|Year : 2018 | Volume
| Issue : 2 | Page : 45-55
Access to diabetes medicines at the household level in eight counties of Kenya
Selam Hailu1, Peter C Rockers1, Taryn Vian1, Monica Onyango1, Richard Laing2, Veronika J Wirtz1
1 Department of Global Health, Boston University School of Public Health, Boston, Massachusetts, USA
2 School of Public Health, Faculty of Community and Health Sciences, University of the Western Cape, South Africa
|Date of Web Publication||10-May-2018|
Dr. Veronika J Wirtz
Crosstown Center, Boston University School of Public Health, Room CT-363, 801 Massachusetts Avenue, Boston, Massachusetts 02118
Source of Support: None, Conflict of Interest: None
Background: In 2016, an estimated 872,000 Kenyans were living with diabetes, a country average of 4%. The study objectives were (1) to describe the sociodemographic and geographic characteristics of the households with individuals diagnosed and on treatment for diabetes (2) to describe the medicines available at the household level, monthly household expenditure on medicines, location of diagnosis and treatment and the associated factors of medicines purchase location. Methods: A household survey in eight countries was conducted asking whether a household member had been diagnosed and treated for a non-communicable disease (NCD). Households with at least one member with diabetes were included in this study. Results: Out of the 142 individuals being diagnosed and treated for diabetes, 68 participants (47.9%) were prescribed single and 74 (52.1%) multiple treatments. While 54.9% of the participants were diagnosed at public hospitals, 50% of individuals purchased their medicines from a private pharmacy/chemist or private hospitals. Purchase of medicines in public facilities was associated with being less wealthy and having more than one NCD. Having medicines not available at home was reported by 26.1% of individuals, mostly because the medicines were too expensive to buy. Conclusions: Affordability of diabetes medicines remains an important barrier to access. In addition, essential medicine list restrictions to offer diabetes medicines at public primary care level limit access. Programs to increase access to NCD medicines need to consider that diagnosis and choice of treatment occurs largely in the public sector whereas medicines purchase most frequently takes place in the private sector.
Keywords: Access, diabetes, household survey, Kenya, medicines
|How to cite this article:|
Hailu S, Rockers PC, Vian T, Onyango M, Laing R, Wirtz VJ. Access to diabetes medicines at the household level in eight counties of Kenya. J Diabetol 2018;9:45-55
| Introduction|| |
Diabetes is the sixth leading cause of death worldwide. According to the 2016 World Health Organization (WHO) global report, 422 million people were living with diabetes in 2014, and an additional 130 million people are expected to be affected by 2030., One in ten people are projected to have diabetes by 2040. An estimated 46.5% of adults living with diabetes are undiagnosed. Type 2 diabetes accounts for about 95% of all diagnosed cases of diabetes while type 1 accounts for only 5%.
Diabetes has been increasing rapidly in low- and middle-income countries (LMIC). The World Bank reported that 4% of the population in Sub-Saharan Africa ages 20–79 had diabetes in 2015. It is estimated that by 2035, 42 million people in Africa will become diabetes patients with 4.9 million deaths every year. The highest growth rate of diabetes in the world occurs in Africa. Risk factors for diabetes include increase in urbanization, population age, sedentary lifestyles and dietary changes.
Kenya is a LMIC with a per capita GDP of $1,376 a year. Compared to 1990, the prevalence of diabetes in Kenya tripled in 2015. Currently, an estimated 872,000 Kenyans are living with diabetes. Gender-based differences in the prevalence have also been reported with more females (4.2%) than males (3.8%) being affected making a country average of 4% in 2016. Large differences were reported between rural and urban as well as ethnic groups. With the prevalence of diabetes on the rise, patients are at greater risk from eye, renal, cardiovascular, diabetic foot ulcer and nerve complications. Many of the patients with these conditions who are referred to specialised and national referral hospitals are patients with diabetes.
Poor socioeconomic status may be associated with higher prevalence of diabetes. Studies suggest that the prevalence of diabetes may be two-fold higher in poor socioeconomic populations when compared with wealthy populations. Poor socioeconomic status complicates the treatment of diabetes in LMIC like Kenya. Most patients are unable to afford treatment even with the government efforts to subsidize diabetes healthcare.
