Pharmacovigilance

In short

Pharmacovigilance (PV) is the science and set of activities concerned with the detection, assessment, understanding and prevention of adverse effects or any other drug-related problem (WHO definition). It exists because pre-marketing trials involve a few thousand patients for a few months, so rare, delayed and interaction-related harms only appear once millions of people are exposed — thalidomide, rofecoxib and the Indian cough-syrup diethylene glycol episodes are the standing examples. In India it is run as the Pharmacovigilance Programme of India (PvPI), with the Indian Pharmacopoeia Commission (IPC), Ghaziabad as National Coordination Centre and CDSCO as the regulator; marketing-authorisation holders carry statutory duties under the New Drugs and Clinical Trials Rules, 2019. It is examined in B.Pharm/Pharm.D papers on pharmacovigilance and pharmacy practice, in GPAT and in NIPER entrance, and is one of the largest graduate recruiters in Indian pharma — case processing, aggregate reporting and medical review roles in Accenture, Cognizant, TCS, Tata Elxsi, IQVIA, Syneos, Parexel, Sun, Cipla, Dr Reddy's and Lupin.

Explained

Definitions and the vocabulary that gets marked

• Adverse event (AE) / adverse experience: any untoward medical occurrence in a patient given a medicine, whether or not related to it. No causality implied.
• Adverse drug reaction (ADR): a response to a medicine which is noxious and unintended and occurs at doses normally used in humans (WHO). Causality is at least suspected. Overdose, abuse and medication error outcomes fall outside the classic WHO definition but are collected anyway.
• Side effect: an unintended effect at normal dose related to the drug's pharmacology — may be harmful, harmless or even useful.
• Adverse drug event (ADE): any injury from a medicine, including from errors — wider than ADR.
• Serious (not "severe"): an event that results in death, is life-threatening, requires hospitalisation or prolongs it, causes persistent or significant disability/incapacity, a congenital anomaly, or is a medically important event requiring intervention. Severity is intensity (mild/moderate/severe) — a severe headache may be non-serious; a brief non-severe arrhythmia may be serious.
• Unexpected/unlisted: nature or severity not consistent with the reference safety information (Investigator's Brochure for investigational drugs, package insert/SmPC for marketed ones).
• Signal: reported information on a possible causal relationship between an event and a drug, previously unknown or incompletely documented, usually from more than one report.
• ICSR: Individual Case Safety Report — the unit of PV data. Valid case needs four elements: an identifiable patient, an identifiable reporter, a suspect drug, and an adverse event/outcome.
• Terms: ADE, AE, ADR, SAE, SUSAR (Suspected Unexpected Serious Adverse Reaction — a trial concept), benefit-risk balance, risk management plan (RMP).

Classification of ADRs

• Rawlins–Thompson (the one asked most):
  • Type A — Augmented: dose-dependent, predictable from pharmacology, common, low mortality. Bleeding with warfarin, hypoglycaemia with insulin, bradycardia with atenolol. Managed by dose reduction.
  • Type B — Bizarre: dose-independent, unpredictable, rare, high mortality; immunological or idiosyncratic. Penicillin anaphylaxis, chloramphenicol aplastic anaemia, halothane hepatitis. Drug must be stopped.
  • Type C — Chronic/Continuous: related to cumulative dose and duration. Analgesic nephropathy, osteonecrosis with bisphosphonates, HPA suppression with steroids.
  • Type D — Delayed: appear long after exposure. Carcinogenesis, teratogenesis (thalidomide phocomelia), tardive dyskinesia.
  • Type E — End-of-use/withdrawal: rebound on stopping. Beta-blocker rebound angina, benzodiazepine withdrawal seizures, steroid withdrawal crisis.
  • Type F — Failure of therapy: often dose-related or interaction-driven. Oral contraceptive failure with rifampicin or phenytoin.
• Other schemes: DoTS (Dose-relatedness, Time-course, Susceptibility); EIDOS (mechanistic); and simple splits — predictable vs unpredictable, immunological vs non-immunological.
• Mechanistic categories: pharmacokinetic (enzyme induction/inhibition, renal impairment), pharmacodynamic (additive effects), immune-mediated (Types I–IV hypersensitivity), genetic/pharmacogenomic (G6PD deficiency and primaquine haemolysis, HLA-B*15:02 and carbamazepine SJS, slow NAT2 acetylators and isoniazid neuropathy).

