Adverse reaction signal detection methodology in pharmacoepidemiology

Adverse reaction signal detection methodology in pharmacoepidemiology European Journal of Epidemiology (2018) 33:507–508 https://doi.org/10.1007/s10654-018-0417-5(0123456789().,-volV)(0123456789().,-volV) COMMENTARY Adverse reaction signal detection methodology in pharmacoepidemiology Bruno H. Stricker Received: 26 May 2018 / Accepted: 28 May 2018 / Published online: 4 June 2018 The Author(s) 2018 In the nineteenth century, the toxicity of chloroform led to expressed as a reporting odds ratio with 95% confidence its withdrawal from clinical use [1] and in the period limits was first proposed in 1992 [9], and as a proportional 1920–1940, hepatic injury by cinchophen [2] and agranu- reporting ratio in 2001 [10]. Of these two effect measures, locytosis by amidopyrine and related agents [3] were rec- the reporting odds ratio has certain advantages [11]. These ognized. But from the point of view of detection of measures are now extensively used by the pharmaceutical important unknown adverse reactions, the thalidomide industry as one of the tools of signal detection, in line with disaster with its thousands of fatal and non-fatal cases of European guidelines [12]. But up till recently, the large congenital malformations was an absolute hallmark [4]. As majority of pharmaceutical marketing authorization hold- a direct consequence, it was made mandatory in the early ers only check their own database which is limited to those sixties of the preceding century to perform extensive tox- drugs which are marketed by that particular company. Only icological, preclinical, and clinical studies before market- some of them also use the WHO Vigibase or the FDA ing of a drug in Western countries, and national Adverse Event Reporting System and since 2018, the spontaneous monitoring systems were set up. These sys- Europeans Medicine Agency’s database Eudravigilance tems in concert with the medical literature, proved to be the can be used. most effective and efficient system for recognizing new A new development in signal detection is to use not only adverse reactions since then [5]. In the years thereafter, adverse reaction reports but complete medical records several drugs were recognized as the cause of serious healthcare databases for this goal. Elsewhere in this jour- disease, such as chronic active hepatitis by oxyphenisatin nal, Hallas et al. [13] describe how a hypothesis-free [6], sclerosing peritonitis by practolol [7] and many more screening of large administrative databases can be used for since then. Such monitoring consists of manual review of recognition of new drug-outcome associations. This is one adverse reaction reports by medical professionals and is of the examples of how the strong increase in computeri- relatively cheap and flexible but suffers from substantial zation in the past decades and the consequent growth of underreporting, potential false-positive reporting and automated healthcare data can be employed to this end. In absence of reliable usage figures. Also, case-by-case current initiatives such as EU-ADR [14] and the Obser- assessments may lead to a loss of overview when large vational health data science and informatics (https://ohdsi. numbers of reports are involved and rests heavily on the org/), networks of administrative databases were built to quality of the professional. identify drug safety issues by data mining, mainly through In 1974, in an attempt to improve adverse reaction a self-controlled design covering data from up to many signal detection, Finney proposed to compare the propor- millions of people. The question whether we should be tion of reports of a certain drug-event association with the happy with such a development is completely irrelevant. In proportion of reports of that event to all other drugs in the human history, any technical development than can be used database and test for significance in a 2 9 2 table [8]. A will be used. And data mining has proven very successful significantly higher proportion comprised a signal. A fur- in genetic research. Genome-wide association studies ther extension of this principle with the magnitude (GWAs) by consortia of population-based cohort studies such as CHARGE were very rewarding in finding new associations between genetic variants and disease [15]. & Bruno H. Stricker Especially in Western countries the combination of risk b.stricker@erasmusmc.nl aversion and legislation is an enormous enforcer to employ Department of Epidemiology, Erasmus Medical Center, such healthcare information for safety research and as long P.O. Box 2040, 3000 CA Rotterdam, The Netherlands 123 508 B. H. Stricker appropriate credit to the original author(s) and the source, provide a as the privacy of patients is guaranteed, there little against link to the Creative Commons license, and indicate if changes were using it. But the consequences are that the number of false- made. positive signals that will have to be tested increases enor- mously. This requires a rigorous process of signal priori- tisation and testing as the number of epidemiological References resources is not endless. Apart from the subject itself, there are a number of important differences between data mining 1. Wade OL, Beeley L. The dawn of concern. In: Wade OL, Beeley L, editors. Adverse reactions to drugs, vol. 1. 