Patients’ Attitudes Towards Involvement of Medical Students in Their Care at Three Selected University Teaching Hospitals in Uganda: A Cross Sectional Study

The acceptance of medical students by patients during care is important and promotes appropriate clinical reasoning, competence and skills development among students. The attitudes and comfort of patients attending care at the medical and obstetrics/gynecology specialties was studied in teaching hospitals of three public universities in Uganda. This was cross sectional study conducted among patients getting care at teaching hospitals for three public universities; of Mbarara University of Science and Technology (MUST), Makerere and Gulu Universities. Logistical regression was used to determine the magnitude of associations between independent and dependent variables. Two-sided p<0.05 was considered statistically signicant.

Direct contact with patients plays a crucial role in the development of clinical reasoning, communication, clinical skills, competences, and professional attitudes. Direct patient student interactions provides real life context like it is in real clinical practice; it obviates abstract imagination when dealing with simulations and virtual patients or models (1). Student patient interactions at the bedside during teaching by the physicians remain the cornerstone of the proper training of future doctors and the practice of medicine. However, some patients may not be comfortable with the presence of medical students during their care. The father of internal medicine, Sir William Osler taught doctors that "it is a safe rule to have no teaching without a patient for a text, and the best teaching is that by the patient him/her self" (2,3).
Most patients often allow medical students to get involved in their care (4,5). Patients who are uncomfortable with involvement of medical students in their care may delay to seek care and/ or not return for follow up visits at university teaching hospitals thus affecting the promptness and continuity of care respectively. The outcome of treatment for such patients are likely to be poor compared to patients who trust the health care system, seek care promptly and adhere to management follow-up care plans.
In both high-and low income countries, majority of patients allow medical students to get involved in patient care. In the United Kingdom a study involving 278 patients attending general practice surgery consultation with and without a medical student present in six general practices in the Oxford area showed that majority of patients had positive attitudes towards involvement of medical students in their care, irrespective of the sex of the medical student. Only eight patients (3%) of all respondents demonstrated discomfort with presence of medical students during their consultations and procedures (4). Majority of patients in a study at a genitourinary facility in Leeds, UK, involving 250 male and 250 female, were comfortable with the involvement of medical students in their care. Only 13% and 15% of women and men respectively expressed discomfort; most of the patients who declined were younger women and men, visiting the clinic for the rst time, and nulliperous women. They were uncomfortable with both male and female medical students (6). Similar ndings were reported from Australia where the majority of patients were comfortable with involvement of medical students and recommend students to be part and parcel of the medical team (7). Other studies have showed that the level of global satisfaction of patients with their care did not vary between patients who consulted with their physicians alone and those who consulted in the presence of medical students (8, 9). A large study in the US, showed a rather low acceptance rate, where only 31% (299/1059) of patients thought students should participate in care of women in the obstetrics/gynecology specialty. Prior experience with medical students was an important determinant of future acceptance of students' participation in a woman's care (5). 68% (n=676) would allow both male and female medical students to participate in their care. While 16% (n=162) would allow only female medical students, one percent (n=10) would allow only male medical students, while 14% (n=141) would not allow medical students to participate in their care (5).
Another study in the US among surgical patients showed that patients' attitudes were favorable regarding participation of medical students in their care. The year of study or experience of the medical students did not signi cantly in uence the decisions of the patients to allow medical students participate in their care. The patients reported that medical students answered their key questions and this improved their satisfaction with care (10). In that study by York et al, most patients stated that they would allow medical students participate in their future hospital care mainly because students provided most of the information they needed for their decision making and were less in a hurry compared to the quali ed healthcare professionals (10). The nding that patients based quite a lot of their decisions on information got from medical students is important for the healthcare professionals during counseling and decision making in case students provided inaccurate information upon which decisions were being made by the patients and their families. A study conducted at various departments in India involving 200 patients showed that the majority (83.5%) of patients were comfortable with the presence of medical students among the hospital care team. Male patients were more positive and welcoming to the students compared to female patients. In regard to specialties, patients in the obstetrics and gynecology wards were more likely to reject student involvement in hospital procedures (11). In a study done in the United Arab Emirates, two hundred sixty-four women (87.1%) accepted student involvement; 158 women (59.8%) preferred female students. Comfort levels were signi cantly lower with male students in all skills that were tested particularly pelvic examination and the discussion of sexual problems (12). In a study involving various specialties in Saudi Arabia, patients were generally acceptable to students' involvement in their care. Refusal rate was only 11 -43%, mainly in the obstetrics/gynecology specialty (13). Similar ndings have been reported from Africa. In Tunisia, it was found that higher acceptance and comfort with medical students' involvement in care was among male patients, patients aged more than 40 years and those employed compared to women, patient aged under 40 and unemployed (14). In Ethiopia, a study involving three hundred and ninety-two inpatients showed that participants were acceptable to involvement of medical students in their care. The level of acceptance varied with the specialty; medical (77.4%), surgical (72.0%) and gynecology ward (69.2%)), although the differences were not statistically signi cant. Less than half (26.8%, n=150) expressed discomfort with presence of medical students during their care (15).