Background on diabetes treatment in Kenya
The health system in Kenya is divided into six levels: Hospitals (central, provincial and district) are level 6, 5 and 4, while health centres and dispensaries are levels 3 and 2. The Kenya Essential Medicines List (KEML) 2016 includes metformin tablet 500 mg, glibenclamide tablet 5 mg, gliclazide tablet 40 mg; insulin, intermediate-acting (human) (70/30) injection 100 IU/ml (10 mL vial) and insulin, soluble (human) injection 100 IU/ml (10 mL vial). The KEML and the Kenya Standard Treatment Guidelines (STGs) are inconsistent because the KEML recommends for the treatment of diabetes at hospitals (level 4) whereas the STGs recommend oral therapy at levels 2 and 3 which includes primary care facilities.,
Studies on health facilities price and availability of diabetes medicines in Kenya provide information on access including affordability of treatment for diabetes and factors associated with such access.,, However, there is a lack of information on the diagnosis and treatment at the household level in Kenya. In 2015, Novartis launched Novartis Access, an initiative to provide low-cost non-communicable disease (NCD) medicines in Kenya. This initiative offers a basket of low-cost NCD medicines to program countries to be delivered through public and non-profit health sectors. Kenya is the first country where this scheme is being implemented. The baseline study for the evaluation of this initiative provided an opportunity to assess diagnosis and treatment of diabetes in Kenya. The objectives of this study were (1) to describe the sociodemographic and geographic characteristics of the households with individuals diagnosed and on treatment for diabetes; and (2) to describe the types of medicines available at the household level, monthly household expenditure on medicines, location of diagnosis and treatment and the associated factors of medicines purchase.
| Methods|| |
The methods used in this survey have previously been described by Rockers et al. in 2016. The households studied are part of a baseline assessment of the Novartis Access initiative launched by Novartis in 2015. Eight hundred households were randomly selected from eight counties using a two-stage sampling method. For the purpose of this study, a household was defined as an individual or a group of individuals living together in the same unit, making common provisions for food or pool their income for the purpose of purchasing food. For the first-stage sampling, ten enumeration areas (EAs) were randomly selected from each county with probability proportional to size based on the data from the 2009 census. In the second stage, ten households that met the eligibility criteria were randomly selected from each EA to be included in the study. To do this, all households in each EA were listed in a random order and enumerators proceeded down the list until ten eligible households were selected. Households with at least one adult (aged 18 or older) who reported having been diagnosed with an NCD (hypertension, asthma, breast cancer or diabetes) and prescribed medication were eligible to participate. All members of the household that met this criterion were recruited. Based on urban/rural as defined by Kenya Statistical Office, 78.1% of the households included in this survey were located in rural areas while 21.8% were from urban areas. Out of these, NCD patients, those patients diagnosed and treated for diabetes, were selected for further study as reported in this paper.
Fifty data collectors were trained in all aspects of NCD diagnosis and treatment relevant to the study. The training covered the use of GPS locators to list households and the data collection software application and the SurveyCTO (SurveyCTO is the name of the survey platform for electronic data-collection, based on Open Data Kit [ODK]). During training, data collectors rehearsed using study instruments (programmed in SurveyCTO on tablets) to collect data with each other to ensure their familiarity with the study instrument. After training, data collectors did a pilot data collection with NCD patients in Nairobi to further enhance their data collection skills. The training and piloting in Nairobi took 2 weeks, after which there was another round of piloting in the eight study counties to test the SurveyCTO data transfer system. Two rounds of pilot testing the instruments and methods were undertaken. Field supervisors reviewed data collected each day in the evenings.
Study procedures and variables
Household data were collected using structured questionnaires. Survey instruments were adapted from previously validated household survey instruments investigating access to medicines.
The household questionnaire included information on household assets, demographics and NCD and other medicines (i.e., whether prescribed medicines were found at the household at the time of the interview, locations where medicines had been purchased, prices paid for medicines converted to US dollars at the August 2016 exchange rate, and overall household expenditures on medicines, health and other goods). We assessed wealth based on household assets. The Kenya DHS (https://dhsprogram.com/Publications/Publication-Search.cfm?ctry_id=20&country=Kenya) was used as input to create the survey instrument asking about household possessions. The wealth index was then calculated based on the information of household's ownership of the selected assets included in the survey (for example, radio, bicycle, cellphone and refrigerator). The household questionnaire was translated into vernacular languages as needed. A total of 7,870 households were visited for the survey out of which 794 households reported diagnosis and treatment for at least one NCD.
