Concomitant Use of Herbal and Conventional Medicines among Patients with Diabetes Mellitus in Public Hospitals of Addis Ababa, Ethiopia: A Cross-Sectional Study

Introduction The majority of the population in developing countries including Ethiopia still relies on traditional medicines (TMs). Patients with chronic illness like diabetes mellitus (DM) are dissatisfied with conventional medicines and thus are more likely to simultaneously use herbal medicines (HMs). However, such practice could result in potential herb-drug interaction. This study aimed to identify the commonly used HMs among patients with DM and determine the magnitude of concomitant use of herbal and conventional antidiabetic medicines. Method A health facility-based cross-sectional study design was employed using both quantitative and qualitative data collection methods to determine the magnitude of concomitant use. Patients with DM and prescribers from four public hospitals were the study population for the quantitative and qualitative study, respectively. Simple descriptive statistics were used to describe variables for the quantitative data, and content analysis had been conducted manually for qualitative data. Result Out of 791 respondents, 409 (51.7%) used traditional medicine at least once in their life time, and 357 (45.1%) used traditional medicine in the last six months prior to data collection. A majority (288 (80.7%)) of the respondents used HMs after starting the conventional antidiabetic medicines within the last six months. Moringa stenopetala, Thymus vulgaris, Trigonella foenum-graecum, Nigella sativa, and Allium sativum were among the frequently mentioned HMs. Prescribers were requesting patients' HM use when they saw sign of liver toxicity and skin disease, and they were not documenting their history in the patient's chart. Conclusion Concomitant use of herbal and conventional antidiabetic medicines was a common practice. Cognizant of its potentially serious herb-drug interactions, efforts should be made to improve awareness and knowledge of healthcare providers about HM potential effects. Further studies on dose, frequency, duration, and modes of interaction are recommended.


Introduction
Traditional medicines (TMs) cover a heterogeneous spectrum of ancient to new-age approaches to prevent or treat disease. ey include the use of herbal medicines (HMs), spiritual healing, and practices such as bone setting [1]. TMs were used for age-related chronic diseases such as hypertension and diabetes mellitus (DM) for which no modern medicine or only palliative therapy is available [2]. e use of HMs, either alone or combined with conventional medicines, has been practiced widely particularly in low-resource setting countries. eir use in patients with chronic disease like DM is very common, owing to its natural origin and perceived lesser side effects or dissatisfaction with conventional medicines [3][4][5].
DM is one of the largest global health emergencies of the 21st century. In 2015, 14.2 million people were estimated to be living with DM in Africa, and this is projected to 34.2 million by 2040 [6]. e prevalence of DM in Ethiopia among adults aged 20-79 was estimated to be 0.2% and 0.28% in 2010 and 2030, respectively, with mean annual increment 0.06% [7]. Literature showed that patients with DM who were dissatisfied with conventional medicine were more likely to use HM to enhance the effect of conventional treatment [8]. However, simultaneous use of herbal and conventional medicines may result in potentially serious herb-drug interactions leading to negative effect on patient outcomes [4]. Mostly, HMs are complex mixtures of more than one active ingredient which increase the possibilities of interactions [9,10]. For instance, treatment with St. John's wort (SJW) significantly increases the apparent clearance of gliclazide [11]; administration of Ginkgo biloba extract increased hepatic clearance of insulin and oral hypoglycemic agents, hence resulting in an increased glucose level [12]. Adverse effects such as hypoglycemia and lactic acidosis were also detected among patients with diabetes who were taking herbal antidiabetic products in China [13]. Such adverse reactions could involve all systems, age groups, and severity [9].
Until recently, however, herb-drug interaction was often unsuspected by healthcare providers. Most trained physicians lack adequate knowledge on HM and its potential for drug interactions [14]. Patients may also not always inform their doctor of the simultaneous use of HM [15].
Ethiopia is not an exception to this worldwide phenomenon as HM is still the main source of healthcare for many millions of people [16]. However, information on most commonly used HM and extent of concomitant use with conventional medicines in DM is scanty. is study aimed to determine the extent of concomitant use of herbal and conventional antidiabetic medicines and also to identify the commonly used HMs among patients with DM.

