Pharmacovigilance and Adverse Drug Reactions in Hospital: Introduction 101

SKS Medication Safety Series, Chapter 10 of 15 · Standard SİY09

Prof. Dr. Mehmet Nurullah Kurutkan, Düzce Üniversitesi İşletme Fakültesi Sağlık Yönetimi Bölümü

← Series hub: all 15 chapters

Chapter at a glance

Even correctly prescribed and administered medicines can harm. Pharmacovigilance is the system for detecting, reporting and learning from such adverse drug reactions. This chapter explains hospital pharmacovigilance, from spontaneous reporting to electronic detection and artificial intelligence. It corresponds to SKS SİY09, which links hospital reporting to the Turkish Pharmacovigilance Centre (TÜFAM) (Sağlık Bakanlığı, 2020).

Evidence base: 502 records of a 4948-record Web of Science corpus on hospital medication safety (Section 7). Reading time: about 25 minutes. Prior knowledge required: none.

Learning objectives

After working through this chapter, the reader should be able to:

  1. define adverse drug reactions and pharmacovigilance;
  2. state the SKS SİY09 reporting requirements;
  3. compare spontaneous reporting with electronic detection methods;
  4. evaluate the limitations of reporting databases.

Key concepts

Contents: How this chapter was produced · 1. Introduction · 2. History and development · 3. Definitions and conceptual schema · 4. Distinguishing it from neighbouring concepts · 5. Types and classifications · 6. Theoretical frameworks and models · 7. Themes and bibliometric profile · 8. The situation in Türkiye · 9. Debates and contradictions in the literature · 10. Measurement problems · 11. Research gaps and agenda · 12. Implications for practice today · Summary · Discussion questions · Further reading · Glossary · References · Source verification and AI use statement · Data and supplementary material

How this chapter was produced

Box 1. Method

Database: Web of Science Core Collection (SCI-EXPANDED, SSCI, ESCI); English; articles and reviews (including early access). Search date: 16 November 2026; all years (1981 to 2027).

Search: (#1 medication safety and error terms AND #2 hospital settings AND #3 fifteen SKS content blocks) NOT #4 animal and preclinical terms; the full string is on the series hub.

Screening: five export files (4955 records) were merged; 7 duplicates were removed by WoS accession number, DOI and normalised title plus year, leaving 4948 records. Records were assigned to this sub-topic when its term set (pharmacovigilance, adverse drug reaction, spontaneous report) appeared in the title or keywords; a record can belong to several sub-topics. Included here: 502 records (249 with the term in the title; 49 reviews).

Limitations: single database; English only; assignment by term matching, checked by reading titles of the most cited and most recent records; claims are checked against abstracts, not full texts.

Web of Science Core Collection, 5 export files
4955 records
↓ duplicates removed: 7
4948 unique records (hospital medication safety corpus)
↓ sub-topic term set applied to title/keywords
Included in this chapter: 502
249 title matches; 49 reviews
Not in this chapter: 4446
other chapters or general records
Figure 1. Flow of records for this chapter. Source: produced by the author from the Web of Science corpus.

1. Introduction

SKS requires defined processes for adverse effect reporting; a pharmacovigilance officer with defined responsibilities whose name, contact details and CV are notified to TÜFAM; reporting of serious and unexpected adverse effects to the officer; and their notification to TÜFAM within 15 days at the latest (Sağlık Bakanlığı, 2020).

The stakes are high. A study of hospital admissions found that many were caused by preventable adverse drug reactions, most often from inadequate monitoring or inappropriate dosing (McDonnell & Jacobs, 2002). In older patients more than 80% of reactions causing admission or occurring in hospital are dose-related (Routledge et al., 2004).

2. History and development

Hospital-level studies in 1999 compared adverse drug event reporting by doctors, nurses and patients (Van Den Bemt et al., 1999), and methods for detection in hospital were reviewed in 2001 (Thürmann, 2001). In 2009 electronic health record data and natural language processing were used for active pharmacovigilance (Wang et al., 2009) and for defining which events to monitor (Trifirò et al., 2009). Since 2020 the field has shifted to analyses of large reporting databases, such as the FDA Adverse Event Reporting System (Zou et al., 2023), and to artificial intelligence (Salas et al., 2022).

  • 1999 Doctors, nurses and patients compared in reporting ADEs (Van Den Bemt et al., 1999).
  • 2001 Methods to detect ADRs in hospitals reviewed (Thürmann, 2001).
  • 2002 Preventable ADRs as a cause of admission (McDonnell & Jacobs, 2002).
  • 2006 Ten-year study of ADRs and their reporting in children (Le et al., 2006).
  • 2009 Active pharmacovigilance using NLP and EHRs (Wang et al., 2009).
  • 2016 Medication errors and ADRs across the Australian patient journey (Roughead et al., 2016).
  • 2022 Systematic review of AI in pharmacovigilance (Salas et al., 2022).
Figure 2. Timeline of milestones in the corpus. Source: produced by the author from the Web of Science corpus.

