Approximate Bayesian Bootstrap procedures to estimate multilevel treatment effects in observational studies with application to type 2 diabetes treatment regimens

Randomized clinical trials are considered as the gold standard for estimating causal effects. Nevertheless, in studies that are aimed at examining adverse effects of interventions, randomized trials are often impractical because of ethical and financial considerations. In observational studies, matching on the generalized propensity scores was proposed as a possible solution to estimate the treatment effects of multiple interventions. However, the derivation of point and interval estimates for these matching procedures can become complex with non-continuous or censored outcomes. We propose a novel Approximate Bayesian Bootstrap algorithm that results in statistically valid point and interval estimates of the treatment effects with categorical outcomes. The procedure relies on the estimated generalized propensity scores and multiply imputes the unobserved potential outcomes for each unit. In addition, we describe a corresponding interpretable sensitivity analysis to examine the unconfoundedness assumption. We apply this approach to examine the cardiovascular safety of common, real-world anti-diabetic treatment regimens for type 2 diabetes mellitus in a large observational database.

[1]  D. Rubin,et al.  Estimation of causal effects of binary treatments in unconfounded studies with one continuous covariate , 2017, Statistical methods in medical research.

[2]  Roderick J A Little,et al.  A Review of Hot Deck Imputation for Survey Non‐response , 2010, International statistical review = Revue internationale de statistique.

[3]  S. Schinner,et al.  Effects of Intensive Glucose Lowering in Type 2 Diabetes , 2009 .

[4]  Adrian E. Raftery,et al.  Model-Based Clustering, Discriminant Analysis, and Density Estimation , 2002 .

[5]  Charles E. Leonard,et al.  Comparative risk of serious hypoglycemia with oral antidiabetic monotherapy: A retrospective cohort study , 2018, Pharmacoepidemiology and drug safety.

[6]  Craig Wilson,et al.  Alogliptin after acute coronary syndrome in patients with type 2 diabetes. , 2013, The New England journal of medicine.

[7]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[8]  S. Gudbjörnsdottir,et al.  Cardiovascular safety of glucose‐lowering agents as add‐on medication to metformin treatment in type 2 diabetes: report from the Swedish National Diabetes Register , 2016, Diabetes, obesity & metabolism.

[9]  D. Levy,et al.  Trends in All-Cause and Cardiovascular Disease Mortality Among Women and Men With and Without Diabetes Mellitus in the Framingham Heart Study, 1950 to 2005 , 2009, Circulation.

[10]  Deepak L. Bhatt,et al.  Saxagliptin and cardiovascular outcomes in patients with type 2 diabetes mellitus. , 2013, The New England journal of medicine.

[11]  R. Holman,et al.  Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34) , 1998, The Lancet.

[12]  E. Tuzcu,et al.  Comparison of pioglitazone vs glimepiride on progression of coronary atherosclerosis in patients with type 2 diabetes: the PERISCOPE randomized controlled trial. , 2008, JAMA.

[13]  Roee Gutman,et al.  Estimation of causal effects of binary treatments in unconfounded studies , 2015, Statistics in medicine.

[14]  M. Pencina,et al.  Effect of Linagliptin vs Placebo on Major Cardiovascular Events in Adults With Type 2 Diabetes and High Cardiovascular and Renal Risk: The CARMELINA Randomized Clinical Trial , 2019, JAMA.

[15]  S. Kaul,et al.  The cardiovascular safety of diabetes drugs--insights from the rosiglitazone experience. , 2013, The New England journal of medicine.

[16]  Rachael Williams,et al.  Pioglitazone and cause-specific risk of mortality in patients with type 2 diabetes: extended analysis from a European multidatabase cohort study , 2018, BMJ Open Diabetes Research & Care.

[17]  P. Ponikowski,et al.  [2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure]. , 2016, Kardiologia polska.

[18]  P. Rose,et al.  Validity of diagnostic coding within the General Practice Research Database: a systematic review. , 2010, The British journal of general practice : the journal of the Royal College of General Practitioners.

[19]  J. Rassen,et al.  Simultaneously assessing intended and unintended treatment effects of multiple treatment options: a pragmatic “matrix design” , 2011, Pharmacoepidemiology and drug safety.

[20]  B. Everitt,et al.  Statistical methods for rates and proportions , 1973 .

[21]  Lane F Burgette,et al.  A tutorial on propensity score estimation for multiple treatments using generalized boosted models , 2013, Statistics in medicine.

[22]  Gary King,et al.  Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference , 2007, Political Analysis.

[23]  Elizabeth A Stuart,et al.  Matching methods for causal inference: A review and a look forward. , 2010, Statistical science : a review journal of the Institute of Mathematical Statistics.

[24]  Anthony D Scotina,et al.  Matching estimators for causal effects of multiple treatments , 2018, Statistical methods in medical research.

[25]  Xiao-Hua Zhou,et al.  Generalized propensity score for estimating the average treatment effect of multiple treatments , 2012, Statistics in medicine.

[26]  R. Ackermann,et al.  Association of Second-line Antidiabetic Medications With Cardiovascular Events Among Insured Adults With Type 2 Diabetes. , 2018, JAMA network open.

[27]  D. Ruppert.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction , 2004 .

[28]  S. Hernández-Díaz,et al.  Matching Weights to Simultaneously Compare Three Treatment Groups: Comparison to Three-way Matching , 2017, Epidemiology.

[29]  A. Azzalini,et al.  Distributions generated by perturbation of symmetry with emphasis on a multivariate skew t‐distribution , 2003, 0911.2342.

[30]  David B. Dunson,et al.  Bayesian Data Analysis , 2010 .

[31]  D. Matthews,et al.  Management of Hyperglycemia in Type 2 Diabetes, 2015: A Patient-Centered Approach: Update to a Position Statement of the American Diabetes Association and the European Association for the Study of Diabetes , 2014, Diabetes Care.

