1 Investigators

Investigator Contact
James Weaver 1,2,3
Patrick B. Ryan 2,3,4
Victoria Strauss 1,2
Marc A. Suchard 3,5
Joel Swerdel 2,3
Daniel Prieto-Alhambra 1,3,6
1Centre for Statistics in Medicine, Nuffield Department of Orthopedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, UK; 2Observational Health Data Analytics, Janssen Research and Development, Titusville, NJ, USA; 3Observational Health Data Sciences and Informatics, New York, NY, USA; 4Department of Biomedical Informatics, Columbia University Medical Center, New York, NY, USA; 5Department of Biostatistics, Fielding School of Public Health, and Department of Biomathematics, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; 6Medical Informatics, Erasmus Medical Centre, Rotterdam, Netherlands

1.1 Disclosures

This study is undertaken within Observational Health Data Sciences and Informatics (OHDSI), an open science collaboration. JW, PR, and JS are employees of shareholders of Janssen R&D (a Johnson & Johnson company). MAS receives grant support from the US National Institutes of Health, US Food & Drug Administration and US Department of Veterans Affairs and contracts from Janssen R&D. VS has no conflicts of interest to declare. DPA’s research group has received grant support from Amgen, Chesi-Taylor, Novartis, and UCB Biopharma. His department has received advisory, consultancy fees from Amgen, Astellas, AstraZeneca, Johnson & Johnson, and UCB Biopharma and fees for speaker services from Amgen and UCB Biopharma. Janssen, on behalf of IMI-funded EHDEN and EMIF consortiums, and Synapse Management Partners have supported training programs organized by DPA’s department and open for external participants organized by his department outside submitted work.

2 Rationale and background

Phenotype error is acknowledged but rarely corrected for in causal effect estimation studies using observational data. Quantitative bias analysis (QBA) is a method for phenotype error correction, but the extent to which it minimizes bias in effect estimates is unclear.

3 Study objectives

  • Empirically evaluate QBA for outcome phenotype error correction in several pharmacoepidemiologic comparative effect estimation scenarios
  • Simulate an analytic space defined by outcome incidence proportions (IP), observed effect estimates, and phenotype measurement errors to determine which QBA input combinations produce valid results.

4 Research methods

4.1 Quantitative bias analysis

QBA is a method for correcting outcome phenotype error that can bias comparative effect estimates[1,2]. Phenotype definition performance characteristics (sensitivity and specificity, positive and negative predictive value) are required inputs for applying QBA. We use QBA for outcome phenotype error correction per the equations in Table 1a and Table 1b.

4.1.1 Table 1a: Observed exposure by outcome contingency table

Outcome T C
O[+] a b
O[-] c d
Total a + c a + d
T: observed target exposures, C: observed comparator exposures, O: observed outcomes

4.1.2 Table 1b Expected exposure by outcome contingency table corrected for outcome phenotype error

Outcome T C
O[+] A = a - (1 - SP1) * (a + c)) / (SN1 - (1 - SP1) B = b - (1 - SP0) * (b + d)) / (SN0 - (1 - SP0)
O[-] C = (a + c) - A D = (b + d) - B
Total A + C B + D
T: expected target exposures, C: expected comparator exposures, O: expected outcomes, SN1: sensitivity in target, SN0: sensitivity in comparator, SP1: specificity in target, SP0 specificity in comparator

4.2 Probabilistic reference standard validation

PheValuator is a method to calculate the performance characteristics of phenotype algorithms, namely, sensitivity, specificity, and positive and negative predictive value[3,4]. It develops a diagnostic predictive model to determine a probabilistic reference standard of patients against which phenotype algorithm performance can be assessed. Table 2 reports the PheValuator confusion matrix and error metric calculations.

4.2.1 Table 2: PheValuator confusion matrix

Diagnostic model output Phenotype algorithm case Phenotype algorithm non-case
Predicted probability from diagnostic predictive model, P(Y) TP = \(\Sigma\)[P(Y | Case)] FP = \(\Sigma\)[1 - P(Y | Case)]
Predicted probability from diagnostic predictive model, P(Y) FN = \(\Sigma\)[P(Y | Non-case)] TN = \(\Sigma\)[1 - P(Y | Non-case)]
TP: true positive, FP: false positive, FN: false negative, TN: true negative, Sensitivity = TP / (TP + FN), Specificity = TN / (TN + FP), PPV = TP / (TP + FP), NPV = TN / (TN + FN)

The following cohorts are required input for a PheValuator validation study:

Detailed cohort definitions for probabilistic reference standard validation are available in Appendix Section A.