3. Definitions and conceptual schema

Table 1. How the corpus defines the concept
SourceYearDisciplineDefinition (paraphrased)Emphasised dimension
Sağlık Bakanlığı, 20202020Accreditation standardA hospital process with a named officer for reporting adverse effects, with serious and unexpected effects sent to TÜFAM within 15 days.Organisation and reporting
McDonnell & Jacobs, 20022002Clinical pharmacyPreventable ADRs as a form of medication error, mostly due to inadequate monitoring or dosing.Preventability
Routledge et al., 20042004Clinical pharmacologyADRs in older people, mostly dose-related (type A).Mechanism
Wang et al., 20092009Medical informaticsActive computerised pharmacovigilance using clinical text and statistics.Detection method
Salas et al., 20222022Pharmaceutical medicineAI-supported pharmacovigilance for case processing and signal detection.Automation
Common core
harmful effects of medicines are detected, reported and analysed so that use can be made safer
PreventabilityDetection methodReporting organisationData sourceAutomation
Figure 3. Conceptual schema: the shared core of the definitions and the dimensions on which they diverge. Source: produced by the author from the Web of Science corpus.

Definitions converge on detection and learning. They diverge on whether ADRs are seen as inherent drug properties (Routledge et al., 2004) or as preventable process failures (McDonnell & Jacobs, 2002).

4. Distinguishing it from neighbouring concepts

Pharmacovigilance overlaps most with reporting systems and interactions (Figure 4): many ADRs are interaction-related and reported through general incident systems.

Reporting/indicators (SİY14)21%
Interactions/compatibility (SİY03)9%
Prescribing (SİY05)7%
High-alert/LASA (SİY10)4%
Rational use/AMS (SİY01.02)3%
Own meds/reconciliation (SİY12)2%
Formulary (SİY02)2%
Preparation/dispensing (SİY06)1%
Controlled drugs (SİY11)1%
Governance (SİY01)1%
Administration (SİY08)1%
Transport/distribution (SİY07)0%
Storage (SİY04)0%
Returns/waste (SİY13, SİY15)0%
Figure 4. Neighbouring-concept matrix: percentage of the 502 records of this chapter that are also tagged with each other SKS sub-topic. Source: produced by the author from the Web of Science corpus.
Table 2. Distinguishing criteria
Neighbouring conceptHow it differs
Incident reporting (SİY14)Covers errors and near misses in general; pharmacovigilance covers drug effects, including those without error (Sağlık Bakanlığı, 2020).
Medication errorAn error may or may not cause harm; an ADR may occur without error (McDonnell & Jacobs, 2002).
Interactions (SİY03)A cause of ADRs; detection methods overlap (Tatonetti et al., 2012).

5. Types and classifications

ADRs are classified as dose-related (type A) or not (type B) (Routledge et al., 2004). Detection methods include spontaneous reporting, chart review, computerised monitors and trigger tools (Thürmann, 2001) and database disproportionality analysis (Hosohata et al., 2019; Ma et al., 2021).

6. Theoretical frameworks and models

The corpus uses an epidemiological surveillance framework, with signal detection statistics applied to report databases (Ma et al., 2021), and an informatics framework for active surveillance of electronic records (Trifirò et al., 2009; Wang et al., 2009).

7. Themes and bibliometric profile

Hospital-based studies are now outnumbered by database studies of single drugs (Figure 5). Database analyses identified drugs most frequently linked to acute kidney injury (Hosohata et al., 2019) and myocarditis signals with immune checkpoint inhibitors (Ma et al., 2021). AI may improve case processing, though its effect on the quality of safety analysis is not yet shown (Salas et al., 2022).

Records per year 2026 (incomplete)

0367319901995200020052010201520202025

Most frequent journals

Frontiers in Pharmacology28
Pharmacoepidemiology and Drug Safety18
British Journal of Clinical Pharmacology16
Expert Opinion on Drug Safety13
Drug Safety12
International Journal of Clinical Pharmacy10

Most frequent countries

USA102
Peoples R China66
Japan50
India34
Italy23
Canada20

Most frequent author keywords: pharmacovigilance (237), adverse drug reactions (81), adverse drug reaction (76), adverse drug events (65), adverse drug event (51), medication safety (50), medication errors (45), patient safety (41), drug safety (38), faers (29).

Figure 5. Bibliometric profile of the 502 records (1999 to 2027). Records dated before 1990 are counted at 1990 and 2027 early-access records at 2026. Source: produced by the author from the Web of Science corpus.