This analysis focused on the subsample of 142 individuals with a reported diagnosis and treatment of diabetes out of the total study sample of 794 households with at least one NCD. Single treatment was defined as one diabetes medicine while multiple treatments were defined as more than one diabetes medicine. Variables were created for age, gender and wealth based on household items observed. The variable 'distance' expressed in kilometres was calculated using the GPS coordinates of households and the nearest public and non-profit facilities. Percentages were calculated for whether participants were taking one versus multiple medicines, the availability of medicines in the home by type of medicine, location of diagnosis and purchase of medicines by provider type (public hospitals or clinics, faith-based hospitals or clinics, private for-profit hospitals and pharmacy/chemist). The responses to the question about why they did not have medications at home were coded by the interviewer into one of the following categories: available but too expensive, ran out of stock at home, managing without medication and lifestyle changes and using herbal medicines instead. Some of the categories where pre-defined, others emerged from the data. Household expenditures for medicines were defined as total spending on medicines in the previous 4 weeks divided by the number of household members (expenditure per person). In case, there was more than one household member with diabetes we excluded these households in the analysis of medicines expenditure. Using a logistic regression model, we explored the relationship between patient characteristics and the probability that the patient bought the diabetes medicines at a public facility. All analysis was done using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).
| Results|| |
Out of the 142 (17.9%) individuals who reported having been diagnosed with diabetes and prescribed antidiabetic medicines, 58.4% were female and 50.7% were aged 60 years and older [Table 1]. Only four respondents (2.8%) were between the ages of 18 and 29 years.
Fifty-seven participants (40.1%) had less than primary school education and 36 participants (25.3%) had completed primary school. Forty-nine participants (34.5%) had education levels of secondary school or higher. Educational attainment varied by county [Annexure 1]: West Pokot county had ten participants (76.9%) with higher than secondary school education while Kwale county had only one participant (8.3%) at this level. Overall, the mean distance from the households to the nearest public health facilities was 11.6 km which was longer than the nearest distance to non-profit private facilities (9.0 km).
Out of the 142 individuals, 68 participants (47.9%) reported being prescribed a single treatment while 74 individuals (52.1%) were receiving multiple treatments [Table 2]. Narok, Embu, West Pokot and Samburu counties represented most of the individuals receiving a single treatment while individuals from Nyeri (64.3%), Kakamega (77.3%) and Makueni (71.4%) counties reported receiving multiple treatments [Annexure 2]. The median household “expenditure” was US$1.60 per person per months (US$1.30 in rural areas and US$2.00 in urban areas) [Table 2].
More than half (54.9%) of the participants were diagnosed at public hospitals while 26.1% were diagnosed at private for-profit hospitals [Figure 1]. Only 9 participants reported being diagnosed at the primary care level in the public sector. Over 50% of participants obtained their medicines from a private pharmacy/chemist or private hospitals while 40.8% received medicines from the public-sector facilities [Figure 1]. About 71% of individuals reported receiving metformin followed by glibenclamide (42.9%) and insulin (10.9%) [Table 3].
|Table 3: Probability of having specific medicines among diabetes patients reporting at least one medicine in the home (n=105)|
Click here to view
Thirty-seven individuals (26.1%) reported they did not have the medicines available at home at the time of the interview [Table 4]. The availability of diabetes treatment varied among counties. [Annexure 3] shows county-specific data. Five of these participants (35.7%) reported medicines were too expensive to buy, three participants (21.4%) said they had run out of medicine and two individuals (14.3%) reported they were managing their diabetes with diet and lifestyle changes [Table 4]. One participant reported using herbal medicines instead of conventional medicines while another said their condition did not require taking medicines, possibly due to gestational diabetes. Nine of the 23 who reported not having any diabetes medicines in their home did not provide an explanation.
Purchase of medicines at public sector facilities was associated with being less wealthy, having been diagnosed with more than one NCD and being male [Table 5]. Education level, urban or rural location of housing, the distance to the facility and number of household members with NCDs were not associated with the location of purchase.
|Table 5: Relationship between patient characteristics and the probability that patient bought medicines at a public facility|
Click here to view
| Discussion|| |
The survey on the prevalence and treatment of diabetes in Kenya set out to describe the sociodemographic and geographic characteristics of the households with individuals diagnosed and on treatment for diabetes. The study also describes the availability and types of medications being used, monthly household expenditure on medicines, location of diagnosis and treatment and reasons for not having medicines available.