2nd ed. London: in genetic epidemiology and in pharmacoepidemiology. William Heinemann Medical Books Ltd; 1976. First, in genome-wide association studies Bonferroni cor- 2. Worster-Draught C. Atophan poisoning. Br Med J. 1923;1:148. rections are used. There are many good arguments against 3. Kracke RR, Parker FP. The etiology of granulocytopenia using Bonferroni corrections whatsoever [16] but in GWAs (agranulocytosis). With particular reference to the drugs con- taining the benzene ring. J Lab Clin Med. 1934;19:799. they are the only workable solutions as using a p value of 4. Taussig HB. A study of the German outbreak of phocomelia. 0.05 as a cut-off would be very impractical in view of the JAMA. 1962;180:1106. abundance of associations when studying millions of single 5. Stricker BH, Psaty BM. Detection, verification, and quantification nucleotide polymorphisms. In data mining with healthcare of adverse drug reactions. Br Med J. 2004;329:44–7. 6. Reynolds TB, Peters RL, Yamada S. Chronic active and lupoid databases the number of associations that can be tested is hepatitis caused by a laxative, oxyphenisatin. N Eng J Med. smaller and Bonferroni corrections are less commonly 1971;285:813. used, maybe also because of a fear of litigations for drug 7. Brown P, Baddeley H, Read AE, et al. Sclerosing peritonitis, an marketing authorization holders for missing associations. unusual reaction to b-adrenergic blocking drug (Practolol). Lan- cet. 1974;2:1477. Second, GWAs in consortia often work with identical 8. Finney DJ. Systemic signaling of adverse reactions to drugs. platforms. Healthcare databases are, however, very Methods Inf Med. 1974;13:1–10. heterogeneous. Not only do they vary between countries 9. Stricker BH, Tijssen JG. Serum sickness-like reactions to cefa- and healthcare systems but also over time changes in clor. J Clin Epidemiol. 1992;45:1177–84. 10. Evans SJW, Waller PC, Davis S. Use of proportional reporting insurance system and disease coding may complicate ratios for signal generation from spontaneous adverse drug consistent analyses. Moreover, hospital-based and general reaction reports. Pharmacoepidemiol Drug Saf. 2001;10:483–6. practitioner’s healthcare information is structured in a 11. Rothman KJ, Lanes S, Sacks ST. The reporting odds ratio and its different way, and mapping towards one analysable dataset advantages over the proportional reporting ratio. Pharmacoepi- demiol Drug Saf. 2004;13:519–23. is a cumbersome challenge which has to be repeated again 12. EMA. Guideline on good pharmacovigilance practices, module and again. Third, and maybe most important, genetic IX—signal management; 2017. GWAs are driven by scientific interest, rather than for 13. Hallas J, Wang SV, Gagne JJ, Schneeweiss S, Pratt N, Pottega ˚rd fulfilling legal obligations. In how far this leads to better A. Hypothesis-free screening of large administrative databases for unsuspected drug outcome associations. Eur J Epidemiol. science remains to be seen. But one conclusion, we can 2018;1:1 (in press). make already now. If we do not improve our ability to 14. Coloma PM, Schuemie MJ, Trifiro ` G, Gini R, Herings R, Hip- distinguish true-positive from false-positive signals in an pisley-Cox J, et al. Combining electronic healthcare databases in efficient way, we might waste epidemiologic resources for Europe to allow for large-scale drug safety monitoring: the EU- ADR Project. Pharmacoepidemiol Drug Saf. 2011;20:1–11. extensive signal-testing as a consequence of our increas- 15. Psaty BM, Hofman A. Genome-wide association studies and ingly demanding society. large-scale collaborations in epidemiology. Eur J Epidemiol 2010;25:525–9. Open Access This article is distributed under the terms of the Creative 16. Greenland S, Rothman KJ. Fundamentals in epidemiologic data Commons Attribution 4.0 International License (http://creative analysis. Modern epidemiology. 2nd ed. Philadelphia: Lippincott- commons.org/licenses/by/4.0/), which permits unrestricted use, dis- Raven; 1998. p. 201–29. tribution, and reproduction in any medium, provided you give http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Journal of Epidemiology Springer Journals

