Access to medicines and diagnostic tests integral in the management of diabetes mellitus and cardiovascular diseases in Uganda: insights from the ACCODAD study

The ACCODAD study was conducted from 15th January 2017 to 28th February 2017 in 22 public hospitals, 23 private hospitals and 100 privately owned pharmacies. The health units were selected from each of the 4 regions of Uganda (central, western, eastern and northern) using random sampling method from the hospital and private pharmacy registries of the Ministry of Health and National Drug Authority (NDA), Republic of Uganda respectively.

The total number of public and private hospitals in Uganda is 155. Two of these are national referral hospitals, 14 are regional referral hospitals and 139 are general hospitals. In terms of ownership, 65 are government owned, 63 private not for profit (PNFP) and 27 are private for profit [22]. The public hospitals offer free medical care to all patients. They procure all their drugs, laboratory tests and medical equipment from one central national procurement institution called the National Medical Stores (NMS). The NMS purchases the essential drugs and diagnostic tests from qualified private suppliers through a locally publicised tender process. Nevertheless, recurrent drug stock outs and unavailability of key diagnostic tests remains a key challenge in the public hospitals. This compels patients to seek medical treatment from the costly privately owned hospitals, clinics and pharmacies. These procure their medicines and diagnostic tests from several private distributors. The NDA registry has a total of 599 registered privately owned retail pharmacies and 90 private hospital pharmacies dealing in human medicines. The majority of the private pharmacies (>70%) are located in the central region [23].

Basing on one of the primary objectives of the ACCODAD study i.e. to determine the availability of the medicines and diagnostic tests of interest, the availability of 4 key medicines in DM and CVD management (intermediate insulin or insulitard®, losartan, simvastatin and isosorbide mononitrate) of 10% as reported by the study performed in Western Cameroon was used as the prevalence (P) [6]. Using the formula: n = Z2P (1-P)/d2 where Z (normal value corresponding to the 95% confidence interval) =1.96, P = 0.1 and d = 0.05 (desired precision of estimation), a sample size of 138 health units (hospitals and private pharmacies) was obtained. The study sample size was however, increased to 145.

For the ACCODAD study, we collected information about 37 medicines and 19 diagnostic tests significant in the management of DM and CVD as highlighted by the Ugandan local guideline and several international guidelines as highlighted below. The CVD of interest were hypertension, coronary artery disease, stroke, dilated cardiomyopathy, rheumatic heart disease, atrial fibrillation, peripheral arterial disease, venous thromboembolism and related complications like heart failure.

The selected medicines of interest are part of the WHO essential medicines list for treatment of chronic diseases in LMICs [24] and are recommended in the management of DM and CVD by the 2012 Uganda clinical guidelines [25], the 2017 American Diabetes Association guidelines of standard of care of DM and related CVD [22] and the recent European Society of Cardiology (ESC) guidelines of management of atrial fibrillation, acute myocardial infarction, acute heart failure [26‐29].

The respective medicine categories and medicines of interest were oral hypoglycaemic agents (OHA) which included metformin 1 g and 500 mg, glibenclamide 5 mg, glimepiride 2 mg and pioglitazone 30 mg, angiotensin II receptor blockers (ARB) which included losartan 50 mg and telmisartan 40 mg, angiotensin converting enzyme inhibitors (ACEI) which included captopril 25 mg, thiazide diuretics (D) which included bendrofluazide 5 mg, ARB-D which included losartan-hydrochlorothiazide 50/12.5 mg and telmisartan-hydrochlorothiazide 40/12.5 mg, loop diuretics which included oral furosemide 40 mg and i.v. furosemide 20 mg, calcium channel blockers (CCB) which included nifedipine 10 mg and 20 mg and amlodipine 5 mg and 10 mg, statins which included simvastatin 20 mg, atovastatin 20 mg and rosuvastatin 10 mg, anti platelet drugs which included cardiac aspirin 75 mg and Clopidogrel 75 mg, low molecular weight heparin which included enoxaparin 60 mg and 80 mg and beta blockers which included atenolol 50 mg, bisoprolol 5 mg, nebivolol 5 mg and carvedilol 6.25 mg. Other surveyed medicines included: parenteral benzathine penicillin, oral digoxin, warfarin, spironolactone, hydralazine, soluble insulin, intermediate insulin, pre mixed insulin, isosorbide mononitrate and unfractionated heparin (UFH).

