Introduction There is evidence of inappropriate medication use, causing unnecessary costs for health systems, particularly those with limited resources. Overprescription is commonly reported and can lead to antibiotic resistance. Prescribing patterns differ between countries; little is known about paediatric prescribing practices in Africa.
Objectives To investigate prescribing practices in children in The Gambia, West Africa.
Method A retrospective survey of prescribing practices in children under 5 years of age based on WHO protocol DAP/93.1 was conducted. Twenty government-run health centres across all six regions in The Gambia were assessed. The first 10 encounters each month in 2010 were recorded. For each encounter, patient demographics, diagnoses and medications were recorded as per protocol.
Results Two thousand and four hundred patient encounters were included. The mean number of medications per encounter was 2.2 (median 2.0, IQR 2.0–3.0). Across different geographical regions within The Gambia antibiotics were prescribed in 63.4% (IQR 62.8–65.8%) and micronutrients in 21.7% (IQR 15.3–27.1%) of patient encounters. There was evidence of high antibiotic prescription in children with cough and coryzal symptoms (54.5%; IQR 35.8–59.0%) and simple diarrhoea without dehydration (44.8%; IQR 36.7–61.3%). 74.8% (IQR 71.8–76.1%) of medications were prescribed generically.
Conclusions The study showed an overprescription of antibiotics and substantial usage of micronutrients despite a lack of international evidence-based guidelines. Cost-effective interventions to improve prescribing practices are called for and more studies with a focus on rational prescribing in paediatrics in low-income settings are urgently required to fill the gap in current knowledge.
- Paediatric Practice
- Tropical Paediatrics
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known on this topic
Irrational prescribing can be costly for health services.
Overprescription of antibiotics is common and can lead to antibiotic resistance.
There is significant variation in prescribing practices between countries.
What this study adds
There is evidence of inappropriate prescribing of antibiotics and micronutrients in The Gambia but lack of published data limits comparison with other settings.
Symptomatic diagnoses are often used, perhaps as a result of limited availability of diagnostic tests, education of nursing staff and availability of doctors.
Identifies issues with poor documentation on drug availability in The Gambia.
The prescription of medication is an important part of medical treatment, particularly so in low-resource countries where facilities for some diagnostic procedures and some surgery are less readily available. The WHO estimates that over 50% of medications are prescribed, dispensed or sold inappropriately.1
Rational prescribing is important to ensure best use of resources, effective therapy and minimisation of potential side effects. Prescription practices have been well studied in high-income countries, but there is disparity among the limited data for resource-poor countries.2
Overuse of antibiotics is commonly reported.3 ,4 As antimicrobial resistance increases, the need for more specific and inevitably more expensive antibiotics arises, placing a growing demand on healthcare.5 There is evidence of irrational prescribing of micronutrients within resource-poor nations,6 ,7 while no guidelines for their use in primary care exist.
The overuse of injections occurs in a number of resource-poor countries;8 ,9 the WHO is working towards minimising this as part of the Safe Injection Global Network.10
In order to uniformly study, monitor and assess prescribing practices, the WHO has created a protocol11 based on recommendations by Hogerzeil et al.12 It recommends focusing on the number of drugs prescribed per patient, whether generic names were used, and the percentage of consultations prescribing antibiotics, injections and the number of drugs prescribed from an Essential Drugs List (EDL). Each EDL is specific to the area in which the study is being carried out.
This study aims to identify current prescribing patterns in the under 5-year-olds in The Gambia, recommend ways to increase rational prescribing and compare results with other studies using the same methodology.
Prescribing practices in children under 5 years old in The Gambia, West Africa were investigated by a retrospective cross-sectional study, using the method outlined in WHO/DAP/93.111 with some amendments.
The Gambia is a small country with a population of 1 729 000 in 2010.13 The under 5-year-old population of The Gambia was 1 93 900 in the 2003 census. The Gambia is working towards meeting the Millennium Development Goals14 and the under 5-year-old mortality rate has been decreasing from 153 per 1000 in 1990 to 103 in 2009. The Gambian government spends 5.5% of its gross domestic product on health.15 Government-run healthcare provision in The Gambia includes 7 hospitals and 39 public health centres (primary and secondary level) as well as basic community and village clinics; private and NGO healthcare provision mainly provided in the west region includes 3 hospitals and 46 clinics. (Personal communication with MC)
Study site selection
Twenty public health centres were surveyed. At least three sites were chosen from all six regions of The Gambia: upper river region (URR), central river region, north bank east (NBE), north bank west, lower river region (LRR) and west region. A fourth site was selected in the two regions with the largest populations: the west region and the URR. Hospital outpatient clinics and non-governmental organisations were excluded. Health centres were not selected randomly within each region but rather chosen to provide as wide a geographical distribution as possible taking feasible logistics into account.
