Pediatric Adverse Drug Reactions: An Observational Cohort Study After Health Care Workers' Training

Abstract

OBJECTIVE

Adverse drug reactions (ADRs) in children are an important but underestimated public health issue. This study describes ADRs in a registered pediatric population of Bologna and demonstrates that ADRs might be better detected after health care personnel training.

METHODS

A prospective cohort was recruited from July 1, 2016, to June 30, 2019, after health care worker sensitization, and compared to a retrospective cohort enrolled from 2013 to 2016. The ADRs are classified by system organ classes and drugs are categorized according to the Anatomical Therapeutic Chemical classification system.

RESULTS

We retrospectively recruited 78 pediatric patients with ADRs in the 2013 to 2016 period, and we prospectively enrolled 127 children in the 2016 to 2019 period. In both periods, most of the ADRs reported were classified as non-serious reactions (68.8%). The most frequent ADRs were general and administration site disorders. During 2013 to 2016 vaccines were the most frequent cause of ADRs (83.3%;) and the main reporters were health care workers other than physicians (84.6%), whereas during the second period, medical doctors become the main signalers (65.4%) and ADRs related to vaccines significantly decreased (55.1%). During the 2016 to 2019 period the number of drug categories was higher than in the 2013 to 2016 period (24 vs 8). Patients with ADRs due to vaccinations present more frequently a favourable outcome (63%).

CONCLUSIONS

This study demonstrates that active pharmacovigilance and health care personnel sensitization are associated with improved ADR detection, providing valuable information about drugs' safety profile in pediatric patients.

Introduction

Pharmacovigilance refers to “the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other drug related problem.”¹ In 2011 the European Medicines Agency defined adverse drug reactions (ADRs) as “A response to a medicinal product which is noxious and unintended. ADRs may arise from use of the product within or outside the terms of the marketing authorization (off-label use, overdose, misuse, abuse and medication errors) or from occupational exposure.”^2,3

Evidence in the literature shows that the incidence of ADRs in hospitalized children is around 9.5%, with serious reactions accounting for 12% of all ADRs.⁴ The frequent off-label drug use in pediatric patients, the use of pharmaceutical forms unsuitable for children, and the need to adjust the dose to body surface area and weight increase the risk of medication errors, thus leading to avoidable ADRs. Despite ADRs in children being about 3 times more frequent than in non-elderly adults,⁵ there is little information regarding the characteristics of ADRs in pediatric population.⁶ Post-marketing experience shows that the safety profile of drugs targeted to children, including vaccines, is often less well known because of smaller sample clinical trials.⁷

The main aim of this study is to demonstrate that raising awareness in health care personnel improves ADR detection. We describe the adverse reactions registered in the study population and we suggest methods and instruments for risk control and prevention.

Materials and Methods

This is a cohort observational study. Starting July 1, 2016, pediatric health professionals, including pediatricians, medical students, and nurses working at the pediatrics department of Maggiore Hospital in Bologna, Italy, have been trained in reporting ADRs. In particular, a hospital pharmacist has been involved and has taken a lecture on the definition of ADRs and medication errors and their correct reporting. Moreover, during the first year after training, the hospital pharmacist actively supported pediatricians in compiling the reporting form. Lastly, we simplified the signaling process at the moment of discharge from the emergency department; in fact, physicians had to check a box on the form for the record to be sent to the pharmacovigilance referent of our hospital.

The study population consists of 205 patients aged 0 to 18 years presenting with ADRs that have been reported to the pharmacovigilance department of Azienda Unità Sanitaria Locale (AUSL) in Bologna. We recruited 127 consecutive, unselective individuals up to 18 years of age presenting with ADRs from July 1, 2016, to June 30, 2019. Then we compared this prospective cohort to a retrospective group of 78 patients presenting ADRs from January 1, 2013, to June 30, 2016.

In case of suspected or confirmed ADRs, physicians perform a full physical examination and collect data regarding dose, route, and timing of administration. All records are reported on the Single Suspected Adverse Reaction Report Form available on the Agenzia Italiana del Farmaco Web site or in the Vigifarmaco online platform.