Adverse reaction signal detection methodology in pharmacoepidemiology

Free
2 pages
Loading next page...
 
/lp/springer_journal/adverse-reaction-signal-detection-methodology-in-pharmacoepidemiology-BfijeUSShj
Publisher
Springer Netherlands
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Epidemiology; Public Health; Infectious Diseases; Cardiology; Oncology
ISSN
0393-2990
eISSN
1573-7284
D.O.I.
10.1007/s10654-018-0417-5
Publisher site
See Article on Publisher Site

Abstract

European Journal of Epidemiology (2018) 33:507–508 https://doi.org/10.1007/s10654-018-0417-5(0123456789().,-volV)(0123456789().,-volV) COMMENTARY Adverse reaction signal detection methodology in pharmacoepidemiology Bruno H. Stricker Received: 26 May 2018 / Accepted: 28 May 2018 / Published online: 4 June 2018 The Author(s) 2018 In the nineteenth century, the toxicity of chloroform led to expressed as a reporting odds ratio with 95% confidence its withdrawal from clinical use [1] and in the period limits was first proposed in 1992 [9], and as a proportional 1920–1940, hepatic injury by cinchophen [2] and agranu- reporting ratio in 2001 [10]. Of these two effect measures, locytosis by amidopyrine and related agents [3] were rec- the reporting odds ratio has certain advantages [11]. These ognized. But from the point of view of detection of measures are now extensively used by the pharmaceutical important unknown adverse reactions, the thalidomide industry as one of the tools of signal detection, in line with disaster with its thousands of fatal and non-fatal cases of European guidelines [12]. But up till recently, the large congenital malformations was an absolute hallmark [4]. As majority of pharmaceutical marketing authorization hold- a direct consequence, it was made mandatory in the early ers only check their own database which is limited to those sixties of the preceding century to perform extensive tox- drugs which are marketed by that particular company. Only icological, preclinical, and clinical studies before market- some of them also use the WHO Vigibase or the FDA ing of a drug in Western countries, and national Adverse Event Reporting System and since 2018, the spontaneous monitoring systems were set up. These sys- Europeans Medicine Agency’s database Eudravigilance tems in concert with the medical literature, proved to be the can be used. most effective and efficient system for recognizing new A new development in signal detection is to use not only adverse reactions since then [5]. In the years thereafter, adverse reaction reports but complete medical records several drugs were recognized as the cause of serious healthcare databases for this goal. Elsewhere in this jour- disease, such as chronic active hepatitis by oxyphenisatin nal, Hallas et al. [13] describe how a hypothesis-free [6], sclerosing peritonitis by practolol [7] and many more screening of large administrative databases can be used for since then. Such monitoring consists of manual review of recognition of new drug-outcome associations. This is one adverse reaction reports by medical professionals and is of the examples of how the strong increase in computeri- relatively cheap and flexible but suffers from substantial zation in the past decades and the consequent growth of underreporting, potential false-positive reporting and automated healthcare data can be employed to this end. In absence of reliable usage figures. Also, case-by-case current initiatives such as EU-ADR [14] and the Obser- assessments may lead to a loss of overview when large vational health data science and informatics (https://ohdsi. numbers of reports are involved and rests heavily on the org/), networks of administrative databases were built to quality of the professional. identify drug safety issues by data mining, mainly through In 1974, in an attempt to improve adverse reaction a self-controlled design covering data from up to many signal detection, Finney proposed to compare the propor- millions of people. The question whether we should be tion of reports of a certain drug-event association with the happy with such a development is completely irrelevant. In proportion of reports of that event to all other drugs in the human history, any technical development than can be used database and test for significance in a 2 9 2 table [8]. A will be used. And data mining has proven very successful significantly higher proportion comprised a signal. A fur- in genetic research. Genome-wide association studies ther extension of this principle with the magnitude (GWAs) by consortia of population-based cohort studies such as CHARGE were very rewarding in finding new associations between genetic variants and disease [15]. & Bruno H. Stricker Especially in Western countries the combination of risk b.stricker@erasmusmc.nl aversion and legislation is an enormous enforcer to employ Department of Epidemiology, Erasmus Medical Center, such healthcare information for safety research and as long