The selected diagnostic tests are part of the WHO minimum workup tests for assessment and management of cardiovascular (CV) risk [24] and are also recommended by the 2017 American Diabetes Association guidelines of standard of care of DM and related CVD [22] and the recent European Society of Cardiology (ESC) guidelines of management of atrial fibrillation, acute myocardial infarction, acute heart failure [26‐29]. These included: lipid profile, glycated haemoglobin (HbA1c), serum uric acid, serum troponin, coagulation profile, thyroid function tests, serum creatinine, serum urea, serum electrolytes, serum ketones, microalbuminuria, complete blood count, serum natriuretic peptides, electrocardiography (ECG), echocardiography (ECHO), chest X ray, liver function tests and urinalysis.

Data was collected using a pre tested questionnaire based on the WHO and HAI standardised methods of assessing medicine prices, availability and affordability in LMIC [21] from 15th January 2017 to 28th February 2017. The data collection team underwent a brief training before commencement of the study to improve quality and standardisation of data.

Information about the availability of diagnostic tests and any medicine in the respective medicine category was obtained. The cost of performing each diagnostic test and the monthly cost of the recommended dose of the available lowest priced generic (LPG) medicine was obtained in Uganda shillings (UgX) and then converted to US dollars (USD) using the existing exchange rate at the time of data collection (1 USD = 3600 UgX). The obtained costs of the medicines were the retail prices charged directly to the patients at the respective pharmacies of the private hospitals and private pharmacies. The cost of the medicines in the public hospitals was not obtained since medical care is offered free of charge.

The obtained unit retail prices of the medicines in USD were converted to a median price ratio (MPR) by dividing the median local price by an international reference price (IRP). The IRP is obtained from the Management Sciences for Health International Drug Price Indicator Guide which reports median prices of high quality multisource medicines offered to LMIC countries by different suppliers. The MPR is used to express how much greater or less the median local medicine price is than the IRP. An MPR of 3 would mean that the local medicine price is three times greater than the IRP. For patients’ medicine prices, MPR ≤ 1.5 were considered reasonable pricing [30].

Affordability was estimated by calculating the number of days’ wages required to purchase a one month course of treatment or pay for a specific diagnostic test using the average salary of the lowest paid government worker in USD. Medicines and diagnostic tests that cost ≤3 days’ wages were considered affordable. The exchange rate of the local currency (UgX) to USD used was the commercial “buy” rate at the time of data collection of 1 USD = 3600 UgX. The lowest paid government worker at the time of the study (scale U8 lower-non formal education teachers) earned a gross salary of 198,793 UgX (USD 55.2). After tax deductions, this translated to a net salary of 139,155 UgX (USD 38.7) per month or 4638.5 UgX (1.3 USD) daily [31].

Low, moderate and high availability was documented in 5 (23.8%), 3 (14.3%) and 13 (61.9%) of the 21 surveyed medicines categories respectively. The availability of the surveyed medicine categories ranged from 20.1% for unfractionated heparin (UFH) to 100% for OHA. High availability was noted for the majority of the key medicine categories in the management of hypertension and cardiac diseases i.e. ARBs, ACEI, D, CCB, statins, anti platelet drugs and beta blockers. None of the insulin types was of high availability. Soluble, intermediate and pre mixed insulin was available in 68.8, 34.7 and 60.1% of all the study sites respectively. Isosorbide nitrate, UFH and LMWH were all of low availability (summarised in Table 1).

With regard to the 19 diagnostic tests of interest, 8 (42.1%) were of low availability, 3 (15.8%) were of moderate availability and 8 (42.1%) were of high availability. The availability ranged from 6.8% for microalbuminuria to 100% for urinalysis. Apart from electrocardiography (ECG) and lipid profile testing which were available in only 54.6 and 65.9% of the study hospitals, the rest of the recommended WHO minimum tests for DM and CVD workup were of high availability (random blood glucose tests-97.7%, serum electrolytes-88.6% and urinalysis-100%). Glycated haemoglobin (HbA1c) tests, a key test in DM diagnosis and monitoring of glycaemic control in diabetes care was available in only 43.2% of the study hospitals. The majority of vital tests in cardiac evaluation (echocardiography, coagulation profile, serum natriuretic peptides and troponin tests) were of low availability (summarised in Table 1).