A systematic review involving sixteen studies (1990 to 2010) showed that patients with emotional problems and those that needed an intimate examination were less likely to allow involvement of medical students. The patients considered participation of medical students important because it contributes to the training of the medical students (altruism), but also because they get more time, a thorough physical examination, and a better patient education while the students are being taught (16). Patients who decline medical students' involvement raised concerns with privacy especially during intimate examinations, inadequate students' supervision, consultations involving emotional problems, and student level of training that may determine their perceived skillsets and competences (16,17). In some studies,women in the obstetrics and gynecology services had concerns with the skills of medical students in certain procedures including delivery and pelvic examinations. Which may in uence their decisions to accept and be comfortable with involvement of medical students in their care or not (17). Patients who negatively perceive involvement of medical students in their healthcare can provide false and inappropriate information simply for the ful llment of the request of the medical students while clerking them (4). False information from patients regarding symptoms and their evolution potentially misleads healthcare professionals including medical students in the evaluation and diagnosis of the patient. In a study of 222 women who accepted medical students and 78 who objected to medical students' involvement in their care, 73% of those who accepted said they do so to support learning of the students, while 61% of those who objected raised concerns with their privacy as the main reasons to refuse medical students' participation in their care (18). In other studies, reasons for refusing involvement of medical students in obstetrics and gynecological care include maintaining privacy and low comfort with pelvic examinations conducted by students (18)(19)(20).
In Uganda, medical students are trained in the regional, tertiary private, and national referral hospitals that are meant for the provision of specialized quality healthcare to the populations (21,22). However, there is limited data from Uganda and sub-Saharan Africa generally on the attitudes and comfort of patients with the presence of medical students during their care. This data is important in informing training of medical students, and provision of quality ethically acceptable medical services to patients. The aim of this study was to evaluate the perceptions, dispositions and willingness of patients attending gynecology, diabetes and hypertension clinics and wards regarding involvement of medical students in patients' care in three selected public university teaching hospitals in Uganda.

Study setting
This study was conducted at the teaching hospitals for three public universities; (Mak), Mbarara University of Science and Technology (MUST), Makerere and Gulu University (GU). Makerere University is the oldest of the three universities, followed by MUST then Gulu . For Makerere University, we collected data from Kiruddu National Referral hospital for the medical specialty patients and Kawempe National Referral hospital for obstetrics -gynecology patients. Medical students in MUST are taught at Mbarara Regional Referral hospital where both medical and obstetrics -gynecology patients receive care. Medical students from Gulu University are taught from St. Mary's hospital Lacor (Private Not for Pro t) and Gulu Regional Referral hospital. Data was collected from Gulu Regional Referral hospital but not from St.
Marys Laco,r to avoid variation in characteristics of the private patients who pay for their care.

Study design
This was a cross sectional study that used questionnaire for data collection.