Description of the Study Area.
e study was conducted in Addis Ababa, the capital city of Ethiopia. Addis Ababa is divided administratively into 10 subcities and has a total population of 3.43 million [17]. 56 hospitals (11 public, 3 army, and 42 private hospitals), 760 private clinics, 96 health centers, and 720 pharmacies were available in the city administration [18].

Study Design.
A health facility-based prospective crosssectional study design was carried out to determine the prevalence of concomitant use of herbal and conventional medicines among patients with DM. Data were collected between April and August, 2016.

Source and Study Population.
e source population for the quantitative data constituted all patients with diabetes attending treatment in outpatient clinics of public hospitals in Addis Ababa City Administration (AACA). Patients with diabetes from four public hospitals in Addis Ababa, two hospitals from AACA (Yekatit 12 Hospital Medical College and Zewditu hospital) and two hospitals from Federal (St. Paul's Hospital Millennium Medical College (SPHMMC) and Tikur Anbessa Specialized Hospital) were the study population. On the contrary, all prescribers working in the diabetic clinics were considered as the source population for the qualitative data.

Inclusion and Exclusion
Criteria. Patients aged ≥18 years and taking antidiabetic medicines were eligible for the survey. Patients with DM who were physically or mentally not capable for the interview and those who were health professionals were excluded from the study. For the qualitative interview, prescribers who were working in diabetic clinics and had no previous diagnosis of DM were recruited.

Sampling and Sample Size Determination.
e sample size was determined using the single population proportion formula [19]. A total of 845 patients with DM were included by considering 50% proportion of concomitant use of herbal and conventional medicine, with 95% confidence interval, 5% margin of error, design effect of 2, and 10% nonresponse rate.
Based on ownership, hospitals were stratified into those administered by AACA and Federal Government. en, from each stratum, two hospitals, Zewditu Memorial Hospital (ZMH) and Yekatit 12 Hospital Medical College, from AACA and Tikur Anbessa Specialized Hospital (TASH) and St. Paul's Hospital Millennium Medical College (SPHMMC) from federal hospitals were selected using simple random sampling technique. e sample size was proportionately allocated considering total patients with diabetes of each hospital of which 336 were from TASH, 206 from SPHMC, 132 from ZMH, and 171 from Yekatit 12 Hospital Medical College. Recruitment of the respondents was achieved through the consecutive sampling approach.
For the qualitative data, selection of participants was based on the purposive sampling method. Key informants were recruited on the basis of their professional role and work experience in the diagnosis and treatment of diabetes mellitus. Selection of participants was continued until theoretical saturation was achieved.

Data Collection Instrument.
For the concomitant use of herbal and conventional medicine assessment, a total of 45item questionnaire was adopted from the previous literature (supplementary 1) [20]. Questions addressing the patients' sociodemographic characteristics (age, sex, average income, marital status, and educational status), traditional medicine use, type of traditional medicine chosen, time of use of traditional medicine (before, after, or during the conventional medicine treatment), reason to use the traditional medicine, perceived effectiveness of the herbal medicine, and adverse effects experienced were included in the assessment questionnaire.
On the contrary, a semistructured open-ended interview guide with flexible probing techniques was employed to explore prescriber's experience of history taking about herbal medicine use during examination of patients with DM.

Data Collection Procedures.
Two trained pharmacists were recruited as data collectors for each hospital. As part of the training, data collectors conducted pretesting of the instrument in Ras Desta Damtew Memorial Hospital under the supervision of the principal investigator, and modifications were done accordingly. All key informant interviews were administered by the principal investigator, who was trained on qualitative research methods. All interviews were recorded, and notes were also taken during interviews to expand later. e interviews were done in Amharic (a local language), and any ambiguities raised from the interviewees were cleared at the time of the interview. Questionnaires prepared in English were translated into Amharic and backtranslated into English to ensure consistency.

Data Analysis.
After the data collection, the principal investigator coded each question, and data entry was made using Epi Info version 3.5.4 and transferred to SPSS version 20 for analysis. Descriptive statistics were used to summarize the data. e qualitative data were analyzed using content analysis. It involved intensive reading and rereading through the data, and content analysis had been conducted manually, focusing on similarities and differences of perspectives between different informants.