Methods of pharmacovigilance

• Spontaneous (voluntary) reporting: the backbone — cheap, covers all drugs and the whole population for the whole life-cycle, and is the only practical way to pick up rare Type B reactions. Limitation: under-reporting (commonly estimated at over 90% of reactions worldwide) and no denominator, so incidence cannot be calculated.
• Stimulated reporting: targeted prompts in a campaign or during a new launch; raises numbers but distorts trend comparisons.
• Intensive monitoring / Prescription Event Monitoring (PEM): patients on a new drug are followed systematically (green-form PEM in the UK); gives a denominator but is costly and drug-specific.
• Cohort event monitoring (CEM): used by WHO for antimalarials and ART programmes; India has used CEM-type designs for public-health programme drugs.
• Record linkage / database studies: electronic health records and claims data; feasible in India mainly through hospital information systems and insurance data (Ayushman Bharat claims).
• Case–control studies: efficient for rare events with a strong suspicion — the classic design that confirmed an association after a signal.
• Cohort studies and registries: pregnancy exposure registries, biologic registries.
• Meta-analysis and RCT safety pooling: best evidence but limited to studied populations.
• Comparison, in one line: spontaneous reporting generates hypotheses (signals); formal epidemiology tests them and quantifies risk.

Causality assessment

• WHO-UMC categories: Certain; Probable/Likely; Possible; Unlikely; Conditional/Unclassified; Unassessable/Unclassifiable. Based on time relationship, plausibility, other explanations, dechallenge and rechallenge.
• Naranjo ADR Probability Scale: 10 questions, score −4 to +13 — ≥9 Definite, 5–8 Probable, 1–4 Possible, ≤0 Doubtful.
• Naranjo's questions cover: previous conclusive reports; timing after the drug; improvement on dechallenge; recurrence on rechallenge; alternative causes; reaction on placebo; toxic blood concentrations; dose–response; previous similar reaction in the same patient; objective confirmation.
• Other algorithms: WHO-UMC (most used in PvPI), Naranjo, Karch–Lasagna, Hartwig & Siegel severity scale (Levels 1–7), Schumock & Thornton preventability scale (definitely/probably/not preventable).
• Dechallenge = reaction abates when the drug is stopped; rechallenge = reaction returns when restarted — strongest evidence but usually unethical to perform deliberately.

The Indian system: PvPI, CDSCO and the law

• PvPI was launched in July 2010 and has been coordinated by IPC, Ghaziabad as National Coordination Centre (NCC) since 2011 (initially under CDSCO/AIIMS). It is a WHO Collaborating Centre for Pharmacovigilance in public health programmes and pharmacovigilance.
• Structure: NCC-PvPI → ADR Monitoring Centres (AMCs) in medical colleges, corporate hospitals and public-health programme sites (several hundred nationally; the number grows, so quote it as "a few hundred AMCs" unless you have the current figure) → reporters (doctors, pharmacists, nurses, dentists, consumers).
• Flow: reporter fills the Suspected ADR Reporting Form (red form) or the Consumer ADR form → AMC does data entry, completeness and causality review → uploaded to VigiFlow → NCC-PvPI signal review panel → CDSCO/DCGI regulatory action → data shared with WHO-UMC's VigiBase in Uppsala, Sweden. India joined the WHO Programme for International Drug Monitoring in 1998.
• Reporting routes: the ADR form (download, post, email to the NCC mailbox), the national toll-free helpline 1800-180-3024, the ADR PvPI mobile app, and the online e-reporting portal. Reporting by healthcare professionals is voluntary; reporting by the industry is mandatory.
• New Drugs and Clinical Trials Rules, 2019 (under the Drugs and Cosmetics Act, 1940) — the statutory PV framework:
  • Every applicant/MAH must run a PV system with a qualified person; the Pharmacovigilance System Master File (PSMF) that documents it is required by the CDSCO PV guidance for MAHs (below) rather than by the text of the Rules.
  • Serious adverse events in clinical trials: report to the licensing authority, ethics committee and sponsor within 24 hours; the sponsor/investigator analysis follows within the further period specified in the Rules; compensation for trial-related injury or death is mandatory, and for death the quantum is decided per the formula notified by the DCGI.
  • Post-marketing: PSUR every six months for the first two years after approval of a new drug, then annually for the next two years, and thereafter as directed by the licensing authority.
  • Expedited reporting of serious unexpected ADRs from the Indian market applies on the 15-day calendar standard internationally (ICH E2A/E2D); confirm the exact Indian timeline against the current CDSCO PV Guidance before quoting it in a regulatory submission, as guidance has been revised.
• CDSCO Guidance Document on Pharmacovigilance Requirements for MAHs (first issued 2018) sets out the PV system, PSMF, qualified person, ICSR submission, PSUR/PBRER and RMP expectations.
• Schedule M (revised, GMP) and Schedule Y (the older trial schedule largely superseded for new drugs by the 2019 Rules) are the neighbouring instruments; Rule 122A/B/D-era numbering belongs to the pre-2019 regime.
• Related Indian vigilance programmes, all coordinated by IPC: Materiovigilance Programme of India (MvPI, 2015) for medical devices; Haemovigilance Programme (HvPI, 2012) for blood and blood products; AEFI (Adverse Events Following Immunisation) surveillance under the Ministry of Health for vaccines; Pharmacovigilance for AYUSH (2018) under the Ministry of Ayush with an All India Institute of Ayurveda node.

