Clinical stories are necessary for drug safety
Post-marketing reports of suspected adverse drug reactions (ADRs) are key elements to prevent patients being harmed by drugs. In addition to the data collected on pre-specified fields, descriptive free-text case stories (narratives) might be crucial when interpreting these reports, which can add to the knowledge of adverse effects and are the basis for regulatory decisions.
Narratives occur in only 11% of the reports in VigiBase,1 the WHO Global Individual Case Safety Report database.2 Usually only minimum information about the drugs and ADRs are listed in the structured data fields. The citation below is a fragment from a published case report for a 13-year-old boy with olanzapine-induced rhabdomyolysis with concomitant lithium-induced electrocardiogram changes.
On day 1 (8 weeks prior to hospitalization) the patient had been admitted to a psychiatric residential care controlled environment facility because of a behavioural disorder and was placed on olanzapine 2.5 mg/d. On day 6, he reported weakness, sore throat, abdominal cramping, myalgias, and diaphoresis, which appeared to be consistent with influenza. On day 14, he experienced increasing weakness and failed to participate in organized activities, which was misinterpreted by the residential care faculty as manifestation of disobedient and oppositional behaviour. Sertraline 100 mg every morning and lithium sustained release 300 mg twice a day were added to olanzapine. On day 27, he fell several times while trying to get out of bed. He was transferred to an outside institution where he was noted to have an elevated creatine phosphokinase, leukocytosis and T-wave inversion of his precordial leads.3
The same case is reported in VigiBase, but the standard report fields only mention patient age and gender, olanzapine and lithium and the ADRs myocardial infarction and rhabdomyolysis. The description of the course of the events provided in the narrative was essential to complete a causality assessment and provided a greater understanding of the clinical picture for rhabdomyolysis in this teenager. Information found exclusively in VigiBase narratives1 includes ADR severity and site, intervention, patient ethnicity and specifications of the underlying disease. Variables having assigned standard fields, regardless of how they are reported exclusively in free-text, are: drug indication, concomitant drugs, dose, onset date of ADR and laboratory findings. Information on withdrawal of drug treatment (dechallenge) or the re-introduction of drug treatment after withdrawal (rechallenge) can also appear exclusively in the narratives.1
Methods
A random sample of 50 reports from VigiBase was evaluated by an expert physician, without access to previously recorded causality assessments. Each narrative was categorised as: (1) crucial to causality assessment of suspected ADRs; (2) considerably affecting the understanding of the clinical course of the reported events; or (3) adding no useful information to the standard report fields.
The length distribution of narratives within the randomised sample ranged from 20 to 3,395 characters, reflecting the general narrative length distribution of reports in VigiBase. In parallel, we also studied longer outlier narratives in VigiBase (defined as reports ranging from 10,000 to 20,000 characters); 50 such reports were evaluated according to the same categorisation.
Results
Within the sample of narratives with general-length distribution, 22% (n=11) contained information that was crucial to the causality assessment. An additional 26% (n=13) considerably affected the understanding of the clinical course of the cases (Table 1). When applying the same categorisation on the sample of 50 long outlier narratives, the proportion of information that was crucial to causality assessment increased to 32% (n=16) and the proportion considerably affecting the clinical understanding to 42% (n=21).
Among 50 reports with narratives of general length, the causality or clinical assessment of almost every second report (48%) was affected when taking the narrative into consideration. The assessment of 50 reports with long narratives (10,000–20,000 characters) was affected in 74% of cases.
Discussion
Our evaluation of two international samples of randomly selected reported clinical stories highlights the importance of detailed descriptions of circumstances under which suspected ADRs occur. Without the case story, crucial misinterpretation of case reports could lead to wrong regulatory decisions and deny clinically useful information to healthcare practitioners.
Footnotes
-
Letters not directly related to articles published in Clinical Medicine and presenting unpublished original data should be submitted for publication in this section. Clinical and scientific letters should not exceed 500 words and may include one table and up to five references.
- © 2014 Royal College of Physicians
References
- ↵
- Karimi G,
- Strandell J,
- Star K,
- et al.
- ↵
- Lindquist M
- ↵
- Rosebraugh CJ,
- Flockhart DA,
- Yasuda SU,
- Woosley RL
Article Tools
Citation Manager Formats
Related Articles
- No related articles found.
Cited By...
- No citing articles found.