ReviewCan causality assessment fulfill the new European definition of adverse drug reaction? A review of methods used in spontaneous reporting
Graphical abstract
Introduction
Pharmacovigilance is the science evaluating the safety profile of medicinal products before and after their marketing authorization in order to ensure an appropriate balance of risk and benefit to patients [1]. Among methods of Pharmacovigilance [2], the spontaneous reporting systems (SRS) is the primary method for collecting post-marketing information on the safety of medicines and detecting safety signals. A safety signal is defined by the European Medicine Agency as ‘an information on a new or known adverse event that may be caused by a medicine and requires further investigation’ [3]. Healthcare professionals and patients can through the SRS report suspected adverse drug reactions (ADRs), which are then evaluated by recognized structures as Pharmacovigilance centers [4]. Two approaches are typically used in the signal detection in an SRS: case-by-case analysis of reports and data mining techniques [2]. The importance in detecting safety signals is strongly related to the well-known role of an ADR as a cause of hospital admissions, morbidity, and mortality [5], [6], [7], [8], [9], [10] as well as a public health burden on society. In this regards, the precocious detection of a suspected ADR is a very important procedure to allow an appropriate management of the patient and to avoid the exposure to an additional drug hazard. Therefore, it is fundamental for the signal detection to establish a causal relationship between the drug and the event (causality assessment). Causality assessment is indeed defined as the likelihood that a particular treatment is the cause of an observed adverse event [11]. Moreover, it permits to distinguish between an ADR and an adverse event which does not necessarily have to have a causal relationship with the pharmacological treatment [12]. Today, it is a worldwide routine practice performed by Pharmacovigilance centers in an SRS during the case-by-case evaluation to minimize ambiguity of the data and the implementation of incorrect conclusions on a correlation drug/event [13]. Establishing causality is, in addition, an important and unavoidable practice to detect preventable ADRs that in last years have drawn the attention of Regulatory Authority to achieve an effective risk minimization in a Pharmacovigilance system [14], [15]. In this regards, several local initiatives were performed to evaluate preventability also in an SRS [16], [17]. Furthermore, the new European Pharmacovigilance legislation included in the definition of an adverse event also effects resulting from abuse, misuse and medication error that are all well-known preventable causes of an ADR [18]. Despite this recent increasing trend towards preventability assessment, biomedical methods to assess causality have an elder introduction. Therefore, it is unknown and needs to be scrutinized if these methods can fulfill the new definition of adverse event and then permit a correct preventability assessment. In light of the crucial role of causality in an SRS, we decided to perform a review to provide an overview of causality assessment methods usable by Pharmacovigilance centers and to assess how these methods can fulfill the new definition of an adverse event.
The World Health Organization (WHO) firstly defined an adverse event as ‘any noxious, unintended, and undesired effect of a drug after doses used in humans for prophylaxis, diagnosis, or therapy’. This definition excludes events onset after intentional and accidental poisoning, drug abuse, or relative to therapeutic failures [19]. Subsequently, with the upcoming of the new European Pharmacovigilance legislation with the Directive 2010/84/EU that amended the previous Directive 2001/83/EC, the definition of adverse event was modified to cover ‘noxious and unintended effects resulting not only from the authorized use of a medicinal product at normal doses, but also from medication errors and uses outside the terms of the marketing authorization, including the misuse and abuse of the medicinal product’. Guidelines on Good Pharmacovigilance Practices were published by the European Medicine Agency to accomplish the new definition and for providing operative definitions of adverse events arise after occupational exposure or a use of the medicine outside the terms of marketing authorization (off-label use, overdose, misuse, abuse, and medication errors). Operative definitions reported in the guidelines are provided below [20]:
‘This refers to the exposure to a medicinal product, as a result of one’s professional or non-professional occupation’.
‘This relates to situations where the medicinal product is intentionally used for a medical purpose not in accordance with the authorized product information’.
‘This refers to the administration of a quantity of a medicinal product given per administration or cumulatively, which is above the maximum recommended dose according to the authorized product information. Clinical judgment should always be applied’.
‘This refers to situations where the medicinal product is intentionally and inappropriately used not in accordance with the authorized product information’.
‘This corresponds to the persistent or sporadic, intentional excessive use of a medicinal product, which is accompanied by harmful physical or psychological effects’.
‘This refers to any unintentional error in the prescribing, dispensing, or administration of a medicinal product while in the control of the healthcare professional or consumer’.
Causality assessment lies on criteria previously described by Sir Bradford Hill to establish causation or association between an environment and a disease [3]. These criteria included assessment of consistency of the association, strength of the association, temporal relationship, biological plausibility, biological gradient (dose response), experimental evidence, specificity, coherence, and judgment of analogy. Among them, a crucial role is covered by the temporal relationship and the presence of a biological mechanism able to explain the causal relationship. However, the biological plausibility depends on the current medical knowledge. In fact, its absence is not necessarily indicative of a non-causal association but rather of a new association.