There were significant differences in the availability of key medicines documented in the public hospitals and the private hospitals and pharmacies. Low availability was noted for these medicine categories in public hospitals compared to the private hospitals and pharmacies: ARBs, ARB-thiazide diuretics, statins, warfarin, LMWH, UFH and nitrates (summarised in Table 2).

With regard to the diagnostic tests, a statistically significant difference in the availability of lipid profile, HbA1c, uric acid, troponin, coagulation profile and thyroid function tests was noted in the surveyed public hospitals compared to the private hospitals. All the documented tests were of low availability in the public hospitals. Tests for serum ketones, microalbuminuria, serum natriuretic peptides and ECHO were of low availability regardless of the study site (summarised in Table 3).

With regard to pricing as reflected by the MPR, the only reasonably priced medicines were parenteral benzathine penicillin (1.2), losartan 50 mg (0.8) and amlodipine 10 mg (1.5). The MPR ranged from 0.8 for losartan 50 mg to 11.1 for simvastatin (summarised in Table 4). With the exception of Glucophage® (metformin) 1 g, Insulitard® (intermediate insulin) and Mixtard® (pre mixed insulin), all of the available originator medicine brands cost more than the available LPG medicine brands. One originator brand (Adalat 30 mg) cost up to 10 times the cost of the available LPG brand (nifedipine 20 mg) (summarised in Table 5).

In our study, low and moderate availability (availability of <80% in all study sites) was reported in 38.1% of the medicines and 57.9% of the diagnostic tests of interest. Several similar studies assessing access to medicines and diagnostic tests of NCDs in LMIC have reported similar findings of low availability of medicines and diagnostic tests especially in the public sector [5‐10].

In the study reported from Western Cameroon, high availability defined as availability ≥80% was only noted with 6 (27%) of the surveyed medicines (parenteral benzathine penicillin 2.4 MU, oral furosemide 40 mg, glibenclamide 5 mg, Actrapid/soluble insulin, metformin 500 mg and Mixtard). The majority of rural study sites had low availability of medicines [6]. In comparison with our study, with the exception of the soluble and pre mixed insulin, high availability of >80% was documented with parenteral benzathine penicillin, oral furosemide, glibenclamide and metformin.

In another study that assessed the availability, pricing and affordability of 5 key cardiovascular medicines (atenolol, captopril, hydrochlorothiazide, losartan and nifedipine) in 36 LMIC (Uganda inclusive), upper middle income and high income countries found an overall poor availability of these medicines. Only 26.3 and 57.3% of the medicines were available in the public and private sector respectively [7]. In another similar multicentre study involving 90 primary care facilities in 8 LMIC, some of the 12 surveyed CVD and DM medicines were not available in some countries. Soluble and long acting insulin was absent in all study sites in Benin, Eriteria, Bhutan and Vietnam. Isosorbide mono nitrate was absent in Benin, Eriteria, Sudan and Bhutan and simvastatin and amlodipine were absent in Eriteria and Bhutan [10]. Low availability of nitrates, intermediate and pre-mixed insulin was also reported by our study (27.8, 34.7 and 60.1% respectively).

With regard to the availability of diagnostic tests, cross sectional studies in LMIC had reported similar findings of low availability [5‐9]. In one of these studies performed in the Western Cameroon in 2012, high availability defined as availability ≥80% was only noted with 50% of the surveyed diagnostic tests (RBG, urinalysis, serum creatinine, serum urea and CBC). Serum electrolytes, lipid profile and uric acid tests, HbA1c tests and ECG were only available in 60, 40, 20 and 10% of all the study sites [6].

Another multicentre study performed in 90 primary care centres of 8 LMIC (Benin, Bhutan, Eritrea, Sri Lanka, Sudan, Suriname, Syria, and Vietnam), lipid profile testing was available only in 33, 25, 20, 14 and 8% of all study sites in Sudan, Benin, Suriname, Syria and Sri Lanka. Low availability of serum creatinine tests was also reported in the majority of the study sites. Of all study sites, serum creatinine tests were available in 58% in Sudan, 33% in Benin, 10% in Suriname and 8% in Sri Lanka. Lipid profile tests were absent in Eriteria and Vietnam while serum creatinine tests were absent in all study sites in Eriteria, Bhutan and Syria. Serum troponin tests were absent in all the countries except Benin and Sudan where it was available in only 8% of the study sites [10].

Comparing with our study, we also reported similar findings of low availability of uric acid, HbA1c, ECG and troponin tests (all <60%).

