An analysis of the trends, characteristics, scope, and performance of the Zimbabwean pharmacovigilance reporting scheme

Abstract We aimed to determine the reporting trends and characteristics of Individual Case Safety Reports (ICSRs) from the Zimbabwean national pharmacovigilance system. ICSRs submitted to VigiBaseTM, the World Health Organisation's ICSR database between January 1993 and December 2017 were retrospectively reviewed with respect to the suspected medicine, System Organ Class (SOC), adverse drug reaction (ADR) type and seriousness, Anatomic Therapeutic Chemical (ATC) group, age, and gender. In total, 4071 ICSRs were submitted to VigiBaseTM from targeted spontaneous reporting (n = 2909; 71.5%), vaccine surveillance (n = 679; 16.7%), and passive spontaneous reporting (n = 483; 11.9%), respectively. The median age, ICSR completeness score and timeliness of reporting were 34.0 years (IQR: 14.0; 43.0), 0.90 (IQR: 0.70; 1.00), and 548.0 days (IQR: 266:1131), respectively. More than half of the ICRS were from female patients (n = 2233; 54.9%). Antiretrovirals, antibiotics, vaccines, and anti‐tubercular medicines were reported in 62.9%, 27.9%, 16.7%, and 13.3% of submitted ICSRs, respectively. The most frequent ADRs involved the skin and subcutaneous systems (n = 1111; 20.5%), nervous system (n = 733; 13.5%), and gastrointestinal disorders system (n = 654; 12.1%). The number of ADRs reported for each patient was significantly related to the reported medicine's ATC category (P = .001. The number of ADRs was significantly related to the use of antiretroviral agents. In conclusion, Zimbabwe has made significant progress in establishing a functional pharmacovigilance system. However, the present system reports on a limited therapeutic spectrum of medicines and potentially underestimates the national ADR burden. Further work is required to strengthen the more sustainable spontaneous reporting system which potentially captures a variety of therapeutic classes.


| INTRODUC TI ON
The safety information on any medicinal product initially derives The ADR burden in most Sub-Saharan Africa (SSA) remains unknown despite training and analytical support from the UMC. 6,7 Developed countries contribute the majority of individual case safety reports (ICSRs) in VigiBase TM , 4 whilst Africa contributes a mere 0.88% of the cumulative global ADR data. 8 Most of these ADRs are related to antibiotics in contrast to the rest of the world where cardiovascular and neurological medicine related ICSRs predominate. 3,8 These observations reflect the unevenly distributed global disease burden, different drug utilization patterns, cultural norms, and medical practices. 2 Furthermore, limited resources, government support and the over-reliance on donor funded public health programmes may explain the preponderance for antibiotic related ICSRs in Africa. 9 Moreover 50% of SSA countries lack the legal mandate to monitor ADRs and/or engage market authorization holders in PMS activities. 8,9 Consequently, the ADR profile of some medicines is inadequately reflected in their summary of product characteristics.
Zimbabwe previously boasted one of Africa's most robust healthcare systems, but over the past few years it has been stagnating due to decreasing government healthcare expenditure. 10,11 HIV/AIDS and tuberculosis have also contributed to the strain on resources necessitating donor funding for service provision. 10,11 However, pharmacovigilance activities for HIV and tuberculosis medicines have disproportionately been supported from donor funded public health programmes for the respective diseases. 12 Non-communicable diseases such as diabetes mellitus and hypertension have been steadily increasing 13 and are projected to increase markedly in people living with HIV. 14 Due to limited pharmacovigilance funding, the majority of any collated ADRs are expected to be from the mandatorily reported ICSRs linked to the funded programmes.
Zimbabwe's national pharmacovigilance scheme was established in 1993, but the country officially joined the PIDM in 1997. 8 It currently utilizes passive spontaneous ADR reporting, vaccine safety surveillance and targeted spontaneous reporting (TSR) to collect unsolicited ADRs, adverse events following immunization (AEFIs) and anti-retroviral and/or anti-tubercular medicine related ADRs respectively. Stimulated reporting through TSR started in 2012 15 after its promulgation by the WHO as a simple, inexpensive PV tool to leverage existing public health and PV programmes. 16 TSR aimed to increase ICSRs by mentoring practitioners in high case-load clinics whilst task shifting ADR reporting to non-physician healthcare practitioners. 16 All collected ICSRs are verified and collated centrally by the Medicines Control Authority of Zimbabwe (MCAZ), the national medicines regulator 17 before causality assessment and subsequent upload into VigiBase TM .
Given the importance of continuous ADR data analysis, 18 we set out to evaluate the trends and characteristics of Zimbabwean derived ICSRs in VigiBase TM . We also aimed to compare the reporting patterns for AEFIs and ADRs across all therapeutic areas; and to determine the relationship between age, gender, drug anatomic therapeutic class (ATC), reporter type and the number of reported ADRs.

