However, proper reporting of methodological issues is sparse in these studies. We aimed to review published papers related to MR studies and identify reporting problems.

Methods

We conducted a systematic review using the clinical articles published between 2009 and 2019. We searched PubMed, Scopus, and Embase databases. We retrieved information from every MR study, including the tests performed to evaluate assumptions and the modelling approach used for estimation. Using our inclusion/exclusion criteria, finally, we identified 97 studies to conduct the review according to the PRISMA statement.

Results

Only 66 (68%) of the studies empirically verified the first assumption (Relevance assumption), and 40 (41.2%) studies reported the appropriate tests (e.g., R2, F-test) to investigate the association. A total of 35.1% clearly stated and discussed theoretical justifications for the second and third assumptions. 30.9% of the studies used a two-stage least square, and 11.3% used the Wald estimator method for estimating IV. Also, 44.3% of the studies conducted a sensitivity analysis to illuminate the robustness of estimates for violations of the untestable assumptions.

Conclusions

We found that incompleteness of the justification of the assumptions for the instrumental variable in MR studies was a common problem in our selected studies. This may misdirect the findings of the studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-022-01504-0.

Protocol for an observational cohort study investigating personalised medicine for intensification of treatment in people with type 2 diabetes mellitus: the PERMIT study

INTRODUCTION

For people with type 2 diabetes mellitus (T2DM) who require an antidiabetic drug as an add-on to metformin, there is controversy about whether newer drug classes such as dipeptidyl peptidase-4 inhibitors (DPP4i) or sodium-glucose co-transporter-2 inhibitors (SGLT2i) reduce the risk of long-term complications compared with sulfonylureas (SU). There is widespread variation across National Health Service Clinical Commissioning Groups (CCGs) in drug choice for second-line treatment in part because National Institute for Health and Care Excellence guidelines do not specify a single preferred drug class, either overall or within specific patient subgroups. This study will evaluate the relative effectiveness of the three most common second-line treatments in the UK (SU, DPP4i and SGLT2i as add-ons to metformin) and help target treatments according to individual risk profiles.

METHODS AND ANALYSIS

The study includes people with T2DM prescribed one of the second-line treatments-of-interest between 2014 and 2020 within the UK Clinical Practice Research Datalink linked with Hospital Episode Statistics and Office of National Statistics. We will use an instrumental variable (IV) method to estimate short-term and long-term relative effectiveness of second-line treatments according to individuals' risk profiles. This method minimises bias from unmeasured confounders by exploiting the natural variation in second-line prescribing across CCGs as an IV for the choice of prescribed treatment. The primary outcome to assess short-term effectiveness will be change in haemoglobin A1c (%) 12 months after treatment initiation. Outcome measures to assess longer-term effectiveness (maximum ~6 years) will include microvascular and macrovascular complications, all-cause mortality and hospital admissions during follow-up.

ETHICS AND DISSEMINATION

This study was approved by the Independent Scientific Advisory Committee (20-064) and the London School of Hygiene & Tropical Medicine Research Ethics Committee (21395). Results, codelists and other analysis code will be made available to patients, clinicians, policy-makers and researchers.

Association of a Novel Index of Hospital Capacity Strain with Admission to Intensive Care Units

Rationale: Prior approaches to measuring healthcare capacity strain have been constrained by using individual care units, limited metrics of strain, or general, rather than disease-specific, populations.Objectives: We sought to develop a novel composite strain index and measure its association with intensive care unit (ICU) admission decisions and hospital outcomes.Methods: Using more than 9.2 million acute care encounters from 27 Kaiser Permanente Northern California and Penn Medicine hospitals from 2013 to 2018, we deployed multivariable ridge logistic regression to develop a composite strain index based on hourly measurements of 22 capacity-strain metrics across emergency departments, wards, step-down units, and ICUs. We measured the association of this strain index with ICU admission and clinical outcomes using multivariable logistic and quantile regression.Results: Among high-acuity patients with sepsis (n = 90,150) and acute respiratory failure (ARF; n = 45,339) not requiring mechanical ventilation or vasopressors, strain at the time of emergency department disposition decision was inversely associated with the probability of ICU admission (sepsis: adjusted probability ranging from 29.0% [95% confidence interval, 28.0-30.0%] at the lowest strain index decile to 9.3% [8.7-9.9%] at the highest strain index decile; ARF: adjusted probability ranging from 47.2% [45.6-48.9%] at the lowest strain index decile to 12.1% [11.0-13.2%] at the highest strain index decile; P < 0.001 at all deciles). Among subgroups of patients who almost always or never went to the ICU, strain was not associated with hospital length of stay, mortality, or discharge disposition (all P ≥ 0.13). Strain was also not meaningfully associated with patient characteristics.Conclusions: Hospital strain, measured by a novel composite strain index, is strongly associated with ICU admission among patients with sepsis and/or ARF. This strain index fulfills the assumptions of a strong within-hospital instrumental variable for quantifying the net benefit of admission to the ICU for patients with sepsis and/or ARF.

