East African Journal of Applied Health Evaluation
Issue 4 | March 2020
East African Journal of Applied Health Monitoring and Evaluation ISSN 2591 - 6769
EDITORIAL BOARD
CO-CHIEF EDITORS Henry Mollel, MSc, PhD (Mzumbe University, Tanzania) Debbie Bain Brickley, MPH, DrPH (University of California, San Francisco)
MANUSCRIPT EDITOR Mackfallen Anasel, MSc, PhD (Mzumbe University)
WEB EDITOR Christen Said, MPH (University of California, San Francisco)
EDITORIAL BOARD Andrew Abaasa, MSc (MRC/UVRI Uganda Research Unit on AIDS) Mackfallen Anasel, PhD, MSc (Mzumbe University, Tanzania) Gershim Asiki, MBChB, MSc, PhD (African Population & Health Research Center, Kenya) Etienne Karita, MD, MSc, MSPH (Projet San Francisco, Rwanda) Claud Kumalija, MSc (Ministry of Health, Community Development, Gender, Elderly and Children, Tanzania) Bunini Manyilizu, MD, MPH, MBA, PhD (Mzumbe University, Tanzania) Lawrencia Mushi, BHSM, MSc.UMD, PhD (Mzumbe University, Tanzania) Roger Myrick, PhD, MA (University of California, San Francisco) Gloria Omosa-Manyonyi, MBChB, MSc (University of Nairobi) George Rutherford, MD, MA (University of California, San Francisco) Jim Todd, MA, MSc (London School of Hygiene and Tropical Medicine)
EDITORIAL ASSISTANTS Sandra Dratler, DrPH (University of California, Berkeley) Susie Welty, MPH (University of California, San Francisco) Karen White, MBA, MPH (University of California, San Francisco)
DESIGN & LAYOUT Mackfallen Anasel (Mzumbe University)
Cover photo by Rose J. Mdami Address all editorial correspondence to Email: [email protected] Website: http://eajahme.com
Copyright 2020 All rights reserved.
Disclaimer: No part of this publication may be reproduced, stored or transmitted in any form or by any means (electronic or otherwise) without prior permission, in writing, of East African Journal of Applied Health Monitoring and Evaluation. Individual authors are responsible for the content of their articles.
Table of Contents
RESEARCH ARTICLES
Factors Associated with Willingness to Pay for Social Health Insurance among 1 Government Employees in Tigrai Region, Northern Ethiopia Ataklti T. Gessesse, Abera A. Berhe, Mulugeta G. Tilahun, Tesfay W. Teklemariam
Knowledge of HIV status among mothers accompanying their infants for immunization 7 in Machakos, Kenya, 2014 Lilly M. Nyagah, Elvis Kirui, Peter W. Young, Joseph Gikunju, Jane Githuku, Sara Lowther, Andrea A. Kim
Trend and predictors for early infant diagnosis by PCR among HIV-exposed infants in 13 Dar es Salaam region, Tanzania, 2014-2016 Michael F. Mboya, Prosper Njau, Jim Todd, Beatrice John Lerayo, Goodluck Wiley Lyatuu, Lameck Machumi, Sia E. Msuya, Michael J. Mahande, Jenny Renju
Ocular symptoms and associated factors among hairdressers in Kinondoni Municipality, 22 Dar es Salaam, Tanzania Olirk Baldwina Tita, Ezra Jonathan Mrema1, Simon Mamuya
Knowledge, attitudes, perceptions and acceptability of Onchocerciasis control through 28 community-directed Treatment with Ivermectin: implications for persistent Transmission in Ulanga district, Tanzania Vivian Mushi, Deodatus Kakoko, Donath Tarimo
Research Article
2020; Issue 4 ISSN 2591 – 6769
Factors Associated with Willingness to Pay for Social Health Insurance among Government Employees in Tigrai Region, Northern Ethiopia Ataklti T. Gessesse1, Abera A. Berhe2, Mulugeta G. Tilahun1, Tesfay W. Teklemariam1
1. Health researcher, Tigrai health research institute, Mekelle, Tigrai, Ethiopia
2. Department of Biostatistics, College health science, Mekelle University, Tigrai, Ethiopia
Correspondence to: Ataklti T. Gessesse; Tel: +251914002439; Email: [email protected]; PO Box 1547; Mekelle, Tigrai, Ethiopia
INTRODUCTION
Developing countries seldom use social health insurance (SHI), and their healthcare finances mostly rely on general revenues and direct out-of-pocket payments. This study investigated the level and factors associated with willingness to pay for SHI among government employees in Tigrai region, North Ethiopia. METHODS An institution-based quantitative cross-sectional study was carried out from June to July 2018 among government employees in Tigrai, Ethiopia. Sample size was determined using single population proportion formula, and multi-stage cluster sampling was used to select the study participants. Data collected using an interviewer-administered questionnaire was analyzed using SPSS Version 20. RESULTS There were 544 (64.5%) respondents who were not willing to pay for SHI. Respondents age older than 39 years were 2.2 times more likely to be willing to pay for SHI, as were those who disagreed with the binding rule of referral system (1.4 times), and with exclusion of periodic medical checkup from the SHI (1.4 times), those who didn’t consider health service quality to be poor (1.6 times), and those who disagreed with the presence of financial insecurity in health institutions (1.7 times). CONCLUSION This study revealed that government employees’ willingness to pay for SHI was low. SHI agencies should publicize the proclamation for SHI and induce employees with SHI referral system, services excluded, and health facilities’ readiness and service quality to increase willingness to pay. In addition, the government should reconsider the implementation of the proclamation for SHI accordingly.
Key words: employee, willingness to pay, social health insurance
INTRODUCTION On the other hand, poor populations in sub-Saharan Africa Globally, high proportions of people suffer and die due to make up approximately 24% of the global disease burden and lack of access to basic healthcare services. Each year in low- account for less than 1% of global health expenditures and middle-income countries, approximately 150 million (Bank.,W., 2011). Similarly, the recent Ethiopian health people suffer a health-related financial catastrophe, and more account showed that the share of total government health than three people are pushed into poverty per second as a spending was not more than 5.6% of the total government result of out-of-pocket (OOP) health expenditures (Bump J, expenditure, and around 34% of this expenditure comes from et al., 2016, Averill C et al., 2013). This signifies that health OOP payment of households (Federal ministry of health service fees are a main barrier to healthcare coverage and [FMoH], 2014). utilization (Dieleman JL et al., 2017). Moreover, many SHI is an agreement that insures individuals from damage, African countries fail to attain the minimum 15% of their illness and death, and transfers risk of an individual's loss budget allocation for health (Spreeuwers, A.M. et al., 2012), (Comfort AB, et al., 2013) by pooling risks with other people which was agreed on through the Abuja Declaration of 2001. (Comfort AB et al., 2013, De Allegri M, et al.,2006). To
1 Research Article East African Journal of Applied Health Monitoring and Evaluation
2 reduce reliance on direct OOP payments and achieve describe selected study variables. Bivariable logistic universal health coverage (UHC), governments encourage regression at p-value <0.2 and multivariate logistic implementation of health insurance schemes (Odeyemi IA., regression at p-value <0.05 was performed to identify the 2014). Several countries have started SHI as the country's factors associated with employees’ willingness to pay for the system of healthcare financing and have attained UHC scheme using odds ratios (OR) at 95% CI. (Carrin G, et al., 2005). However, low- and middle-income countries seldom use SHI and mostly rely on general All procedures performed in this study were in accordance revenues and direct OOP payments as sources of healthcare with the ethical standards of the institutional and national financing (Hsiao W.C & Shaw. R.P). As of 2008, Ethiopia research committee and with the 1964 Helsinki Declaration had only 1.1% of any kind of insurance with 1% of and its later amendments or comparable ethical standards. government health expenditures spent on insurance activities Participants were not provided with any incentives or (FMoH, 2008). payment to take part in the study and were given the full right Even though Ethiopia is preparing to start SHI, to refuse from participating or to withdraw from the study at willingness to pay for the scheme is not well researched any time. Ethical clearance was obtained from Tigrai Health (Agago TA et al., 2014, MoH,2010). Studies have been Research Institute Institutional Review Board, and an official limited to a particular facility and city with small sample support letter was obtained from Tigrai Regional Health sizes. This study aims to investigate the level and factors Bureau. Consent was obtained from each study participant, associated with willingness to pay among government and confidentiality of data and scientific honesty was employees in Tigrai, which can help individuals to make maintained. informed choices and provide evidence for policymakers and program implementers to set equitable level of premiums. RESULTS Socio-demographic characteristics METHOD A total of 843 government employees participated in the The study used an institution-based quantitative cross- study. More than half of the respondents (432; 51.2%) were sectional study design carried out from June to July 2018 male, and 439 (52.1%) had 1-3 family members. The mean among government employees in Tigrai region, North age of the respondents was 35 years. More than half (466; Ethiopia. Sample size was determined using single 55.3%) of the respondents were less than 35 years of age and population proportion formula using 47% prevalence of 377 (44.7%) of the respondents were 35 years or older. With willingness to pay for SHI (Tesfamichael A.A, et al.,2014), regard to educational status of the respondents, 347 (41.2%) standard score Z /2 =1.96, margin of error 0.05% at 95% were diploma and lower levels and 496 (58.8%) were first confidence level, design effect of 2, and 10% non-response degree and above levels. One hundred and thirty-four rate. Multi-stage cluster sampling was used to select the (15.9 %) of the respondents were earning a monthly salary study participants. In the first stage, 10 institutions were of lower than $81 USD, 631 (74.8%) were earning $81-315 sampled out of the 42 using simple random sampling. In the USD, and 78 (9.3%) were earning more than $315 USD a second stage, 843 employees out of the total 100,172 in the month (Table 1). institutions’ sampling frame proportionally allocated to each institution were selected using simple random sampling with Willingness to pay and health service-related variables replacement technique. The inclusion criteria was permanent Among the respondents, 299 (35.5%) were willing to pay the employees who worked for more than six months; the 3% premium for SHI, and 544 (64.5%) of the respondents exclusion criterion was employees who were included in any were not willing. The main reasons mentioned for not being health insurance scheme. willing were non-affordability, exclusion of certain services from SHI, and poor health service quality. Furthermore, the Data was collected using a structured, interviewer- monthly health service expenditure of 831 (98.6%) of the administered questionnaire. The questionnaire had two parts: respondents was <20%, and that of 12 (1.4%) respondents sociodemographic characteristics and health service-related was >20% of their monthly income. Seven hundred and variables. The dependent variable was willingness to pay for thirty nine (87.7%) of the respondents had ever heard about the 3% premium set for SHI (measured using a ‘Yes or No’ SHI, while 104 (12.3%) had not ever heard about SHI. question), and the independent variables were socio- Nearly half (49.7%) of the respondents perceived quality of demographic characteristics and health service-related health service in SHI to be good, whereas the remainder variables. A pretest was conducted on 5% of the sample size (50.3%) perceived that quality of health service in SHI was on government employees of non-sampled institutions to not good. More than two-thirds (67.9%) of the respondents make necessary corrections to the questionnaire and time of agreed with the fixed referral chain system in SHI, while the data collection. remaining (32.1%) disagreed (Table 2).