Industry pharmacovigilance: the case life-cycle

1. Case receipt — spontaneous reports, literature, solicited sources (studies, patient support programmes), regulators, licence partners, social media/digital.
2. Triage — validity check (four elements), seriousness, expectedness, due date clock starts at Day 0 = first receipt by any employee of the company.
3. Data entry — into a safety database (Argus Safety, ARISg, Veeva Vault Safety, LSMV); coding with MedDRA (events — SOC/HLGT/HLT/PT/LLT hierarchy) and WHO Drug Dictionary (medicines).
4. Medical review and causality — company assessment plus reporter assessment; narrative writing; labelling assessment against the reference safety information.
5. Quality review and submission — in E2B(R3) XML via gateways to regulators; ICH E2B is the electronic transmission standard.
6. Follow-up — chase missing data for serious and unexpected cases.
7. Aggregate reporting — PSUR/PBRER (ICH E2C(R2)), DSUR (ICH E2F) for drugs under development, addendum reports, ad hoc line listings.
8. Signal management (ICH E2E / GVP Module IX) — detection (data mining, literature), validation, confirmation, analysis, prioritisation, action.
9. Risk management — RMP with the safety specification, PV plan and risk minimisation measures: routine (label changes, pack size, Dear Healthcare Professional letters) and additional (educational material, controlled access, pregnancy prevention programmes).
10. Audit and inspection readiness — PSMF, SOPs, CAPA, training records; inspections by CDSCO, EMA and the US FDA.
11. Timelines commonly examined (ICH/EU/US practice): serious unexpected (expedited) ICSRs 15 calendar days; non-serious 90 calendar days where required; US FDA 15-day alert reports; trial SAEs to Indian authorities 24 hours.
12. Key ICH guidelines: E2A (clinical safety definitions and standards for expedited reporting), E2B (electronic ICSR transmission), E2C (PBRER), E2D (post-approval safety data), E2E (pharmacovigilance planning), E2F (DSUR), E6(R2/R3) GCP, M1 MedDRA.

Signal detection, disproportionality and benefit–risk

• A spontaneous report database has no denominator, so risk is judged by comparing a drug–event pair's share of reports with all other drug–event pairs — disproportionality analysis.
• Measures: PRR (proportional reporting ratio), ROR (reporting odds ratio), IC/IC025 (Bayesian information component used by WHO-UMC), EBGM/EB05 (empirical Bayes, used by the US FDA), chi-square support.
• A common PRR signal threshold used in practice: PRR ≥ 2 with at least 3 cases and chi-square ≥ 4. Thresholds are conventions, not law.
• Regulatory actions available after a confirmed signal: label/package insert change with a boxed or special warning, restriction of indication or population, dose change, prescriber/pharmacy restrictions, Dear Doctor letter, batch recall, suspension, and withdrawal or ban under section 26A of the Drugs and Cosmetics Act, 1940 — the power the Government of India used for fixed-dose combination bans and for drugs such as rofecoxib, cisapride and phenylpropanolamine.
• Limitations of PV data: under-reporting, reporting bias (new drugs, media attention — the Weber effect: reports peak in the second year after launch then decline), notoriety bias, duplicate reports, incomplete narratives, no incidence estimation, confounding by indication.