Ethical Consideration.
Ethical approval was obtained from the Ethics Review Committee of the School of Pharmacy, Addis Ababa University, and AACAHB Ethics Review Board. e study was conducted after obtaining permission from the respective hospitals. Participants of the study were also asked for verbal consent before participating in the study. Participants were assured about confidentiality of the information obtained in the course of the study, and there was no use of personal identifiers.

Demographic Characteristics of Respondents.
From 845 respondents, 791 completed the interview, making up a response rate of 93.6%. Males and females accounted almost similar proportion of sample size of 402 (50.8%) and 389 (49.2%), respectively. e age of the respondents ranged from 18 to 83 with the mean age of 49.07 ± 14.76 years. Among the respondents, one-third of them were housewives, and two-thirds of respondents had attended at least elementary school. A majority (667 (84.3%)) of the respondents were reported as residents of Addis Ababa (Table 1).

Clinical Condition of Respondents.
At the time of the study, 210 (26.5%) of patients were on insulin injection, 499 (63.1%) on oral hypoglycemic medicines, and 82 (10.4%) on both. A sizeable number of the patients (333 (42.1%)) were diagnosed as diabetic before 5 to 10 years. More than half (58%) of them had other comorbid conditions, of which 313 (39.5%) of patients had hypertension (Table 2).

Traditional Medicine Use Practice of Patients with DM.
More than half (409 (51.7%)) of the respondents revealed their use of TM at least once in their life time, and 357 (45.1%) used TM in the last six months. Of these, the majority (218 (61.1%)) of them was used to treat only diabetes mellitus. From the total TM users within the last six months, 288 (80.7%) used HMs.
Patients with diabetes mentioned various reasons for using TM. e most common one was patient's belief that TM is more potent than conventional antidiabetic medicines which was mentioned by 240 (58.7%) of the respondents. A sizable number of respondents (96 (23.5%)) also believed conventional medicine is expensive and reported to use TM (Table 3).

Source, Duration, and Part of the Herb
Used. e majority (260 (90.3%)) of patients remembers the name of the herbal medicine they used, and 211 (73.3) of them reported using the leaf of the herb. Market 217 (75.3%), garden 49 (17%), and traditional healers 31 (10.8%) were the major sources for obtaining herbal medicine. From the total herbal medicines, 111 (38.5%) respondents used herbal medicine once, and 40 (13.9%) respondents used for more than 30 days (Table 4).

Commonly Used Herbal
Medicines. Moringa stenopetala was used by 144 (50%) of the patients with DM. Similarly, ymus vulgaris, Trigonella foenum-graecum, Nigella sativa, and Allium sativum were reported by 50 (17.4%), 21 (7.3%), 19 (6.6%), and 16 (5.6%) of patients with DM, respectively. A considerable amount of herbal medicines was prepared in crushed form as cited by 194 (67.4%) of patients, and 243 (84.4%) of them expressed that the herb was available in dried form. Oral administration was the major route of herbal medicines used among patients with DM (Table 5).

Source of Information and Decision on Herbal Medicines Used.
Of 288 herbal medicine users, 263 (91.3%) respondents tried to get more information about the HM related to its effectiveness and toxicity they used (Table 6). Of those, 164 (62.4%) of them obtained the information from the community, and only 3 (1.2%) respondents obtained it from pharmacists. Of the total 288 concomitant users of herbal and antidiabetic conventional medicines, only 103 (35.8%) of them informed their healthcare providers that they were using HMs. Being afraid of the doctor and prevailed perception that HM has no problem were the frequently mentioned reasons for failing to inform their      (Table 6). Among those who informed their use of HM to health professionals, 74 (71.8%) of them were counseled to continue the herbal medicine, while 28 (27.2%) were advised to discontinue (Table 6).