The study employed nine trained diploma-level data collectors and four supervisors to ensure data completeness and validity. Data were entered and cleaned using EPI-info version 7 and analyzed using SPSS version 20. Descriptive analysis using frequency and percentage was performed to 2 Research Article East African Journal of Applied Health Monitoring and Evaluation
3 Table 1. Socio demographic characteristic of variables such as gender, educational status, marital status and family size were not significantly associated with respondents (n=843, Tigray region, 2018) willingness to pay for SHI. Respondent characteristics N (%) Age group Table 2. Willingness to pay and health service-related variables <24 91 (10.8%) (n=843, Tigrai region, 2018) 25-29 197 (23.4%) Respondent characteristics N (%)
30-34 178 (21.1%) Willingness to pay 3% premium for SHI 35-39 119 (14.1%) Yes 299 (35.5%) >39 258 (30.6%) No 544 (64.5%) Sex Reasons for unwillingness to pay 3% premium: Male 432 (51.2%) Not enough money to buy SHI Card 88 (16.2%) Female 411 (48.8%) No illness/ healthy/ always in good health 10 (1.8%) Marital status Complicated SHI administration 21 (3.9%) Single 283 (33.6%) Poor healthcare quality 30 (5.5%) Married 505 (59.9%) The scheme does not cover all needed services 87 (16.0%) Widowed/divorced 55 (6.5%) The cutoff point (3%) premium is unaffordable 308 (56.6%) Family size Suggested amount of % premium 1-3 439 (52.1%) <2% 344 (40.8%) 4-6 366(43.4%) 2% 193 (22.9%) >7 38(4.5%) >3% 7 (0.8%) Education level Percent of health service expenditure 1-12 grade 93(11.1%) 0% 146 (17.3%) Diploma 254 (30.1%) <20% 685 (81.3%) First Degree 447 (53.0%) >20% 12 (1.4%) Second degree 49 (5.8%) Perceived affordability of health service expenditure Total regular monthly income (USD)* Affordable 435 (51.6%) < 81 134 (15.9) Unaffordable 408 (48.4%) 81-315 631 (74.8) Ever heard about SHI > 315 78 (9.3) Yes 739(87.7%) Income per capita (USD)* No 104(12.3%) < 55 427(50.7%) Willingness to join 55.1-93 187(22.2%) Yes 588(69.8%) 93.1-160 139(16.5%) No 255(30.2%) >160 90(10.7%) Worrying about future health service cost Work experience (years) Yes 470 (55.8%) 1-10 508 (60.3%) No 373 (44.2%) 11-20 177 (21.0%) Exclusion of periodic medical checkup >20 158 (18.7%) Agree 233 (27.6%) *Commercial bank of Ethiopia, 30 July 2018 exchange rate Disagree 610 (72.4%) Perceived good quality of health service in SHI Factors associated with willingness to pay Yes 419 (49.7%) The bivariate logistic regression analysis (at P-value 0.2 cut off point) indicated that age, periodic medical checkup No 424 (50.3%) unrelated to illness, rule of referral system to get higher level Fixed referral chain system in SHI of services, poor financing mechanism of government health Agree 572 (67.9%) facilities, and perceived poor quality of health facilities were statistically significant variables. The sociodemographic Disagree 271 (32.1%)
3
Research Article
2020; Issue 4 ISSN 2591 – 6769
Table 3. Bivariate and multivariable analysis of factors associated with willingness to pay (n=843, Tigrai region, 2018) Willingness to pay OR (95%CI) Variables Category Yes (%) No (%) COR AOR <24 39(42.9) 52(57.1) Ref Ref 25-29 73(37.1) 124(62.9) 1.3(0.77, 2.1) 1.3(0.78,2.2) Age 30-34 80(44.9) 98(55.1) 0.92(0.55, 1.53) 1.1(0.62,1.8) 35-39 39(32.8) 80(67.2) 1.54(0.88, 2.7) 1.7(0.96,3.1) >39 68(26.4) 190(73.6) 2.1(1.3,2.5)*** 2.2(1.34,3.74)*** Presence of abiding rule of referral system to get Agree 219(38.3) 353(61.7) Ref Ref higher level health services Disagree 80(29.5) 191(70.5) 1.5(1.1, 2)** 1.4(1.02,1.94)** Agree 102(43.8) 131(56.2) Ref Ref Exclusion of periodic medical checkup from SHI Disagree 197(32.3) 413(67.7) 1.6(1.2, 2.2)** 1.4(1.02,1.95)** Service quality in government health institutions Yes 175(41.8) 244(58.2) Ref Ref Perceived to be poor No 124(29.2) 300(70.8) 1.74(1.3,2.3)**** 1.6(1.2, 2.1)*** Presence of financial insecurity in government Agree 90(26.3) 252(73.7) Ref Ref health institutions Disagree 44(26.67) 121(73.33) 2(1.5, 2.7)**** 1.7(1.23,2.31)*** * Significant at P<0.05, ** significant at P<0.01,*** significant at P<0.001, **** significant at P<0.000, OR= Odds Ratio, COR=crude odds ratio, AOR=Adjusted Odds Ratio
In the multivariable analysis, age, periodic medical checkup DISCUSSION unrelated to illness, rule of referral system to get higher level This study was conducted to find out the level of willingness of services, poor financing mechanism of government health to pay for SHI and the factors related to willingness to pay institutions, and perceived poor quality of health institutions among government employees in Tigrai region in north were statistically significantly associated with willingness to Ethiopia. The result of the study showed that willingness to pay for SHI (Table 3). Respondents older than 39 years were pay for SHI was low (35.5%), and this was is lower than the found to be more than twice as likely (AOR= 2.24, 95%CI 67% finding in Addis Ababa, Ethiopia, (LasebewY,M, et [1.34, 3.74]) to be willing to pay compared to respondents al,2017), the 74.4% in Wolaita-Sodo, Ethiopia (Tesfamichael younger than 24 years. Those who disagreed with the binding A.A, et al.,2014) and the 67.8% in Juba, South Sudan rule of referral system were 1.4 times (AOR= 1.4 [1.02, 1.94]) (Basaza et al,2017). more likely to be willing to pay the 3% premium for SHI than This study found that age, periodic medical checkup those who agreed with the binding rule of referral system. unrelated to illness, rule of referral system to get higher level Those who disagreed with the exclusion of periodic medical of services, poor financing mechanism of government health checkup from the SHI were 1.4 times (AOR=1.41, 95%CI institutions, and perceived poor quality of health institutions [1.02, 1.95]) more likely to be willing to pay the 3% premium were the factors statistically significantly associated with for SHI than those who agreed with the exclusion of periodic willingness to pay for SHI. The study also indicated that medical checkup from the SHI. Those who didn’t consider older respondents (age greater than 39 years) were more the service quality in government health institutions to be likely to be willing to pay for SHI than younger respondents poor were 1.6 times (AOR=1.6, 95%CI [1.2, 2.1]) more (age less than 24 years), which is comparable with a study likely to be willing to pay the 3% premium for SHI than those conducted in South Africa where respondents age 35 and who considered the service quality in government health above more willing to pay for SHI than younger respondents institutions to be poor. Those who disagreed with the (Jane Goudge1, et al., 2018). Also, in St. Vincent and presence of financial insecurity in government health Grenadines, people age 31-45 were more willing to pay for institutions were 1.7 times (AOR= 1.7, 95%CI [1.23, 2.31]) SHI than younger respondents (age less than 30 years) more likely to be willing to pay the 3% premium for SHI than (Rosmond Adams et al, 2012). those who agreed with the presence of financial insecurity in Our study found that those who disagreed with the government health institutions. exclusion of periodic medical checkup from the SHI were more likely to be willing to pay for SHI than those who
4 Research Article East African Journal of Applied Health Monitoring and Evaluation
5 agreed with the exclusion of periodic medical checkup from and push them down to the poverty line. However, 685 the SHI. About 72.4% of study participants disagreed to pay (98.3%) respondents had expended in line with and below a 3% premium SHI due to prohibition of periodic medical the 20% WHO recommendation (WHO, 2010). However, checkup unrelated to illness. A similar study highlighted that about 408 (48.4%) of the study participants declared that respondents were willing to pay for SHI scheme with no OOP expenditure was not affordable and 470 (55.8%) were exclusions of certain services like dialysis and dental care worried about their future health service expenditure; this is over exclusions, and in terms of coverage, respondents were similar to a study done in Kosovo (Fatime Arenliu Qosaj, et more willing to pay for SHI to get full coverage of tests over al. 2018). On the other hand, 462 (54.8%) of the study partial coverage (Amarech Obse, et al., 2016). Furthermore, participants were concerned about the transparency and respondents who did not consider service quality in accountability of the insurance system in terms of government health institutions to be poor were more likely susceptibility to bias and misuse of SHI, which is similar to to be willing to pay for SHI than others. A similar study a study in Addis Ababa where 56.5% showed fear of SHI indicated that respondents who believed SHI scheme being exposed to corruption (LasebewY,M.et al, 2017). improve quality of health services were more likely to be willing to pay than those who did not believe the scheme to The overall mean level of negative perception towards improve quality (LasebewY,M, et al, 2017). However, referral chain, prohibition of periodic medical checkup studies done in Ghana and South Africa showed that once unrelated to illness, financial insecurity of government people are insured, they tend to perceive health services health facilities and compulsory payment of employed quality to be poor compared to the uninsured during health husband and wife to 3% premium of SHI in this study was service provision (Jane Goudge1, et al., 2018 and Kwasi S, 53% (95%CI [51%, 55%]), which is much higher than the 17% O., et al., 2018). Furthermore, those who disagreed with the finding of a study done in St. Paul's hospital in Addis Ababa, binding rule of referral system more likely to be willing to Ethiopia. This indicated that there is a high level of pay for the SHI than those who agreed with the binding rule unwillingness to pay for the newly proposed system of SHI, of referral system. and this may be due to socioeconomic differences. Our study also found that those who disagreed with the presence of financial insecurity in government health CONCLUSION institutions were more likely to be willing to pay for the SHI This study revealed that government employees’ willingness than those who agreed with the presence of financial to pay for SHI was low, and this could be due to poor health insecurity. Furthermore, 40.6% of the study participants service quality and readiness, insured referral system, service disagreed to pay the SHI premium due to fear of financial exclusion, periodic medical checkup unrelated to illness, insecurity of government health facilities, which is financial insecurity of health facility, and poor trust of SHI. analogous to a study done in Saudi Arabia where over two- Hence, SHI agencies should familiarize employees with the thirds (69.6%) of the respondents were willing to participate proclamation for SHI and the referral policy and improve in SHI if health facilities were well financed (Mohammed employees’ trust on the scheme to increase willingness to pay. Khaled Al, et al., 2018). Similarly, a study in northwest SHI agencies should also improve health facilities’ readiness Ethiopia affirmed that 77.5% of participants had poor and service quality. Finally, the government should attitudes on SHI, and 64.3% of respondents did not have trust reconsider the implementation of the federal proclamation in the ability of government’s SHI agency to present intended for SHI No 691/2010 accordingly. benefit packages (Yeshiwas S, et al.,2018). ACKNOWLEDGEMENTS The study findings also indicated that 87.7% of The authors would like to acknowledge Tigrai Health respondents had heard about the proposed SHI scheme, of Research Institute for funding the study. We extend our which 69.8% were willing to join, and this finding was gratitude to Tigrai Regional Health Bureau, the organizations, comparable with the 67% finding of a study in Addis Ababa study participants and the data collectors for their (LasebewY,M.et al,2017). Moreover, of all respondents of cooperation during the study time. our study, 32.1% disagreed to follow referral chain of health system. This could suggest that insured individuals should be CONFLICTS OF INTEREST allowed to choose which level to get care. Even though The authors declare that they have no competing interests. income did not have a significant association with willingness to pay, about 90.7% of respondents were in the AUTHORS' CONTRIBUTIONS lower income category, which is less than 2,280.85 Ethiopian AG conceived the study idea, designed tools and performed Birr (104 USD) per month. This inability to pay for SHI the analysis and the write-up. All the authors AB, MT & TT could be one reason for the low level of willingness to pay. have substantially contributed to the study’s design, data As far as level of health service expenditure is concerned, management, performance, write-up and reviewing of the 697 (82.7%) of the respondents visited health facilities, and manuscript. The authors agree to be accountable for all 12 (1.7%) of them had expended high amount of OOP aspects of the work related to the accuracy or integrity of any payment for HSU in the previous 12 months. This is beyond part of the work. All authors have read and approved the the WHO recommendation, which can cause people to manuscript. become impoverished due to catastrophic health expenditure REFERENCE 5 Research Article East African Journal of Applied Health Monitoring and Evaluation
6 Bump J, Cashin C, Chalkidou K, et al.(2016). Health Economics and Implementing pro-poor universal health coverage. Lancet Policy;2(4):159166http://www.sciencepublishinggroup.co Global Health.;4(1):e14–6. m/j/hepdoi:10.11648/j.hep.2017020 4.13 Averill C, Marriott A.(2013). Universal health coverage: Basaza et al. (2017). Willingness to pay for National why health insurance schemes are leaving the poor behind. Health Insurance Fund among public servants in Juba City, Oxford: Oxfam International. South Sudan: a contingent evaluation. International Journal Dieleman JL. Campbell M, Chapin A, Eldrenkamp E, for Equity in Health 16:158 DOI 10.1186/s12939-017- Fan VY, Haakenstad A, Kates J, Li Z, Matyasz T, Micah A, 0650-7 Reynolds A.(2017). Future and potential spending on health Jane Goudge1, et al. (2018). Social health insurance 2015–2040: development assistance for health, and contributes to universal coverage in South Africa, but government, prepaid private, and out-of-pocket health generates inequities: survey among members of a spending in 184 countries. Lancet.;389(10083):2005–30. government employee insurance scheme. International Spreeuwers, A.M., and G.J. Dinant. (2012). Success and Journal for Equity in Health, 17:1. Failure in Social Health Insurance in Sub-Saharan Africa: Rosmond Adams, Yiing-Jenq Chou and Christy Pu. What Lessons can be Learnt?’, Global Medicine, Official (2012). Willingness to participate and Pay for a proposed Magazine of IFMSA-NL. national health insurance in St. Vincent and the grenadines: Bank, W.(2011). International Finance Corporation, a cross-sectional contingent valuation approach Healthy Partnerships: How Governments Can Engage the Amarech Obse, et al. (2016). Eliciting preferences for Private Sector to Improve Health in Africa, Washington DC social health insurance in Ethiopia: a discrete choice Federal ministry of health. (2014). Improving health experiment. Health Policy and Planning, care financing in Ethiopia Jane Goudge1, et al. (2018). Social health insurance Comfort AB, a. Peterson LA, and Hatt LE. (2013). Effect contributes to universal coverage in South Africa, but of Health Insurance on the Use and Provision of Maternal generates inequities: survey among members of a Health Services and Maternal and Neonatal Health government employee insurance scheme. International Outcomes: A Systematic Review. J Health Popul Nutr; 31(4 Journal for Equity in Health, 17:1. Suppl 2): S81-105. Kwasi S, O., et al. (2018). Perceptions of healthcare De Allegri M and a.S.R. Sanon M. (2006). "To enroll or quality in Ghana: Does health insurance status matter? not to enroll?” A qualitative investigation of demand for PLoS ONE 13(1):e0190911. health insurance in rural West Africa. Soc Sci Med; Mohammed Khaled Al, et al. (2018). Investigating the 62:1520-1527. Willingness to Pay for a Contributory National Health Odeyemi IA.(2014). Community-based health insurance Insurance Scheme in Saudi Arabia: A Cross-sectional Stated programs and the national health insurance scheme of Preference Approach. Appl Health Econ Health Policy Nigeria: challenges to uptake and integration. Int J Equity 16:259–271 Health. 13(1):20. Yeshiwas S, et al. (2018). Civil servants' demand for Carrin G, James C. (2005). Social health insurance: key social health insurance in Northwest Ethiopia. BMC factors affecting the transition towards universal coverage. WHO. (2010). World health report. Health systems Int Soc Security Rev. 58:45–64 financing: the path to universal coverage. 1st ed. Geneva: Hsiao W.C and Shaw R.P. (2007). Social health World Health Organization; insurance for developing nations. Washington: the Fatime Arenliu Qosaj, et al. (2018). Catastrophic international bank for reconstruction and development. The expenditures and impoverishment due to out-of-pocket World Bank; 31-4. health payments in Kosovo. Federal Ministry of Health. (2008). Planning and Programming Department. Health Insurance Strategy. Addis Ababa; 13-7. Agago TA and a.O.S. Woldie M. (2014). Willingness to Join and Pay for the Newly Proposed Social Health Insurance among Teachers in Wolaita Sodo Town, South Ethiopia. Ethiop J Health Sci.; 24(3):195-202. Federal Democratic Republic of Ethiopia Ministry of Health, 2010/11-2014/15. (2010) Health Sector Development Program. Ministry of health Ethiopia; 33-7. Tesfamichael A.A, Mirkuzie W., and Shimeles O. (2014). Willingness to join and pay for the newly proposed social health insurance among teachers in Wolaita Sodo town, south Ethiopia Ethiop J Health Sci., Vol. 24, No. 3 LasebewY, M. and AbdelmenanS. (2017). Willingness to Pay for the Newly Proposed Social Health Insurance among Health Workers at St. Paul’s Hospital Millennium Medical College, Addis Ababa, Ethiopia. International Journal of 6
Research Article
2020; Issue 4 ISSN 2591 – 6769
Knowledge of HIV status among mothers accompanying their infants for immunization in Machakos, Kenya, 2014
Lilly M. Nyagah1, 2, Elvis Kirui3, Peter W. Young4, Joseph Gikunju2, Jane Githuku1, Sara Lowther5, Andrea A. Kim4
1 – Ministry of Health, Field Epidemiology and Laboratory Training Program, Nairobi, Kenya 2 – Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya 3 – National Public Health Laboratory, Kenya 4 – US Centers for Disease Control and Prevention, Division of Global HIV/AIDS & TB, Nairobi, Kenya 5 – US Centers for Disease Control and Prevention, Division of Global Health Protection, Nairobi, Kenya Correspondence to: Lilly M. Nyagah (MBCHB, [email protected], +254722326725)
Background Elimination of mother-to-child transmission of HIV is achievable through prevention of mother-to-child transmission (PMTCT) interventions. For HIV-positive mothers to access PMTCT services, they must be identified through HIV testing and counseling. We sought to measure the proportion of mothers with unknown HIV status and to determine associated factors.