4.3 Empirical example

4.3.1 Design

Active comparator, new user comparative cohort study to estimate the risk of ischemic stroke among patients with hypertension initiating:

  • Angiotensin-converting enzyme inhibitors (ACE) vs angiotensin receptor blockers (ARB)
  • Angiotensin-converting enzyme inhibitors (ACE) vs Thiazide/thiazide-like diuretics (THZ)

4.3.2 Exposure cohort definitions

Detailed exposure cohort definitions for 3 class-level hypertension treatments are in Appendix Section B.

4.3.2.1 ACEI new users with prior hypertension

  • First use of ACEI on or after January 1, 2010 with ≥365 days of prior continuous database observation
    • ≥1 condition occurrence of hypertension between 365 and 0 days relative to first use
    • exactly 1 exposure to hypertension medications between 0 and 7 days relative to first use
    • no prior exposure to hypertension medications

4.3.2.2 ARB new users with prior hypertension

  • First use of ARB on or after January 1, 2010 with ≥365 days of prior continuous database observation
    • ≥1 condition occurrence of hypertension between 365 and 0 days relative to first use
    • exactly 1 exposure to hypertension medications between 0 and 7 days relative to first use
    • no prior exposure to hypertension medications

4.3.2.3 THZ new users with prior hypertension

  • First use of THZ on or after January 1, 2010 with ≥365 days of prior continuous database observation
    • ≥1 condition occurrence of hypertension between 365 and 0 days relative to first use
    • exactly 1 exposure to hypertension medications between 0 and 7 days relative to first use
    • no prior exposure to hypertension medications

4.3.3 Outcome definition

  • Inpatient or emergency room visits on or after January 1, 2010
    • ≥1 condition occurrence of ischemic stroke
    • exactly 0 condition occurrences of ischemic stroke between -365 and -1 days relative to inpatient or emergency room visit with ischemic stroke

The detailed outcome definition for inpatient ischemic stroke is in Appendix Section C.

4.3.4 Data sources

The study will be executed against 4 US adminstrative healthcare claims and 1 US electronic health record databases.

  • Optum® de-identified Clinformatics® Datamart - Date of Death (optum_extended_dod)
  • Optum® Electronic Health Record (optum_ehr)
  • IBM MarketScan® Commercial Database (truven_ccae)
  • IBM MarketScan® Multi-State Medicaid (truven_mdcd)
  • IBM MarketScan® Medicare Supplemental Beneficiaries (truven_mdcr)

The database descriptions are in Appendix Section D.

4.3.5 Outcome definition

See Appendix Section B for detailed exposure definitions.

4.3.6 Time-at-risk

  • 1 day to 365 days relative to exposure start
  • 1 day to 730 days relative to exposure start

4.3.7 Analyses

  • Calculate database-level (i.e., non-differential) and exposure-level (i.e., differential) ischemic stroke phenotype definition sensitivity and specificity using probabilistic reference standard validation studies in each data source
  • Estimate comparative treatment effect using logistic regression (odds ratio [OR] with 95% confidence intervals [CI]) for ACE vs ARB and ACE vs THZ under the following analysis specifications:
    • Unadjusted
    • Non-differential QBA adjustment
    • Differential QBA adjustment
    • 1:1 propensity score (PS) matched
    • 1:1 PS matched with non-differential QBA
    • 1:1 PS matched with differential QBA
  • Execute 5 databases x 2 comparisons x 2 TARs x 6 analyses] = 120 analyses

4.3.8 Evaluation metrics

QBA performance evaluated by bias difference, relative bias, squared error, and precision difference between analyses that did vs did not include QBA.

  • Bias difference: log(OR) - log(ORQBA)
  • Relative bias: (OR - ORQBA) / OR * 100
  • Precision difference: 1 / (SE(log(OR))2 – 1 / (SE(log(ORQBA))2
  • Relative precision: (1 / (SE(log(OR))2) – (1 / (SE(log(ORQBA))2) / (1 / (SE(log(OR))2) * 100
  • Squared error: (log(OR) - log(ORQBA)2

4.4 Grid space simulation

4.4.1 Inputs

Create grid space of all combinations of 4 input parameters:

  • 5 outcome incidence proportions (IP) [10-1 , 10-2 , 10-3 , 10-4 , 10-5]
  • 6 uncorrected odds ratios (OR) [1, 1.25, 1.50, 2, 4, 10]
  • 20 non-differential sensitivity values [0.05 to 1.00 by 0.05]
  • 20 non-differential specificity values [1 - prevalence to 1.00 by 5%ile]

The complete grid space consists of 12,000 2x2 contingency tables, each with 1,000,000 target and 1,000,000 comparator exposures and associated inputs.

4.4.2 Analysis

For each IP-OR combination, compute a distribution of QBA-corrected ORs with 95% CIs across combinations of sensitivity and specificity values and plot contours across the complete IP by OR grid space.

