Prevalence of, and harm caused by medication error in antiretroviral programs in Uganda

Sub-Saharan Africa is home to two thirds of all people living with HIV. The World Health Organisation promotes a 'public health' approach to HIV treatment in countries with limited resources, meaning that many African countries have rapidly increased availability of antiretrovirals (ARVs) in order to treat as many patients as possible. Due to rapid service growth, healthcare workers have been trained to undertake tasks traditionally provided by doctors. This 'task shifting' has been important in successful scale up of HIV services in countries such as Uganda. Patients with HIV may also take medicines for other conditions, and some medicines are not compatible when given together. ARVs and other drugs can affect the way some medicines work, leading to side effects or decreased effectiveness. Ineffective ARV treatment can lead to resistance of the virus to one or more drug, limiting future treatment options. Doctors require a full medication list for each patient in order to manage these drug interactions. Medication errors occur at 4 main stages of treatment: medication recording in clinical notes, prescribing, dispensing, and also how patients take medication. Prescribing errors and drug interactions involving ARVs are common in Europe, but have not been fully studied in ARV programs in Africa. A UK study found that HIV physicians only recognised a third of drug interactions involving ARVs. Harm caused by medication errors or drug interactions may reduce effectiveness of treatment, cause toxic effects, or permanently reduce treatment options. Additional financial cost to patients and healthcare systems may result. No data currently exist on the harm caused to patients from ARV drug interactions or medication errors.
A study will be conducted in 3 different outpatient clinics in Uganda. Patients will be asked to detail all medicines they take, and if they have missed any doses of ARVs in the previous month. The medicine list will be checked for drug interactions using a smart phone application developed by the University of Liverpool, which can be used offline in clinics. A summary of each consultation will be recorded in patients' clinic notes for doctors to see before treating the patient. Prescribing and dispensing errors will be assessed, and whether patients have experienced any adverse reactions, or side effects which may be related to an error. A second study will evaluate patients who are taking ARVs and are admitted to hospital. Inpatients will be evaluated by a pharmacist to determine whether the hospital admission is related to a medication error or drug interaction. Adverse reactions resulting from errors or drug interactions will be assessed for the extent of harm caused to patients. The primary outcomes of the project will be the prevalence, type and severity of medication errors and drug interactions; the patients most at risk, and the health system factors which may cause or protect from errors. Also, the extent of clinical harm arising from medication error in patients taking ARVs admitted to hospital; the proportion of patients taking medication that their HIV doctors were not aware of, and the extent to which prescribers recognise drug interactions.
This project will demonstrate the scale of the problem caused by medication error and drug interactions in ARV programs in Sub-Saharan Africa, and lead to development of specific interventions which will reduce patient harm an increase the effectiveness of ARV treatment, providing benefit to both patients and the wider health system.

The project aims to quantify the prevalence of and harm caused by medication error and unmanaged drug-drug interactions (DDIs) in antiretroviral (ARV) programs in Uganda. In an outpatient study, a medication history and adherence assessment will be taken for 1000 patients across diverse sites. Medication lists will be screened for DDIs aided by an offline Android application. Information will be recorded in the clinical notes prior to scheduled clinic appointments. Prescribing and dispensing errors will be assessed. Inpatient studies will identify the prevalence, type, and severity of prescribing errors and DDIs in hospital settings. Hospital admissions will be evaluated for adverse drug events and causality in relation to medication errors and unmanaged DDIs. Primary endpoints are: the prevalence, type and severity of medication errors and DDIs; the patients most at risk of medication error in terms of CD4 count, WHO stage, number of medicines, age, gender (via multivariate analysis); and health systems factors precipitating or protecting from errors. Also, the extent of clinical harm arising from medication error and DDIs in patients taking ARVs admitted to hospital; the proportion of patients taking medication that their HIV doctors are not aware of; and the extent of DDI recognition by prescribers.
Semi-structured interviews and focus groups will explore attitudes of healthcare workers to the causes of and barriers to medication errors and the acceptability of potential interventions. healthcare utilisation resulting from medication error and unmanaged DDIs will be described and reported. Recommendations to reduce and prevent harm from errors will be developed. Studies of medication error in African ARV programs are lacking, and no data exist on the extent of harm caused by error and unmanaged DDIs, although impact on patients and programs is likely to be significant.

Direct patient benefit is expected during the project. Specifically, harms may be averted, for example resulting from un-managed drug interactions or administration error. If left undetected, these kinds of errors may precipitate irreversible viral resistance to one or more antiretroviral, significantly reducing future treatment options for patients. Full medication histories and details of drug interactions will be filed in the clinical notes of all outpatients. Long term benefits may arise from increased awareness and vigilance for drug interactions and prescribing errors in all settings.

Local institutions will be made aware of local system factors which may promote or protect from prescribing errors. This will inform staff training and allow systems to be improved and developed. Awareness and vigilance for medication error, drug interactions and adverse drug reactions will be promoted. This may encourage an environment of assessing practice and local procedure and internal audit to inform improvement of services. I have discussed the proposed research with the Head of Prevention, Care and Treatment Program at the Infectious Diseases Institute, Kampala, who is keen to support medication safety research and if successful, disseminate information to national policy makers. I have discussed with the Head of Information Services at the Infectious Diseases Institute the future potential to develop systems which reduce medication error, such as electronic prescribing and alerts.
Local pharmacists will be involved in data collection, and gain research skills. Clinicians will be asked to give feedback and opinion on the utility of a medication safety intervention, and their views and input will inform recommendations for development and optimisation of interventions.

The national HIV program is expected to benefit, as appropriate management of drug interactions, and adherence support for patients who need it, will result in reduced treatment failure, less resistance to therapy, and therefore fewer patients progressing to second line therapy at greater cost. The proposed research is expected to inform HIV policy, and its interaction with other major health programs for example family planning, infectious and non-infectious diseases. Engaging government agencies is important for this work, in order to tackle issue of medication error openly. This represents an important advance for capacity building, as in many countries the issue of medication error and quality of care is not acknowledged or prioritised.

The national pharmacovigilance program will be strengthened, as all adverse drug reactions observed during the study will be reported to the Ugandan National Drug Authority (NDA) as per national policy. Long term sustainable benefits are expected, as the study will promote adverse reaction reporting among local clinicians and health centres. Post analysis, data sharing with the NDA will ensure that they have data on adverse events occurring in antiretroviral programs, and those associated with medication error and drug interactions. Initial meetings with the drug information and pharmacovigilance sections of the NDA have confirmed support for the proposed project. I am in contact with the Uppsala Monitoring Committee, particularly UMC-Africa who specialise in pharmacovigilance in Africa and specific cohorts such as patients taking ARVs.