Methods We recruited into a cross-sectional study 400 mothers accompanying their infants for routine immunization at mother-child health clinic (MCH) at Machakos Hospital, Kenya, collected information on HIV testing from their antenatal records, and offered opt- out HIV testing to those with unknown HIV status.
Results Overall, 304 (76.0%) mothers had unknown status at MCH, of whom 25 (8.6%) tested HIV-positive. HIV positivity was three times higher among mothers who never tested during pregnancy/delivery versus those whose last negative HIV test was >3 months prior to study enrolment (18% versus 6%). Women living more than three kilometers away from the health facility had 2.7-fold greater odds of having unknown status compared to those who lived less than three kilometers (adjusted odds ratio [AOR] 2.7, 95% CI 1.3–5.6). Married women had 3.2-fold greater odds of having unknown status as compared to those who were single (AOR 3.2, 95% CI 1.4–7.0). Those with education had a tenfold reduction in odds of having unknown status compared to those with none (AOR 0.1, 95% CI 0.0–0.2).
Conclusion Interventions are needed to improve women’s ANC attendance, uptake of HIV testing, and disclosure of HIV status during pregnancy to achieve the elimination of MTCT.
Keywords: HIV, PMTCT, Unknown HIV status, MCH
INTRODUCTION Sirengo et al., 2015). Mother-to-child transmission of HIV accounts for 90% of Kenya is implementing the UNAIDS strategy aimed at infections in children (Joint United Nations Programme on eliminating mother-to-child transmission of HIV, which is HIV/AIDS (UNAIDS), 2012). In 2017, there were 180,000 defined as achieving a rate of HIV transmission below 5% children who acquired HIV infection, 90% of whom live in among infants born to HIV-positive mothers (UNAIDS, sub-Saharan Africa. In the same year, Kenya was one of 22 2015). This reduction will be accomplished through several priority countries that had the highest burden of HIV-positive interventions, including screening of women for HIV pregnant women in need of prevention of mother-to-child infection during pregnancy, delivery and postnatal care; transmission. In Kenya it was estimated that 8,000 children initiating highly active antiretroviral therapy (HAART) for became newly infected with HIV in 2017 (Joint United all HIV-positive pregnant and lactating mothers; and Nations Programme on HIV/AIDS (UNAIDS), 2017, 2018; maintaining these mothers on treatment for life.
7 Research Article East African Journal of Applied Health Monitoring and Evaluation
2 HIV-infected pregnant and lactating women need to learn born outside a health facility, they should be taken to a health their HIV-positive status in order to access PMTCT facility to receive the birth oral polio vaccine. All mothers interventions. The World Health Organization (WHO) accompanying their infants for either the first dose of recommends that pregnant mothers be tested at their first pentavalent vaccine or birth oral polio vaccine were eligible antenatal care (ANC) visit and those with HIV-negative and recruited consecutively into the study. After obtaining results be retested three months later. All women not written consent from participants, we conducted face-to-face previously diagnosed with HIV should be retested at delivery. interviews using a structured questionnaire to collect self- Breastfeeding mothers with unknown HIV status are also reported information on the mother’s obstetric history, access tested (WHO, 2015). Targeting HIV testing for PMTCT to antenatal services and delivery. HIV testing history for during the antenatal period in Kenya should be an effective antenatal and labor and delivery was abstracted from the way to diagnose HIV-positive pregnant women because 96% ANC mother-baby booklet. The interview was conducted by of all pregnant women attend at least one ANC visit, and the a clinical officer working in Machakos Hospital who was proportion of pregnant women who accept HIV testing and trained by the investigator. Quality was assured through counseling services during ANC is also high, at 94% (Kenya supervision and record review by the investigator. National Bureau of Statistics (KNBS), 2014; National AIDS and STI Control Programme (NASCOP), 2012b). However, The PMTCT Cascade the remaining 4% of women who do not attend ANC during The PMTCT cascade begins at pregnancy, where all pregnant their pregnancy and 6% of ANC attendees who do not accept women with unknown HIV status should be offered opt-out HIV testing and counseling during pregnancy represent a testing at the first ANC visit, and a repeat HIV test (after 3 group of women with missed opportunities for PMTCT, if months) in the third trimester should be offered to all women. they are HIV-infected. Pregnant and breastfeeding women Women who decline HIV testing at the first antenatal visit are also at increased risk of HIV infection and of transmitting should be offered HIV testing again in subsequent visits. the virus to their children due to high viral load, hence the Women presenting in labor without documented HIV testing importance of retesting women who test HIV-negative at should have opt-out testing done urgently. HIV counseling regular intervals during pregnancy and breastfeeding. and testing should be offered to all women with unknown The mother-to-child transmission rate in Kenya was high HIV status at each postnatal visit. For HIV-negative mothers in 2018 at 11.5% (National AIDS Control Council (NACC), not screened previously during the post-natal period, 2018b, 2018a), despite adopting and implementing the WHO screening at first child immunization is also recommended. guidelines since 2012 (NASCOP, 2012a). Identifying factors If a woman is diagnosed with HIV at any of these stages then that impede access to PMTCT services among HIV-positive she is linked to HIV care and initiated on ART (National women is an important step towards implementing AIDS and STI Control Programme (NASCOP), 2012a). appropriate interventions and eliminating mother-to-child transmission of HIV. Although studies have assessed HIV Measures testing practices during the perinatal and maternity period, it We defined an opportunity for HIV testing as occurring at is not known how many HIV-positive mothers have unknown each point in a woman’s pregnancy when she should have HIV status when they present at MCH for their infants’ been tested for HIV as per national guidelines (National immunization visits. Therefore, we sought to estimate the AIDS and STI Control Programme (NASCOP), 2012a). proportion of mothers with unknown HIV status presenting Women were considered to have a known HIV status if they for routine infant vaccination. We also sought to identify the had medical record documentation of HIV-positive status time points where the opportunity to test was missed and to prior to their first ANC visit, were newly-diagnosed with identify factors associated with unknown HIV status. HIV infection during their ANC visits or during delivery, or had HIV-negative status based on tests conducted within METHODS three months prior to enrolment in the study. Mothers with Study setting unknown HIV status were defined as those women who did In 2017, Machakos County, Kenya, had an estimated not know their HIV status because they self-reported that population of 1.2 million and a HIV prevalence rate of 3.8%. they had never been tested for HIV or had their last HIV- During the same period, it was estimated that 1000 people negative test more than three months prior to enrolment in were infected annually, 18.5% of whom were children the study. The time from HIV test was computed as the (NACC, 2018b). According to the Machakos County HIV difference between the recorded date of the last HIV test and profile, only half of all pregnant women attended the date of enrolment into the study. recommended four ANC visits, and 31.9% of all HIV- positive pregnant women delivered in a hospital in the same Sampling procedures county (National AIDS Control Council (NACC), 2016). At the time of the survey, 46% of adults in Machakos County had never been tested for HIV. A minimum sample size of Study design 381 mothers was chosen to achieve a 95% confidence We conducted a facility-based cross-sectional study in interval of +/- 5% for the proportion of women never tested Machakos County Referral Hospital MCH services from July for HIV (46%, based on estimated proportion of adults never 2014 to January 2015. The first dose of pentavalent vaccine tested for HIV in Machakos County). Mothers were enrolled is usually given to infants at six weeks of age. If an infant is consecutively until the desired sample size was achieved. 8 Research Article East African Journal of Applied Health Monitoring and Evaluation
3 Laboratory methods Table 1: Socio – demographics (baseline) characteristics of Service providers who routinely offer HIV testing at MCH mothers accompanying infants for immunization at MCH in conducted HIV testing and counseling for all women enrolled in the study who had either never been tested or had Machakos County Referral Hospital, Kenya, 2014 a HIV test more than three months prior to enrolment; testing All mothers Mothers with unknown Select variables was offered on an opt-out basis. Rapid HIV testing followed (N400) HIV status (N=304) national guidelines and was conducted using the Determine test kit as the screening assay (Abbott Diagnostics, Abbott Age (years) n (%) n (%) Park, IL). If the screening assay was reactive, a confirmation 15-19 53 (13.3) 37 (12.2) test was performed using the Unigold test kit (Trinity Biotech 20-24 157 (39.3) 120 (39.5) PLC, Ireland). A blood specimen was classified as HIV- positive when it was reactive on both the screening and 25-29 90 (22.5) 72 (23.7) confirmatory assays. A blood specimen was classified as 30 and above 100 (25) 75 (24.7) HIV-negative if non-reactive on the screening assay. For Marital Status specimens with discrepant screening and confirmatory test Single 90 (22.5) 89 (23.4) results, dry blood spots were collected for confirmatory testing by polymerase chain reaction (PCR). As required in Married 310 (77.5) 291 (76.6) Kenya’s national HIV testing guidelines, we provided pre- Level of Education test and post-test counseling for all clients, and those who No school 18 (4.5) 5 (1.6) tested HIV-positive were linked to HIV care and HIV- exposed infant follow-up services (Ministry of Health Primary school 154 (38.5) 115 (37.8) National AIDS and STI Control Program, 2013). Secondary 137 (34.3) 118 (38.8) College/University 91 (22.8) 66 (21.7) Data analysis We described study participants by age, marital status, level Occupation of education, total number of children, occupation, and Housewife/No employment 217 (54.3) 158 (52.0) distance of residence from health facility. Bivariate analysis Self-employed/Casual 65 (16.3) 50 (16.5) compared demographic and clinical characteristics of Formal 62 (15.5) 53 (17.4) enrolled women with unknown HIV status to those who knew their HIV status. Variables that were found to be Student 36 (9) 33 (10.9) significant (p<0.25) in bivariate analysis or considered to be Self-employed/Formal 20 (5) 10 (3.3) potential confounders were included in a stepwise logistic Distance from health facility regression model to identify independent factors associated with missed opportunity. Variables that remained significant 1km 55 (13.8) 50 (16.5) (p<0.05) in the model were considered to be statistically 2km 51 (12.8) 42 (13.8) significant. All statistical analyses were performed using 3km 68 (17) 53 (17.4) STATA version 13 (StataCorp, Texas). 4km 50 (12.5) 30 (9.9) Ethical considerations More than 5 km 176 (44.0) 129 (42.3) The respondents provided written informed consent for study No of children procedures prior to enrolment in the study. The study 1 209 (52.3) 159 (52.3) protocol was reviewed and approved by Kenyatta National Hospital Institutional Review Board in Nairobi, Kenya. 2 92 (23) 72 (23.7) Institutional approval for the study was granted by the 3 53 (13.3) 35 (11.5) Machakos County Referral Hospital administration. >3 46(11.5) 38 (12.5)
HIV status at the time of infant RESULTS A total of 400 mothers were enrolled in the study. No mother immunization* refused to participate in the study. Half of mothers were aged HIV-positive 14 (14.5) N/A below 25 years and 77% were married (Table 1). Seventy-six percent of all mothers had an unknown HIV status at MCH, HIV-negative 62 (64.6) N/A 19% knew their HIV status, 5% declined to disclose their Missing¶ 20 (20.8) N/A HIV status. Eighty percent of the mothers with unknown * Refers to HIV status just prior to study enrolment. † Women who had never status were in the category because they had their last HIV previously tested for HIV or whose last HIV-negative test was > 3 months test more than three months prior to the study, while the remaining 20% had not been tested through the cascade of prior. ¶ Women who refused to disclose their HIV status at time of study PMTCT. enrolment
9 Research Article East African Journal of Applied Health Monitoring and Evaluation