8. The situation in Türkiye

Five records have Turkish authors, including a study of antibiotic-associated adverse drug events in hospitalised children (Guezeloglu & Karaci, 2022), a machine learning detection system for medication errors in neonatal intensive care (Yalçin et al., 2023) and a study of potential interactions in a medical intensive care unit (Gülçebi Idriz Oglu et al., 2016).

9. Debates and contradictions in the literature

Passive or active surveillance? Spontaneous reporting is cheap but under-detects; electronic methods find more events but need validation (Wang et al., 2009).

Who should report? In children, reporting by health professionals was infrequent and follow-up limited (Le et al., 2006); patients themselves reported more reactions to new drugs than doctors or nurses (Van Den Bemt et al., 1999).

Database signals or clinical truth? Disproportionality analyses generate hypotheses, not causal estimates (Zou et al., 2023).

10. Measurement problems

Incidence of ADRs depends on definitions and detection methods (Thürmann, 2001). National estimates combine heterogeneous sources; in Australia about 230,000 medication-related admissions per year were estimated (Roughead et al., 2016). Under-reporting makes report counts poor indicators of safety.

11. Research gaps and agenda

  1. How complete is hospital reporting to TÜFAM within the SKS 15-day rule? (no study found)
  2. Can hospital EHR data support active pharmacovigilance in Türkiye? (Wang et al., 2009)
  3. How can paediatric ADR reporting be improved? (Guezeloglu & Karaci, 2022; Le et al., 2006)
  4. Does AI improve the quality, not only speed, of safety analysis? (Salas et al., 2022)
  5. Which preventable ADRs should be prioritised for hospital monitoring? (McDonnell & Jacobs, 2002)

12. Implications for practice today

Make the SKS pharmacovigilance officer visible, simplify reporting forms, feed back results to reporters, and use triggers or electronic queries to supplement voluntary reports (Thürmann, 2001; Wang et al., 2009). Focus monitoring on dose-related reactions in older adults (Routledge et al., 2004).

Summary

  • ADRs can occur without error but many are preventable.
  • SKS SİY09 requires a pharmacovigilance officer and TÜFAM reporting within 15 days.
  • Spontaneous reporting under-detects; electronic methods add sensitivity.
  • Database studies dominate recent output.
  • Turkish studies focus on paediatric and ICU settings.

Discussion questions

  1. Distinguish an ADR from a medication error.
  2. Why are spontaneous reports insufficient for estimating incidence?
  3. Design a trigger-based ADR detection process for your hospital.
  4. What could motivate physicians to report to TÜFAM?
  5. What can and cannot be concluded from a FAERS signal?

Further reading

Glossary

TurkishEnglishShort definition
Advers ilaç reaksiyonuAdverse drug reactionHarmful unintended response to a medicine
FarmakovijilansPharmacovigilanceDetection and prevention of adverse effects
Spontan bildirimSpontaneous reportingVoluntary reporting of suspected reactions
Sinyal saptamaSignal detectionStatistical search for new drug-event associations

References

Guezeloglu, E., & Karaci, M. (2022). Antibiotic-associated Adverse Drug Events in Hospitalized Children. Journal of Pediatric Infection, 16(3), E198-E204. https://doi.org/10.5578/ced.20229714

Gülçebi Idriz Oglu, M., Küçükibrahimoglu, E., Karaalp, A., Sarikaya, Ö., Demirkapu, M., Onat, F., & Gören, M. Z. (2016). Potential drug-drug interactions in a medical intensive care unit of a university hospital. Turkish Journal of Medical Sciences, 46(3), 812-819. https://doi.org/10.3906/sag-1504-147

Hosohata, K., Inada, A., Oyama, S., Furushima, D., Yamada, H., & Iwanaga, K. (2019). Surveillance of drugs that most frequently induce acute kidney injury: A pharmacovigilance approach. Journal of Clinical Pharmacy and Therapeutics, 44(1), 49-53. https://doi.org/10.1111/jcpt.12748

Le, J., Nguyen, T., Law, A. V., & Hodding, J. (2006). Adverse drug reactions among children over a 10-year period. Pediatrics, 118(2), 555-562. https://doi.org/10.1542/peds.2005-2429

Ma, R. L., Wang, Q. Z., Meng, D. Y., Li, K., & Zhang, Y. (2021). Immune checkpoint inhibitors-related myocarditis in patients with cancer: an analysis of international spontaneous reporting systems. BMC Cancer, 21(1), Article 38. https://doi.org/10.1186/s12885-020-07741-0

McDonnell, P. J., & Jacobs, M. R. (2002). Hospital admissions resulting from preventable adverse drug reactions. Annals of Pharmacotherapy, 36(9), 1331-1336. https://doi.org/10.1345/aph.1A333