The ADRs were categorized according to the System Organ Classification⁸ (Supplemental Table S1) and drugs were categorized according to the Anatomical Therapeutic Chemical (ATC)⁹ classification system. In order to simplify the comparison, we classified as drugs all the medications that were different from vaccines. The reporting sources were classified as medical doctors (MDs), pharmacists, other health care workers (OHWs; such as nurses), citizens, caregivers, patients, and lawyers. Patients were divided into 4 subgroups, according to age (<28 days, 28 days to 2 years, 2–12 years, >12 years). The age categories were based on the European Medicines Agency guideline on clinical investigation of medicinal products in the pediatric population.¹⁰

The ADRs are defined as serious, according to World Health Organization (WHO), when they result in death, require hospital admission or prolongation of existing hospital stay, result in persistent or significant disability or incapacity, are life-threatening, or resul in cancers, congenital anomalies, or birth defects. Regarding ADR outcomes, we distinguish between favorable outcome or resolution and unfavorable outcome, which includes permanent damage, lack of clinical improvement, clinical situation unchanged or worsened, and death.

All calculations were performed with the commercially available program IBM SPSS 20.0 for Windows (Statistical Package for Social Science, Chicago, IL). Qualitative variables were reported as absolute frequencies and percentages and were analyzed using Pearson χ² test in case of frequencies more than 5 cases or Fisher exact test in case of smaller frequencies. Quantitative data are expressed as mean values ± SD and were analyzed using a 2-sided Student test. A p value <0.05 is considered as significant for each test.

Results

We found 205 children with reported ADRs from 2013 to 2019. Considering all ADRs reported in the Bologna AUSL, the prevalence in children during the whole study period was 7.9% (205 of 2598). Overall, 96 (46.8%) were female and 107 (52.2%) were male, considering missing data in 1 patient for each 3-year period. Mean age was 6 ± 5.9 years in the whole group. Table 1 shows comparison of mean age between subgroups (i.e., period of onset, sex, seriousness, outcome, and type of medication).

Table 1.

Comparison of Median Age Between Subgroups

	3-Year Period		Sex		Vaccine/Other Drug		Severity		Outcome
	2013–2016 (n = 78)	2016–2019 (n = 127)	Male (n = 96)	Female (n = 107)	Vaccine (n = 135)	Drug (n = 70)	Serious (n = 62)	Non-serious (n = 137)	Favorable (n = 111)	Unfavorable (n = 94)
Age, mean ± SD, yr	5.2 ± 6	6.6 ± 5.9	5.6 ± 5.8	6.4 ± 6	4.4 ± 5.1	9.2 ± 6.2	6.9 ± 5.7	5.7 ± 6.6	5.1 ± 5.7	7.2 ± 6.1
p value	NS		NS		0.001		0.005		0.05

NS, not significant

Overall, 68.8% (137) of ADRs were classified as non-serious reactions, whereas 31.2% (62) were classified as serious reactions. Most of the ADRs (54.1%) resolved without any sequelae. Table 2 describes the correlation between type of medication and seriousness/outcome. Vaccines caused non-serious ADRs in most cases (75.2%; 100 of 133), whereas the prevalence of non-serious ADRs due to drugs was significantly lower (56.1%; 37 of 66; p < 0.01). Moreover, patients with ADRs due to vaccinations presented more frequently with a favourable outcome (63%; 82 of 135; p < 0.0001), regardless of ADR severity. Consequently, the mean age of patients with severe adverse reactions was significantly higher than the patients' age with non-serious ADRs (6.9 ± 5.7 vs 5.7 ± 6.6 years; p < 0.005); at the same time, unfavorable outcome was associated with higher age (7.1 ± 6.1 vs 5.1 ± 5.7 years; p < 0.05).

Table 2.

Correlation Between Type of Medication, Outcomes, and Severity ^*

	Outcomes		Severity
	Favorable	Unfavorable	Serious	Non-serious
Vaccine	88	47	33	100
Drug	23	47	29	37
Total	111	94	62	137
p value	0.0001†		0.01†

^* Data are expressed as absolute values.