4 When retesting was done, 5.8% of the mothers without a eligible for HIV testing. One in five of those mothers with recent HIV test had an HIV-positive result, while 17.8% of unknown HIV status in MCH had missed an opportunity for the mothers who had been missed tested HIV-positive HIV testing at every point in the PMTCT cascade, either (Figure 1). because they did not attend ANC and maternity, or they were not tested for HIV at either ANC or maternity. This could be Figure 1. HIV status at enrolment and after testing at MCH an indication of inconsistent implementation of testing among women attending routine child immunization visits, guidelines in PMTCT (Rogers et al., 2016). A similar study Machakos County Referral Hospital, Kenya, 2014 (n=400) conducted in South Africa found that a third of mothers missed at least one step in the PMTCT cascade (Woldesenbet et al., 2015). Eight percent of the mothers in MCH with unknown HIV status and eligible for HIV testing at vaccination tested HIV- positive. HIV positivity among those mothers who had a HIV test greater than three months prior was 6% while it tripled for those who had been missed completely through the cascade. This is similar to the findings in a study conducted in western Kenya that found that pregnant women who had never attended ANC were about 6 times more likely to newly test HIV-positive compared to those who had attended ANC (Ndege, Washington, Kaaria, & Meara, 2016). Although acceptance for HIV testing is high in Kenya, failure to attend ANC remains a barrier to accessing PMTCT services (Delva et al., 2010). Of the HIV-positive women who were enrolled in our study, 44% were diagnosed at the time of routine infant In total, there were 62 (15.5%) mothers who missed HIV vaccination. We cannot rule out that some of these women testing at every point during their ANC or maternity care. might have already known that they were HIV-infected but Two thirds of these mothers were missed either at ANC or did not disclose their HIV status during ANC or maternity maternity while 27% did not attend ANC but were missed at care visits in health facilities. Failure to disclose HIV status maternity (Figure 1). In the bivariate analysis, among the 304 has been found to be higher among HIV-infected mothers and mothers who had an unknown HIV status at time of infant has been associated with poor use of PMTCT services immunization, 129 (42.4%) were older than 25 years, 299 (Kinuthia et al., 2018; Spangler, Onono, Bukusi, Cohen, & (98.4%) had at least primary school education, 231 (76.0%) Turan, 2014). Irrespective of reason, the unavailability of had fewer than 2 children, 159 (52.3%) were living more than HIV status among these HIV-infected mothers in a clinical 3 kilometers away from the health facility, and 292 (96.0%) setting prevents access to PMTCT interventions to reduce the had made at least one ANC visit. risk of perinatal transmission to their babies. Decreasing In unadjusted analysis, having unknown HIV status was stigma and enhancing health education opportunities for associated with being married, no education and living a long pregnant women may help improve disclosure. Antenatal and distance from the health facility (Table 2). After adjusting for maternity care is often one of the busiest sectors of the health potential confounders, women who lived more than three facility, resulting in short consultations with minimum time kilometers away from the health facility had 2.7 greater odds available for counseling. These insights provide useful of having unknown status compared to those who lived less suggestions for improving interventions targeted at than three kilometers away (AOR 2.7, 95% CI 1.3–5.6). eliminating MTCT. Women who were married had 3.2 greater odds of having The study found that living more than three kilometers unknown status as compared to those who were single (AOR away from a health facility was associated with unknown 3.2, 95% CI 1.4–7.0). Finally, women who were educated HIV status, and may relate to ease of access to HIV testing had a ten-fold reduction in odds of having unknown status and counseling. Distance has been shown to hinder access to compared to those with no education (AOR 0.1, 95% CI 0.0– various health services in the region (Asweto, Aluoch, 0.2). Obonyo, & Ouma, 2014; Golub, Herman-roloff, Hoffman, Jaoko, & Bailey, 2017; Hanson et al., 2015). Addressing the DISCUSSION issue of access to a health facility by bringing services closer Although PMTCT interventions have been shown to to the community may help bridge the gaps. Kenya has significantly reduce transmission of HIV to infants, benefits introduced mobile clinics to address this issue. In contrast to take effect only once the mother has been diagnosed with a similar study conducted in Kampala (Namara-Lugolobi EC HIV infection and receives treatment. We found that three et al, 2015), our study did not find significant associations quarters of the mothers attending MCH clinic in Machakos between unknown HIV status with age and marital status. Hospital had an unknown HIV status and therefore were
10 Research Article East African Journal of Applied Health Monitoring and Evaluation
5 Table 2. Factors associated with unknown HIV status among women accompanying their infants for immunization at MCH services in Machakos County Referral Hospital, Kenya, 2014 Unknown HIV status Known HIV status Select variables All (N = 380) * Unadjusted Adjusted (N=304) (N=76)
Age n (%) n (%) n (%) OR (95% CI) OR (95% CI) ≥25 years 213 (47.1) 175 (57.6) 38 (50.0) 1
˂25 years 167 (52.9) 129 (42.4) 38 (50.0) 1.3 (0.8- 2.2) Marital Status Living Alone 89 (23.4) 79 (26.0) 10 (13.2) 1 Living with a Partner 291 (76.6) 225 (74.0) 66 (86.8) 2.31 (1.3-4.7) 3.2 (1.4-7.0) Level of Education No education 17 (4.5) 5 (1.6) 12 (15.8) 1 Educated 363 (95.5) 299 (98.4) 64 (84.2) 0.1 (0.0-0.3) 0.1 (0.0-0.2) Occupation Employed 141 (37.1) 113 (37.2) 28 (36.8) 1 Unemployed 239 (62.9) 191 (62.8) 48 (63.2) 1.0 (0.6-1.7) Distance from health facility Less than 3 KM 167 (43.9) 145 (47.7) 22 (28.9) 1 Above 3 KM 213 (56.1) 159 (52.3) 54 (71.1) 2.2 (1.3-3.8) 2.7 (1.3-5.6) No of children More than 2 94 (24.7) 73 (24.0) 21 (27.6) 1 Fewer than 2 286 (75.3) 231 (76.0) 55 (72.4) 0.8 (0.5-1.5) ANC Visits None 19 (5) 12 (4.0) 7 (9.2) 1 At least one visit 361 (95) 292 (96.0) 69 (90.8) 0.4 (0.2-1.1)
* 20 women who refused to disclose their HIV status are excluded
LIMITATIONS of HIV status during pregnancy should be used to optimize Some of the information collected from mothers was self- PMTCT outcomes in order to accelerate the attainment of the reported and vulnerable to recall and social desirability bias. ambitious UNAIDS 90-90-90 goals where we aim to identify Associations with unknown HIV status were limited to a 90% of all people living with HIV, put on treatment 90% of small number of demographic and clinical variables those, and achieve viral suppression. These efforts will collected and the relatively small sample size. These findings accelerate elimination of mother-to-child transmission of cannot be generalized beyond the population of pregnant HIV. women that attend ANC, maternity, or MCH in Machakos County. ACKNOWLEDGEMENTS The authors would like to thank the respondents from CONCLUSION Machakos Hospital who participated in the study and We identified a number of HIV-infected mothers who had Thomas Muthusi, the clinical officer/ HIV counsellor in unknown HIV status at the time of infant vaccination. Machakos Hospital, for providing the services to the mothers. Maternal child health clinics provide an important avenue for We also wish to acknowledge the technical support from identifying mothers who were missed and linking them and Alfred Musekiwa and Dorothy Southern. their babies to appropriate care. With adoption of highly active antiretroviral therapy for life, innovative testing and COMPETING INTERESTS counseling methods need to be used to encourage mothers to The authors declare that they have no competing interests. test for HIV infection or disclose prior knowledge of HIV- positive status in order to benefit from PMTCT interventions. AUTHORS’ CONTRIBUTIONS Community and peer-based interventions can also improve LMN: conceptualized study, drafted the protocol, drafted uptake of testing and disclosure. manuscript; EK: analyzed data; Joseph G, Jane G & AAK: Interventions to improve ANC attendance, uptake of HIV reviewed and commented on the protocol. PWY, Joseph G, testing services across the PMTCT cascade, and disclosure Jane G, SL and AAK revised manuscript. All authors
11 Research Article East African Journal of Applied Health Monitoring and Evaluation
6 provided substantial input/edits to the manuscript and have Ndege, S., Washington, S., Kaaria, A., & Meara, W. P. read and approved the final manuscript. (2016). HIV Prevalence and Antenatal Care Attendance among Pregnant Women in a Large Home-Based HIV REFERENCES Counseling and Testing Program in Western Kenya. 995, 1– Asweto, C. O., Aluoch, J. R., Obonyo, C. O., & Ouma, J. 10. https://doi.org/10.1371/journal.pone.0144618 O. (2014). Maternal Autonomy, Distance to Health Care Rogers, A. J., Weke, E., Kwena, Z., Bukusi, E. A., Oyaro, Facility and ANC Attendance: Findings from Madiany P., Cohen, C. R., & Turan, J. M. (2016). Implementation of Division of Siaya County, Kenya. 2(4), 153–158. repeat HIV testing during pregnancy in Kenya: a qualitative https://doi.org/10.12691/ajphr-2-4-5 study. BMC Pregnancy and Childbirth, 16(1), 151. Delva, W., Yard, E., Luchters, S., Chersich, M. F., Muigai, https://doi.org/10.1186/s12884-016-0936-6 E., Oyier, V., & Temmerman, M. (2010). A Safe Motherhood Sirengo, M., Muthoni, L., Kellogg, T. A., Kim, A. A., project in Kenya: assessment of antenatal attendance, Katana, A., Mwanyumba, S., … Study, K. (2015). Mother- service provision and implications for PMTCT. 15(5), 584– to-Child Transmission of HIV in Kenya: Results From a 591. https://doi.org/10.1111/j.1365-3156.2010.02499.x Nationally Representative Study. 66, 66–74. Golub, G., Herman-roloff, A., Hoffman, S., Jaoko, W., & Spangler, S. A., Onono, M., Bukusi, E. A., Cohen, C. R., Bailey, R. C. (2017). The relationship between distance and & Turan, J. M. (2014). HIV-positive status disclosure and use post-operative visit attendance following medical male of essential PMTCT and maternal health services in rural circumcision in Nyanza Province, Kenya. 20(11), 2529–2537. Kenya. Journal of Acquired Immune Deficiency Syndromes https://doi.org/10.1007/s10461-015-1210-z.The (1999), 67 Suppl 4(Suppl 4), S235–S242. Hanson, C., Cox, J., Mbaruku, G., Manzi, F., Gabrysch, S., https://doi.org/10.1097/QAI.0000000000000376 Schellenberg, D., … Foundation, M. G. (2015). Maternal UNAIDS. (2015). A progress report on the Global Plan mortality and distance to facility-based obstetric care in rural towards the elimination of new HIV infections among southern Tanzania: a secondary analysis of cross-sectional children by 2015 and keeping their mothers alive. census data in 226 000 households. The Lancet Global Woldesenbet, S., Jackson, D., Lombard, C., Dinh, T.-H., Health, 3(7), e387–e395. https://doi.org/10.1016/S2214- Puren, A., Sherman, G., … Team, S. A. P. E. (SAPMCTE). 109X(15)00048-0 (2015). Missed Opportunities along the Prevention of Joint United Nations Programme on HIV/AIDS Mother-to-Child Transmission Services Cascade in South (UNAIDS). (2012). Global report: UNAIDS report on the Africa: Uptake, Determinants, and Attributable Risk (the global AIDS epidemic SAPMTCTE). PloS One, 10(7), e0132425–e0132425. Joint United Nations Programme on HIV/AIDS https://doi.org/10.1371/journal.pone.0132425 (UNAIDS). (2017). Ending AIDS. World Health Organization. (2015). Consolidated Joint United Nations Programme on HIV/AIDS guidelines on HIV testing services. (UNAIDS). (2018). UNAIDS data. Kenya National Bureau of Statistics (KNBS). (2014). Kenya Demographic and Health Survey (KDHS). Kinuthia, J., Singa, B., McGrath, C. J., Odeny, B., Langat, A., Katana, A., … John-Stewart, G. (2018). Prevalence and correlates of non-disclosure of maternal HIV status to male partners: a national survey in Kenya. BMC Public Health, 18(1), 671. https://doi.org/10.1186/s12889-018-5567-6 Ministry of Health National AIDS and STI Control Program. (2013). HIV testing guidelines. Namara-Lugolobi EC et al. (2015). Prevalence and predictors of unknown HIV status among women delivering in Mulago National Referral Hospital, Kampala, Uganda. Afr Health Sci. 2017;17(4):963–973. Doi:10.4314/Ahs.V17i4.3, 124(3), 235–239. https://doi.org/10.1016/j.ijgo.2013.08.011.Predictors National AIDS and STI Control Programme (NASCOP). (2012a). Guidelines for Prevention Of Mother To Child Transmission (PMTCT ) of HIV / AIDS in Kenya. National AIDS and STI Control Programme (NASCOP). (2012b). Kenya AIDS Indicator Survey (KAIS). National AIDS Control Council (NACC). (2016). Kenya HIV County Profiles 2016. National AIDS Control Council (NACC). (2018a). Kenya AIDS response progress report 2018. National AIDS Control Council (NACC). (2018b). Kenya HIV estimates. 12
Research Article
2020; Issue 4 ISSN 2591 – 6769
Trend and predictors for early infant diagnosis by PCR among HIV-exposed infants in Dar es Salaam region, Tanzania, 2014-2016
Michael F. Mboya, MD, MSc1,3*; Prosper Njau, MD, MSc1,3; Jim Todd, PhD1,2; Beatrice John Lerayo, MSc1; Goodluck Wiley Lyatuu, MD, MPH4; Lameck Machumi, MD4; Sia E. Msuya; MD, PhD1; Michael J. Mahande, PhD1; Jenny Renju, PhD1,2
1Department of Epidemiology & Biostatistics, Institute of Public Health, Kilimanjaro Christian Medical University College, Moshi, Tanzania 2Department of Population Health, London School of Hygiene & Tropical Medicine, London, UK 3Ministry of Health, Community Development, Gender, Elderly and Children, Dodoma, Tanzania 4Management and Development for Health, Dar Es Salaam, Tanzania
Email address: [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected] *Corresponding author Michael Fred Mboya, P.O Box 2240 Moshi, Tanzania, [email protected]
Background Early infant diagnosis (EID) of HIV and timely initiation of antiretroviral therapy reduces morbidity and mortality in HIV- infected infants and children. Although the WHO had a global target for 2018 that 90% of HIV-exposed infants (HEI) should be tested by the age of 8 weeks, by 2015 only half were being tested. This study describes trends and predictors for HIV first testing in HEI in Dar es Salaam, Tanzania, from 2014 to 2016.
Methods We conducted a cross-sectional study of HEI attending HIV services using secondary data collected at health facilities and collated at the national level. We estimated odds ratios and 95% confidence intervals for factors affecting uptake of HIV testing.
Results Of 12,117 HEI, the proportion tested for HIV by age 8 weeks increased from 53.2% in 2014 to 69.2% in 2016; 2.3% were HIV- positive. Replacement feeding (aOR=2.94, 95% CI 2.31 – 3.74) and receiving nevirapine prophylaxis (aOR=1.55, 95% CI 1.28- 1.88) were predictors for EID testing uptake. HEI born to mothers with WHO stage II (aOR=0.53, 95% CI 0.41 – 0.67), stage III (aOR=0.64, 95% CI; 0.52 – 0.79) and stage IV (aOR=0.58, 95% CI 0.34 – 0.99) were less likely to be tested than those born to mothers with WHO stage I disease progression.
Conclusion There was an increasing trend in the uptake of HIV testing of infants at age 8 weeks during the study period. However, it is still below the global target. Efforts to promote EID testing are still needed.