Key points

• Pharmacovigilance — detection, assessment, understanding and prevention of adverse effects and other drug-related problems (WHO).
• AE vs ADR — AE needs no causal link; ADR implies suspected causality at normal therapeutic doses.
• Serious vs severe — serious is a regulatory outcome list (death, life-threatening, hospitalisation, disability, congenital anomaly, medically important); severe is only intensity.
• Valid ICSR — identifiable patient + identifiable reporter + suspect drug + adverse event; any one missing makes it invalid.
• Rawlins–Thompson types — A augmented, B bizarre, C chronic, D delayed, E end-of-use, F failure of therapy.
• Type A vs Type B — A is dose-dependent, predictable, common, treat by dose reduction; B is dose-independent, unpredictable, rare, high mortality, stop the drug.
• Spontaneous reporting — cheapest and widest method, detects rare reactions, but under-reports heavily and gives no denominator.
• WHO-UMC scale — Certain, Probable, Possible, Unlikely, Conditional, Unassessable; PvPI's working tool.
• Naranjo score — ≥9 definite, 5–8 probable, 1–4 possible, ≤0 doubtful (range −4 to +13).
• PvPI — launched July 2010; NCC at IPC, Ghaziabad; AMCs feed VigiFlow, which feeds WHO's VigiBase (Uppsala); India in the WHO programme since 1998.
• Mandatory vs voluntary — industry reporting is a legal duty; healthcare-professional reporting in India is voluntary.
• New Drugs and Clinical Trials Rules, 2019 — PV system (PSMF per CDSCO guidance), trial SAE in 24 hours, PSUR 6-monthly for 2 years, then yearly for 2 years, then as directed.
• Expedited ICSR — serious unexpected cases on the 15-calendar-day clock from Day 0 (first company receipt).
• MedDRA and WHO-DD — standard dictionaries for coding events and drugs; E2B(R3) is the electronic submission format.
• Disproportionality — PRR, ROR, IC/IC025, EBGM; a frequently used cut-off is PRR ≥ 2, cases ≥ 3, χ² ≥ 4.
• Section 26A, Drugs and Cosmetics Act, 1940 — the legal power to prohibit or restrict a drug in the public interest.
• Sister programmes — MvPI (devices, 2015), HvPI (blood, 2012), AEFI (vaccines), Ayush PV (2018), all with IPC/Ministry nodes.
• Weber effect — ADR reporting for a new drug peaks around the second year of marketing, then falls.

Formulas and examples

Example 1 — Naranjo scoring. A 42-year-old on ciprofloxacin develops a photosensitive rash on day 3; the rash settles within 4 days of stopping; no rechallenge; no other new drug or sun-exposure change; the reaction is well documented in the literature; no drug levels measured.

• Previous conclusive reports: +1; event appeared after the drug: +2; improved on dechallenge: +1; rechallenge: 0 (not done); alternative causes: +2 (none apparent); placebo: 0; toxic levels: 0; dose–response: 0; previous similar reaction: 0; objective confirmation (rash seen and documented by the physician): +1.
• Total = 1 + 2 + 1 + 2 + 1 = +7 → Probable ADR.

Example 2 — PRR, ROR and χ². A safety database holds 200,000 reports. For Drug D: 40 reports of acute pancreatitis and 1,960 reports of other events. For all other drugs: 2,000 reports of pancreatitis and 196,000 other events.