Prescribing indicators data collection
Data was obtained from Health Care Centre Registry Books, which are mandatory government records of all primary healthcare consultations in The Gambia. A 12-month period was studied: January to December 2010 inclusive. Starting from the first day of each month, the first 10 eligible patients under 5 years of age were identified. Patients were discounted if they were attending for routine child welfare appointments (including childhood vaccinations), referred on to hospital, duplicate entries or the record was illegible. In these cases, the next eligible patient was used.
Data on each patient was recorded using a unique anonymised identifier; no patient names were transcribed. For each patient the following was recorded: age to the nearest month (calculated from date of birth if present), sex, date of consultation, diagnosis, number, name and type of drugs prescribed, number of drugs prescribed generically, and the number of drugs on our EDL. A generic prescription contains non-proprietary names, allowing any suitable drug to be dispensed unrestricted to brand. The WHO/DAP/93.1 stipulates that the EDL for this study should contain about 10–15 drugs, however Gambia has an extensive government drug list. An EDL was compiled based on recommendations in the WHO protocol,11 communication with the Department of Public and Environmental Health, The Gambia and a pilot study reviewing prescriptions in two healthcare centres in Kiang West, lower river region, The Gambia, thereby taking local disease patterns and requirements into account. The EDL consisted of 17 medications: amoxicillin, benzoic acid ointment, calamine lotion, chlorphenamine, chloramphenicol eye ointment, coartem, co-trimoxazole, erythromycin, ferrous sulfate with folic acid, mebendazole, metronidazole, multiple vitamins, nystatin oral suspension, paracetamol, salbutamol, tetracycline eye ointment, vitamin B complex.
Two persons (RR, HN) were involved in all the data collection supported by a senior nurse employed by the Medical Research Council, The Gambia Unit, International Nutrition Group.
All data were manually entered into the computer and analysed using Excel 2007 as stipulated in the WHO recommendations.11 Using the six regional values, means and SD or medians and IQRs were calculated depending on whether the data was normally distributed or not. For the number of drugs prescribed per encounter, the average, median and IQR were calculated from combined data from all 20 sites.
Antibiotics were defined according to the WHO protocol11 which suggests only antibacterials, anti-infective dermatological agents, anti-infective ophthalmological agents and antidiarrhoeals are counted as antibiotics. The British National Formulary (BNF) was used to confirm or disprove whether medications were prescribed generically. A drug was only considered an injection if this was specified in the Health Care Centre Registry Book. In addition to the prescribing indicators suggested by the WHO, data on micronutrient prescriptions were also collected. Diagnoses were recorded using International Classification of Diseases-10 codes.16
At the time of visiting each health centre the diagnostic tests available, the number and grade of staff working, and the presence and edition of a BNF was recorded by asking the nurse in charge or laboratory staff. The BNF is commonly used in The Gambia alongside the standard Gambian drug treatment manual. Nursing qualifications are graded according to level and length of training. In The Gambia state registered nurses and state registered midwives, and state-enrolled nurses and state-enrolled midwives attend a 4-year course or a 3-year course, respectively. Community health nurses and community health midwives train for 3 years focusing on public health issues and community nurse attendants train at a local level of every facility.
Across the healthcare centres surveyed, drug availability from the pharmacies was not reliably recorded as such but requests for more drugs from the regional health teams were documented in requisition books. Recorded requests from 2010 were used as surrogate markers for drug availability, making the assumption that requests were made when drugs ran out. Drugs out of stock at the time of the survey visit were also recorded for each health centre.
Two thousand and four hundred patient encounters for children under 5 years of age were reviewed across 20 health centres with a total of 2991 diagnoses. The mean age was 1.71 years (SD 1.17) with a female: male ratio of 1:1.3. In 1896 encounters, only a single diagnosis was given. The most common diagnosis was diarrhoea (16.3% of all diagnoses given), 5.5% of which had accompanying dehydration. ‘Cough and cold’ was diagnosed in 14.1% and pneumonia in 8.7%. The number of ‘symptomatic diagnoses’ was high, with 328 (10.6%) as ‘fever’ and a further 325 (10.6%) as ‘nausea and vomiting’ or ‘cough’ or ‘abdominal pain’. Other symptomatic diagnoses included ‘chest pain’, ‘general body pain’ and ‘difficulty breathing’.