Keywords: HIV early infant diagnosis, perinatal transmission, Tanzania
INTRODUCTION Nations General Assembly Special Session (UNGASS) and In 2010, the World Health Organization (WHO) published a Millennium Development Goals (WHO, 2010). Without strategic vision for prevention of mother-to-child antiretroviral therapy (ART), one third of all HIV-infected transmission (PMTCT) of HIV, which aimed to reduce the children are likely to die during the first 18 months of life and rate of mother-to-child transmission (MTCT) to 5% or less half before they reach 24 months (UNAIDS, 2016). In among breastfeeding women, and to 2% or less among non- Tanzania, the prevalence of HIV was 5.6% among pregnant breastfeeding women by 2015, consistent with United women attending antenatal care in 2011 (Manyahi et al.,
13 Research Article East African Journal of Applied Health Monitoring and Evaluation
2 2017). According to the UNAIDS progress report on the attending HIV services in the Dar es Salaam region between global plan for 21 priority countries, the estimated prevalence January 2014 and December 2016 were used in the study. Dar of HIV among HIV-exposed infants (HEI) at 6 weeks of age es Salaam region is the largest city in Tanzania, with a was 3% in 2014 (UNAIDS, 2015). population of over 4.4 million people (National Bureau of Statistics, 2013) and 248 health facilities providing PMTCT Testing HIV-exposed infants (HEI) early (within 4-6 weeks) services. Management and Development for Health (MDH) is referred to as early infant diagnosis (EID), and it enables manages 192 of the 248 PMTCT facilities and collects and the timely initiation of ART and dramatically reduces collates electronic data for patient monitoring. morbidity and mortality in those infants found to be HIV- infected (Aledort et al.,2006; National PMTCT Guideline, Study population 2013). Tanzania adopted the WHO guideline for EID in 2013, All HIV-infected mothers and their exposed infants up to 18 which recommends a first HIV test with DNA PCR for months of age registered at the 192 MDH-supported facilities exposed children at 4 to 6 weeks postpartum or soon were included in the study. Participants were excluded if they thereafter (National PMTCT guideline, 2013). In the same were missing dates of birth. All patients received the national year, Tanzania published a national target to ensure that by recommended counselling prior to HIV testing of the infant, 2018, 90% of all HEI were tested between 4 to 8 weeks of age with information on the benefits for the infant including free and received their results within 4 weeks (TACAIDS, 2013). ART treatment for infants who test HIV-positive. Since then, services to prevent MTCT, including EID testing, have been integrated into routine reproductive and child Data sources health services alongside routine childhood immunisation. The study data were extracted from the MDH electronic The integration of services aims to increase accessibility and patient-level database at the 192 PMTCT facilities. The thereby promote the uptake of testing, whilst also providing electronic database contained data for the mother and the HEI the opportunity to provide health education to mothers and who each had unique identifying numbers, with a link caregivers (National PMTCT guideline, 2013). between mother and infant pairs. The information from the mother’s treatment card and child’s HEI diagnostic card were Globally, the uptake of HIV testing at 4 to 6 weeks remains extracted from the patient-level electronic database. We also low. In 2015, only half (50%) of HEI were reported to have extracted potential predictor variables including mother’s age, been tested by 8 weeks old (UNICEF, 2015). Whilst uptake marital status, date started on ART, CD4 count, and WHO of EID has been increasing in some countries, such as South stage, and the HEI’s date of birth, birth weight, sex, type of Africa where it increased from 31.4% in 2008 to 54.7% in prophylaxis given, cotrimoxazole initiation, and the date 2010 (Sherman et al., 2014) and Thailand where it increased when HIV test results were reported. from 35.2% in 2008 to 50.5% in 2011 (Naiwatanakul et al., 2016), rates remain below global targets. A study conducted Study variables in Tanzania in 2013 found a decreasing trend in the median The main outcome variable was uptake of EID. EID was age at first DBS test (Chiduo et al., 2013). Various factors defined as a DNA PCR HIV test that was performed within have been reported to influence uptake of EID services. These eight weeks following birth. HIV infection was defined as a include infant feeding options, knowledge of caregivers on positive HIV test result using DNA PCR. when EID should be done, stigma and fear of disclosure of mothers’ status, travel distance to health facilities, out-of- Data analysis pocket costs at the health facility, and long waiting times We performed data analysis using STATA (StataCorp. College (Hassan et al., 2012; Cook et al., 2012; Ramaiya et al.,2016). Station, Texas 77845 USA) version 13.0. We summarised This study builds on the limited information in Tanzania on descriptive statistics using frequencies for categorical the proportion of HEI tested by EID by PCR, and advances variables and central tendency and dispersion for continuous our understanding of what variables predict optimal uptake of variables. We calculated odds ratios (OR) and 95% EID. It also improves our understanding of the current confidence interval (CI) for the factors affecting uptake of intervention progress to achieve the WHO targets, reduce HIV testing using a multivariable logistic regression model. MTCT to less than 2%, and identify the gaps in the Adjusted odds ratios (aOR) and 95% CI are shown for all programmes which are prohibiting the achievement of the predictors included in the final model with a p-value of less testing targets. This study aimed to describe the trend and than 5%, which was considered statistically significant. predictors for EID testing of HEI in Dar es Salaam from 2014 to 2016. Ethical considerations We obtained ethical clearance (989) from Kilimanjaro METHODS Christian Medical University College and permission from Study design and setting the MoHCDGEC and MDH to use the data of Dar es Salaam We designed a cross-sectional study using secondary data region. To ensure anonymity, only identity numbers were routinely collected at health facilities and collated at the used to represent the subjects. national level by the Ministry of Health, Community Development, Gender, Elderly and Children (MoHCDGEC). RESULTS Data on all HEI infants born to HIV-infected mothers Characteristics of the study participants 14 Research Article East African Journal of Applied Health Monitoring and Evaluation
3 A total of 12,417 HEI were enrolled into the PMTCT Table 1. More than half (6,345; 52.3%) were female; the programme in the health facilities managed by MDH during median age was 1.4 (IQR 1.7-3.1) months. 7,670 (63.3%) of the period of 2014-2016. Of these, 300 were missing date of the infants initiated nevirapine syrup; 9,883 (81.5%) received birth and excluded from the analysis (Figure 1). Of the total cotrimoxazole treatment; and 1,806 (14.9%) were on sample of 12,117 HEI over the four years, 2,030 (17%) had replacement feeding. Maternal characteristics of the study tested for antibody at first test (aged above 9 months) and participants are shown in Table 2. The mothers’ mean age was 10,087 (83%) received a DNA PCR HIV test. Of the 10,087, 30.8 (SD, 6.0) years; 6,925 (57.2%) were between 25 and 35 2,347 did not have the HIV results recorded in the database years. Only 1,102 (9.1%) had been on ART for more than 3 and were excluded from prevalence calculations. Out of 7,740 years; 7,394 (61.0%) were in WHO stage 1 at the time of with a DNA PCR test result, 177 (2.3%) were HIV-positive. enrolment into EID care. The mothers’ median CD4 was 548 cells/µL (IQR, 382-655) with 6.4% (774) having CD4 cell The characteristics of the 12,117 infants are presented in count less than 200 cells/µL.
Table 1: Characteristics of HIV-exposed infants, Dar es Salaam, Tanzania, 2014-2016 (N=12,117) Variable Total 2014 2015 2016 N (%) N (%) N (%) N (%) 12,117 4,660 4,037 3,420
Sex Female 6,345 (52.3) 2,341 (50.2) 2,146 (53.2) 1,858 (54.3) Male 5,772 (47.7) 2,319 (49.8) 1,891 (46.8) 1,562 (45.7) Age of child at test (months) <2 7,160 (59.1) 2,481 (53.2) 2,313 (57.3) 2,366 (69.2) ≥2-18 4,957 (40.9) 2,179 (46.8) 1,724 (42.7) 1,054 (30.8) Median age (Range) 1.4 (1.7 - 3.1) Birth weight of infant (kg) < 2.5 3,224 (26.6) 1,126 (24.2) 1,540 (16.3) 558 (16.3) 2.5 and above 8,622 (71.2) 3,463 (74.3) 2,439 (60.4) 2,720 (79.5) Missing 271 (2.2) 71 (1.5) 58 (1.4) 142 (4.2) Median birth weight (Range) 2.70 (2.4 - 3.3) Feeding practice at testing Exclusive breast feeding 6,149 (50.7) 1,790 (38.4) 1,766 (43.8) 2,593 (75.8) Mixed feeding 1,753 (14.5) 764 (16.4) 816 (20.2) 173 (5.1) Replacement feeding 1,806 (14.9) 838 (18.0) 732 (18.1) 236 (6.9) Missing 2,409 (19.9) 1,268 (27.2) 723 (17.9) 418 (12.2)
Nevirapine No 1,606 (13.3) 693 (14.9) 715 (17.7) 198 (5.8) Yes 7,670 (63.3) 2,923 (62.7) 2,518 (62.4) 2,229 (65.2) Missing 2,841 (23.4) 1,044 (22.4) 804 (19.9) 993 (29.0) Cotrimoxazole No 1,148 (9.5) 460 (9.9) 418 (10.4) 270 (7.9) Yes 9,883 (81.5) 3,907 (83.8) 3,195 (79.1) 2,781 (81.3) Missing 1,086 (9.0) 293 (6.3) 424 (10.5) 369 (10.8)
Trend in the uptake of early infant diagnosis among HEI Predictors associated with uptake of EID at less than 8 Of the 12,117 infants and children in our study population, weeks of age 7,160 (59.1%) children tested for DNA PCR at less than 8 The factors associated with uptake of EID in crude and adjusted weeks age. The proportion of HEI testing before 8 weeks of analyses are shown in Table 3. The unadjusted analysis shows age increased over time from 53.2% (2,481/4,660) in 2014 to that male infants had lower odds (OR=0.91, 95% CI 0.88 – 0.98) 57.3% (2,314/4,037) in 2015 and 69.2% (2,366/3,420) in of being tested within 8 weeks compared to female infants. 2016 (Table 3). Infants who had not been exclusively breastfed had lower odds
15 Research Article East African Journal of Applied Health Monitoring and Evaluation
4 (OR=0.63, 95% CI 0.57 – 0.70) of HIV testing compared to Infants born to mothers aged above 35 years had higher odds infants who were exclusively breastfed. Infants born to mothers (OR=1.18, 95% CI 1.06 – 1.32) of being tested compared to with higher WHO stages were less likely to be tested when those who were born to mothers age less than 25 years. compared to those who were born to mothers with WHO stage Furthermore, infants born to mothers who were married had 1 [(OR=0.24, 95% CI 0.21 – 0.26), (OR=0.61, 95% CI 0.55 – higher odds (OR= 1.19, 95% CI 1.08 – 1.31) of being tested 0.68), (OR=0.55, 95% CI 0.43 – 0.71) for WHO stage 2, 3 and compared to those born to unmarried mothers. 4, respectively]. Infants weighing more than 2.5 kg at birth had higher odds (OR=1.18, 95% CI 1.09 – 1.28) of being tested compared to their counterparts who were born at <2.5kg.
Table 2: Characteristics of HIV-infected mothers, Dar es Salaam, Tanzania, 2014-2016 (N=12,117) Variable Total 2014 2015 2016 N (%) N (%) N (%) N (%) 12,117 4,660 4,037 3,420
Age of mother (years) Less 25 2,208 (18.2) 939 (20.2) 756 (18.7) 2,207 (18.2) 25-35 6,925 (57.2) 2,741 (58.8) 2,286 (56.6) 6,925 (57.2) 35-above 2,984 (24.6) 980 (21.0) 995 (24.7) 1,009 (29.5) Mean age (SD) 30.8 (6.0) Marital status Single/cohabiting 2,355 (16.4) 928 (19.9) 873 (21.6) 554 (16.2) Divorced/widowed 625 (5.2) 206 (4.4) 224 (5.6) 195 (5.7) Married 4,511 (37.2) 1,951 (41.9) 1,595 (39.5) 965 (28.2) Missing 4,626 (38.2) 1,575 (33.8) 1,345 (33.3) 1,706 (49.9) Duration on ART (years) <1 2,813 (23.2) 1,666 (35.8) 822 (20.4) 325 (9.5) 1-3 4,480 (37.0) 1,591 (34.1) 1,579 (39.1) 1,310 (38.3) >3 1,102 (9.1) 290 (6.2) 380 (9.4) 4332 (12.6) Missing 3,722 (30.7) 1,113 (23.9) 1,256 (31.1) 1,353 (39.6) Median duration on ART (years) WHO stage 1 7,394 (61.0) 2,746 (58.9) 2,435 (60.3) 2,213 (64.7) 2 1,898 (15.7) 776 (16.7) 680 (16.8) 680 (442) 3 1,643 (13.6) 673 (14.4) 639 (15.8) 331 (9.7) 4 254 (2.1) 112 (2.4) 88 (2.2) 54 (1.6) Missing 928 (7.7) 353 (7.6) 195 (4.8) 380 (11.1) CD4 count (cells/L) ≤200 774 (6.4) 282 (6.1) 267 (6.6) 225 (6.6) 200-500 3,244 (26.7) 1,228 (26.4) 1,240 (30.7) 776 (22.7) >500 5,933 (49.0) 2,325 (49.9) 1,877 (46.5) 1,731 (50.6) Missing 2,166 (17.9) 825 (17.7) 653 (16.2) 688 (20.1)
In multivariate analysis, replacement feeding (aOR=2.94, 95% CI 0.41 – 0.67), stage 3 (aOR=0.64, 95% CI; 0.52 – 0.79), CI 2.31 – 3.74), nevirapine prophylaxis (aOR=1.55, 95% CI stage 4 (aOR=0.58, 95% CI 0.34 – 0.99). 1.28-1.88) and maternal WHO stage 1 were found to be independently associated with EID before 8 weeks of age. DISCUSSION Infants born to mothers with higher WHO stages of disease We found that the coverage of EID at less than 8 weeks of age had lower odds for HIV testing compared to those born to in Dar es Salaam, Tanzania, increased over the three-year mothers with WHO stage 1 of disease; stage 2 (aOR=0.53, 95% period 2014 to 2016 from 53.2% to 69.2%.