Study population, sample size and sampling procedure The study population included patients with hypertension and diabetes attending the medical clinics and wards, and patients with gynaecological disorders attending care at the gynaecology clinics and wards of those selected university teaching hospitals. The study participants were women with gynecologic disorders, and men and women with established diagnosis of diabetes and hypertension attending care in the designated aforementioned teaching hospitals during the study period; all participants(patients) were aged 18years and above, of sound mind, and willing to provide written informed consents. Sample size was calculated based on the Kish Leslie formula for survey. The proportion of outcome of interest was estimated at 50% since there were no similar studies from the region. The calculated sample size was based on allowable error of 5% and a two sided level of signi cance at 5%, and alpha value of 1.96 was 384. We applied a design effect of 2 to account for inter-cluster variation which otherwise would lead to erroneous and/or inclusive ndings. Estimated sample size was therefore 384 x 2 = 768. An additional 12% (92) was considered for non-response and incomplete data. 860 participants were recruited; 287 from each of the public universityhospitals and half (430) from each of medical and obstetricsgynaecology specialty. Systematic random sampling approach was used to select participants for the study. The number of patients coming into the hospitals had reduced during the period of data collection because of travel restrictions related to the coronavirus 2019 (COVID-19) pandemic. Therefore every third patient registered on the clinic days for the given hospital was sampled, and every third patient among the new admissions. A list of registered patients was made as the patients reported to the clinics on the speci c clinic days. The research assistants then selected every third patient on the list for inclusion into the study. The research assistants also visited the wards early morning (Monday to Friday) and made a list of all new patients. The rst participant was identi ed by simple random sampling of the rst 3 patients. Thereafter, every 3rd patient on the list was approached for inclusion. If a selected patient declined, the next patient on the list would be considered. Thereafter, the interval of every 3 rd patient was resumed. This process was continued till the sample size was achieved for each teaching hospital.

Data collection and management
Data was collected during June through September 2020 when the Uganda National Council for Sciences and Technology (UNCST) allowed research involving patients to restart after a ban due to Covid-19 was lifted . The tool for data collection was based on experience and literature as well as adapting questions from questionnaires used in earlier studies in the high-income countries. Questions were adjusted to make them relevant to the circumstances in Uganda. For example, questions regarding booking of appointments and keeping a given number of minutes for the appointment were removed as they do not apply to healthcare consultations in Uganda where patients walk in and meet any healthcare professionals assign to manage them without necessarily making a speci c personal appointment. The tool was piloted with ve patients from each of medical and obstetrics -gynaecology departments at the Makerere University teaching hospitals , then reviewed and adjusted accordingly before use in the main data collection. The 10 patients/participants involved in the pilot phase were not included in the main study. The nal version of the tool (supplementary 1) was translated into Luo and Luganda to help the research assistants communicate the same thing to participants who did not understand English. No selected participants declined to participate.
Data collection was conducted by trained ,experienced graduate research assistants (RA), The selected Research Assistants (RAs) were trained for two days on the study tool, objectives of the study, and consenting procedures; they were also provided a brief background on issues regarding patient doctor relationships and how the interactions can in uence health-seeking including delay and advanced disease at diagnos is. These were to ensure quality data were collected using android phones loaded with the Open Data Kit (ODK) software. Each of the eight RA interviewed patients independently. two (both female) in Gulu University, four (two female) in Makerere University, and two (one female) in Mbarara University of Science and Technology. Data from the ODK system of each RA was uploaded to an excel spreadsheet and reviewed by a biostatistician and GW. After data collection was completed, the biostatistician reviewed data from 12% of randomly selected participants to ensure data quality. There were no signi cant inconsistencies in the nal dataset used for analyses.

Data analysis
Univariate analysis to describe the demographic characteristics of participants, attitudes and dispositions of patients towards presence and involvement of medical students in their care was used.
The results were reported as frequencies and percentages. Bivariate analysis using Chi square tests were conducted to determine associations between the binary outcome measures with socio-demographic correlates. Multivariate logistic regression models were applied on categorical variables with binary outcomes (Yes and No) to determine the magnitudes of associations between the binary outcome measures (e.g. questions 112 and 108) with selected independent variables. Statistical signi cance was set at two-sided p < 0.05. Effect measures reported were the odds ratios with their accompanying 95% con dence intervals. Question 112 in the questionnaire was used as the main outcome measure; "How do you feel about medical students being present while you are talking to the doctor about your problem?", and the outcome was categorized as "Yes" and "No". We also used question 113 as proxy and conducted sensitivity analyses using it. Question 112 was, "Would you allow medical student(s) to be present while you are talking to the doctor about your problem?" Question 108 was also used as an outcome measure regarding ability of participants to recognize medical students and differentiate them from quali ed doctors. The question was; "When doctors come to see you for the problems that have brought you to the hospital, would you be able to tell whether some of the doctors are medical students?"

Results
Characteristics of the participants Patients' attitudes towards presence of medical students during consultations Most participants (70.1%; n=599) could recognize and differentiate medical students from quali ed physicians; more than half (64.8%, n=554) had ever had medical students present during earlier consultations. more than half (68.5%; n=586) would not mind, while 13.8% (n=118) would be eager/very eager with medical students during their consultations ( Table 2).