[32]  D. Rubin.  2 Statistical Inference for Causal Effects, With Emphasis on Applications in Epidemiology and Medical Statistics , 2007 .

[33]  Chunhao Tu,et al.  Comparison of clustering algorithms on generalized propensity score in observational studies: a simulation study , 2013 .

[34]  Ju-Young Kim,et al.  Differential Cardiovascular Outcomes after Dipeptidyl Peptidase-4 Inhibitor, Sulfonylurea, and Pioglitazone Therapy, All in Combination with Metformin, for Type 2 Diabetes: A Population-Based Cohort Study , 2015, PloS one.

[35]  D. Rubin,et al.  Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction , 2016 .

[36]  M. Fischereder,et al.  Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes. , 2016, The New England journal of medicine.

[37]  Sophie Chen,et al.  How Big is a Big Odds Ratio? Interpreting the Magnitudes of Odds Ratios in Epidemiological Studies , 2010, Commun. Stat. Simul. Comput..

[38]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[39]  Kosuke Imai,et al.  Causal Inference With General Treatment Regimes , 2004 .

[40]  Ofer Harel,et al.  Inferences on missing information under multiple imputation and two-stage multiple imputation , 2007 .

[41]  G. Imbens.  The Role of the Propensity Score in Estimating Dose-Response Functions , 1999 .

[42]  G. Imbens,et al.  On the Failure of the Bootstrap for Matching Estimators , 2006 .

[43]  Anthony D Scotina,et al.  Matching algorithms for causal inference with multiple treatments , 2018, Statistics in medicine.

[44]  A. Gelman,et al.  Multiple Imputation with Diagnostics (mi) in R: Opening Windows into the Black Box , 2011 .

[45]  Robert A Greevy,et al.  Comparative Effectiveness of Sulfonylurea and Metformin Monotherapy on Cardiovascular Events in Type 2 Diabetes Mellitus , 2012, Annals of Internal Medicine.

[46]  Lena Osterhagen,et al.  Multiple Imputation For Nonresponse In Surveys , 2016 .

[47]  Fan Li.  Propensity Score Weighting for Causal Inference with Multi-valued Treatments , 2018, 1808.05339.

[48]  Shu Yang,et al.  Propensity score matching and subclassification in observational studies with multi‐level treatments , 2015, Biometrics.

[49]  Y. Jang,et al.  Standards of Medical Care in Diabetes-2010 by the American Diabetes Association: Prevention and Management of Cardiovascular Disease , 2010 .

[50]  M. Pencina,et al.  Effect of Sitagliptin on Cardiovascular Outcomes in Type 2 Diabetes. , 2015, The New England journal of medicine.

[51]  Neil Skolnik,et al.  Pharmacologic Therapy for Type 2 Diabetes: Synopsis of the 2017 American Diabetes Association Standards of Medical Care in Diabetes , 2017, Annals of Internal Medicine..

[52]  Dylan S. Small,et al.  The use of bootstrapping when using propensity-score matching without replacement: a simulation study , 2014, Statistics in medicine.

[53]  Raul Cano.  On The Bayesian Bootstrap , 1992 .

[54]  D. Rubin,et al.  Multiple Imputation for Interval Estimation from Simple Random Samples with Ignorable Nonresponse , 1986 .

[55]  Chen-Ying Hung,et al.  Glitazones and alpha-glucosidase inhibitors as the second-line oral anti-diabetic agents added to metformin reduce cardiovascular risk in Type 2 diabetes patients: a nationwide cohort observational study , 2018, Cardiovascular Diabetology.

[56]  D. Rubin.  The design versus the analysis of observational studies for causal effects: parallels with the design of randomized trials , 2007, Statistics in medicine.

[57]  A. Hoes,et al.  Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005): The Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology. , 2005, European heart journal.

[58]  R. Gutman,et al.  Monitoring methods in the surgical arena: assumptions and potential pitfalls. , 2013, European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology.

[59]  M. Petticrew.  Diagoras of Melos (500 BC): an early analyst of publication bias , 1998, The Lancet.

[60]  Michael J. Lopez,et al.  Estimation of causal effects with multiple treatments: a review and new ideas , 2017, 1701.05132.

[61]  D. Rubin,et al.  Testing treatment effects in unconfounded studies under model misspecification: Logistic regression, discretization, and their combination , 2009, Statistics in medicine.

[62]  P W Lavori,et al.  A multiple imputation strategy for clinical trials with truncation of patient data. , 1995, Statistics in medicine.

[63]  D. V. Lindley,et al.  Randomization Analysis of Experimental Data: The Fisher Randomization Test Comment , 1980 .

[64]  Erland Erdmann,et al.  Secondary prevention of macrovascular events in patients with type 2 diabetes in the PROactive Study (PROspective pioglitAzone Clinical Trial In macroVascular Events): a randomised controlled trial , 2005, The Lancet.

[65]  Donald B. Rubin,et al.  Bayesian Inference for Causal Effects: The Role of Randomization , 1978 .

[66]  W. Hiatt,et al.  Two new drugs for homozygous familial hypercholesterolemia: managing benefits and risks in a rare disorder. , 2013, JAMA internal medicine.

[67]  R. Guthrie.  Canagliflozin and cardiovascular and renal events in type 2 diabetes , 2018, Postgraduate medicine.

[68]  W. Conover.  Statistical Methods for Rates and Proportions , 1974 .

[69]  Rajeev Dehejia,et al.  Propensity Score-Matching Methods for Nonexperimental Causal Studies , 2002, Review of Economics and Statistics.

[70]  N. Skolnik,et al.  Management of Hyperglycemia in Type 2 Diabetes , 2012 .