A semiparametric instrumental variable approach to optimal treatment regimes under endogeneity

There is a fast-growing literature on estimating optimal treatment regimes based on randomized trials or observational studies under a key identifying condition of no unmeasured confounding. Because confounding by unmeasured factors cannot generally be ruled out with certainty in observational studies or randomized trials subject to noncompliance, we propose a general instrumental variable approach to learning optimal treatment regimes under endogeneity. Specifically, we establish identification of both value function E [ Y D ( L ) ] for a given regime D and optimal regimes arg max D E [ Y D ( L ) ] with the aid of a binary instrumental variable, when no unmeasured confounding fails to hold. We also construct novel multiply robust classification-based estimators. Furthermore, we propose to identify and estimate optimal treatment regimes among those who would comply to the assigned treatment under a monotonicity assumption. In this latter case, we establish the somewhat surprising result that complier optimal regimes can be consistently estimated without directly collecting compliance information and therefore without the complier average treatment effect itself being identified. Our approach is illustrated via extensive simulation studies and a data application on the effect of child rearing on labor participation.

The medical care costs of obesity and severe obesity in youth: An instrumental variables approach

This paper is the first to use the method of instrumental variables to estimate the impact of obesity and severe obesity in youth. on U.S. medical care costs. We examine data from the Medical Expenditure Panel Survey for 2001-2015 and instrument for child BMI using the BMI of the child's biological mother. Instrumental variables estimates indicate that obesity in youth raises annual medical care costs by $907 (in 2015 dollars) or 92%, which is considerably higher than previous estimates of the association of youth obesity with medical costs. We find that obesity in youth significantly raises costs in all major categories of medical care: outpatient doctor visits, inpatient hospital stays, and prescription drugs. The costs of youth obesity are borne almost entirely by third-party payers, which is consistent with substantial externalities of youth obesity, which in turn represents an economic rationale for government intervention.

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.

Sensitivity analysis and power for instrumental variable studies

In observational studies to estimate treatment effects, unmeasured confounding is often a concern. The instrumental variable (IV) method can control for unmeasured confounding when there is a valid IV. To be a valid IV, a variable needs to be independent of unmeasured confounders and only affect the outcome through affecting the treatment. When applying the IV method, there is often concern that a putative IV is invalid to some degree. We present an approach to sensitivity analysis for the IV method which examines the sensitivity of inferences to violations of IV validity. Specifically, we consider sensitivity when the magnitude of association between the putative IV and the unmeasured confounders and the direct effect of the IV on the outcome are limited in magnitude by a sensitivity parameter. Our approach is based on extending the Anderson-Rubin test and is valid regardless of the strength of the instrument. A power formula for this sensitivity analysis is presented. We illustrate its usage via examples about Mendelian randomization studies and its implications via a comparison of using rare versus common genetic variants as instruments.

Re-evaluating associations between the Supplemental Nutrition Assistance Program participation and body mass index in the context of unmeasured confounders

OBJECTIVE

To evaluate the association between participation in the Supplemental Nutrition Assistance Program (SNAP) and body mass index (BMI) in the presence of unmeasured confounding.

METHODS

We applied new matching methods to determine whether previous reports of associations between SNAP participation and BMI were robust to unmeasured confounders. We applied near-far matching, which strengthens standard matching by combining it with instrumental variables analysis, to the nationally-representative National Household Food Acquisition and Purchasing Survey (FoodAPS, N = 10,360, years 2012-13).

RESULTS

In ordinary least squares regressions controlling for individual demographic and socioeconomic characteristics, SNAP was associated with increased BMI (+1.23 kg/m2, 95% CI: 0.84, 1.63). While propensity-score-based analysis replicated this finding, using instrumental variables analysis and particularly near-far matching to strengthen the instruments' discriminatory power revealed the association between SNAP and BMI was likely confounded by unmeasured covariates (+0.21 kg/m2, 95% CI: -3.88, 4.29).

CONCLUSIONS

Previous reports of an association between SNAP and obesity should be viewed with caution, and use of near-far matching may assist similar assessments of health effects of social programs.

A tutorial on the use of instrumental variables in pharmacoepidemiology

PURPOSE

Instrumental variable (IV) methods are used increasingly in pharmacoepidemiology to address unmeasured confounding. In this tutorial, we review the steps used in IV analyses and the underlying assumptions. We also present methods to assess the validity of those assumptions and describe sensitivity analysis to examine the effects of possible violations of those assumptions.