4.4.3 Evaluation metrics

The grid space simulation analysis will be evaluated by bias difference and relative bias between the unadjusted OR and the 25%ile, 50%ile, 75%ile, and maximum of the QBA-corrected distribution of estimates. Report the overall, IP-stratified, OR-stratified, and IP-OR-stratified proportion of the total grid space that produces valid (i.e., non-zero) QBA-corrected counts.

  • Bias difference: log(OR) - log(ORQBA)
  • Relative bias: (OR - ORQBA) / OR * 100

5 Strengths and limitations

5.1 Strengths

  • 1
  • 2
  • 3

5.2 Limitations

  • Empirical example uses simple and multidimensional QBA only, no probabilistic QBA or multiple bias modeling
  • Only uses sensitivity and specificity approach, no use of PPV and NPV
  • Logistic regression outcome model assumes constant risk, discards survival information
  • Assumes probabilistic reference validation metrics are accurate
  • Validation study within exposure-indication populations will have incomplete overlap with restricted study populations

6 Protection of human subjects

This work does not involve human patient research. It uses de-identified patient-level data collected during routine healthcare provision. Confidentiality of patient records will be maintained. Study reports will contain aggregate data only and will not identify individual patients of care providers.

References

1
Lash TL, Fox MP, Fink AK. Applying quantitative bias analysis to epidemiologic data. Springer New York 2009.
2
Lash TL, Fox MP, MacLehose RF, et al. Good practices for quantitative bias analysis. Int J Epidemiol 2014;43:1969–85.
3
Swerdel JN, Hripcsak G, Ryan PB. PheValuator: Development and evaluation of a phenotype algorithm evaluator. J Biomed Inform 2019;97:103258.
4
Swerdel JN, Schuemie M, Murray G, et al. PheValuator 2.0: Methodological improvements for the PheValuator approach to semi-automated phenotype algorithm evaluation. J Biomed Inform 2022;135:104177.

Appendix

A Probabilistic reference standard validation cohort definitions

A.1 xSpec validation cohort

A.1.1 Cohort Entry Events

People may enter the cohort when observing any of the following:

  1. visit occurrences of any visit, starting on or after January 1, 2010.

Restrict entry events to having at least 2 condition occurrences of ‘[QBA eval] Cerebral infarction NC’, starting in the 1 days prior to cohort entry start date.

A.1.2 Cohort Exit

The cohort end date will be offset from index event’s start date plus 365 days.

A.1.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

A.2 xSens validation cohort

A.2.1 Cohort Entry Events

People may enter the cohort when observing any of the following:

  1. visit occurrences of any visit, starting on or after January 1, 2010.

Restrict entry events to having at least 1 condition occurrence of ‘[QBA eval] Cerebral infarction NC’, starting in the 91 days prior to cohort entry start date.

A.2.2 Cohort Exit

The cohort end date will be offset from index event’s start date plus 365 days.

A.2.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

A.3 Prevalence validation cohort

A.3.1 Cohort Entry Events

People enter the cohort when observing any of the following:

  1. condition occurrences of ‘[QBA eval] Cerebral infarction NC’, starting on or after January 1, 2010.

Limit cohort entry events to the earliest event per person.

A.3.2 Cohort Exit

The person also exists the cohort at the end of continuous observation.

A.3.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

A.4 Database population evaluation cohort

A.4.1 Cohort Entry Events

People enter the cohort when observing any of the following:

  1. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; having no condition occurrences of ‘[QBA eval] Cerebral infarction NC’.

  2. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; having at least 1 condition occurrence of ‘[QBA eval] Cerebral infarction NC’, starting between 0 days before and all days after ‘Inpatient or Inpatient/ER visit’ start date and starting anytime on or before ‘Inpatient or Inpatient/ER visit’ end date.

A.4.2 Cohort Exit

The cohort end date will be offset from index event’s start date plus 365 days.

A.4.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

A.5 ACEI new users evaluation cohort

A.5.1 Cohort Entry Events

People enter the cohort when observing any of the following:

  1. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; with all of the following criteria:

  2. having no condition occurrences of ‘[QBA eval] Cerebral infarction NC’.

  3. having at least 1 drug exposure of ‘[QBA eval] ACEIs’, starting anytime on or before ‘Inpatient or Inpatient/ER visit’ start date; having at least 1 condition occurrence of ‘[QBA eval] hypertension’, starting anytime on or before ‘[QBA eval] ACEIs’ start date.

  4. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; with all of the following criteria:

  5. having at least 1 condition occurrence of ‘[QBA eval] Cerebral infarction NC’, starting between 0 days before and all days after ‘Inpatient or Inpatient/ER visit’ start date and starting anytime on or before ‘Inpatient or Inpatient/ER visit’ end date.

  6. having at least 1 drug exposure of ‘[QBA eval] ACEIs’, starting anytime on or before ‘Inpatient or Inpatient/ER visit’ start date; having at least 1 condition occurrence of ‘[QBA eval] hypertension’.