Perception of Effectiveness and the Method of Side Effect
Aversion. From 288 respondents, 40 (13.9%) of them had encountered side effects from HM use. Among these, constipation, vomiting, and headache were the most frequently reported side effects as mentioned by 14 (24.6%), 8 (20%), and 8 (20%) of the participants. Of those who had side effects, 24 (60%) discontinued herbal medicine use by themselves, and 10 (25%) consulted health professionals to avert the side effects (Table 7).

Qualitative Findings.
A total of 8 physicians (4 general practitioners and 4 consultants) were interviewed. Five of them were males, and their mean age was 44.2 (ranged from 29 to 54 years). Participants' work experience in diabetic clinics was from 2 to 12 years. ree major themes were defined in qualitative data analysis: prescriber's perceived knowledge about herb-drug interactions, herbal medicine history taking practice, and commonly used herbal medicines.

Prescribers' Perceived Knowledge about Drug-Herb
Interaction.
e majority (n � 6) of key informants agreed on the importance of knowing most commonly used herbal medicines and their effect when used concomitantly with conventional medicine. Nevertheless, all (n � 8) of them revealed that they had no adequate knowledge about herb-drug interactions. ey also reported that much emphasis was not given to the herb-drug interaction course in medical schools. As a result, low attention was given to taking history of herbal medicine use. One general practitioner strengthened this as "If I ask the patient whether he/she has been using herbal medicine and found that he/she was using any herb, I wouldn't take any decision towards using herb because I don't have any knowledge about a specific herb and the conventional anti-diabetic drug interactions."

Herbal Medicine History Taking
Practice. All (n � 8) key informants mentioned that documentation of herbal medicine information for patients taking conventional medicines was not usual practice. On the contrary, a majority (n � 5) of them claimed of asking patients about their herbal medicine experience when they saw perceived side effects like sign of liver toxicity. Lack of knowledge about effects of concomitantly using herbal and conventional medicines and high patient load were ascribed as major reasons for this. is was strengthened by one of the participants as follows: "Usually, I don't check patients' history of herbal medicine use practice. I haven't learnt it at school and I don't have experience on how to handle it. But even if I want to do it, there are a lot of patients waiting for service so that I don't have time to ask my clients about their herbal medicine use practice but occasionally when I suspect signs of liver toxicity, I would ask whether they took herbal medicine or not. But, I would not record it on the patients' chart."

Commonly Used Herbal Medicines.
Most of the key informants (n � 7) agreed that Moringa stenopetala is the most commonly used herb by patients with diabetes as it was believed to have hypoglycemic effect. Respondents perceived that concomitant use of herbal-conventional medicines was a prevalent practice, but they found it difficult to remember the other types of herbs used mentioned by the patient. Only one physician had encountered patients who were using Coffea arabica extract for diabetic treatment.
One general practitioner also explained that there were patients who told him as they were using herbal medicines for the treatment of diabetes mellitus. He expressed his experience as follows: " ere were plants used by patients with diabetes that I didn't remember the name and type of the herbs since I did nothing with that. I only advised them to discontinue using herbal medicine."