METHODS

Observational studies based on regression or propensity score analyses rely on the untestable assumption that there are no unmeasured confounders. IV analysis is a tool that removes the bias caused by unmeasured confounding provided that key assumptions (some of which are also untestable) are met.

RESULTS

When instruments are valid, IV methods provided unbiased treatment effect estimation in the presence of unmeasured confounders. However, the standard error of the IV estimate is higher than the standard error of non-IV estimates, e.g., regression and propensity score methods. Sensitivity analyses provided insight about the robustness of the IV results to the plausible degrees of violation of assumptions.

CONCLUSIONS

IV analysis should be used cautiously because the validity of IV estimates relies on assumptions that are, in general, untestable and difficult to be certain about. Thus, assessing the sensitivity of the estimate to violations of these assumptions is important and can better inform the causal inferences that can be drawn from the study. Copyright © 2017 John Wiley & Sons, Ltd.

Strengthening Instrumental Variables Through Weighting

Instrumental variable (IV) methods are widely used to deal with the issue of unmeasured confounding and are becoming popular in health and medical research. IV models are able to obtain consistent estimates in the presence of unmeasured confounding, but rely on assumptions that are hard to verify and often criticized. An instrument is a variable that influences or encourages individuals toward a particular treatment without directly affecting the outcome. Estimates obtained using instruments with a weak influence over the treatment are known to have larger small-sample bias and to be less robust to the critical IV assumption that the instrument is randomly assigned. In this work, we propose a weighting procedure for strengthening the instrument while matching. Through simulations, weighting is shown to strengthen the instrument and improve robustness of resulting estimates. Unlike existing methods, weighting is shown to increase instrument strength without compromising match quality. We illustrate the method in a study comparing mortality between kidney dialysis patients receiving hemodialysis or peritoneal dialysis as treatment for end-stage renal disease.

Beyond Mendelian randomization: how to interpret evidence of shared genetic predictors

OBJECTIVE

Mendelian randomization is a popular technique for assessing and estimating the causal effects of risk factors. If genetic variants which are instrumental variables for a risk factor are shown to be additionally associated with a disease outcome, then the risk factor is a cause of the disease. However, in many cases, the instrumental variable assumptions are not plausible, or are in doubt. In this paper, we provide a theoretical classification of scenarios in which a causal conclusion is justified or not justified, and discuss the interpretation of causal effect estimates.

RESULTS

A list of guidelines based on the 'Bradford Hill criteria' for judging the plausibility of a causal finding from an applied Mendelian randomization study is provided. We also give a framework for performing and interpreting investigations performed in the style of Mendelian randomization, but where the choice of genetic variants is statistically, rather than biologically motivated. Such analyses should not be assigned the same evidential weight as a Mendelian randomization investigation.

CONCLUSION

We discuss the role of such investigations (in the style of Mendelian randomization), and what they add to our understanding of potential causal mechanisms. If the genetic variants are selected solely according to statistical criteria, and the biological roles of genetic variants are not investigated, this may be little more than what can be learned from a well-designed classical observational study.

A review of instrumental variable estimators for Mendelian randomization

Instrumental variable analysis is an approach for obtaining causal inferences on the effect of an exposure (risk factor) on an outcome from observational data. It has gained in popularity over the past decade with the use of genetic variants as instrumental variables, known as Mendelian randomization. An instrumental variable is associated with the exposure, but not associated with any confounder of the exposure–outcome association, nor is there any causal pathway from the instrumental variable to the outcome other than via the exposure. Under the assumption that a single instrumental variable or a set of instrumental variables for the exposure is available, the causal effect of the exposure on the outcome can be estimated. There are several methods available for instrumental variable estimation; we consider the ratio method, two-stage methods, likelihood-based methods, and semi-parametric methods. Techniques for obtaining statistical inferences and confidence intervals are presented. The statistical properties of estimates from these methods are compared, and practical advice is given about choosing a suitable analysis method. In particular, bias and coverage properties of estimators are considered, especially with weak instruments. Settings particularly relevant to Mendelian randomization are prioritized in the paper, notably the scenario of a continuous exposure and a continuous or binary outcome.

Are all biases missing data problems?

Estimating causal effects is a frequent goal of epidemiologic studies. Traditionally, there have been three established systematic threats to consistent estimation of causal effects. These three threats are bias due to confounders, selection, and measurement error. Confounding, selection, and measurement bias have typically been characterized as distinct types of biases. However, each of these biases can also be characterized as missing data problems that can be addressed with missing data solutions. Here we describe how the aforementioned systematic threats arise from missing data as well as review methods and their related assumptions for reducing each bias type. We also link the assumptions made by the reviewed methods to the missing completely at random (MCAR) and missing at random (MAR) assumptions made in the missing data framework that allow for valid inferences to be made based on the observed, incomplete data.