The findings of this study show wealth have an effect on diagnosis and treatment of diabetes. The majority of individuals from Nyeri, Kakamega and Makueni counties reported receiving multiple diabetes medicines which might be associated with the higher wealth status of these counties.
Availability of medicines for the treatment of diabetes at the household level is an issue in the eight counties. About one in four individuals (26.1%) did not have diabetes medicines available at home. The most frequently used antidiabetic medicines in Kenya are metformin tablets 500 mg and glibenclamide tablets 5 mg which conforms with Kenya STGs. The highest availability of insulin at home was found in Narok (21.1%) and Nyeri (14.3%) counties which could be associated with the wealth status of these counties.
One challenge for diabetes treatment is that Kenyan standard treatment guidelines (STGs) contain inconsistencies and do not always adhere to international best practice. First, there is a discrepancy between the KEML (2016) and the Kenyan Clinical Guidelines for management and Referral of Common Conditions (2009) at levels 2–3: primary Care. The STGs incorporates the use of oral hypoglycaemic agents and insulin for the management of diabetes at primary care level while the KEML 2016 sets the level of diabetes medicine supply at level 4 (hospitals) of the health system.,, [Annexure 4],[Annexure 5],[Annexure 6],[Annexure 7].
The discrepancies between the EML and STGs do not appear to have affected how people are treated as most type 2 diabetes patients receive the oral hypoglycaemic medicines metformin and glibenclamide. However, the finding that most of the patients are diagnosed in the public sector (54.9%) while they obtain their medicines in the private sector may be affected by the KEML level setting for these medicines being at level 4 (hospitals). To facilitate access to these key oral medicines, the KEML should be changed to match the Kenya STGs and make diabetes medicines available at level 2 and 3. However, we do not recommend adding the many outdated medicines on the Kenya STGs (e.g., chlorpropamide or tolbutamide) to the KEML. This list should be guided by the WHO model EML. Most of the patients diagnosed in the public sector reported diagnosis at hospital level. The reason for the low number of patients diagnosed at primary care facilities may be due to the lack of NCD diagnostic capacity at this level of care. However, the latest Service Availability and Readiness Assessment Mapping (SARAM) conducted in 2012 did not assess the diagnostic capacity of diabetes at primary care level (SARAM, 2013).
Our findings are similar to the previous studies which document that access to diabetes treatment in Kenya is higher through the private sector. Novartis Access aims to increase availability and decrease costs of medicines in the public sector. In order for Novartis Access to have an effect, it is necessary that stock availability in the public sector to be improved and that patients switch from private sector to public sector purchase.
The household monthly expenditure on medicines per person per month varied from $1.30 in rural areas to $2.00 in urban areas. While this spending would include expenditure on other medicines such as analgesics and antimalarial, this regular spending can impoverish families. Affordability is mentioned most frequently as a reason for not having the medicines available at households, as noted in other studies such as the previous pricing surveys done by HAI which indicate similar results. Affordability of diabetes medicines, especially those in the private sector, is clearly an obstacle.
The household survey focuses on those individuals that have been diagnosed with at least one NCD and prescribed medication. The methodology excludes a segment of the population that had not been diagnosed. Therefore, data on the prevalence of diabetes cannot be acquired from this survey. In addition, the number of non-respondents for the qualitative survey that set out to understand why patients did not have antidiabetic medicines at home was relatively high. As a result, information regarding factors associated with not having medicines available at home might be missing in this report.
| Conclusions and Recommendations|| |
Diabetes is affecting more people every day, especially in LMIC. This study reports the availability of antidiabetic medicines in eight counties in Kenya. Barriers to treatment include affordability and discrepancies in guidelines which may be limiting access. The KEML should be revised to be in line with the WHO Model List of Essential Medicines and to make diabetes treatment available at the primary care level, while the National Clinical Guidelines for the management for diabetes need to be updated to match current international treatment guidelines and the Kenyan EML. Offering glibenclamide through Novartis Access may enhance access to available and affordable medicines in the future.
We would like to thank Ela Fadli for her support in the data analysis.
Financial support and sponsorship
VJW and ROL are the principal investigators of Novartis Access which is funded by Sandoz International GmbH. PCR, TV and MO are coinvestigators of the study.
Conflicts of interest
SH declares no conflict of interest.
| References|| |
Sagna Y, Tieno H, Guira O, Yanogo DAR, Benon LE, Zida S,et al
. Prevalence and associated risk factors of diabetes and impaired fasting glucose in urban population; a study from Burkina Faso. J Diabetol 2014;2:4.