The main considerations patients make in deciding whether or not medical students be involved in their care included perceived severity of illness (55.2%; n=472) and sex (78.4%; n=670) of the medical students. However, patients did not take into account prior experience with medical students (64.8%; n=554), duration of consultation (70%; n=598) and their own religious beliefs (91.5%; n=782) ( Table 2).
Perceived importance of involving medical students in patients' care Most participants (95.7%; n=818) considered involvement of medical students in patients' care as an essential (very important and important) ingredient of training of future doctors. The majority of the participants would even prefer that medical students are trained in the public hospitals where they go for care (79.9%; n=683) as opposed to the minority (23.7%; n=203) who would prefer students to be trained in separate designated university teaching hospitals (Table 3). Ability of patients to recognize medical students Upon adjusting for age, sex, educational attainment and employment status, participants who were single/never married were 68% less likely to recognize and differentiate medical students (AOR = 0.32, 95%CI: 0.22 -0.53) from other members of the healthcare team as compared with the married participants. On adjusting for the same socio-demographic factors above, participants from the newest of the three universities (Gulu University teaching hospital) were two and half times more likely to recognize and differentiate medical students from other members of the healthcare team (AOR = 2.51, 95%CI: 1.65 -3.80) as compared to participants from Makerere University.     (Table 8).
However, when asked as to whether medical students be trained in the tertiary public hospitals as opposed to separate university teaching hospitals, most participants (79.9%; n=683) answered to the a rmative. On adjusting for the socio-demographic factors, participants from both GU and MUST teaching hospitals had 2 -8 folds higher odds (AOR = 2.27 -8.14, 95%CI: 1.50 -14.50) compared to those from Mak teaching hospitals prefer training of medical students to be done in the tertiary public hospitals (Table 9).  Table 9: Medical students to be trained in the regional and national referral hospitals Majority of the participants had previous experiences with medical students, and did neither did they mind the involvement of medical students in their care nor were they bothered with the increased consultation time because the physicians were teaching medical students during consultations.
Participants prefer to attend care in the university teaching hospitals where medical students are trained than in equivalent level hospitals where students are not. The sex of the medical student was not an important consideration in deciding whether or not a medical student should be involved in a patient's care. The patients' level of comfort with medical students did not signi cantly vary between patients attending care at the medical and obstetrics-gynecology departments. Highly educated patients from the medical department were unlikely to be comfortable with the involvement of medical students in their care. The few participants who were uncomfortable with involvement of medical students in patient care were concerned about breach of privacy by the unquali ed medical students. The patients from Makerere University teaching hospitals situated in the capital city were signi cantly more likely to be uncomfortable with involvement of medical students in patients' care.
In this study, majority of participants were young, aged less than 40 years, female, married and with some formal employments. Age of the patients did not signi cantly in uence their comfort with involvement of medical students in their care. The age distribution of participants in this study is similar to other studies in this area research. For example, a study in Canada that involved 625 patients from various specialties had a mean age of 39 years, with the majority aged 30 -65 years. Majority of the patients (62%) in that study were female (23). The US, Midwest study involving 213 obstetric gynecology patients had a mean age of 34.9 years (20). In this study, age of participants was not statistically associated with being comfortable with involvement of medical students in patient care. While it would be expected that older patients would be uncomfortable with involvement of medical students (most are young) in their care, our data do not show the same. Earlier studies showed that both outpatients and older hospitalized patients have positive attitudes towards medical students' involvement in their care (24). Although more context speci c data are needed, our nding that age of patients does not determine patients' comfort and acceptance of medical students' involvement in patient care means that the deployment of medical students shall therefore not be restricted by the age of the patients. Students can therefore interact with older patients as it is with younger patients and therefore learn geriatrics care without need for models.
Most participants in this study were able to recognize and differentiate medical students from the quali ed medical doctors. Majority of them had ever had medical students present in their previous consultations and healthcare. The participants who were married, with higher educational attainment and those formally employed were more likely to recognize medical students. Our nding coheres with most studies in which majority of patients recognize medical students. However, a varying proportion of patients don't know how to differentiate quali ed healthcare professionals from medical students (10,14). For example, in Tunisia, up to 78% of patients did not realize that medical students were involved in their care (14). In Australia, a study among women attending antenatal care reported that more than half of the women (54%; N=625) had challenges differentiating medical students from other health professional cadres (25). It is important that patients are told that medical students are involved in their care and their roles. The patients should consciously consent to involvement of medical students in their care.