A.5.2 Cohort Exit

The cohort end date will be offset from index event’s start date plus 365 days.

A.5.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

A.6 ARB new users evaluation cohort

A.6.1 Cohort Entry Events

People enter the cohort when observing any of the following:

  1. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; with all of the following criteria:

  2. having no condition occurrences of ‘[QBA eval] Cerebral infarction NC’.

  3. having at least 1 drug exposure of ‘[QBA eval] ARBs’, starting anytime on or before ‘Inpatient or Inpatient/ER visit’ start date; having at least 1 condition occurrence of ‘[QBA eval] hypertension’, starting anytime on or before ‘[QBA eval] ARBs’ start date.

  4. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; with all of the following criteria:

  5. having at least 1 condition occurrence of ‘[QBA eval] Cerebral infarction NC’, starting between 0 days before and all days after ‘Inpatient or Inpatient/ER visit’ start date and starting anytime on or before ‘Inpatient or Inpatient/ER visit’ end date.

  6. having at least 1 drug exposure of ‘[QBA eval] ARBs’, starting anytime on or before ‘Inpatient or Inpatient/ER visit’ start date; having at least 1 condition occurrence of ‘[QBA eval] hypertension’.

A.6.2 Cohort Exit

The cohort end date will be offset from index event’s start date plus 365 days.

A.6.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

A.7 THS new users evaluation cohort

A.7.1 Cohort Entry Events

People enter the cohort when observing any of the following:

  1. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; with all of the following criteria:

  2. having no condition occurrences of ‘[QBA eval] Cerebral infarction NC’.

  3. having at least 1 drug exposure of ‘[QBA eval] THZs’, starting anytime on or before ‘Inpatient or Inpatient/ER visit’ start date; having at least 1 condition occurrence of ‘[QBA eval] hypertension’, starting anytime on or before ‘[QBA eval] THZs’ start date.

  4. visit occurrences of ‘Inpatient or Inpatient/ER visit’, starting on or after January 1, 2010; with all of the following criteria:

  5. having at least 1 condition occurrence of ‘[QBA eval] Cerebral infarction NC’, starting between 0 days before and all days after ‘Inpatient or Inpatient/ER visit’ start date and starting anytime on or before ‘Inpatient or Inpatient/ER visit’ end date.

  6. having at least 1 drug exposure of ‘[QBA eval] THZs’, starting anytime on or before ‘Inpatient or Inpatient/ER visit’ start date; having at least 1 condition occurrence of ‘[QBA eval] hypertension’.

A.7.2 Cohort Exit

The cohort end date will be offset from index event’s start date plus 365 days.

A.7.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

B Exposure cohort definitions

B.1 ACEI new users with prior hypertension

B.1.1 Cohort Entry Events

People with continuous observation of 365 days before event may enter the cohort when observing any of the following:

  1. drug exposure of ‘[QBA eval] ACEIs’ for the first time in the person’s history, starting on or after January 1, 2010.

Limit cohort entry events to the earliest event per person.

Restrict entry events to with all of the following criteria:

  1. having no drug exposures of ‘[QBA eval] hypertension drugs’, starting anytime prior to cohort entry start date.
  2. having at least 1 condition occurrence of ‘[QBA eval] hypertension’, starting between 365 days before and 0 days after cohort entry start date.
  3. having exactly 1 distinct standard concepts from drug era of ‘[QBA eval] hypertension drugs’, starting between 0 days before and 7 days after cohort entry start date.

B.1.2 Cohort Exit

The cohort end date will be based on a continuous exposure to ‘[QBA eval] ACEIs’: allowing 30 days between exposures, adding 0 days after exposure ends, and using days supply and exposure end date for exposure duration.

B.1.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

B.1.4 Concept: [QBA eval] hypertension

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
316866 Hypertensive disorder 38341003 SNOMED NO YES NO