16
Research Article
2020; Issue 4 ISSN 2591 – 6769
Table 3: Crude and adjusted predictors associated with EID testing at less than 8 weeks, Dar es Salaam, Tanzania, 2014-2016 (N= 12,117) Number tested within 8 weeks Variable Total (%) OR [95% CI] aOR [95% CI] Year 2014 4,660 2,481 (53.2) 1 1 2015 4,037 2,314 (57.3) 1.18 [1.08 – 1.28] 0.95 [0.79 – 1.16] 2016 3,420 2,366 (69.2) 1.97 [1.79 – 2.16] 1.04 [0.84 – 1.30] Sex Female 6,345 3,819 (53.3) 1 1 Male 5,772 3,341 (46.7) 0.91 [0.88 – 0.98] 0.96 [0.82 – 1.13] Birth weight of infant (kg) < 2.5 3,224 1,793 (25.9) 1 1 2.5 and above 8,622 5,142 (74.1) 1.18 [1.09 – 1.28] 1.14 [0.96 – 1.34] Feeding practice at testing Exclusive breast feeding 6,149 3,997 (65.4) 1 1 Mixed feeding 1,753 945 (15.4) 0.63 [0.57 – 0.70] 0.87 [0.71 – 1.05] Replacement feeding 1,806 1,173 (19.2) 0.99 [0.89 – 1.11] 2.94 [2.31 – 3.74] Nevirapine prophylaxis No 1,606 1,016 (18.6) 1 1 Yes 7,670 4,436 (81.4) 0.79 [0.71 – 0.89] 1.55 [1.28 – 1.88]
Age of mother (years) Less than 25 2,208 1,265 (17.7) 1 1 25-35 6,925 4,065 (56.8) 1.06 [0.96 – 1.17] 1.12 [0.90 – 1.39] Above 35 2,984 1,830 (25.5) 1.18 [1.06 – 1.32] 1.13 [0.88 – 1.45] Marital status Single/cohabiting 2,355 1,255 (30.0) 1 1 Divorced/widowed 625 336 (8.0) 1.02 [0.85 – 1.22] 0.93 [0.64 – 1.36] Married 4,511 2,597 (62.0) 1.19 [1.08 – 1.31] 0.88 [0.74 – 1.05] Duration of mothers on ART (years) <1 2,813 1,759 (34.4) 1 1 1-3 4,480 2,685 (52.5) 0.90 [0.81 – 0.98] 0.89 [0.75 – 1.07] >3 1,102 667 (13.1) 0.92 [0.796 – 1.05] 0.87 [0.67 – 1.13] WHO stage 1 7,394 4,877 (75.1) 1 1 2 1,898 599 (9.2) 0.24 [0.21 – 0.26] 0.53 [0.41 – 0.67] 3 1,643 892 (13.7) 0.61 [0.55 – 0.68] 0.64 [0.52 – 0.79] 4 254 131 (2.0) 0.55 [0.43 – 0.71] 0.58 [0.34 – 0.99] CD4 count (cells/µL) ≤200 774 486 (8.1) 1 1 200-500 3,244 1,986 (33.3) 0.94 [0.79 – 1.10] 0.93 [0.67 – 1.29] >500 5,933 3,49617 (58.6) 0.85 [0.72 – 0.99] 0.97 [0.70 – 1.33]
Research Article
2020; Issue 4 ISSN 2591 – 6769
The uptake of EID was associated with infant feeding practice maternal age was independently associated with EID. We also at time of testing, infant nevirapine prophylaxis and lower report that older mothers were more likely to test on time; maternal WHO stage. The study found an overall HIV however, this was not statistically significant. It is possible prevalence of 2.3% in HEI. This study highlights the positive that younger women are more likely to drop out of care due steps that the PMTCT programme in Dar es Salaam is making to psychosocial and mental challenges (Ramaiya et al., 2016; to reach national and international targets to eliminate MTCT Izudi et al., 2017). Findings from studies in South Africa, of HIV. Uganda and Kenya found that maternal knowledge about EID was independently associated with timely EID testing Our findings of increasing EID at 8 weeks of age are (Woldesenbet et al., 2015; Goggin et al., 2016; Izudi et al., consistent with studies from South Africa in 2014 and 2017). Moreover, stigma, discrimination and lack of Thailand in 2016 that showed an increasing trend (Sherman education were associated with low EID uptake (Woldesenbet et al., 2014; Naiwatanakul et al., 2016). The findings are also et al., 2015; Goggin et al., 2016). consistent with findings from a study in Ethiopia which found 66.7% of the exposed infants had EID at 8 weeks of age In our study, a high number of participants had missing (Olana et al., 2016). Chiduo and colleagues from Tanzania in information for the date of birth and HIV status of the child. 2013 also found an increasing trend of EID from 77.2% in These missing data could result from a recording issue at the 2009 to 97.8% in 2011. However, this study captured infant health facility, such as not recording measurements and test testing up to 18 months of age (Chiduo et al., 2013), and results or not updating registers in time for collation at the further highlighted the missed opportunity for testing and national level. The missing data could also be due to errors in subsequently treating HIV-positive infants below 18 months administering the HIV test or to clients moving between of age. While the increasing trend in the uptake of EID is facilities pre- or post-testing. Prior to 2014, individual-level clearly positive for the three-year period, efforts to reach the electronic EID data in routine care was virtually non-existent; national target of 90% coverage for those under 8 weeks of the shift to individual-level electronic data collection may age by 2018 were not realised (TACAIDS, 2013). have impacted data quality, including providers’ omissions or errors in data entry. Without accurate data, it is not possible Infants of mothers who were on replacement feeding and to evaluate the PMTCT programme success, nor effectively nevirapine prophylaxis were more likely to undergo EID, and target activities to increase EID uptake and reduce they may represent mothers who were better engaged in care. transmission rates. In Tanzania, EID services (as part of PMTCT) were integrated into reproductive and child health clinics in 2013 (National The use of routine data enabled the study to obtain a large PMTCT Guideline, 2013). Further research is needed to number of study participants from all facilities managed by understand how this integration is taking place in practice and MDH in the study area. The large sample increased the to highlight the potential gaps. In this setting, post-natal generalisability of the results and the ensured the study was reproductive and child health clinics are already time- adequately powered to assess the uptake and predictors of consuming and often require mother and baby pairs to wait testing with precision. However, the study also had several long hours, visit multiple providers and receive various limitations; in addition to challenges of using routine data that services. The integration of EID into this service makes sense have already been mentioned, the study included only women logistically but may overburden the mothers (Ramaiya et al., who registered their baby at PMTCT and where the baby was 2016). It is likely that additional efforts are needed to reach tested. Women who were HIV-infected and did not bring their the “harder to reach” mothers and infants. The predictors of children to be tested were not included in the analysis, uptake of EID in this study included replacement feeding, the potentially resulting in an overestimation of the uptake of HEI taking nevirapine and the mother having a lower WHO testing. All facilities used by the study were operated by stage at enrolment. The first two factors may be indicative of MDH and were not randomly sampled. This may have biased mothers who know their child is at risk, have taken steps to selection, and hence the estimates obtained might not be true reduce this risk, and are already engaged in care. Women who representations for the region. Moreover, because the study have not yet reached this stage may need additional support used existing data, we were not able to measure associations to enable them to also increase their uptake of these services. with variables identified as significant by other studies. Finally, a large proportion of tested infants had missing test Our study was not able to identify some predictors results, which may have resulted in an incorrect estimation of highlighted in other studies (for example travel distances, out- uptake. of-pocket costs at the health facility and long waiting times), and we report some differences in findings. This study used CONCLUSION secondary data routinely collected in health facilities and the We found that there was an increasing trend over three years analysis was therefore restricted to the variables collected as in the uptake of EID test at less than 8 weeks of age. However, part of the routine data. A study by Izudi found that increased the proportion of infants testing fell below the national target, 18 Research Article East African Journal of Applied Health Monitoring and Evaluation
7 and more efforts are needed to ensure a 90% testing coverage Goggin K, Wexler C, Nazir N, Staggs S V, Gautney B, at 8 weeks of age. The sub-optimal uptake may undermine Okoth V. 2016. Predictors of Infant Age at Enrollment in efforts to reduce HIV/AIDS-related morbidities and mortality Early Infant Diagnosis Services in Kenya. AIDS Behaviour among children, as the majority of children are diagnosed at (2016) 20:2141–2150. a late age (missed opportunity for early diagnosis) and hence Global Report, 2013. UNAIDS report on the Global AIDS delayed initiation of treatment. Infants who were initiated on epidemic nevirapine were more likey to test for EID, and this calls for Izudi J, Auma S, Alege J B, G., 2017. Early Diagnosis of advocating caretakers to take their infants to post-natal HIV among Infants Born to HIV-Positive Mothers on Option- services. B Plus in Kampala, Uganda, Hindawi AIDS Research and Treatment Volume 2017, COMPETING INTERESTS https://doi.org/10.1155/2017/4654763, pp.3–7. The authors declare that they have no competing interests. Kpeltzerhsrcacza, K.P. & Mlambo, G., 2010. Factors determining HIV viral testing of infants in the context of AUTHORS’ CONTRIBUTIONS mother-to-child transmission. 6(6), pp.590–596. MM designed the study, participated in statistical data Kristine F, Finocchario-KesslerBrad S, Okotth G V, analysis and wrote the manuscript. PN and BJL participated Goggin K (2016). Progress Toward Eliminating Mother to in statistical data analysis; JT participated in designing the Child Transmission of HIV in Kenya: Review of Treatment study, and advised on the analysis; GL and LM participated Guideline Uptake and Pediatric Transmission at Four in conception of the study; SM and MJM participated in Government Hospitals Between 2010 and 2012. AIDS and conception and study design; JR participated in study design, Behavior, pp.2602–2611. reviewed the manuscript for intellectual content, and The United Republic of Tanzania, 2014. Global AIDS statistical analysis. All authors reviewed the manuscript and Response Country Progress Report read and approved the final manuscript. Hassan A S, Sakwa E M, Nabwera H M, Taegtmeyer M M, Kimutai R M, Sander E J, Ken K. A, Mutinda M N, Molyneux, ACKNOWLEDGEMENTS C.S, Berkley J.A., (2012). Dynamics and Constraints of Early The first author would sincerely like to thank the Department Infant Diagnosis of HIV Infection in Rural Kenya., pp.5–12. of Epidemiology & Biostatistics, Institute of Public Health, Manyahi J, Jullu B,Abya M, Juma J, Kilama B, Sambu V Kilimanjaro Christian Medical College, Moshi, Tanzania for (2017). Decline in the prevalence HIV among pregnant facilitating the study and for the encouragement, with special women attending antenatal clinics in Tanzania, 2001-2011. appreciations to MM, SM, BJL, JT and JR who are supported Tanzania Journal of Health Research Volume 19, Number 2, by DELTA/THRIVE under DEL 15-011/07742/Z/15/Z. The pp 3-6. authors wish to also thank the team from the PMTCT unit (Dr. Naik N M, Bacha J, Gesase A E, Barton T, Schutze G E, Deborah Kajoka, Dr. Michael Msangi, Dr. Mukome Wanless R S, Minde M M, Mwita L F, Tolle M A (2016). Nyamhagata, Dr. Amir Juya, Levina Lema, Elizabeth Sallu Antiretroviral Therapy in Children Less Than 24 Months of and Thomas Sanga) of MoHCDGEC and the team from Age at Pediatric HIV Centers in Tanzania: 12-Month Clinical regional and municipal councils for dedicating their time and Outcomes and Survival. supporting the study. Last but not least, the first author would Naiwatanakul T, Voramongko N, Punsuwan N, Lolekha R, like to thank the team of Management and Development for Gass R, Thaisri H, Leechanachai P, Wolfe M, Boonsuk Health (MDH) for supporting the dataset and reviewing of the S,Bhakeecheep S (2016). Uptake of early infant diagnosis in results. Thailand’ s national program for preventing mother-to-child HIV transmission and linkage to care.Journal of the REFERENCES International AIDS Society, 19:20511 , pp.1–9. Aledort J.E, Allan, Ronald, Sylvie M, Le BlancqRenee R., National PMTCT Guideline, 2013. National Guidelines for et al. (2006). Reducing the burden of HIV / AIDS in infants: Comprehensive Care Services of Prevention of Mother to the contribution of improved diagnostics. Child Transmission of HIV and Keeping Mothers Alive.The Ciaranello L A, Ji-Eun P, Ramirez-Avila L, Freedberg K A, United Republic of Tanzania. Rochelle P W and Valeriane (2011). Early infant HIV-1 National Bureau of Statistics. (2013). 2012 Population and diagnosis programs in resource-limited settings: housing census; population distribution by adminstrative opportunities for improved outcomes and more cost-effective areas. National Bureau of Statistics. interventions. BMC medicine, 9(1), p.59. Newell ML & Cortina-Borja M. (2004). Mortality of Cook R.E, Ciampa P.J, Sidat M, Meridith B, Burlison J, infected and uninfected infants born to HIV-infected mothers Mario A.D. et al. (2012). Predictors of successful early infant in Africa: a pooled analysis. Lancet, vol 364, pp. 1236-43. diagnosis of HIV in a rural district hospital in Zambézia, Nuwagaba-Biribonwoha, H, Werq-Semo B,Abdallah Mozambique, J Acquir Immune Defic Syndr56(4), pp.1–14. A,Cunningham A, Gamaliel C J, Coovadia H M, Nigel C R,Bland R, Kirsty L, Coutsoudis Mtunga S,Nankabirwa V, Malisa I, Gonzalez L F, A, Michael L B, Marie-Louise N Massambu C, Nash D, (2007). Mother-to-child transmission of HIV-1 infection Justman J, Abrams E J (2010). Introducing a multi-site during exclusive breastfeeding in the first 6 months of life: an program for early diagnosis of HIV infection among HIV- intervention cohort study. Lancet (London, England), exposed infants in Tanzania. BMC pediatrics, 10, p.44. 369(9567), pp.1107–16. PMTCT & TWG. 2015. PMTCT annual program report – 19 Research Article East African Journal of Applied Health Monitoring and Evaluation
8 TANZANIA mainland.Unpublished. Olana T, Bacha T, Walelign W and Tadesse B, 2016. Early Ramaiya, M.K, Kristen A, Sullivan, O’ Donnel K, infant diagnosis of HIV infection using DNA‑PCR at a Cunningham C K, Shayo A M, Mmbaga B T, Dow D E (2016). referral center: an 8 years’ retrospective analysis. AIDS Res A Qualitative Exploration of the Mental Health and Ther (2016) 13:29. Psychosocial Contexts of HIV-Positive Adolescents in UNAIDS, 2015. Progress Report on the Global Plan. Tanzania., pp.1–13. UNAIDS, 2016. Global AIDS Update. Sherman, G.G, MB BCh, Lilian R R,Bhardwa S,Candy S, UNICEF, 2016. For Every Child End AIDS Seventh Barron P, (2014). Prevention OF Mother-To-Child Stocktaking Report Transmission, Laboratory information system data UNICEF, 2015. Children AIDS 2015 statistical update. demonstrate successful implementation of the prevention of Woldesenbet S, Jackson D, Goga A, Crowley S, Crowley T, mother- to-child transmission programme in South Africa., Mogashoa M et al. 2015. Missed Opportunities for Early 104(3), pp.1–3. Infant HIV Diagnosis: Resultsof A National Study in South The United Republic of Tanzania, 2013. Global Aids Africa. Acquir Immune Defic Syndr 2015;68: e26–e32. Response Country Progress Report WHO, 2010. PMTCT strategic vision 2010–2015: TACAIDS, 2013. Tanzania Third National Multi-Sectoral preventing mother-to-child transmission of HIV to reach the Strategic Framework for HIV and AIDS 2013/2014- UNGASS and Millennium Development Goals. World Health 2017/2018. Prime Minister’s Office Tanzania. Organization, pp.1–40. Tanzania Elimination of Mother to Child Transmission of WHO, 2010. Rapid advice: Use of antiretroviral drugs for HIV Plan, 2012-2015 (2013). treating pregnant women and preventing HIV infection in Townsend, C.L, Cortina-Borjaa M, Peckhama C S, de infants. Geneva: WHO, (April), pp.1–117. Ruiterb A, Lyallc, Tookeya P A, (2008). Low rates of mother- Zvanyadza, G.F, 2008. Pediatric HIV Testing Challenges in to-child transmission of HIV following effective pregnancy Resource Limited Settings. interventions in the United Kingdom and Ireland, 2000-2006. AIDS (London, England), 22(8), pp.973–81.
20
Research Article
2020; Issue 4 ISSN 2591 – 6769
Figure 1: Flow chart of HIV testing results for HIV-exposed infants enrolled in Dar es Salaam from January 2014 to December 2016.
21
Research Article
2020; Issue 4 ISSN 2591 – 6769
Ocular symptoms and associated factors among hairdressers in Kinondoni Municipality, Dar es Salaam, Tanzania
Olirk Baldwina Tita1,2; Ezra Jonathan Mrema1; Simon Mamuya1 1. Department of Environmental and Occupational Health, School of Public Health and Social Sciences, Muhimbili University of Health and Allied Sciences, Dar es Salaam - Tanzania.