IDF Diabetes Atlas. 7th
ed. Belgium: International Diabetes Federation. Available from: http://www.diabetesatlas.org/
. [Last accessed on 2017 Feb 08].
Mwavua SM, Ndungu EK, Mutai KK, Joshi MD. A comparative study of the quality of care and glycemic control among ambulatory type 2 diabetes mellitus clients, at a tertiary referral hospital and a regional hospital in central Kenya. BMC Res Notes 2016;9:12.
Kassahun CW, Mekonen AG. Knowledge, attitude, practices and their associated factors towards diabetes mellitus among non diabetes community members of bale zone administrative towns, South East Ethiopia. A cross-sectional study. PLoS One 2017;12:e0170040.
Azevedo M, Alla S. Diabetes in sub-Saharan Africa: Kenya, Mali, Mozambique, Nigeria, South Africa and Zambia. Int J Diabetes Dev Ctries 2008;28:101-8.
Christensen DL, Friis H, Mwaniki DL, Kilonzo B, Tetens I, Boit MK,et al
. Prevalence of glucose intolerance and associated risk factors in rural and urban populations of different ethnic groups in Kenya. Diabetes Res Clin Pract 2009;84:303-10.
Amin A, Karagu A, Abdikamal A, Bhatti L, Kendagor A, Melanie C,et al
. MoH-Kenya. Kenya STEPwise Survey for Non-Communicable Diseases Risk Factors 2015 Report. 1st
ed. Kenya, Nairobi: 2015. p. 10-1.
Rabi DM, Edwards AL, Southern DA, Svenson LW, Sargious PM, Norton P,et al
. Association of socio-economic status with diabetes prevalence and utilization of diabetes care services. BMC Health Serv Res 2006;6:124.
Ministry of Health. Kenya Essential Medicines List. Nairobi, Kenya: Ministry of Health; 2016. p. 57.
Clinical Management and Referral Guidelines: Clinical Guidelines for Management and Referral of Common Conditions at Levels 2–3: Primary Care (Republic of Kenya). Vol. 2. Ministry of Medical Services and Ministry of Public Health and Sanitation; 2009. Available from: http://www.apps.who.int/medicinedocs/en/d/Js20999en/
. [Last accessed on 2017 Mar 23].
Clinical Management and Referral Guidelines: Clinical Guidelines for Management and Referral of Common Conditions at Levels 4–6: Hospitals (Republic of Kenya). Vol. 3. Ministry of Medical Services and Ministry of Public Health and Sanitation; 2009. Available from: http://www.apps.who.int/medicinedocs/en/d/Js21000en/
. [Last accessed on 2017 Mar 23].
Ministry of Health. Monitoring Medicine Prices and Availability in Kenya, Summary Report. Ministry of Health; 2007. p. 5-6.
Database of Medicine Prices, Availability, Affordability and Price Components. Health Action International; 2004. Available from: http://www.haiweb.org/MedPriceDatabase/
. [Last accessed on 2017 Mar 03].
Access to Essential Medicines in Kenya, A Health Facility Survey. Ministry of Medical Services and Ministry of Public Health & Sanitation; 2009. p. 32-3.
Rockers PC, Wirtz VJ, Vian T, Onyango MA, Ashigbie PG, Laing R,et al
. Study protocol for a cluster-randomised controlled trial of an NCD access to medicines initiative: Evaluation of novartis access in Kenya. BMJ Open 2016;6:e013386.
Manual for the Household Survey to Measure Access and Use of Medicines (Draft). World Health Organization; 2008.
Kenya – Demographic and Health Survey 2014. Government of Kenya – GovKEN – Funded the study United States Agency for International Development – USAID – Funded the study United Nations Population Fund – UNFPA – Funded the Study United Kingdom Department for International Development; 2014. Available from: http://www.microdata.worldbank.org/index.php/catalog/2544
. [Last accessed on 2017 Apr 05].
Ministry of Public Health and Sanitation. National Clinical Guidelines for the Management of Diabetes. 1st
ed. Nairobi, Kenya: Ministry of Public Health and Sanitation; 2010.
World Health Organization. Prevention and Control of Non-Communicable Diseases: Guidelines for Primary Health Care in Low Resource Settings. Geneva: World Health Organization; 2012.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]