• a = 40, b = 1,960, c = 2,000, d = 196,000; N = 200,000.
• PRR = (40/2,000) / (2,000/198,000) = 0.0200 / 0.010101 = 1.98.
• ROR = ad/bc = (40 × 196,000)/(1,960 × 2,000) = 7,840,000/3,920,000 = 2.00.
• SE of ln(ROR) = √(1/40 + 1/1,960 + 1/2,000 + 1/196,000) = √(0.025 + 0.000510 + 0.000500 + 0.0000051) = √0.026015 = 0.1613.
• 95% CI = exp(ln 2.00 ± 1.96 × 0.1613) = exp(0.6931 ± 0.3161) = 1.46 to 2.74 — the lower bound is above 1, so the disproportionality is statistically supported.
• χ² = 200,000 × (40 × 196,000 − 1,960 × 2,000)² / (2,000 × 198,000 × 2,040 × 197,960). Numerator difference = 7,840,000 − 3,920,000 = 3,920,000; squared = 1.53664 × 10¹³; × 200,000 = 3.07328 × 10¹⁸. Denominator = 2,000 × 198,000 = 3.96 × 10⁸; 2,040 × 197,960 = 4.03838 × 10⁸; product = 1.59920 × 10¹⁷. χ² = 19.2 (≫ 4).
• Verdict: cases ≥ 3, PRR ≈ 2 and χ² ≥ 4 — the pair qualifies as a signal for validation, not proof of causation.

Example 3 — incidence, RR, AR and NNH. In a cohort study, severe hypoglycaemia occurs in 36 of 1,200 patients on Drug X and 12 of 1,200 on standard therapy over one year.

• Incidence(exposed) = 36/1,200 = 3.0%; incidence(unexposed) = 12/1,200 = 1.0%.
• RR = 3.0/1.0 = 3.0; AR = 3.0% − 1.0% = 2.0% (0.02).
• NNH = 1/0.02 = 50 — one extra severe hypoglycaemic episode for every 50 patients treated for a year.

Example 4 — reporting rate and under-reporting. A company's product sold 2.4 million DDDs in a year in India (≈ 6,575 patient-years at 365 DDDs per patient-year). It received 18 ICSRs of hepatotoxicity.

• Reporting rate = 18/6,575 = 0.00274 per patient-year ≈ 2.7 per 1,000 patient-years.
• If under-reporting is assumed at 94% (only 6% captured), true incidence ≈ 0.00274/0.06 = 0.0456 per patient-year ≈ 4.6 per 100 patient-years — a reminder that spontaneous rates are floors, not estimates.

Example 5 — seriousness and submission clock. A patient on a newly approved Indian product is hospitalised with Stevens–Johnson syndrome; the company's medical information line takes the call on 1 March (Day 0); the insert does not list SJS.

• Serious? Yes — hospitalisation. Unexpected? Yes — not in the reference safety information. Hence expedited.
• 15-calendar-day due date = 16 March; the same case also appears in the next PSUR line listing and triggers a labelling assessment. (If the product were an investigational drug in a trial, the SAE route to the licensing authority and ethics committee is 24 hours.)

Where it is used

• PV service and CRO industry: drug safety associates and case processors in Accenture, Cognizant, TCS, Capgemini, Tata Elxsi, IQVIA, Parexel, Syneos, Fortrea and Navitas — the largest single entry route for B.Pharm graduates in India, working in Argus/ARISg on E2B submissions for global sponsors.
• Indian pharma MAHs: corporate PV departments at Sun, Cipla, Dr Reddy's, Lupin, Zydus, Torrent, Glenmark, Alkem and Serum Institute — PSMF ownership, PSUR/PBRER writing, RMPs, EU-QPPV coordination for export markets, and CDSCO/EMA/FDA inspection readiness.
• Hospital and clinical pharmacy: ADR Monitoring Centres in AIIMS, PGIMER, JIPMER, CMC Vellore and state medical colleges; clinical pharmacists run ADR reporting, causality and preventability assessment, antimicrobial stewardship and medication-error review as part of NABH quality requirements.
• Public-health programmes: ADR and AEFI surveillance in the national TB programme (shorter and bedaquiline-containing regimens need active monitoring), the national AIDS control programme, malaria programmes and the universal immunisation programme.
• Regulatory and academic roles: CDSCO/state drug control, IPC-PvPI technical associates, signal review panels, and research publishing from AMC data; MvPI and HvPI roles for device and blood-bank safety.
• Medical writing and auditing: narrative writing, aggregate report authoring, literature surveillance (Embase/Medline), PV audits and CAPA, and PV system training.