Discussion
is study has been carried out with the purpose of determining the prevalence of herbal medicine use and identifying the commonly used herbs by patients with DM. e study estimated that, from total patients with DM, 288 (36.4%) took herbal medicine which was higher than a study in black South African communities, where 28 (21%) used HM to treat hypertension [21]. is difference could be due to the clinical difference of patients included in the study. On the contrary, our finding was lower than the prevalence in Gucha district, Kenya (68.9%), and southeast region of Morocco (80%) [22,23]. e difference in magnitude may be due to difference in the study population. In this survey, 231 (69.6%) of the respondents used HMs along with prescribed therapy for the last six months which was higher than in South Sudan (58%) [24]. e higher concomitant use in this study might be due to the difference in sociodemographic characteristics of the study populations between the two studies. Similarly, a higher prevalence was also obtained from India, where 71% of patients with DM used HMs along with the conventional therapy [8]. e higher use might be due to better integration of HMs with the modern healthcare system in India [25].
Even though the quantitative study indicated high proportion of patients with diabetes used HMs concomitantly with conventional antidiabetic medicines, finding on exploration of HM history taking practice of prescribers during physical examination of patients with diabetes showed low. e contradictory results from the quantitative and qualitative data could be due to prescriber's low knowledge on herb-drug interaction and work load.
Moringa stenopetala, ymus vulgaris, Trigonella foenum-graecum, Nigella sativa, and Allium sativum were the most frequently used HMs as claimed by the participants in this study. A similar study in Oriental Morocco also indicated that the most five HMs used by Morocco population were Salvia officinalis, Trigonella foenum-graecum, Olea europaea, Artemisia herba-alba, and Origanum vulgare [26]. Another similar study in Turkey revealed that the most frequently used HMs by the patients are green tea, garlic, and ginger [27]. Different HMs might be used in different areas with different levels as a result of difference in accessibility of the herb and information dissemination among patients with diabetes. Moringa stenopetala and Coffea arabica extract were also identified as the commonly used HMs by diabetic patients in the qualitative study.
A majority of patients with DM (240 (58.7%)) in this study used HM since they believed that TM was more potent and affordable than conventional antidiabetic medicine which was the major reason to use HM. Similarly, studies in North Sudan [24] and Kenya [28] indicated patients' belief in herbal medicine efficacy, easy accessibility, and lower cost have encouraged patients with diabetes to use them.
is study also revealed that 69.1% of the HM users perceived herbal medicines they used were effective. Qualitative data also indicated patients with diabetes were benefited from Moringa stenopetala. Scientific experimental studies on Moringa stenopetala leaves [29], Allium sativum [30], yme vulgaris [31], Trigonella foenum-graecum [32], and Nigella sativa [33] had showed significant decrease in blood glucose level. Another study also revealed that fasting blood glucose was reduced by Zingiber officinale to normal as effectively as glibenclamide [34] and Allium sativum as compared to metformin [35].
is study also revealed that 40 (13.9%) of HM users reported they experienced side effects. is might be due to an additive and/or synergetic effect of the concomitant use of these herbs with conventional antidiabetic medicines. Like hypoglycemic effect, experimental studies on Trigonella foenum-graecum [36,37], Nigella sativa [38], and Moringa stenopetala [39] identified different toxic effects.
is study revealed that the very common method of remedy preparation was in dried and crushed form which was taken by mixing with water. However, another study in Ethiopia indicated more than half of the HMs were used in fresh form [40], and decoction was the most common method of preparations [41]. Even though oral administration was the major route and leaf of the herb was the most frequently used 211 (73%) parts of the herb in this study, different route of administration and various parts of the herb in different preparation were also used. Using different route of administration, preparation and parts of the herb may yield different active ingredients of a herb which might result in different effect [42].
is study found that market and community were the major sources and sources of information for the HMs, respectively. e societies' belief on antidiabetic effect of HMs might affect adherence to the conventional medicine.
is survey also indicated that a small proportion of respondents obtained information about the HMs from healthcare providers. is result was supported from Ubon Ratchathani province, ailand [43]. Lack of free communication between prescribers and patients with diabetes might be the reason.

Conclusion
According to the findings of this study, herbal medicines are used concomitantly with conventional antidiabetic medicines for the treatment and management of DM by patients attending treatment in four selected hospitals of Addis Ababa. More than eighty percent of patients with diabetes used HMs, and 231 (69.6%) of them used HMs concomitantly with conventional antidiabetic medicines. Combined use of HMs and conventional antidiabetic medicines could potentially lead to serious toxic effects due to herb-drug interactions. Moringa stenopetala, ymus vulgaris, Trigonella foenum-graecum, Nigella sativa, Allium sativum, Linum usitatissimum, Zingiber officinale, Ocimum lamiifolium, Brassica nigra, Ruta chalepensis, and Camellia sinensis were the most commonly used herbs by patients with diabetes.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
SG contributed to the conceptualization of the study, supervised the data collection and analysis processes, and Evidence-Based Complementary and Alternative Medicine 7 wrote the first draft of the manuscript. MY contributed to the write-up of the manuscript. GB and ZT cosupervised the analysis processes and played a key role in the write-up and submission of the manuscript. TG led the statistical analysis carried out in this study and contributed to the write-up.