B.1.5 Concept: [QBA eval] hypertension drugs

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
1319998 acebutolol 149 RxNorm NO YES NO
1317967 aliskiren 325646 RxNorm NO YES NO
991382 amiloride 644 RxNorm NO YES NO
1332418 amlodipine 17767 RxNorm NO YES NO
1314002 atenolol 1202 RxNorm NO YES NO
40235485 azilsartan 1091643 RxNorm NO YES NO
1335471 benazepril 18867 RxNorm NO YES NO
1322081 betaxolol 1520 RxNorm NO YES NO
1338005 bisoprolol 19484 RxNorm NO YES NO
932745 bumetanide 1808 RxNorm NO YES NO
1351557 candesartan 214354 RxNorm NO YES NO
1340128 captopril 1998 RxNorm NO YES NO
1346823 carvedilol 20352 RxNorm NO YES NO
1395058 chlorthalidone 2409 RxNorm NO YES NO
1398937 clonidine 2599 RxNorm NO YES NO
1328165 diltiazem 3443 RxNorm NO YES NO
1363053 doxazosin 49276 RxNorm NO YES NO
1341927 enalapril 3827 RxNorm NO YES NO
1309799 eplerenone 298869 RxNorm NO YES NO
1346686 eprosartan 83515 RxNorm NO YES NO
1353776 felodipine 4316 RxNorm NO YES NO
1363749 fosinopril 50166 RxNorm NO YES NO
956874 furosemide 4603 RxNorm NO YES NO
1344965 guanfacine 40114 RxNorm NO YES NO
1373928 hydralazine 5470 RxNorm NO YES NO
974166 hydrochlorothiazide 5487 RxNorm NO YES NO
978555 indapamide 5764 RxNorm NO YES NO
1347384 irbesartan 83818 RxNorm NO YES NO
1326012 isradipine 33910 RxNorm NO YES NO
1386957 labetalol 6185 RxNorm NO YES NO
1308216 lisinopril 29046 RxNorm NO YES NO
1367500 losartan 52175 RxNorm NO YES NO
1305447 methyldopa 6876 RxNorm NO YES NO
907013 metolazone 6916 RxNorm NO YES NO
1307046 metoprolol 6918 RxNorm NO YES NO
1309068 minoxidil 6984 RxNorm NO YES NO
1310756 moexipril 30131 RxNorm NO YES NO
1313200 nadolol 7226 RxNorm NO YES NO
1314577 nebivolol 31555 RxNorm NO YES NO
1318137 nicardipine 7396 RxNorm NO YES NO
1318853 nifedipine 7417 RxNorm NO YES NO
1319880 nisoldipine 7435 RxNorm NO YES NO
40226742 olmesartan 321064 RxNorm NO YES NO
1327978 penbutolol 7973 RxNorm NO YES NO
1373225 perindopril 54552 RxNorm NO YES NO
1345858 pindolol 8332 RxNorm NO YES NO
1350489 prazosin 8629 RxNorm NO YES NO
1353766 propranolol 8787 RxNorm NO YES NO
1331235 quinapril 35208 RxNorm NO YES NO
1334456 ramipril 35296 RxNorm NO YES NO
970250 spironolactone 9997 RxNorm NO YES NO
1317640 telmisartan 73494 RxNorm NO YES NO
1341238 terazosin 37798 RxNorm NO YES NO
942350 torsemide 38413 RxNorm NO YES NO
1342439 trandolapril 38454 RxNorm NO YES NO
904542 triamterene 10763 RxNorm NO YES NO
1308842 valsartan 69749 RxNorm NO YES NO
1307863 verapamil 11170 RxNorm NO YES NO

B.1.6 Concept: [QBA eval] ACEIs

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
1335471 benazepril 18867 RxNorm NO YES NO
1340128 captopril 1998 RxNorm NO YES NO
1341927 enalapril 3827 RxNorm NO YES NO
1363749 fosinopril 50166 RxNorm NO YES NO
1308216 lisinopril 29046 RxNorm NO YES NO
1310756 moexipril 30131 RxNorm NO YES NO
1373225 perindopril 54552 RxNorm NO YES NO
1331235 quinapril 35208 RxNorm NO YES NO
1334456 ramipril 35296 RxNorm NO YES NO
1342439 trandolapril 38454 RxNorm NO YES NO

B.2 ARB new users with prior hypertension

B.2.1 Cohort Entry Events

People with continuous observation of 365 days before event may enter the cohort when observing any of the following:

  1. drug exposure of ‘[QBA eval] ARBs’ for the first time in the person’s history, starting on or after January 1, 2010.

Limit cohort entry events to the earliest event per person.

Restrict entry events to with all of the following criteria:

  1. having no drug exposures of ‘[QBA eval] hypertension drugs’, starting anytime prior to cohort entry start date.
  2. having at least 1 condition occurrence of ‘[QBA eval] hypertension’, starting between 365 days before and 0 days after cohort entry start date.
  3. having exactly 1 distinct standard concepts from drug era of ‘[QBA eval] hypertension drugs’, starting between 0 days before and 7 days after cohort entry start date.

B.2.2 Cohort Exit

The cohort end date will be based on a continuous exposure to ‘[QBA eval] ARBs’: allowing 30 days between exposures, adding 0 days after exposure ends, and using days supply and exposure end date for exposure duration.