Different methods for causality assessment are available with a large inter-, and intra-variability in their implementation. Today, they are categorized into expert judgment/global introspection, probabilistic methods (Bayesian approaches), and algorithms [21].
Expert judgment/global introspection is a diagnostic process using all available and relevant data to achieve a causal conclusion. Therefore, it bases on the clinical experience and knowledge of the expert for establishing a causal relationship between a drug and an event without excluding limits as subjectivity and lack of standardization that can implicate high variability as well as poor reproducibility [22], [23], [24], [25]. However, these limitations can be overcome when the assessment is made based on a consensus of senior experts [26].
Probabilistic methods are characterized by several advantages [27] as the respect of the basic rule of probability theory that leads to the absence of influences on the causality estimation also when relevant information and arguments are lacking [28]. Moreover, they can provide a precise estimation (as probability or odds) of the causal link between a medicine and an event on a continuous scale. Despite these advantages, this approach is not devoid of trouble in a routine use. In fact, it requires precisely quantified information to model probability distributions for each parameter involved, even if sometimes some assumptions can be made [28].
Algorithms are structured and standardized methods to assess with an operational procedure the identification of an ADR. Generally, they base on parameters such as time to onset of the event or temporal sequence, previous drug/adverse reaction history, other etiological causes, dechallenge, and rechallenge. Typically, they consist in the fulfillment of these parameters presented as a binary or multiple choice that will lead to a final score expressed as a sum of scores or decision tree. Algorithms are useful, easier to use and allows of decreasing inter- or intra-variability and increasing reproducibility. However, sometimes they can require a clinical opinion to arrive at a causal conclusion [29], [30]. Moreover, the achievement of the final conclusion is influenced, for each specific algorithm, by the weight of each criterion that is arbitrarily established by authors of the method [28].
Expert judgment and probabilistic methods are for the aforementioned limitations poorly used in an SRS. On the contrary, algorithms are for their operational procedure and easier applicability one of the most commonly used biomedical methodologies for causality assessment by Pharmacovigilance centers [31]. Therefore, we reported below the fulfillment of the new ADR definition focusing on algorithmic methods.
Section snippets
Methods
Algorithmic methods for causality assessment were identified through searches in Pubmed using the search terms ‘algorithm’, or ‘causality’ in association with the search term ‘adverse drug reaction reporting system’. For the research strategy, we considered also algorithmic methods described in systematic reviews. Only articles published in the English language were considered. All articles deemed eligible for inclusion based on titles and abstracts were read in full. Algorithmic methods were
Focus on algorithmic methods
Twenty-two algorithms were identified comprising their updated versions (Table 1). The first algorithm was developed by Irey in 1976 and was composed of six criteria (exclusion, dechallenge, rechallenge, the singularity of drug, pattern, and quantification of drug level) to link a medicine with an adverse event [32]. The exclusion criterion means that a drug event couple is excluded from causality evaluation whenever an implausible time to event is reported; the dechallenge criterion strengthen
Considerations on algorithmic methods for causality assessment
Different algorithmic methods were published over the years, with the most of them referring to the old definition of ADR and with the intentional purpose of avoiding causality assessment in particular clinical scenario not considered by the terms of marketing authorization [25], [33], [40], [41], [43], [45], [46], [48], [56]. However, some of them allow exceptions in the evaluation of overdose cases as in the Venulet algorithm [45], [46], or highlight the importance of the dosage in the
Conclusion
No algorithm seems perfectly fulfill the new criteria of adverse event definition although some of them come close for the evaluation of abuse/overdose cases. In our opinion, to assess a causal relationship drug/event in an SRS, algorithmic methods should be updated in order to guarantee a correct evaluation of the new definition of adverse event. In fact, an algorithm should always focus on the classic topics as temporal relationship and biological plausibility, which are very important to
Funding
This review did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflicts of interest
The authors state no conflicts of interest.
References (60)
- et al.
The Delphi technique as a forecasting tool: issues and analysis
Int. J. Forecast.
(1999) - et al.
Causality assessment in pharmacovigilance: the French method and its successive updates
Therapie
(2016) - et al.
Validation and reproducibility of the updated french causality assessment method: an evaluation by pharmacovigilance centres & pharmaceutical companies
Therapie
(2015) - et al.
Causality assessment of adverse reactions to drugs–I. A novel method based on the conclusions of international consensus meetings: application to drug-induced liver injuries
J. Clin. Epidemiol.
(1993) - et al.