Roughead, E. E., Semple, S. J., & Rosenfeld, E. (2016). The extent of medication errors and adverse drug reactions throughout the patient journey in acute care in Australia. International Journal of Evidence-Based Healthcare, 14(3), 113-122. https://doi.org/10.1097/XEB.0000000000000075

Routledge, P. A., O’Mahony, M. S., & Woodhouse, K. W. (2004). Adverse drug reactions in elderly patients. British Journal of Clinical Pharmacology, 57(2), 121-126. https://doi.org/10.1046/j.1365-2125.2003.01875.x

Sağlık Bakanlığı. (2020). Sağlıkta kalite standartları: Hastane (Sürüm 6.1). Sağlık Hizmetleri Genel Müdürlüğü, Sağlıkta Kalite, Akreditasyon ve Çalışan Hakları Dairesi Başkanlığı. https://dosyamerkez.saglik.gov.tr/Eklenti/41258/0/skshastane-seti-s-61–09082021pdf.pdf

Salas, M., Petracek, J., Yalamanchili, P., Aimer, O., Kasthuril, D., Dhingra, S., Junaid, T., & Bostic, T. (2022). The Use of Artificial Intelligence in Pharmacovigilance: A Systematic Review of the Literature. Pharmaceutical Medicine, 36(5), 295-306. https://doi.org/10.1007/s40290-022-00441-z

Tatonetti, N. P., Fernald, G. H., & Altman, R. B. (2012). A novel signal detection algorithm for identifying hidden drug-drug interactions in adverse event reports. Journal of the American Medical Informatics Association, 19(1), 79-85. https://doi.org/10.1136/amiajnl-2011-000214

Thürmann, P. A. (2001). Methods and systems to detect adverse drug reactions in hospitals. Drug Safety, 24(13), 961-968.

Trifirò, G., Pariente, A., Coloma, P. M., Kors, J. A., Polimeni, G., Miremont-Salamé, G., Catania, M. A., Salvo, F., David, A., Moore, N., Caputi, A. P., Sturkenboom, M., Molokhia, M., Hippisley-Cox, J., Acedo, C. D., Van Der Lei, J., & Fourrier-Reglat, A. (2009). Data mining on electronic health record databases for signal detection in pharmacovigilance: which events to monitor? Pharmacoepidemiology and Drug Safety, 18(12), 1176-1184. https://doi.org/10.1002/pds.1836

Van Den Bemt, P. M. L. A., Egberts, A. C. G., Lenderink, A. W., Verzijl, J. M., Simons, K. A., Van Der Pol, W. S. C. J. M., & Leufkens, H. G. M. (1999). Adverse drug events in hospitalized patients – A comparison of doctors, nurses and patients as sources of reports. European Journal of Clinical Pharmacology, 55(2), 155-158. https://doi.org/10.1007/s002280050611

Wang, X. Y., Hripcsak, G., Markatou, M., & Friedman, C. (2009). Active Computerized Pharmacovigilance Using Natural Language Processing, Statistics, and Electronic Health Records: A Feasibility Study. Journal of the American Medical Informatics Association, 16(3), 328-337. https://doi.org/10.1197/jamia.M3028

Yalçin, N., Kasikci, M., Çelik, H. T., Allegaert, K., Demirkan, K., Yigit, S., & Yurdakök, M. (2023). Development and validation of a machine learning-based detection system to improve precision screening for medication errors in the neonatal intensive care unit. Frontiers in Pharmacology, 14, Article 1151560. https://doi.org/10.3389/fphar.2023.1151560

Zou, F., Zhu, C. Y., Lou, S. Y., Cui, Z. W., Wang, D., Ou, Y. Y., Wang, L., Chen, J. Y., & Lan, Y. B. (2023). A real-world pharmacovigilance study of mepolizumab in the FDA adverse event reporting system (FAERS) database. Frontiers in Pharmacology, 14, Article 1320458. https://doi.org/10.3389/fphar.2023.1320458

Source verification and AI use statement

Source verification. All 17 references were generated programmatically from the Web of Science records of the corpus, so authors, year, title, journal and DOI match the records exactly; every in-text citation was cross-checked by code against the reference list. Claims were checked against the record abstracts; full texts were not read. The SKS Hospital Set (Version 6.1) is the only source outside the corpus; it was accessed and read for this chapter.

AI use statement. Claude (Anthropic) was used to merge and de-duplicate the Web of Science exports, assign records to sub-topics, compute the bibliometric indicators and figures, and draft the text under the author’s instructions. The author is responsible for the final content.

Data and supplementary material

Supplementary file S1 (Excel) lists the 502 records of this chapter (authors, year, title, journal, DOI, title match flag), together with the full screening log of the 4948-record corpus and the citation verification report. The list is not reproduced here; it is available from the author on request.

← Chapter 9: Safe Medication AdministrationChapter 11: High-Alert and Look-Alike Sound-Alike Medications →