B.2.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

B.2.4 Concept: [QBA eval] hypertension

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
316866 Hypertensive disorder 38341003 SNOMED NO YES NO

B.2.5 Concept: [QBA eval] hypertension drugs

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
1319998 acebutolol 149 RxNorm NO YES NO
1317967 aliskiren 325646 RxNorm NO YES NO
991382 amiloride 644 RxNorm NO YES NO
1332418 amlodipine 17767 RxNorm NO YES NO
1314002 atenolol 1202 RxNorm NO YES NO
40235485 azilsartan 1091643 RxNorm NO YES NO
1335471 benazepril 18867 RxNorm NO YES NO
1322081 betaxolol 1520 RxNorm NO YES NO
1338005 bisoprolol 19484 RxNorm NO YES NO
932745 bumetanide 1808 RxNorm NO YES NO
1351557 candesartan 214354 RxNorm NO YES NO
1340128 captopril 1998 RxNorm NO YES NO
1346823 carvedilol 20352 RxNorm NO YES NO
1395058 chlorthalidone 2409 RxNorm NO YES NO
1398937 clonidine 2599 RxNorm NO YES NO
1328165 diltiazem 3443 RxNorm NO YES NO
1363053 doxazosin 49276 RxNorm NO YES NO
1341927 enalapril 3827 RxNorm NO YES NO
1309799 eplerenone 298869 RxNorm NO YES NO
1346686 eprosartan 83515 RxNorm NO YES NO
1353776 felodipine 4316 RxNorm NO YES NO
1363749 fosinopril 50166 RxNorm NO YES NO
956874 furosemide 4603 RxNorm NO YES NO
1344965 guanfacine 40114 RxNorm NO YES NO
1373928 hydralazine 5470 RxNorm NO YES NO
974166 hydrochlorothiazide 5487 RxNorm NO YES NO
978555 indapamide 5764 RxNorm NO YES NO
1347384 irbesartan 83818 RxNorm NO YES NO
1326012 isradipine 33910 RxNorm NO YES NO
1386957 labetalol 6185 RxNorm NO YES NO
1308216 lisinopril 29046 RxNorm NO YES NO
1367500 losartan 52175 RxNorm NO YES NO
1305447 methyldopa 6876 RxNorm NO YES NO
907013 metolazone 6916 RxNorm NO YES NO
1307046 metoprolol 6918 RxNorm NO YES NO
1309068 minoxidil 6984 RxNorm NO YES NO
1310756 moexipril 30131 RxNorm NO YES NO
1313200 nadolol 7226 RxNorm NO YES NO
1314577 nebivolol 31555 RxNorm NO YES NO
1318137 nicardipine 7396 RxNorm NO YES NO
1318853 nifedipine 7417 RxNorm NO YES NO
1319880 nisoldipine 7435 RxNorm NO YES NO
40226742 olmesartan 321064 RxNorm NO YES NO
1327978 penbutolol 7973 RxNorm NO YES NO
1373225 perindopril 54552 RxNorm NO YES NO
1345858 pindolol 8332 RxNorm NO YES NO
1350489 prazosin 8629 RxNorm NO YES NO
1353766 propranolol 8787 RxNorm NO YES NO
1331235 quinapril 35208 RxNorm NO YES NO
1334456 ramipril 35296 RxNorm NO YES NO
970250 spironolactone 9997 RxNorm NO YES NO
1317640 telmisartan 73494 RxNorm NO YES NO
1341238 terazosin 37798 RxNorm NO YES NO
942350 torsemide 38413 RxNorm NO YES NO
1342439 trandolapril 38454 RxNorm NO YES NO
904542 triamterene 10763 RxNorm NO YES NO
1308842 valsartan 69749 RxNorm NO YES NO
1307863 verapamil 11170 RxNorm NO YES NO

B.2.6 Concept: [QBA eval] ARBs

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
40235485 azilsartan 1091643 RxNorm NO YES NO
1351557 candesartan 214354 RxNorm NO YES NO
1346686 eprosartan 83515 RxNorm NO YES NO
1347384 irbesartan 83818 RxNorm NO YES NO
1367500 losartan 52175 RxNorm NO YES NO
40226742 olmesartan 321064 RxNorm NO YES NO
1317640 telmisartan 73494 RxNorm NO YES NO
1308842 valsartan 69749 RxNorm NO YES NO

B.3 THZ new users with prior hypertension

B.3.1 Cohort Entry Events

People with continuous observation of 365 days before event may enter the cohort when observing any of the following:

  1. drug exposure of ‘[QBA eval] THZs’ for the first time in the person’s history, starting on or after January 1, 2010.

Limit cohort entry events to the earliest event per person.

Restrict entry events to with all of the following criteria:

  1. having no drug exposures of ‘[QBA eval] hypertension drugs’, starting anytime prior to cohort entry start date.
  2. having at least 1 condition occurrence of ‘[QBA eval] hypertension’, starting between 365 days before and 0 days after cohort entry start date.
  3. having exactly 1 distinct standard concepts from drug era of ‘[QBA eval] hypertension drugs’, starting between 0 days before and 7 days after cohort entry start date.