Savings in Medical Expenditures Associated with Reductions in Body Mass Index Among US Adults with Obesity, by Diabetes Status

BACKGROUND

The prevalence of obesity has more than doubled in the USA in the past 30 years. Obesity is a significant risk factor for diabetes, cardiovascular disease, and other clinically significant co-morbidities. This paper estimates the medical care cost savings that can be achieved from a given amount of weight loss by people with different starting values of body mass index (BMI), for those with and without diabetes. This information is an important input into analyses of the cost effectiveness of obesity treatments and prevention programs.

METHODS

Two-part models of instrumental variables were estimated using data from the Medical Expenditure Panel Survey (MEPS) for 2000-2010. Models were estimated for all adults as well as separately for those with and without diabetes. We calculated the causal impact of changes in BMI on medical care expenditures, cost savings for specific changes in BMI (5, 10, 15, and 20 %) from starting BMI levels ranging from 30 to 45 kg/m(2), as well as the total excess medical care expenditures caused by obesity.

RESULTS

In the USA, adult obesity raised annual medical care costs by $US3,508 per obese individual, for a nationwide total of $US315.8 billion (year 2010 values). However, the relationship of medical care costs over BMI is J-shaped; costs rise exponentially in the range of class 2 and 3 obesity (BMI ≥35). The heavier the obese individual, the greater the reduction in medical care costs associated with a given percent reduction in BMI. Medical care expenditures are higher, and rise more with BMI, among individuals with diabetes than among those without diabetes.

CONCLUSIONS

The savings from a given percent reduction in BMI are greater the heavier the obese individual, and are greater for those with diabetes than for those without diabetes. The results provide health insurers, employers, government agencies, and health economists with accurate estimates of the change in medical care expenditures resulting from weight loss, which is important information for calculating the cost effectiveness of interventions to prevent and treat obesity.

Instrumental variable methods for causal inference

A goal of many health studies is to determine the causal effect of a treatment or intervention on health outcomes. Often, it is not ethically or practically possible to conduct a perfectly randomized experiment, and instead, an observational study must be used. A major challenge to the validity of observational studies is the possibility of unmeasured confounding (i.e., unmeasured ways in which the treatment and control groups differ before treatment administration, which also affect the outcome). Instrumental variables analysis is a method for controlling for unmeasured confounding. This type of analysis requires the measurement of a valid instrumental variable, which is a variable that (i) is independent of the unmeasured confounding; (ii) affects the treatment; and (iii) affects the outcome only indirectly through its effect on the treatment. This tutorial discusses the types of causal effects that can be estimated by instrumental variables analysis; the assumptions needed for instrumental variables analysis to provide valid estimates of causal effects and sensitivity analysis for those assumptions; methods of estimation of causal effects using instrumental variables; and sources of instrumental variables in health studies.

Case Definition and Design Sensitivity

In a case-referent study, cases of disease are compared to non-cases with respect to their antecedent exposure to a treatment in an effort to determine whether exposure causes some cases of the disease. Because exposure is not randomly assigned in the population, as it would be if the population were a vast randomized trial, exposed and unexposed subjects may differ prior to exposure with respect to covariates that may or may not have been measured. After controlling for measured pre-exposure differences, for instance by matching, a sensitivity analysis asks about the magnitude of bias from unmeasured covariates that would need to be present to alter the conclusions of a study that presumed matching for observed covariates removes all bias. The definition of a case of disease affects sensitivity to unmeasured bias. We explore this issue using: (i) an asymptotic tool, the design sensitivity, (ii) a simulation for finite samples, and (iii) an example. Under favorable circumstances, a narrower case definition can yield an increase in the design sensitivity, and hence an increase in the power of a sensitivity analysis. Also, we discuss an adaptive method that seeks to discover the best case definition from the data at hand while controlling for multiple testing. An implementation in R is available as SensitivityCaseControl.

The Impact of Cigarette Quitting during Pregnancy on Other Prenatal Health Behaviors

Several economic studies have evaluated the effects of cigarette smoking and quitting on other health behaviors such as alcohol use and weight gain. However, there is little research that evaluates the effects of cigarette quitting during pregnancy on other health behaviors such as caloric intake, alcohol consumption, multivitamin use, and caffeine intake. In this paper, we evaluate these effects and employ a genetic variant that predicts cigarette quitting to aid in identification. We find some evidence that cigarette quitting during pregnancy may increase multivitamin use and caloric intake and reduce caffeine consumption.