^† p value for the total numbers.

Regarding outcomes, patients with non-serious ADRs presented with a favorable outcome in most cases (82 of 137; 59.9%); on the contrary, serious ADRs were mainly related to unfavorable outcome (35 of 62; 56.5%; p < 0.05; Figure).

Figure.

Correspondence between ADRs seriousness and outcome (p < 0.05)

We enrolled 127 children in the 3-year period from July 1, 2016, to June 30, 2019, whereas 78 ADRs were registered in the 3-year period from July 1, 2013, to June 30, 2016. Considering all ADRs reported in the Bologna AUSL, the prevalence in children was 8.4% (78 of 927 cases) in the first period (2013–2016) and 7.6% (127 of 1671 cases) in the second period (2016–2019). Table 3 describes all qualitative variables as frequencies and percentage compared between the 2 periods. In the first 3-year period vaccines were the most frequently reported (83.3% vs 16.7%; p < 0.0001), patients were younger (5.2 ± 6 vs 6.6 ± 5.9 years; p value not significant), and the group age from 28 days to 2 years was the most involved (53.8% vs 29.1%; p < 0.001). The OHWs were the main reporters in the first period, whereas MDs became the main signalers after the training period (p < 0.0001). During the first 3-year period the percentage of patients with a favourable outcome was significantly higher than in the second period (64.1% vs 48%; p < 0.05).

Table 3.

Demographic and Clinical Data During the Whole Period and Comparison Between the Two 3-Year Periods

	Total	2013–2016	2016–2019	p value
Sex, n				NS
Male	96	39	57
Female	107	38	69
Missing data*	2	1	1
Age, n				0.001
<1 mo	5	0	5
1 mo to 2 yr	79	42	37
2–12 yr	67	17	50
>12 yr	54	19	35
Reporters, n				0.0001
Physician	90	7	83
OHW	873	66	21
Pharmacist	3	2	1
Citizen	23	3	20
Lawyer	2	0	2
Cause of ADR, n				0.0001
Vaccine	135	65	70
Drug	70	12	57
Seriousness				NS
Serious	62	20	42
Non-serious	137	54	83
Missing data*	6	4	2
Outcome				0.05
Favorable	111	50	61
Unfavorable	94	28	66
Missing data*	46	10	36

ADR, adverse drug reaction; NS, not significant; OHW other health care worker

^* Missing data were not considered in the statistical analysis.

Supplemental Tables 1 and 2 describe ADRs according to system organ classification and ATC, comparing the two 3-year periods. In both periods, the most frequent ADRs were general and administration site disorders (67 of 219 system organ classifications registered in the first period, 73 of 332 in the second period), followed by skin and subcutaneous disorders (43 of 219 system organ classifications in the first period, 71 of 332 in the second period). Excluding adverse reactions due to vaccines, during 2013 to 2016 the most frequent ADRs were psychiatric disorders followed by skin and subcutaneous disorders; during 2016 to 2019 the most frequent ADRs were skin and subcutaneous disorders, followed by general and administration site disorders and neurologic disorders. Vaccines were mainly involved in ADRs in both periods (54 vaccines and 78 patients in the first period, and 64 vaccines and 126 patients in the second period). The drugs mostly involved in ADRs were those active in the nervous system in the first period (7 of 78) and antibiotics in the second period (39 of 126). During 2016 to 2019, in this 3-year period the number of ADRs increased and, consequently, the number of drug categories involved was higher than in the previous period (24 vs 8).

Discussion

Our study confirms that pharmacovigilance is an important tool to better understand drugs' safety profile among children, considering the limited clinical trials involving pediatric patients. In fact, because of differences in pharmacokinetics and pharmacodynamics between adults and infants or children, studies about drugs' safety profile regarding general population cannot be applied to pediatric patients. Moreover, this study underlines how adequate training may increase awareness about pharmacovigilance among health care workers.