2. Department of Environment, Muhimbili National Hospital, Dar es Salaam - Tanzania. Correspondence to: Olirk B, Department of Environmental and Occupational, School of Public Health and Social Sciences, Muhimbili University of Health and Allied Sciences, P.O. Box 65015, Dar es Salaam, Tanzania. Mob: +255766477824; e-mail: [email protected]
INTRODUCTION Hairdressers are exposed to chemicals from hair products such as hair dyes, glues, sprays, and relaxants, which increase risks for ocular diseases. There are limited studies assessing ocular diseases among hairdressers in Tanzania. The aim of this study was to estimate the prevalence of ocular symptoms and associated factors among hairdressers in Kinondoni Municipality, Dar es Salaam, Tanzania.
METHODS This study employed an analytical cross-sectional study design. Data were collected from 300 hairdressers from 157 salons selected from ten wards of Kawe and Magomeni divisions of Kinondoni Municipality. Data on participants’ sociodemographic characteristics, years of work experience, occupational activities, use of protective equipment, hazard awareness, training on chemicals use, types of hair products used, and self-reported ocular symptoms were gathered by using a structured questionnaire. Descriptive statistics, partial correlation and logistic regression were performed to understand the prevalence of ocular symptoms and association with other study variables.
RESULTS The study enrolled a total of 300 participants who were primarily female (92%). The mean age (±SD) of the participants was 28 (±7) years. The prevalence of ocular symptom was 78.7%. The most prevalent symptoms reported were headache (68.7%) and itching (60.0%). There was a statistically significant association between years of work experience (r = 0.314, p < 0.0001), lack of training on chemical use [AOR = 0.0203 (0.01 - 0.054)], lack of hazard awareness [AOR = 0.05 (0.02 - 0.094)] and ocular symptoms.
CONCLUSION Prevalence of ocular symptoms was high among the study participants. The variables associated with ocular symptoms were lack of training on chemical use, inadequate hazard awareness and years of work experience. The Occupational Health and Safety Authority (OSHA) of Tanzania should facilitate and conduct health promotion to hairdressers on protective measures and frequent breaks to reduce chemical exposure.
Keywords: Ocular symptoms; Ocular diseases; Hairdressers; Salons; Chemical
INTRODUCTION lead to health problems such as ocular diseases, occupational The hairdressing occupation is associated with exposure to rhinitis, cancer, asthma, fertility problems and respiratory different kinds of health hazards such as chemical, biological, diseases (Ferreira, 2013; Johansson et al., 2014; Kim et al., physical, psychological and ergonomic hazards. Chemical 2016; Kumah et al., 2017). exposure includes hair products such as shampoos, There have been limited studies conducted in Africa permanent wave solutions, dyes, hair sprays, bleach, hair assessing ocular diseases among hairdressers. A study in treatment and relaxants (Kim et al., 2016; Tsigonia, Lagoudi, Ghana reported high prevalence (66.8%) of ocular symptoms Chandrinou & Linos, 2010). Exposure to these hazards may among hairdressers; tearing was the most prevalent symptom
22 Research Article East African Journal of Applied Health Monitoring and Evaluation
2 (89.1%), followed by eye redness (88.1%) and painful eyes hairdressing salons with their contact information was (81.2%). The high prevalence of ocular symptoms was due obtained from the department of business in Kinondoni to exposure of hairdressers to various chemicals that are Municipality. Ten wards with a large number of hairdressing potentially harmful to the eye and the visual system (Kumah salons were selected: six wards from Magomeni Division et al., 2017). In Tanzania, a study assessing the occupational and four wards from Kawe Division. Eighty-seven (87) and health hazards of cosmetologists (those who are doing 70 hairdressing salons were randomly selected from treatment of skin, hairdressing and nail treatment), showed Magomeni and Kawe divisions, respectively. To avoid that workers suffered a number of health problems including recruiting hairdressers with a brief period of chemical musculoskeletal disorders, physical injuries, dermatitis, exposure, the study was restricted to 319 participants who fungal infection, respiratory problems and infectious had worked as hairdressers for at least two years. diseases (Bigambo, 2016). However, the study did not assess Participation was voluntary. Hairdressers were informed ocular diseases and symptoms. Another study in Tanzania about the purpose of the study and their consent was assessed compliance to occupational and public health requested. In order to have a representative sample, two requirements among barber shops and hair salons (Nshunju, hairdressers were interviewed at each salon; all were 2012). The study showed that workers who were between 18 and 47 years old. knowledgeable about Occupational Health and Safety and Public Health requirements had higher compliance than Data collection tools and procedures those who had no such knowledge. A face-to-face interview with hairdressers was conducted The Occupational Health and Safety (OHS) Act 2003 is using a structured questionnaire that was pilot-tested in four the main legislation governing occupational health and hairdressing salons in Magomeni and Kawe Divisions. The safety practices in Tanzania, but it omits the self-employed pilot testing was done to check the validity of the and informal sectors like hairdressing (Mrema et al, 2015). questionnaire. The questionnaire gathered information on Much of the legislation addresses safety issues such as lifting socio-demographic characteristics, work experience, equipment and steam boilers. The legislation does not occupational activities, use of personal protective equipment incorporate all working populations, as it provides OHS (PPE), hazard awareness, training on the use of chemicals, services to enterprises that can afford services. This leaves types of hair products used, reported ocular symptoms and out informal sectors because of their low income (Mrema, et health-seeking behaviour. The questionnaire was written in al., 2015). English and translated into Swahili. A checklist was used to Given the lack of data from Tanzania, we conducted a record the working conditions in each hairdressing salon. study to assess work-related ocular symptoms and associated These conditions include presence and type of ventilation (as factors among hairdressers in Kinondoni Municipal, Dar es per Tanzania Public Health Act 2009, which considers Salaam. The findings from this study will enable policy openings such as windows and air conditioners to be makers to understand the extent of the problem and adequate for ventilation in salons) and the presence of PPE encourage them to review the existing policies in order to (the recommended equipment by Tanzania Occupational implement appropriate interventions to improve health and Health and Safety Act 2003, such as eye goggles for eye promote safety among this occupation. protection, gloves for protection of the hands and masks for protecting the nose). Prior to data collection, research METHOD assistants were trained and familiarized with the study Study design and setting objectives and the data collection tools. At each hairdressing This analytical cross-sectional study was conducted to salon, we conducted observation and interviewed a estimate prevalence of ocular symptoms and related factors maximum of two hairdressers. Data collection was done on among hairdressers. The study was conducted in Kinondoni weekdays from morning to afternoon hours. Municipality in Dar es Salaam region, Tanzania, which has a total area of 522.3 km2 and a population of 1,775,049 Data Analysis (Tanzania Demographic Census 2012). The municipality has Variables and measures two divisions, Magomeni and Kawe, with eleven and nine The dependent variable, ocular symptoms (i.e. blurry vision, wards, respectively, from which the study participants were headache, burning sensation, painful eye, discharge, tearing), recruited. The Kinondoni Municipality was selected for this were measured as categorical variables where 1 is the study because it has a large number of hairdressing salons presence of any ocular symptom and 2 was the absence of (304) compared to other districts. symptoms. The prevalence was calculated by the formula developed by Bhopal, (2002). Prior to statistical analysis, the Study subjects and sampling methodology internal consistency of the variables used to obtain total The study was conducted from April to June 2018. The scores on ocular symptoms was investigated by using method used for sample size estimation was probability Cronbach’s alpha test. Results were 0.8 value for Cronbach’s sample size estimation developed by Fox, & Hunn, 2007. alpha, which is above the threshold of 0.7, suggesting one The sampling procedure was a multistage sampling from can proceed with the analysis since the variables are municipality to division to ward then to salon and finally to internally consistent (Pallant, 2003). For each participant, hairdresser. To account for the design effect, the sample size each ocular symptom reported was tallied to generate a total was multiplied by 1.5. The sampling frame of all 304 score, which was later used to provide a benchmark for 23 Research Article East African Journal of Applied Health Monitoring and Evaluation
3 establishing a relationship between years of work experience Table 1: Socio-demographic characteristics of hairdressers in and the symptoms in the correlation analysis. Each form of Kinondoni Municipality (N = 300) ocular symptom was scored as 1. Independent variables included hazard awareness, training Characteristics Frequency (n) Percentage (%) on the use of chemicals, the use of PPE, presence of adequate Age (years) ventilation, and length of working experience in years. ≤30 194 64.7 Potential confounders such as age, education, smoking, previous ocular disease, and conditions such as diabetes and >30 106 35.3 hypertension were controlled for in the multiple logistic Gender regression model. Male 24 8
Statistical analyses Female 276 92 The data was analyzed by using IBM Statistical Package for Level of education Social Sciences (SPSS) for Windows version 20 (IBM Corp., No formal education 13 4.3 Armonk, New York, USA), applying both descriptive and Primary education 126 42 inferential statistical approaches. Inferential statistical approaches applied were chi-square test, logistic regression Secondary education 153 51 and correlation test with p-values less than 0.05 considered Higher education 8 2.7 statistically significant. Chi-square test was used to show the Smoking status relationship between categorical variables (ocular symptoms) and independent variables (hazard awareness, training on the Smoker 1 0.3 use of chemicals and the use of protective gears such as Non smoker 298 99.3 goggles). Logistic regression was used to determine factors Previous Smoker 1 0.3 which strongly affected the dependent variable while correcting for confounders (age, education, smoking, history Diabetic Status of ocular diseases, diabetes and blood pressure). Correlation Diabetic 24 8 test was used to determine the relationship between Non diabetic 276 92 numerical variables (total scores of ocular symptoms and Blood pressure status 17.5 length of work experience). Partial correlation was also used to correct for confounders. High blood pressure 24 8 Normal blood pressure 276 92 Ethical Considerations Ethical approval was obtained from Muhimbili University of Very few participants (8.3%) had training on the use of Health and Allied Sciences (MUHAS) research and chemicals. Work experience for hairdressers differed among publication committee. The permission to conduct the study respondents; the minimum reported was two years and the at the Kinondoni Municipality was obtained from the District maximum was 28 years. The mean number of working hours Medical Officer (DMO) of Kinondoni Municipality. for hairdressers was 81 ± 4 hours per week. Air-conditioning Participants were given written informed consent forms was rarely used, and circulatory fans were used occasionally. before consenting to participate. The study participants who Assessment of the availability and usage of protective were found to have ocular symptoms were advised to seek devices showed that 60 respondents (20.0%) used gloves, 15 health care from the nearest health care facility. (5.0%) used goggles and 60 (20.0%) used nose devices (Table 2). RESULTS Background characteristics of study participants Ocular symptoms A total of 319 hairdressers were contacted, and 300 The prevalence of ocular symptoms was 78.7%. About 64.3% hairdressers from 157 salons participated in the study, (193 hairdressers) reported to have multiple symptoms. resulting in a participation rate of 94%. The mean age was Headache was the most prevalent symptom reported (68.7%), 28 ± 7 years, and most participants had received either followed by itching (60.0%). About 40% of participants primary (42.0%) or secondary (51.0%) education. Most complained about tearing and redness of eyes (Table 3). study participants were non-smokers (99.3%), non-diabetic (92.0%), and 92.0% had normal blood pressure (Table 1). Relationships between ocular symptoms and independent variables Distribution of risk factors This study shows statistical evidence of an association A total of 220 respondents (73.3%) had received on-the-job between prevalence of ocular symptoms and lack of hazard training, but only 58 participants (19.3%) had their training awareness; crude OR = 0.031 [0.039 - 0.172], adjusted OR = in recognized vocational institutes. About 63% of the 0.05 [0.02 - 0.094]. The study revealed a statistically respondents were aware of hazards that they could encounter significant association between ocular symptoms and lack of at their work place. training on chemical use; crude OR = 0.028 [0.014 - 0.059],
24 Research Article East African Journal of Applied Health Monitoring and Evaluation
4 adjusted OR = 0.02 [0.01 - 0.054]. The study also showed a Table 3: Distribution of ocular symptoms reported by statistical association between ocular symptoms and non-use hairdressers in Kinondoni Municipality (N = 300) of eye protection; crude OR 0.117 [0.038 - 0.386] but not for the adjusted OR. A partial correlation test yielded a fair Ocular symptoms Frequency (n) Percentages (%) positive correlation (r = 0.314, p < 0.0001) between the Multiple symptoms 193 64.3 ocular symptoms and years of work experience. However, Headache 206 68.7 there was no evidence of a statistically significant difference in prevalence of ocular symptoms among hairdressers who Itching 180 60 received vocational skills and those who did not (Table 4). Redness 121 40.3 Tearing 116 38.7 Table 2: Distribution of risk factors among hairdressers in Painful eye 105 35 Kinondoni Municipality (N = 300) Foreign body sensation 90 30 Risk factor Frequency (n) Percentage (%) Discharge 69 23 Job training Blurry vision 51 17 Had job training 220 73.3 Burning sensation 27 9 Had no job training 80 26.7 Swollen eyelids 26 8.7 Place of training Vocational institutes 58 19.3 Non-vocational institutes 162 54 Table 4: Determinants of ocular symptoms Hazard awareness Crude OR Adjusted OR Variable Aware of hazards 189 63 OR (95%CI) p-value OR (95%CI) p-value Unaware of hazards 111 37 Hazard 0.031(0.039 - 0.05(0.02 - 0.0001 0.0001 Chemical use training awareness 0.172) 0.094) Trained on chemical use 25 8.3 Training on 0.028(0.014 - 0.02(0.01 - 0.0001 0.0001 Not trained on chemical use 275 91.7 chemicals 0.059) 0.054) Length of work experience Use of eye 0.117(0.038 - 0.0001 1 0.969 1-5 years 167 55.7 wear 0.386) 6-10 years 90 30 Vocational 1 0.9007 - - 11-15 years 21 7 skills 16-20 years 18 6 Length of r = 0.455 0.0001 r = 0.314 0.0001 > 20 years 4 1.3 work Use of protective devises Used any protective devices 135 45 DISCUSSION This study investigating ocular symptoms among Did not use any protective 165 55 hairdressers of Kinondoni Municipality found that the devices overall prevalence of ocular symptoms was 78.7%, with the Types of protective devices used most prevalent symptoms being headache (68.7%) and itching (60.0%). Symptoms were associated with lack of Gloves 60 20 hazard awareness, lack of training on chemical use, and Eye goggles 15 5 length of service. Nose devises 60 20 The prevalence of ocular symptoms in our study was lower compared to those reported by Kumah, et al. (2016, 2017) Type of ventilation (157 salons) and Omoti1(2008) from other occupations such as farming, Circulatory fans 98 salons 62.4 mechanics and butchers. This difference may be due to Air conditioning 3 salons 1.9 differences in chemical use. However, the prevalence was higher compared to a similar study done by Kumah et al. (2017). A plausible explanation for higher prevalence in our study could be due to differences in age, sample size and in exposure time. Results from direct questioning revealed that hairdressers were working for an average of 81 ± 4 hours per week compared to 54.6 hours in the Kumah et al. (2017) study. The maximum length of working experience among