The majority of participants were comfortable with involvement of medical students in their care. They did neither feel that presence of medical students would adversely affect the quality of care, nor were they concern with the longer duration of consultation when medical students are present. This positive nding regarding acceptance of medical students by patients in university teaching hospitals is quite encouraging. This nding is similar to results from other studies in both the high(HIC) -and low income countries and low-and -middle income countries (LMIC) where patients across specialties have shown acceptance for medical and other healthcare students to be involved in patients' care as part of the students' training. Data from the high-income countries show that medical students are highly accepted (55 to 95% acceptance) in accident and emergency services but not as much among pregnant women, especially during intrapartum care (26)(27)(28)(29). In addition, data from the HIC show that most patients would consult with medical students in their future hospital visits (20,28). Similarly, data from the LMIC also show high level of acceptance of medical students' involvement in patients' care. For example, in Ethiopia, 69.2% -77.4% accept medical students to participate in their care (14,15). Similarly high level of acceptance have been reported from the Middle East; for example, from a study in Saudi Arabia, patients from various specialties were generally acceptable to students' involvement in their care. Refusal rate was only 11% -43%, mainly in the obstetrics-gynecology specialty (13). Acceptance of medical students is higher for non-invasive contacts including reading patients le, observing doctors during ward rounds, and taking history than with intimate procedures including digital rectal exams, vaginal deliveries ,episiotomy repairs, and pelvic examinations (13,14). Patients accept medical students to participate in their care because they want to contribute to the learning of the students and making of future doctors, companionship, and because they learn quite signi cantly about their own health states from the medical students who often give signi cant time to explain to patients their health conditions. Patients also feel that they learn more when the doctors are teaching medical students during consultations (17,20,23,25,28,30). Majority of patients concur that patients-medical students' interactions is a critical factor in training of competent future doctors (23). Participation of medical students in patients' care was considered a worthwhile learning experience for the students (29). Similar ndings were reported from Australia, where 96% of patients (N=248) acknowledged the importance of students' involvement in patients' care as part of their training (7). Most patients are often comfortable with involvement of medical students in their care (16, [31][32][33]. A systematic review that included sixteen studies showed that majority of patients are comfortable with involvement of medical students in their care. The patients thought that medical students add value to their care including through being readily available and There are characteristics of patients and medical students that have been shown to in uence patients' disposition to accept or reject involvement of medical students in their care. It was noted that participants with higher education standards attending care at the medical specialty clinics and wards were less comfortable with involvement of medical students in their care. However, demographic characteristics of participants including age, sex, marital and employment status are not signi cantly associated with attitudes and comfort with involvement of medical students in patients care. Our ndings are different from that of a study from Tunisia, where it was found that higher acceptance and comfort with medical students' involvement in care was among male patients, patients aged more than 40 years and those employed compared to women, patient aged under 40 and unemployed patients (14). Our ndings also differs from that of Hartz et al which showed that patients' education level did not in uenced their decision to allow involvement of medical students in their care (20). However, they reported that level of education signi cantly in uenced acceptance and comfort level with medical students during intimate examinations including pelvic examinations and performance of Pap smear among women. Higher educated women are more willing to accept medical students' involvement than the less educated women (20). Our ndings also differ from a study in Australia, which showed that obstetrics and gynecology patients (n=255) aged less than 40 years and those who were inpatients were signi cantly more likely to be satis ed with involvement of medical students in their care. Satisfaction was higher among patients seen by female medical students (86%) compared with male students (74%) (7). Majority of medical students maybe young and that is why they would be accepted easily by younger patients aged less than 40 years. In this study, majority of patients were young, with median age 38 years (54% younger than 40 years). This explains why age did not come out as a signi cant factor in determining patients' comfort with medical students' involvement. There was no signi cant difference by specialty regarding the level of acceptance and comfort with medical students except for education attainment.
Passaperuma et al also found no inter-specialty differences regarding patients' comfort and acceptance of medical students' involvement in their care (23). However a large study involving 932 participants from 14 teaching hospitals in Kuwait showed signi cant inter-specialty difference in acceptance of medical students' involvement in patients' care. While acceptance was highest in the pediatrics specialty, refusal was highest in the obstetrics gynecology specialty (34). more studies in sub Saharan Africa across various specialties should be done to shine more light on the effects of subspecialties and patients' comfort with medical students.