Interview questions

• Q: Define pharmacovigilance and state its aims. A: The science and activities relating to detection, assessment, understanding and prevention of adverse effects or any other drug-related problem; aims are patient safety, rational use, early signal detection and an evidence-based benefit–risk judgement across a product's life-cycle.
• Q: Difference between an adverse event and an adverse drug reaction? A: An AE is any untoward occurrence during treatment regardless of cause; an ADR implies at least a suspected causal link at normally used doses.
• Q: Is a severe adverse event always serious? A: No. Severity is intensity; seriousness is defined by outcome — death, life-threatening, hospitalisation or its prolongation, disability, congenital anomaly or a medically important event.
• Q: What four elements make a valid case? A: An identifiable patient, an identifiable reporter, a suspect medicinal product and an adverse event or outcome.
• Q: When does Day 0 start and what is the expedited timeline? A: Day 0 is the date any employee of the company first receives the minimum information for a valid case; serious unexpected cases follow the 15-calendar-day expedited clock, and clinical-trial SAEs in India must reach the licensing authority and ethics committee within 24 hours.
• Q: Explain Rawlins–Thompson classification with one example each. A: A augmented (warfarin bleeding), B bizarre (penicillin anaphylaxis), C chronic (analgesic nephropathy), D delayed (thalidomide teratogenicity), E end-of-use (beta-blocker rebound), F failure (OC failure with rifampicin).
• Q: Who coordinates PvPI and where does Indian ADR data finally go? A: The Indian Pharmacopoeia Commission, Ghaziabad, as National Coordination Centre; AMC data entered into VigiFlow is shared with WHO-UMC's VigiBase in Uppsala, Sweden.
• Q: What is a signal, and does a signal prove causation? A: Information suggesting a possible new or incompletely documented causal association, usually from more than one report; it is a hypothesis requiring validation and confirmation, not proof.
• Q: Why do we use PRR instead of incidence in spontaneous databases? A: Spontaneous data have no exposure denominator, so absolute risk cannot be computed; PRR compares the proportion of a given event for one drug with that for all other drugs in the same database.
• Q: What is the Weber effect and why does it matter in signal detection? A: ADR reporting for a new drug rises to a peak around the second marketing year and then falls, so a drop in report numbers may reflect reporting behaviour rather than improved safety.
• Q: What does an RMP contain and name a routine and an additional risk-minimisation measure? A: The safety specification, the pharmacovigilance plan and risk-minimisation measures; routine — a contraindication or warning in the package insert; additional — a pregnancy prevention programme or controlled-access scheme.
• Q: Under which legal provision can the Government of India ban a drug on safety grounds? A: Section 26A of the Drugs and Cosmetics Act, 1940, used for several fixed-dose combinations and for withdrawn drugs such as rofecoxib and cisapride.
• Q: What are the PSUR submission intervals for a new drug in India? A: Every six months for the first two years after approval, annually for the next two years, and thereafter as the licensing authority directs, under the New Drugs and Clinical Trials Rules, 2019.

Common mistakes

• Using "serious" and "severe" interchangeably — the single most penalised error in both exams and PV interviews.
• Writing that a spontaneous reporting system gives ADR incidence; it cannot, because there is no exposure denominator.
• Treating causality assessment as proof of causation, or claiming "Certain" without dechallenge, rechallenge or objective confirmation.
• Misstating Naranjo bands (quoting "definite >9" instead of ≥9) or adding item scores with the wrong sign for "alternative causes present".
• Confusing the regulator with the coordinator: CDSCO/DCGI takes regulatory action; IPC Ghaziabad runs PvPI and collates data.
• Submitting or accepting a case without a reporter or without an identifiable patient and still calling it an ICSR.
• Classifying a withdrawal reaction as Type A, or a dose-dependent interaction failure as Type B — map the mechanism first.
• Quoting Schedule Y timelines for new drugs that are now governed by the New Drugs and Clinical Trials Rules, 2019.

Related topics

Clinical research and trial phases · Drugs and Cosmetics Act and CDSCO · Regulatory affairs basics · Good Manufacturing Practice (GMP) · Quality control and quality assurance · Clinical pharmacy and therapeutic drug monitoring