| Study design
A retrospective descriptive analysis of anonymized ICSR data from Zimbabwe collected during the period 1 January 1993 to 31 December 2017 was conducted. The ICSR data was extracted from VigiBase TM using VigiLyze®, the database's search and analysis software tool on 2018-10-06 (dataset date: 2018-09-30). We included

| Classification of ICSRs
Adverse drug reaction preferred terms (PTs) were classified accord- Chemical (ATC) Classification system at level 2 for the suspected medicines. 23 In this system, drugs are divided into five different levels based on the system or organ on which they act; their chemical, pharmacological and therapeutic properties. 24 These classification criteria are similar to those used by Ozcan et al and de Vries et al 24,25

| Statistical analysis
The de-duplicated, MedDRA version 21.0 coded ADR data were exported into a Microsoft Office Excel TM package for further analysis (Microsoft Corporation, Redmond, WA, USA). The timeliness of reporting was calculated by subtracting the date of onset of ADR from the date of VigiBase TM entry. 26 Where the date for the onset of the ADR was incomplete, the first day of the month was used, otherwise the ICSR was excluded from timeliness calculations. 26 VigiGrade TM completeness scores for ICSRs were obtained from VigiLyze® measuring the completeness of time-to-onset, age, gender, indication, outcome, report type, dose, country, primary reporter, and comments. 27 Descriptive statistical methods were used to analyse the surveil- for the statistical analyses and for graphing the analyses respectively. All statistical tests were done at the 5% significance level.

| Ethical considerations
Ethical exemption for the study was granted by the Medical Research

Council of Zimbabwe (MRCZ Ref: MRCZ/E/207). The exemption
was granted because ICSR data collection is a routine surveillance programme which uses anonymized data.

| Characteristics of the pharmacovigilance system
A total of 4126 ICSRs were extracted from VigiBase TM for the pe-

| Demographic characteristics
The demographic, reporter and reporting characteristics are shown in Table 1 67.0%), while 72 (1.8%) in adults older than 65 years of age.

| Reporting trends
The annual number of ICSRs gradually increased during the study

| Characteristics of the adverse drug reactions
The frequency of at least one serious ADR was 30 (30); both ADR and suspect drug not specified (9) were the most frequent followed by significant disability/incapacita- 13.5%), and gastrointestinal disorders (n = 654; 12.1%) as shown in

| Characteristics of the drugs
The median number of recorded drugs per ICSR was 3.00 (IQR: 3.00; 4.00). A total of 14 medicines were suspected in the majority

| D ISCUSS I ON
Our analysis shows that the total number of collected ICSRs has to the total number of submitted ICSRs. 29 The latter has mostly been driven by strict legal reporting requirements which are not present in most African nations. 25,29 The markedly high reporting by nurses possibly shows a success of the TSR programme because its main purpose was to task shift ICSR reporting to non-physician cadres. 16 In addition, the nurse driven ICSR submission reflects the greater dependency of healthcare delivery in district hospitals on nurses as observed for the TSR programme and in Togo. 15,31 The most frequently reported SOCs were skin and subcutaneous tissue disorders, gastrointestinal disorders, general disorders and administration site conditions, and the nervous system disorders.