In this study, majority of participants said they would put into considerations the sex of the medical student (78%) and the seriousness of the disease at the time of consultation. They would not decide based on their own religious and cultural beliefs, nor length of time of consultations. Majority of the participants also said they would not be in uenced by the quality of their previous experiences with medical students. These ndings make it easy to deploy medical students to interact with patients without the need to rst sort patients on the basis of certain characteristics. Regarding sex of the students, it is important for medical educators and attending physicians to explain to the patients beforehand the need for both male and female medical students gaining the required skills. Our ndings on sex of students, previous experiences with medical students and consultation time is similar to results from other studies. For example, York et al found no signi cant differences between attitudes of patients who had previous experiences with medical students and those who had not (10).
Another factor that in uenced acceptance of medical students in our study included the university teaching hospitals. Participants from the oldest university teaching hospital located in the capital city were less likely to be comfortable with medical students when compared with the other two universities which are younger and located away from the capital city. Patients at the oldest university teaching hospitals were signi cantly less comfortable with the involvement of medical students in their care. Our nding regarding previous students' patients interactions differ from some other previous studies which show that previous experience with medical students positively in uence attitudes and acceptance of medical students in subsequent patient student interactions (4,16,31,33,35). Positive earlier encounter with medical students tend to increase acceptance of medical students in future consultations (10).
Negative experience tend to reduce chance of the patients accepting participation of medical students in their care (36). It remains unclear why acceptance of students were relatively lower at Makerere University teaching hospitals. Makerere university medical school started way back in 1923, and it is probable that most patients who come to that hospital are aware that there are medical students among the healthcare team. The other reason in uencing the attitude of patients at Makerere university teaching hospital could relate more to the location in the city center and the cosmopolitan nature of the population than an intrinsic factor within the university teaching hospitals themselves. However, it could also be that most of the patients included had never had prior contact with medical students even if they are attending care at the oldest university teaching hospital in Uganda. Future studies need to categorize the patients based on their frequency of future contacts with students. In the US, patients who had had fewer contacts with medical students during their care were more likely to decline medical students' participation in their care (20). In addition, qualitative studies exploring quality of care in previous student involvement in care may elucidate the way in which previous student involvement in uence future acceptance and comfort with medical students.

Limitations
This study has some limitations inherent in the design. This was a cross sectional study; we can only appreciate associations between the socio-demographic and health systems' factors with patients' acceptance without asserting causality. Secondly the results could be in uenced by social desirability bias because data collection was conducted in the hospital setting and patients could have responded in a manner that would be socially desirable. We minimized this bias by not involving the doctors in the data collection process. Data were collected by masters of public health students and graduate research assistants who explained to the patients their status and encouraged them to provide appropriate responses without fear of any retributions.

Conclusions
Patients are generally comfortable with medical students' involvement in their care; and prefer to seek care in university teaching hospitals where medical students contribute to care.besides, clinical teachers in university teaching hospitals need to provide patients with some protected time in the absence of medical students so that matters that could not be addressed in the presence of medical students do come out. There is also need for disseminating clear messages to the public regarding the presence and roles of medical students in the university teaching hospitals so that people who come their know in advanced and make deliberate choices to come there and accept or decline participation of medical students in their care. Patients need to be encouraged to provide feedback regarding the involvement of medical students in their care. feedback can inform decisions of medical educators and preceptors on how to best train medical students without infringing on patients' autonomy.

Declarations
Ethical approval and consent to participate This study protocol was presented to and approved by the School of Biomedical Sciences Research Ethics Committee (SBSREC), Ref: SBS-REC-761. We obtained written permissions to access the study hospitals from the respective hospital directors. The clinic and ward in charges held meetings with the research assistants and planned the recruitment based on the appropriate days for clinic and ward round days in each of the hospitals. All prospective participants were informed about the purpose of the study, their rights as potential participants, and the con dentiality of their information during and after the interviews. Participants were informed that participation in the study was voluntary, that they did not have to answer all questions, and that they could voluntarily stop participation if they needed to without any retributions. Written informed consents were obtained from every participant. Data collection tools were designed in English and two local languages -Luganda and Acoli/Luo. In order to maintain privacy, study numbers were used instead of participants' names and other identifying information. A transport refund of USD 2.8 was provided for to every participants after the interviews.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.