Overall 5277 medications were prescribed, the mean number of drugs prescribed per encounter was 2.2 (median 2.0, IQR 2.0–3.0, range 0–5) with only 34 patients receiving nothing. In 1.15% (IQR 0.9–1.4%) of consultations injectables were used although this was higher in the NBE region (4.2%). Generic names were used in three quarters of all drugs prescribed (74.8%; IQR 71.8–76.1%) with up to 83.9% in the URR region. The most common trade prescription was co-trimoxazole, prescribed as Septrin.
A large proportion of prescriptions of antibiotics (63.4% of encounters; IQR 62.8–65.8%) was observed with 54.5% (IQR 35.8–59.0%) of patients with coryzal symptoms only and 44.8% (IQR 36.7–61.3%) with simple diarrhoea without dehydration receiving an antibiotic (table 1). Across all regions a consistently high number of drugs (95.4%; IQR 95.0–96.5%) prescribed were part of the EDL.
Micronutrients were frequently prescribed in the health centres surveyed (21.7% of encounters; IQR 15.3–27.1%) with 9.7% observed in the central river region and 30.6% in LRR. Of all micronutrient prescriptions, multivitamins accounted for 80.7% and haematinics for 10.8%, the latter mainly as ferrous sulfate with folic acid. Vitamin A accounted for 1.6%, vitamin B complex (containing a number of B vitamins) for 5.5%, vitamin C for 1.3% and one patient was prescribed vitamin E.
Overall, paracetamol was prescribed in 88.8% and co-trimoxazole was the most commonly used antibiotic (40.2% of all prescriptions) particularly for respiratory and skin infections (table 2).
The number of times a drug was out of stock within the year of the survey was 6.1 (IQR 4.3–10.2). Unless specified in the EDL all routes of drugs were considered separately. Records on the day of visit showed 5.0 out of 17.0 drugs (from the EDL) being out of stock (IQR 3.0–7.0). None of the centres had tetracycline eye ointment at the time of the survey but paracetamol was available at all survey sites.
Ninety per cent of healthcare facilities had a BNF for reference; the two centres without a BNF used the Gambian standard drug treatment manual.17 All facilities had rapid diagnostic tests or were able to perform a blood film in order to diagnose malaria. Other diagnostic tests available varied across the health centres surveyed. Only two facilities did not have a state registered nurse or a state-enrolled nurse, one of those was led by a state registered midwife and the other only by a community health nurse. Doctors onsite were only present in two facilities.
The current study, using data collected from 20 health centres across The Gambia showed significant variation between geographical regions for all indicators (p<0.05) possibly related to drug availability, differences in education of nursing staff and high variability between health centres within each region. For the criteria investigated in this study, no established or recommended reference ranges were found to enable appropriate comparison. Differing disease burden between countries and the lack of literature on current prescribing practices make establishing reference ranges a difficult task. Several studies following the WHO protocol11 were reviewed (tables 3 and 4). Only two of those studies in low-income and lower-middle-income settings focused on paediatrics: one was limited to the under 5-year-olds in Tanzania (2006)18 and another study in India (2010)19 focused on children under 12 years. This highlights a gap in the current literature in paediatric prescribing.20
The average of 2.2 drugs prescribed per patient consultation in The Gambia was similar to paediatric studies in Tanzania18 and India,19 the latter including only prescribing encounters. A multicentre study in high-income and upper-middle-income settings showed higher and lower averages (1.3 in Barcelona, Spain and 2.9 in Smolensk, Russia)21 (table 4).
The Gambia's median injection rate across sites (1.15%) is low in comparison with the Tanzanian study (26.2%),18 although similar to that observed in India (1.6%). Factors contributing to the low rate could include differences in the education level of healthcare workers and availability of injectable drugs. Drug availability may have contributed to the observed difference between regions in The Gambia with NBE using 4.2%. Although not specified for injectables in particular, average drug availability in NBE was better (5.9) compared with The Gambia overall (7.5).
The high percentage of drugs (95.6%) from the EDL is perhaps more indicative of the relatively large EDL used in this study (17 medications).