System organ class
Most commonly reported preferred term prominence of blood, lymphatic system, and immune system disorders in the latter. The observed differences may be due to the varied ADR reporting practices in addition to differences in the most administered ATC groups within these countries. 4 Moreover most

Blood and lymphatic system disorders
ICSRs were reported for medicines in the ATC group J with a significantly higher proportion of anti-retrovirals as previously shown in other low income countries such as Togo and Nigeria. 4,8,31,34,35 In contrast, cardiovascular and nervous system medications were more frequently reported in developed countries, such as France and Germany. 29,32 The ADRs in Colombia and Portugal were mostly attributed to anti-infectives for systemic use followed by nervous and cardiovascular system medicines. 33,36 The quality of ICSR reporting as indicated by ICSR timeliness and completeness is comparable to the global average but lags be- The ATC code had an influence on the number of reported ADRs whilst patient age, gender and the number of prescribed medicines had no influence as previously observed elsewhere. 40 The number of ADRs was significantly related to the use of antiretroviral agents.
ICSRs were observed to be more frequent in females regardless of the male to female gender ratio in the population as observed in Nigeria, Israel, Italy and the USAs. 34,35,[40][41][42][43] However, ICSRs on deaths and significant disability/incapacitation were more common in males as previously observed in Italy and Sweden. 41,44 In contrast to the general trend, the male to female ratio was higher in children under 5 years of age and even higher in infants as expected from previous studies. 41 In addition, it was noted that ICSRs were more common in the 20-65 year age group possibly due to multiple drug therapy secondary to anti-retroviral and anti-tubercular drugs co-prescription. 34,41,45 This is supported by similar observations in Nigeria where comparatively higher HIV prevalence was noted in this age group. 34,35 The overall predominance of female ICSRs could be due to underlying physiological differences, females' higher medical-care seeking behavior and use of more prescriptions compared to males. 46,47 Despite the notable growth in the Zimbabwean pharmacovigilance scheme, mandatory reporting by market authorization holders should be considered in order to increase ICSR reporting. In addition, decentralization of ICSR collection to provincial and district healthcare facilities and the utilization of mobile phone ADR reporting platforms can strengthen the collation and completeness of ADR data. The latter measures may reduce barriers to reporting and extend the reach and availability of ADR reporting platforms to all relevant stakeholders including patients as observed in Kenya. 48,49 Furthermore, it is critical to provide relevant feedback to healthcare practitioners to provide meaning and an appreciation of ICSR reporting. 30 This could subsequently help stimulate ICSR reporting, thereby increasing the numbers and spectrum of the submitted ICSRs. It is important to regularly review the performance of the scheme as a basis for informing regulatory measures. 29 The spectrum of ICSRs could also be increased by encouraging the submission of ADRs from other therapeutic areas beyond anti-infectives.
The inherent limitations of a study based on spontaneous ADR reports include under-reporting and the inability to calculate incidence rates because of the unavailability of exposure/denominator data. 32,42 While ICSR data is voluntarily submitted, stimulated reporting through TSR introduced bias especially regarding the observed ADR, ATC, and SOC profiles. The ICSRs were skewed towards anti-retroviral drugs, anti-tubercular medicines and vaccines as would be expected given the dominance of public health programmes targeting HIV, tuberculosis, and immunization. The observed profiles may also indicate the prescription patterns, prevalent diseases and acute, severe and/or well-known ADRs. 50 The major strength of the current study is in highlighting areas in need of improvement and how the current set-up compares to international regulatory pharmacovigilance schemes.
In conclusion, significant progress has been made in establishing a functional pharmacovigilance system. However, the present pharmacovigilance scheme is based on a limited therapeutic spectrum of medicines and may underestimate the ADR burden. In addition, there is need to improve the timeliness of ICSRs to enable further case reviews and timely signal detection.

CO N FLI C T O F I NTE R E S T S
There are no competing interests to declare.

Data from the WHO Collaborating Centre for International Drug
Monitoring were used. The information is not homogeneous at least with respect to origin or likelihood that the pharmaceutical product caused the adverse reaction. The information in this paper does not represent the opinion of the World Health Organization.

AUTH O R CO NTR I B UTI O N S
JM conceived and designed the study and contributed to data collection, data analysis, and interpretation of the results. SK contributed to data analysis and interpretation. Both authors reviewed the manuscript and approved the final version the final version of the manuscript.

DATA S H A R I N G A N D ACCE SS I B I LIT Y
The data that support the findings of this study are available from the corresponding author upon reasonable request.

O RCI D
Josiah Tatenda