The majority of surveyed medicines and diagnostic tests have also been reported to be unaffordable in most studies in LMIC, similar to the findings of our study. Our study findings documented that only 27% of the surveyed medicines and 32% of the surveyed diagnostic tests were affordable in Uganda.

In comparison, only 7 (32%) of the studied medicines (oral aspirin 500 mg, hydrochlorothiazide 50 mg, furosemide 40 mg, nifedipine 10 and 20 mg, glibenclamide 5 mg and metformin 500 mg) were affordable according to the study definition of affordability i.e. monthly cost of a medicine of ≤1 days’ wages of a lowest paid public servant in the study performed in Western Cameroon [6]. The most unaffordable medicines in this study were Mixtard, simvastatin 20 mg and heparin 5000 IU costing 18.7 days’ wages, 30.5 days’ wages and 182.36 days’ wages respectively [6]. In our study, only 6 (16%) medicines cost less than a days’ wages (parenteral benzathine penicillin, oral furosemide, glibenclamide, bendrofluazide, atenolol, and cardiac aspirin).

In another multicentre study by Mendis S et al. in 6 LMIC (Bangladesh, Malawi, Sri Lanka, Brazil, Nepal and Pakistan) assessing availability and affordability of 32 medicines essential in CVD, DM, asthma, glaucoma and palliative care, anti hypertensive management using hydrochlorothiazide monotherapy and glycaemic therapy using either glibenclamide or metformin was affordable in all the 6 countries (cost ≤1 days’ wages). Using intermediate insulin for DM management was unaffordable with a monthly cost equivalent to 2.8 days’ wages in Brazil, 4.7 days’ wages in Pakistan, 6.1 days’ wages in Sri Lanka, 7.3 days’ wages in Nepal and 19.6 days’ wages in Malawi. Secondary prevention of CVD with an oral aspirin, statin, beta blocker and ACEI in this study was also unaffordable in some countries. This combination using a generic medicine would cost ≤1.6 days’ wages in Sri Lanka and Bangladesh, ≤ 6.1 days’ wages in Brazil, Pakistan and Nepal and 18.4 days’ wages in Malawi. The cost of treatment in Malawi would increase to 48.8 days’ wages if an innovator statin brand was used [5]. In our study, the monthly cost of combination therapy used in secondary CVD prevention using the cheapest beta blocker, ACEI, statin and cardiac aspirin was 14.1 days’ wages; almost similar to the cost in Malawi.

Regarding diagnostic tests, only urinalysis and random blood glucose tests cost less than 1.5 days’ wages in the study by Jingi A et al. in Western Cameroon. Other key tests recommended by the WHO in optimal DM and CVD workup in this study like lipid profile tests cost 3.1–3.6 days’ wages and ECG cost 10.7 days’ wages. Glycated haemoglobin (HbA1c), an important test in diagnosis and monitoring of glycaemic control in diabetes care cost 12.6 days’ wages [6]. Comparing to our study findings, serum electrolytes, lipid profile test, HbA1c test and ECG cost 5.3 days’ wages, 7.5 days’ wages, 8.6 days’ wages and 10.7 days’ wages respectively.

There are several plausible explanations for the low availability of medicines and diagnostic tests especially in the public hospitals in our study. Some of these key medicines are not included in the 2012 national essential medicine list and the 2016 Uganda Clinical guidelines of management of NCDs. Inadequate allocation of funds to the health sector to enable procurement of most essential medicines and diagnostic tests, poor stock maintenance, forecast inaccuracy and inefficient distribution systems from the national central procurement institution could also explain the low availability in the public and private sectors.

The study finding of the majority of medicines and diagnostic tests being unaffordable in the private sector could be explained by the lack of local legislation to regulate the maximum retail prices of medicines and a vibrant local pharmaceutical industry sector to produce cheap quality generic medicines for chronic diseases.

Being a point in time study, variations in availability, pricing and affordability of the medicines and diagnostic tests was not put into consideration. We were unable to obtain the procurement prices from the central procurement institution, NMS to obtain the prices of the medicines and diagnostic tests in the public hospitals. Using the daily wage of the lowest paid unskilled government to calculate affordability of medicines and diagnostic tests has its limitations because a significant proportion of the Ugandan population earns less than this amount. Despite these limitations, the study has its strengths. The standardised WHO/HAI methodology that was used has been widely validated.