B.3.2 Cohort Exit

The cohort end date will be based on a continuous exposure to ‘[QBA eval] THZs’: allowing 30 days between exposures, adding 0 days after exposure ends, and using days supply and exposure end date for exposure duration.

B.3.3 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

B.3.4 Concept: [QBA eval] hypertension

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
316866 Hypertensive disorder 38341003 SNOMED NO YES NO

B.3.5 Concept: [QBA eval] hypertension drugs

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
1319998 acebutolol 149 RxNorm NO YES NO
1317967 aliskiren 325646 RxNorm NO YES NO
991382 amiloride 644 RxNorm NO YES NO
1332418 amlodipine 17767 RxNorm NO YES NO
1314002 atenolol 1202 RxNorm NO YES NO
40235485 azilsartan 1091643 RxNorm NO YES NO
1335471 benazepril 18867 RxNorm NO YES NO
1322081 betaxolol 1520 RxNorm NO YES NO
1338005 bisoprolol 19484 RxNorm NO YES NO
932745 bumetanide 1808 RxNorm NO YES NO
1351557 candesartan 214354 RxNorm NO YES NO
1340128 captopril 1998 RxNorm NO YES NO
1346823 carvedilol 20352 RxNorm NO YES NO
1395058 chlorthalidone 2409 RxNorm NO YES NO
1398937 clonidine 2599 RxNorm NO YES NO
1328165 diltiazem 3443 RxNorm NO YES NO
1363053 doxazosin 49276 RxNorm NO YES NO
1341927 enalapril 3827 RxNorm NO YES NO
1309799 eplerenone 298869 RxNorm NO YES NO
1346686 eprosartan 83515 RxNorm NO YES NO
1353776 felodipine 4316 RxNorm NO YES NO
1363749 fosinopril 50166 RxNorm NO YES NO
956874 furosemide 4603 RxNorm NO YES NO
1344965 guanfacine 40114 RxNorm NO YES NO
1373928 hydralazine 5470 RxNorm NO YES NO
974166 hydrochlorothiazide 5487 RxNorm NO YES NO
978555 indapamide 5764 RxNorm NO YES NO
1347384 irbesartan 83818 RxNorm NO YES NO
1326012 isradipine 33910 RxNorm NO YES NO
1386957 labetalol 6185 RxNorm NO YES NO
1308216 lisinopril 29046 RxNorm NO YES NO
1367500 losartan 52175 RxNorm NO YES NO
1305447 methyldopa 6876 RxNorm NO YES NO
907013 metolazone 6916 RxNorm NO YES NO
1307046 metoprolol 6918 RxNorm NO YES NO
1309068 minoxidil 6984 RxNorm NO YES NO
1310756 moexipril 30131 RxNorm NO YES NO
1313200 nadolol 7226 RxNorm NO YES NO
1314577 nebivolol 31555 RxNorm NO YES NO
1318137 nicardipine 7396 RxNorm NO YES NO
1318853 nifedipine 7417 RxNorm NO YES NO
1319880 nisoldipine 7435 RxNorm NO YES NO
40226742 olmesartan 321064 RxNorm NO YES NO
1327978 penbutolol 7973 RxNorm NO YES NO
1373225 perindopril 54552 RxNorm NO YES NO
1345858 pindolol 8332 RxNorm NO YES NO
1350489 prazosin 8629 RxNorm NO YES NO
1353766 propranolol 8787 RxNorm NO YES NO
1331235 quinapril 35208 RxNorm NO YES NO
1334456 ramipril 35296 RxNorm NO YES NO
970250 spironolactone 9997 RxNorm NO YES NO
1317640 telmisartan 73494 RxNorm NO YES NO
1341238 terazosin 37798 RxNorm NO YES NO
942350 torsemide 38413 RxNorm NO YES NO
1342439 trandolapril 38454 RxNorm NO YES NO
904542 triamterene 10763 RxNorm NO YES NO
1308842 valsartan 69749 RxNorm NO YES NO
1307863 verapamil 11170 RxNorm NO YES NO

B.3.6 Concept: [QBA eval] THZs

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
1395058 chlorthalidone 2409 RxNorm NO YES NO
974166 hydrochlorothiazide 5487 RxNorm NO YES NO
978555 indapamide 5764 RxNorm NO YES NO
907013 metolazone 6916 RxNorm NO YES NO

C Outcome cohort definition

C.1 Ischemic stroke events during inpatient or emergency room visits

C.1.1 Cohort Entry Events

People may enter the cohort when observing any of the following:

  1. condition occurrences of ‘Cerebral infarction’, starting on or after January 1, 2010.

Restrict entry events to having at least 1 visit occurrence of ‘Inpatient or Inpatient/ER visit’, starting anytime on or before cohort entry start date and ending between 0 days before and all days after cohort entry start date.