Antibiotic prescription in children under 5 years of age was high (63.4%), more than twice of that seen in public primary healthcare clinics in Tanzania (30.5%)18 and rather similar to what was observed in paediatric private outpatient clinics in India (79%).19 For patients with coryzal symptoms, 54.5% were prescribed antibiotics and 0.4% received an injectable compared with none with a diagnosis of pneumonia, suggesting significant overusage of antibiotics and identifies an area for specific focus on practice (table 1). Similarly, of those with simple diarrhoea, 44.8% received an antibiotic and 0.6% received an injectable. Huschler et al suggested a number of factors including the knowledge and behaviour of medical professionals, the knowledge and behaviour of patients and the social-cultural context that could account for the overprescription of antibiotics.22 Increasing prescribers’ knowledge of antibiotic prescribing would be beneficial and it should be an integral part of the new undergraduate medical and nursing teaching programmes in The Gambia. Overprescription of antibiotics and injectables in coryza may stem from the inexperience of establishing the correct clinical findings and relating them to the appropriate diagnosis. The surveillance of antimicrobial use is recommended by the WHO to evaluate antibiotic prescribing and highlight areas for improvement, with the aim of limiting the development of antibiotic resistance and reducing long-term costs through the avoidance of inappropriate prescribing.23
Das et al7 found liberal prescription of micronutrients by consultants in Karachi, Pakistan to individuals who do not necessarily require them (table 3). The benefit of multiple micronutrients in a clinical setting is debatable.24 ,25 In our study, the high rates of micronutrient prescriptions found were similar to a Nigerian study, where 32.3% of paediatric patients were prescribed vitamin C and 21.3% received multivitamin supplementation.26 Observed differences in prescribing practices for micronutrients between geographical regions may again be related to drug availability or differences in knowledge of the prescribers. Interestingly, in 58% of encounters (318/548), where a micronutrient was prescribed, no antibiotic was given. Micronutrients may have been used as a substitute for antibiotics where the patient demands medication but guidelines discourage antibiotic treatment.
Limitations of the current study include the non-random selection of health centres (51% of all public health centres) surveyed although their geographical location and population served are representative of the remaining centres. The study is retrospective in nature, relying on written records exemplified by the poor documentation on medication stock. Our drug availability data was based on the assumption that those drugs requested were out of stock at the time and all orders were fulfilled. More accurate information was impossible to obtain, due to missing or incomplete documentation at some sites; occasionally stock availability was not recorded when placing their drug orders and if some documentation was missing it was impossible to ascertain how long each drug was out of stock. The accuracy of diagnoses could not be verified. Only prescribing practices relative to the diagnosis were analysed, not the child's clinical picture. However, with over 600 symptomatic diagnoses, it suggests that if the diagnosis was uncertain then the main symptom was recorded. The high number of symptomatic diagnoses given could be due to the lack of diagnostic facilities available at the healthcare centres visited. At the time of visiting the majority of minor and major health facilities in The Gambia, most diagnostic decisions were made by nurses. Extending facilities and focusing the education of nursing staff on diagnostic skills may improve reaching definitive diagnoses hence providing a basis for appropriate prescribing methods.
This study shows an overprescription of antibiotics in The Gambia, particularly for coryzal symptoms and simple diarrhoea. Reiteration of the dangers of inappropriate antibiotic use is recommended. Prescription of micronutrients needs to be reviewed across more resource-limited countries and international evidence-based prescription guidelines in primary healthcare settings should be established. Documentation of drug availability requires attention to monitor the impact on appropriate prescribing. More studies on paediatric prescribing practices in resource-limited countries need to be undertaken using the WHO protocol to reliably compare settings and set standards.
The authors are grateful to Musa Sambou, head nurse at MRC Keneba, The Gambia for helping with the data collection; Lauren Maddy (University of Sheffield) for helping with the pilot work and other staff members involved at MRC Keneba. The authors are also grateful to Professor Andrew Prentice for comments on an earlier draft of the paper. HN would like to thank the Royal Medical Society, Edinburgh for the travel and study grant they awarded him. The authors would also like to thank Musa Sowe, (Ministry of Health and Social Welfare, The Gambia) for helping with the data on health facilities in The Gambia.
RR and HN contributed equally.
Contributors SU conceived the original research idea and supervised the study. MC and SEM contributed to the study design. AB undertook pilot work for this project and contributed to the initial literature review. RR and HN developed the study protocol, collected and analysed the data. RR, HN and SU drafted the manuscript. All authors contributed to and approved the final version of the manuscript.
Funding This work was supported by the Medical Research Council (MRC) UK core funding to the MRC International Nutrition Group (MC-A760-5QX00). The source of funding did not influence study methods or interpretation of the results.
Competing interests None.
Ethics approval Ethical approval was obtained from the joint Gambian Government/MRC Unit The Gambia Ethics Committee (SCC 1271v2). Government officials and regional health directors were informed of the study and their consent obtained.
Provenance and peer review Not commissioned; externally peer reviewed.