Development and validation of a clinical scale for the diagnosis of drug-induced hepatitis
Hepatology
(1997) Validity study of a new diagnostic scale for drug-induced liver injury in Japan–comparison with two previous scales
Hepatol. Res.
(2004)- et al.
A quantitative approach of using genetic algorithm in designing a probability scoring system of an adverse drug reaction assessment system
Int. J. Med. Inform.
(2008) - et al.
Risk of gastrointestinal complications associated to NSAIDs, low-dose aspirin and their combinations: results of a pharmacovigilance reporting system
Pharmacol. Res.
(2016) - et al.
Pharmacovigilance in Europe
Transplantation
(2015) - et al.
Pharmacovigilance: methods, recent developments and future perspectives
Eur. J. Clin. Pharmacol.
(2008)
Guideline on Good Pharmacovigilance Practices (GVP) Module IX ?Signal Management
Improvement of patient adverse drug reaction reporting through a community pharmacist-based intervention in the Campania region of Italy
Expert Opin. Drug Saf.
Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies
JAMA
Reporting of fatal adverse drug reactions
Can. J. Clin. Pharmacol.
Drug-related emergency department visits and hospital admissions
Am. J. Hosp. Pharm.
Admissions caused by adverse drug events to internal medicine and emergency departments in hospitals: a longitudinal population-based study
Eur. J. Clin. Pharmacol.
Dipeptidyl peptidase (DPP)-4 inhibitor-induced arthritis/arthralgia: a review of clinical cases
Drug Saf.
New era in treatment options of chronic hepatitis C: focus on safety of new direct-acting antivirals (DAAs)
Expert Opin. Drug Saf.
The Use of the WHO-UMC System for Standardised Case Causality Assessment Why Causality Assessment?
ICH Harmonised Tripartite Guideline Clinical Safety Data Management: Definitions and Standards For Expedited Reporting E2A Clinical Safety Data Management: Definitions and Standards for Expedited Reporting
Causal or casual? The role of causality assessment in pharmacovigilance
Drug Saf.
The monitoring medicines project: a multinational pharmacovigilance and public health project
Drug Saf.
Corporate Induction: Part 2 Identification of Preventable Adverse Drug Reactions from a Regulatory Perspective EMA Workshop on Medication Errors
Campania Region (Italy) spontaneous reporting system and preventability assessment through a case-by-case approach: a pilot study on psychotropic drugs
Expert Opin. Drug Saf.
Campania preventability assessment committee: a focus on the preventability of the contrast media adverse drug reactions
Expert Opin. Drug Saf.
European Parliament and of the Council, Directive 2010/84/EU of the European Parliament And Of The Council
Meeting on the Role of the Hospital in International Drug Monitoring (1968: Geneva, W.H., Organization, International drug monitoring: the role of the hospital, report of a WHO meeting [held in Geneva from 18 to 23 November 1968]
Guideline on Good Pharmacovigilance Practices (GVP) Module VI –Management and Reporting of Adverse Reactions to Medicinal Products (Rev 1)
Methods for causality assessment of adverse drug reactions: a systematic review
Drug Saf.
Difficulties in assessing the adverse effects of drugs
Br. J. Clin. Pharmacol.
Cited by (41)
Mirtazapine-associated acute pancreatitis in a patient with insomnia and co-occurring psychiatric disorders
2022, Journal of the National Medical AssociationCitation Excerpt :At a six-month follow-up, the patient did not report any recurrence of pancreatitis. The Naranjo Adverse Drug Reaction Probability Scale total score was 5, which indicated a probable ADR.14 Our case is noteworthy for the complexity of the patient's presentation.
Toward Kidney-Specific Causality Assessment Tool
2022, Clinical TherapeuticsCitation Excerpt :However, none perfectly fulfill the criteria of the ADR definition. In the study by Mascolo et al58 evaluating 22 causality assessment algorithms, parallel findings have been highlighted: the authors write that the algorithmic methods should be updated to guarantee a correct evaluation of the definition of AE for assessing a causal relationship to a drug/event. The results obtained from this screening have been used to suggest valuable parameters for a KSCAT (Table II).
Pharmacovigilance – The next chapter
2019, TherapieCitation Excerpt :Once these elements are gathered, any number of causality assessment methods, also called imputation or imputability can be applied. Over the years, several dozen methods to assess causality in individual case reports have been developed, and more continue to be invented regularly [15–37]. Different causality methods may give different results, probably because of different objectives [21,24,25,30].
Time series causal relationships discovery through feature importance and ensemble models
2023, Scientific ReportsEstimating Culprit Drugs for Adverse Drug Reactions Based on Bayesian Inference
2023, Clinical Pharmacology and Therapeutics
- 1
Annamaria Mascolo and Cristina Scavone serve as co-first authors for this manuscript and that Liberata Sportiello and Annalisa Capuano serve as co-lead authors.