P.O. Box 2040, 3000 CA Rotterdam, The Netherlands 123 508 B. H. Stricker appropriate credit to the original author(s) and the source, provide a as the privacy of patients is guaranteed, there little against link to the Creative Commons license, and indicate if changes were using it. But the consequences are that the number of false- made. positive signals that will have to be tested increases enor- mously. This requires a rigorous process of signal priori- tisation and testing as the number of epidemiological References resources is not endless. Apart from the subject itself, there are a number of important differences between data mining 1. Wade OL, Beeley L. The dawn of concern. In: Wade OL, Beeley L, editors. Adverse reactions to drugs, vol. 1. 2nd ed. London: in genetic epidemiology and in pharmacoepidemiology. William Heinemann Medical Books Ltd; 1976. First, in genome-wide association studies Bonferroni cor- 2. Worster-Draught C. Atophan poisoning. Br Med J. 1923;1:148. rections are used. There are many good arguments against 3. Kracke RR, Parker FP. The etiology of granulocytopenia using Bonferroni corrections whatsoever [16] but in GWAs (agranulocytosis). With particular reference to the drugs con- taining the benzene ring. J Lab Clin Med. 1934;19:799. they are the only workable solutions as using a p value of 4. Taussig HB. A study of the German outbreak of phocomelia. 0.05 as a cut-off would be very impractical in view of the JAMA. 1962;180:1106. abundance of associations when studying millions of single 5. Stricker BH, Psaty BM. Detection, verification, and quantification nucleotide polymorphisms. In data mining with healthcare of adverse drug reactions. Br Med J. 2004;329:44–7. 6. Reynolds TB, Peters RL, Yamada S. Chronic active and lupoid databases the number of associations that can be tested is hepatitis caused by a laxative, oxyphenisatin. N Eng J Med. smaller and Bonferroni corrections are less commonly 1971;285:813. used, maybe also because of a fear of litigations for drug 7. Brown P, Baddeley H, Read AE, et al. Sclerosing peritonitis, an marketing authorization holders for missing associations. unusual reaction to b-adrenergic blocking drug (Practolol). Lan- cet. 1974;2:1477. Second, GWAs in consortia often work with identical 8. Finney DJ. Systemic signaling of adverse reactions to drugs. platforms. Healthcare databases are, however, very Methods Inf Med. 1974;13:1–10. heterogeneous. Not only do they vary between countries 9. Stricker BH, Tijssen JG. Serum sickness-like reactions to cefa- and healthcare systems but also over time changes in clor. J Clin Epidemiol. 1992;45:1177–84. 10. Evans SJW, Waller PC, Davis S. Use of proportional reporting insurance system and disease coding may complicate ratios for signal generation from spontaneous adverse drug consistent analyses. Moreover, hospital-based and general reaction reports. Pharmacoepidemiol Drug Saf. 2001;10:483–6. practitioner’s healthcare information is structured in a 11. Rothman KJ, Lanes S, Sacks ST. The reporting odds ratio and its different way, and mapping towards one analysable dataset advantages over the proportional reporting ratio. Pharmacoepi- demiol Drug Saf. 2004;13:519–23. is a cumbersome challenge which has to be repeated again 12. EMA. Guideline on good pharmacovigilance practices, module and again. Third, and maybe most important, genetic IX—signal management; 2017. GWAs are driven by scientific interest, rather than for 13. Hallas J, Wang SV, Gagne JJ, Schneeweiss S, Pratt N, Pottega ˚rd fulfilling legal obligations. In how far this leads to better A. Hypothesis-free screening of large administrative databases for unsuspected drug outcome associations. Eur J Epidemiol. science remains to be seen. But one conclusion, we can 2018;1:1 (in press). make already now. If we do not improve our ability to 14. Coloma PM, Schuemie MJ, Trifiro ` G, Gini R, Herings R, Hip- distinguish true-positive from false-positive signals in an pisley-Cox J, et al. Combining electronic healthcare databases in efficient way, we might waste epidemiologic resources for Europe to allow for large-scale drug safety monitoring: the EU- ADR Project. Pharmacoepidemiol Drug Saf. 2011;20:1–11. extensive signal-testing as a consequence of our increas- 15. Psaty BM, Hofman A. Genome-wide association studies and ingly demanding society. large-scale collaborations in epidemiology. Eur J Epidemiol 2010;25:525–9. Open Access This article is distributed under the terms of the Creative 16. Greenland S, Rothman KJ. Fundamentals in epidemiologic data Commons Attribution 4.0 International License (http://creative analysis. Modern epidemiology. 2nd ed. Philadelphia: Lippincott- commons.org/licenses/by/4.0/), which permits unrestricted use, dis- Raven; 1998. p. 201–29. tribution, and reproduction in any medium, provided you give

Journal

European Journal of EpidemiologySpringer Journals

Published: Jun 4, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off