C.1.2 Additional Inclusion Criteria

I. has no events in prior ‘clean window’ - 365 days

Entry events having no condition occurrences of ‘Cerebral infarction’, starting in the 365 days prior to cohort entry start date; allow events outside observation period; having at least 1 visit occurrence of ‘Inpatient or Inpatient/ER visit’, starting anytime on or before ‘Cerebral infarction’ start date and ending between 0 days before and all days after ‘Cerebral infarction’ start date.

C.1.3 Cohort Exit

The cohort end date will be offset from index event’s start date plus 1 day.

C.1.4 Cohort Eras

Entry events will be combined into cohort eras if they are within 0 days of each other.

C.1.5 Concept: Inpatient or Inpatient/ER visit

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
262 Emergency Room and Inpatient Visit ERIP Visit NO YES NO
9201 Inpatient Visit IP Visit NO YES NO

C.1.6 Concept: Cerebral infarction

Concept ID Concept Name Code Vocabulary Excluded Descendants Mapped
443454 Cerebral infarction 432504007 SNOMED NO YES NO
40479572 Infarct of cerebrum due to iatrogenic cerebrovascular accident 441526008 SNOMED YES YES NO
4046360 Lacunar infarction 230698000 SNOMED YES YES NO
372435 Periventricular leukomalacia 230769007 SNOMED YES NO NO
377254 Multi-infarct dementia, uncomplicated 70936005 SNOMED YES NO NO
379778 Multi-infarct dementia 56267009 SNOMED YES NO NO
443790 Multi-infarct dementia with delusions 25772007 SNOMED YES NO NO
443864 Multi-infarct dementia with depression 14070001 SNOMED YES NO NO
444091 Multi-infarct dementia with delirium 10349009 SNOMED YES NO NO
4046089 Vascular dementia of acute onset 230285003 SNOMED YES NO NO
4046090 Mixed cortical and subcortical vascular dementia 230287006 SNOMED YES NO NO
4129534 Pituitary apoplexy 237701005 SNOMED YES NO NO

D Data sources

Data source Short name Description
Optum(c) de-identified Electronic Health Record Dataset optum_ehr Optum(c) de-identified Electronic Health Record Dataset is derived from dozens of healthcare provider organizations in the United States (that include more than 700 hospitals and 7,000 Clinics treating more than 103 million patients) receiving care in the United States. The medical record data includes clinical information, inclusive of prescriptions as prescribed and administered, lab results, vital signs, body measurements, diagnoses, procedures, and information derived from clinical Notes using Natural Language Processing (NLP).
Optum(c) de-Identified Clinformatics<ae> Data Mart Database <96> Date of Death optum_dod Optum(c) De-Identified Clinformatics(c) Data Mart Database is an adjudicated administrative health claims database for members with private health insurance, who are fully insured in commercial plans or Medicare Advantage. The population is primarily representative of US commercial claims patients (0-65 years old) with some Medicare (65+ years old) however ages are capped at 90 years. It includes data captured from administrative claims processed from inpatient and outpatient medical services and prescriptions as dispensed, as well as results for outpatient lab tests processed by large national lab vendors who participate in data exchange with Optum. Optum DOD also provides date of death (month and year only) for members with both medical and pharmacy coverage from the Social Security Death Master File (however after 2011 reporting frequency changed due to changes in reporting requirements) and location information for patients is at the US state level.
IBM MarketScan Commercial Claims and Encounters Database truven_ccae IBM MarketScan Commercial Claims and Encounters Database (CCAE) is a US employer-based private-payer administrative claims database. The data include adjudicated health insurance claims (e.g., inpatient, outpatient, and outpatient pharmacy) as well as enrollment data from large employers and health plans who provide private healthcare coverage to employees, their spouses, and dependents. Additionally, it captures laboratory tests for a subset of the covered lives. This administrative claims database includes a variety of fee-for-service, preferred provider organizations, and capitated health plans.
IBM MarketScan MultiState Medicaid Database truven_mdcd IBM MarketScan Multi-State Medicaid Database (MDCD) contains adjudicated US health insurance claims for Medicaid enrollees from multiple states and includes hospital discharge diagnoses, outpatient diagnoses and procedures, and outpatient pharmacy claims as well as ethnicity and Medicare eligibility. Members maintain their same identifier even if they leave the system for a brief period; however, the dataset lacks lab data.
IBM MarketScan Medicare Supplemental and Coordination of Benefits Database truven_mdcr IBM MarketScan Medicare Supplemental and Coordination of Benefits Database (MDCR) represents health services of retirees in the United States with primary or Medicare supplemental coverage through privately insured fee-for-service, point-of-service, or capitated health plans. These data include adjudicated health insurance claims (e.g., inpatient, outpatient, and outpatient pharmacy). Additionally, it captures laboratory tests for a subset of the covered lives.