Adverse reactions in the pediatric population are a significant public health issue because they may result in hospital admission, prolonged hospitalization, permanent disability, or even death^11,12 We believe that ADRs are not properly reported because of lack of awareness and adequate training among health care workers. In Italy it is mandatory to report every suspected ADR within 36 hours in the case of vaccines and biologic agents and within 48 hours for other drugs.

Incidence rates reported in pediatric patients are extremely heterogeneous and there are many difficulties in comparing ADR incidence rates from different observational studies. These studies differ in clinical settings and sample characteristics (outpatients or inpatients, group age), drugs included (some studies exclude vaccines), and methodologies used while evaluating suspected ADRs.^5,13 Our study shows a general prevalence of ADRs of 8.3% in the first period (2013–2016) and 7.5% in the next period (2016–2019). Ferrajolo et al¹⁴ reported a prevalence of 6.8% in an Italian pediatric population analyzed from 2001 to 2012. Star et al¹⁵ used the WHO Global Individual Case Safety Report database (VigiBase) and reported an ADR rate of 7.7% in 268,145 children from 0 to 17 years of age during a 40-year period. Both studies do not consider vaccines. Excluding methodology, ADR underreporting might be one of the reasons we found a lower prevalence than previous studies.

In the whole study period vaccines were the major reason for reporting ADRs, suggesting that awareness regarding ADR reporting is widespread among health care workers who administer vaccines. However, after health care personnel sensitization, reported adverse reactions due to drugs other than the vaccines and the number of drug categories involved increased. Accordingly, in the second period patients were older and the group age from 1 months to 2 years was less involved than during the previous period, likely because vaccines are mainly administered in the first 2 years of life, according to Italian vaccination campaigns.

According to our data, good outcome and non-severe ADRs are significantly associated with vaccinations and consequently to younger mean age. These data underline how the risk-benefit ratio is clearly in favor of vaccines. During the second period the amount of drugs other than vaccines significantly rose; consequently, unfavorable outcomes were significantly higher than in the first period. However, severity was not affected in the same way; in fact, we noticed that the gap between serious and non-serious reactions was reduced in the second period. Nevertheless the difference between the two periods was not significant. This is probably due to the fact that unfavorable outcome includes all patients without complete resolution, and therefore patients with non-serious reactions.

Antibiotics, psycholeptics and anticonvulsants, and anesthetics are the drug classes most frequently associated with ADRs in our sample, excluding vaccines, as described by other authors.^13–16 In the 2016 to 2019 period, we noticed an increase in reports made by MDs, suggesting that adequate training may significantly improve pharmacovigilance.

Raising awareness about reporting ADRs and simplifying the signaling process can encourage health care workers to report ADRs. Collaboration between pediatricians, nurses, and hospital pharmacists might improve pharmacovigilance activity and therefore the safe use of medicines, especially in children.

Our study has some limitations; in particular, it cannot take into account the trends in the prescription of drugs with a lower therapeutic index, which may be related to more ADRs. Moreover, training may not be the only reason ADR reporting increased in the second 3-year period. We have not considered the effect of changes in Italian ADR reporting legislation made in April 2015 that may have increased awareness about pharmacovigilance among health care workers. Lastly, we did not take into consideration the Naranjo score, because data were collected before the application of the score by Agenzia Italiana del Farmaco. In our department, to prevent ADRs due to medical mistakes, a group of pharmacists, nurses, and pediatricians has created a handbook with information about drugs' routes of administration and formulations. Secondly, we organize annual courses to update health care personnel about drug management. Finally, to prevent drug administration errors made by parents, our department has created simplified brochures describing the proper use of different drugs; these tools are targeted especially to foreign parents who struggle to understand the Italian language.

Conclusion

Active post-marketing pharmacovigilance provides valuable information about drugs' safety profile, especially in children. Adequate training about pharmacovigilance is associated with a major effect on ADR reporting made by physicians, especially on medications other than vaccines.

ABBREVIATIONS

ADRs

adverse drug reactions

ATC

Anatomical Therapeutic Classification

AUSL

Azienda Unità Sanitaria Locale

MD

medical doctor

OHW

other health care workers

WHO

World Health Organization