25 Research Article East African Journal of Applied Health Monitoring and Evaluation
5 respondents was five years (10.4%) in Kumah et al. (2017) Development) via NORHED Project Tan:1300646-12 (North compared to our study (28 years), with 44% working more - South - South collaboration). We would also like to thank than six years. This shows that hairdressers in our study had hairdressers of Kinondoni Municipality who spared their longer exposure time to chemicals in hair products and heat precious time to participate in our study. from dryers and irons, which could account for the difference in ocular symptoms among them. The longer the exposure CONFLICT OF INTERESTS period the higher the exposure doses and hence health effects The authors declare that there is no conflict of interests (Common Wealth, 2002). regarding the publication of this paper. Also, in the study of Kumah et al. (2017), 64 participants were above thirty years old compared to 106 participants in AUTHORS' CONTRIBUTIONS our study. It is believed that vision capacity decreases as one OBT coordinated the study, designed the questionnaire, is aging (Varma, Vajaranant, Wu, & Torres, 2016). The conducted field investigations and performed the statistical higher prevalence of ocular symptoms in our study is also analysis and writing of the paper. She is the main author of likely to be due to differences in environmental conditions of the manuscript. EJM contributed to the formulation of the the selected salons. Most of the salons in our study had poor study, participated in data collection and writing of the working/environmental conditions compared to those manuscript. MS contributed to the inception of the study, reported by Kumah et al. (2017), in which 93.5% of the data analysis and writing of the paper. All authors read and hairdressing salons were tidy and well-ventilated. The higher approved the final manuscript prevalence may also be due to lack of training on the use of chemicals, lack of hazard awareness, non- use of eye REFERENCES protection, and longer working experience. Eye protection is Commonwealth of Australia 2002. Environmental Risk important as it acts as a barrier between the eyes and Assessment: Guidelines for assessing human health risks contaminated air (Peate, 2007). Research suggests that eye from environmental hazards. Australia: Health Council. protective devices, training of workers on handling Retrieved from chemicals, medical regular checkups and improvement of http://www.health.gov.au/pubhlth/strateg/envhlth/risk/pdf. working environment might reduce chemical exposure (Abu Abu, E. K., Bert, S., Kusi, B., Opuni, P. Q., Kyei, S., Ansah, et al., 2016; Kumah et al., 2017; Robinson et al, 2011). A. O., et al. (2016). Ocular Health and Safety Assessment Our study had several limitations. All variables were self- among Mechanics of the Cape Coast Metropolis, Ghana. reported by participants, meaning that any variable could Journal of Ophthalmic and Vision Research, 11(1), 78-83. have been over-reported or under-reported. In addition, the https://doi.org/10.4103/2008-322X.158890. hairdressers may have felt the need to underreport symptoms Bhopal, R. 2002. Concepts of Epidemiology. New York: or deny having ocular problems because they liked their Oxford University Press Inc. work. Finally, the data were collected in only two divisions Bigambo, F. M. (2016). Occupational Health Risks of one municipality in Tanzania, and so the results may not Among Cosmetologist: A Case of Kinondoni Municipality be generalizable beyond this area or outside of Tanzania. Dar Es Salaam, Tanzania. Journal of Multidisciplinary These potential sources of bias could have affected our Engineering Science Studies, 2(8), 766–773. results either by underestimating or overestimating the true Ferreira, A. P. (2013). An assessment of occupational symptoms that hairdressers might have. health risks in female hairdressers forefront to xenobiotics. Environmental Occupational Science, 55(21), 190–198. CONCLUSION Fox, N., Hunn, A., & Mathers, N. (2007). Sampling and The study revealed a high prevalence of ocular symptoms Sample Size Calculation Authors. National institute for among hairdressers in one municipality of Tanzania. Lack of Health Reasrch for the East Midlands Retrieved from http:// hazard awareness, lack of training on the use of chemicals, www.rds-eastmidlands.nihr.ac.uk/pdf. and length of service were associated with ocular symptoms Johansson, G. M., Jönsson, A. G., Axmon, A., Lindh, C. in this population. Hairdressers should work in adequately H., Lind, M., Gustavsson, M., et al. (2014). Exposure of ventilated salons, use protective gear and have periodic eye hairdressers to ortho- and meta-toluidine in hair dyes. and health check-ups to screen for ocular symptoms and Occupuation Environimental Medicine, 72(1), 57–63. other health effects. Future studies should build on these https://doi.org/10.1136/oemed-2013-101960. findings by assessing work factors associated with ocular Kim, D., Yeol, M., Choi, S., Park, J., Hye-Ji, L., & Kim, symptoms and by measuring ocular diseases and their E. A. (2016). Reproductive disorders among cosmetologists severity. and hairdressers. International Archives of Occupational and Environmental Health, 89(5), 739–753. ACKNOWLEDGEMENT https://doi.org/10.1007/s00420-016-1112-z. We acknowledge Dr. Vera Ngowi, Mr. Alan Lwanga and Dr. Kumah, B. D., Abdul-kabir, M. Aidoo, F., Kuutiero, I. W., Hussein Mwanga for their support towards accomplishment Ablordeppey, R. K., Sarpong, S., et al. (2017). Prevalence of of this study. The authors are grateful for the financial Ocular Conditions among Hairdressers in the Kumasi support provided by NORAD through the NORHED- Metropolis, Ghana. BAOJ Opthalmology, 1(3), 1–6. programme (Norwegian Programme for Capacity Kumah, D. B., Mohammed, A., Bisiw, J. D., Kabir, M. A., Development in Higher Education and Research for & Osae, E. A. (2016). Ocular conditions among singed – hide 26 Research Article East African Journal of Applied Health Monitoring and Evaluation
6 butchers at The Kumasi abattoir. Epidemiological Research, -ed_protect/---protrav/-- 2(1), 1–24. https://doi.org/10.5430/jer.v2n1p20. ilo_aids/documents/legaldocument/wcms_151286.pdf Mrema, E. J., Ngowi, A. V, & Mamuya, S. H. (2015). Robinson, E. N., & Arcury, T. A. (2011). Eye Health and Status of Occupational Health and Safety and Related Safety Among Latino Farmworkers. Journal of Challenges in Expanding Economy of Tanzania. Annals of Agromedicine, 16(2), 143–152. Global Health, 81(4), 538–547. https://doi.org/10.1080/1059924X.2011.554772.Eye. https://doi.org/10.1016/j.aogh.2015.08.021. Tsigonia, A., Lagoudi, A., Chandrinou, S., & Linos, A. Nshunju, R. K. (2012). Compliance to Occupational and (2010). Indoor Air in Beauty Salons and Occupational Health Public Health Requirements and Associated Factors in Exposure of Cosmetologists to Chemical Substances. Barbershops and Hair Dressing Salons, A case of Kinondoni International Journal of Environmental Research and Public Municipality , Dar es Salaam, Tanzania. [Master thesis] Health, 7, 314–324. https://doi.org/10.3390/ijerph7010314. Muhimbili University of Health and Allied Sciences. Varma, R., Vajaranant, T., Wu, S., & Torres, M. (2016). Retrieved from ihi.eprints.org. Visual Impairment and Blindness in Adults in the United Occupational Health and Safety Act; United Republic of States: Demographic and Geographic Variations From 2015 Tanzania. 2003. Retrieved from to 2050. American Medical Association Opthamology, https://www.lrct.go.tz/download/laws_2003/05 134(7):802-809. doi:10.1001/jamaophthalmol.2016.1284. 2003_The%20Occupational%20Health%20and%20Safety% Volquind, D., Bagatini, A., Massaro, G., Monteiro, C., 20Act,%202003.pdf Londero, J. R., & Benvenutti, G. D. (2013). Occupational Omoti1 AE, Waziri-Erameh, J. M., & Enock, M. (2008). Hazards and Diseases Related to the Practice of Ocular disorders in a petroleum industry in Nigeria. Eye Anesthesiology, Brazillian Journal of Anasthesiology, 63(2), (London), 22(7), 925–929. 227–232. https://doi.org/10.1038/sj.eye.6702772. Pallant, J. (2003). Spss survival manual (2nd edn). Australia. Retrieved from www.allenandunwin.com/spss.htm/pdf. Peate, W. F. (2007). Work-Related Eye Injuries and Illnesses. American Family Physician, 75(7), 1017–1020. Public Health Act; United Republic of Tanzania. 2009. Retrieved fromhttps://www.ilo.org/wcmsp5/groups/public/--
27
Research Article
2020; Issue 4 ISSN 2591 – 6769
KNOWLEDGE, ATTITUDES, PERCEPTIONS AND ACCEPTABILITY OF ONCHOCERCIASIS CONTROL THROUGH COMMUNITY-DIRECTED TREATMENT WITH IVERMECTIN: IMPLICATIONS FOR PERSISTENT TRANSMISSION IN ULANGA DISTRICT, TANZANIA VIVIAN MUSHI1, DEODATUS KAKOKO2, DONATH TARIMO1 1. Department of Parasitology and Medical Entomology, Muhimbili University of Health and Allied Science, Tanzania. 2. Department of Behavioral Science, Muhimbili University of Health and Allied Science, Tanzania.
Correspondence to: Vivian Mushi, +255789458561/+255653942973; [email protected]; P.O.Box 35869, Dar es salaam, Tanzania
Background Ulanga, an onchocerciasis-endemic district in Tanzania, has received preventive chemotherapy for onchocerciasis by community-directed treatment with ivermectin (CDTI). With CDTI, interruption of transmission could occur after 14-17 years, yet transmission persists even after 20 years. This may be due to high baseline endemicity and other factors that can lower coverage of CDTI, including poor knowledge of the disease and negative attitudes and perceptions towards CDTI. This study examined community knowledge, attitudes and perceptions towards onchocerciasis control and acceptability of the CDTI program in Ulanga.
Methods A mixed methods cross-sectional study was carried out from June to July 2018. A quantitative questionnaire was administered to 422 household members in the community. This was complemented with qualitative in-depth interviews and focus group discussions. Results
A majority (94.1%) of community members had heard of onchocerciasis. Only 15.4% had a high level of knowledge about onchocerciasis; about half (49.2%) had a low level of knowledge on the disease and CDTI. Negative attitudes and perceptions towards CDTI prevailed in nearly half of participants (46.2% and 44%, respectively). A majority (83.4%) of respondents agreed that they will take ivermectin, and 74.4% were willing to comply with treatment for 12 to 15 years to eliminate onchocerciasis. In focus group discussions, participants reported mistrust of the method of dose calculation, that affects the acceptability and use of ivermectin treatment. Conclusion Inadequate levels of knowledge, negative attitudes and perceptions have the potential to affect participation in CDTI. This may result in continual transmission of onchocerciasis.
Key words: CDTI, onchocerciasis knowledge, attitudes, perceptions, acceptability, Ulanga district.
INTRODUCTION resulting in visual impairment and blindness (WHO, 2017). Onchocerciasis is a parasitic disease of the eye and skin Onchocerciasis is endemic in Africa, the Americas and the caused by the filarial worm Onchocerca volvulus. The Arabian Peninsula. Globally, in 2017, 198 million people in disease is transmitted to humans by repeated bites of an 36 countries lived in endemic areas that put them at risk of infected black fly that belongs to Simulium spp (Crump et al., onchocerciasis, and over 40 million people were infected 2012). People who live near rapidly flowing rivers and (WHO, 2017). Among the infected people, more than 6.5 streams are at high risk of acquiring onchocerciasis because million suffered from severe itching or dermatitis, 800,000 black flies breed near fast flowing rivers. Infected people had visual impairment, and 300,000 people were blind due may have a skin disease in the form of nodules under the skin to infection (CDC, 2017; WHO, 2017). The African region is and debilitating itching; they may also have eye disease highly affected compared to other regions, as 99% of 28 Research Article East African Journal of Applied Health Monitoring and Evaluation
2 infected people live in sub-Saharan Africa. The burden of Sampling onchocerciasis has led to the introduction of control For the quantitative component, a multistage sampling interventions such as vector control (simulium control) and technique was used to obtain a representative sample. This community-directed treatment with ivermectin (CDTI). This was done in three stages using a cluster-sampling technique. has relieved 40 million people from infection, prevented The first stage involved simple random selection of one 600,000 people from going blind and ensured more than 18 endemic ward from the list of all 29 onchocerciasis-endemic million children were born free from the risk of acquiring the wards; Isongo ward was selected. The second stage was disease and becoming blind (WHO, 2017). random selection of two representative villages from Isongo In Tanzania, 6 million people are at risk of onchocerciasis, ward’ Isongo and Uponera villages were selected from five mainly in Morogoro, Tanga, Iringa, Lindi, Mbeya and villages. The third stage was simple random sampling of one Ruvuma regions. The main vector for transmission of representative village unit from the list of all village units in Onchocerca volvulus in Tanzania is Simulium damnosum each of the two villages, which together had 422 households. sensu lato (Mweya et al., 2007; Maegga et al.,. 2011). One household member was randomly selected from each Morogoro region, where Ulanga is located, was the first to household, resulting in a total of 422 household members be identified as hyperendemic, with microfilariae prevalence who met the inclusion criteria and were interviewed using of around 60% and nodule prevalence of 95% (Mwaiko et questionnaires. al.,1990). This led to the introduction of an onchocerciasis Participants for in-depth interviews (IDIs) and focus group control program. Tanzania, in collaboration with the African discussions (FGDs) were sampled purposively. Five Programme for Onchocerciasis Control, introduced CDTI in community drug distributors (CDDs) were recruited for IDIs 1997 (NTDCP, 2016). and eight community members were recruited for each of the The operational prediction is that interruption of two FGDs. The inclusion criteria for purposive sampling transmission would be reached within 14-17 years with were community members and community drug distributors annual CDTI coverage of at least 80% (Colebunders et al., older than 15 years and residents of their respective villages 2018). Following two decades of annual CDTI, a for more than 12 years. transmission assessment survey carried out in 2017 in the Mahenge with a focus of Ulanga district found the Measurements prevalence of onchocerciasis to be 2.1%, indicating The main measurements were knowledge, attitudes, persistent transmission (NTDCP 2016; NTDCP 2017). This perceptions and acceptability of onchocerciasis control. A could plausibly be explained by several factors. CDTI series of questions were asked to measure knowledge and coverage in Ulanga district was below the 80% optimal acceptability. Questions of knowledge were measured as coverage (Colebunders et al., 2018), at less than 65% awareness and familiarity on cause, symptoms, mode of between 1997 to 2002 and average of 76% between 2003 to transmission and prevention strategies of onchocerciasis. 2017 (NTDCP, 2017). Higher baseline endemicity levels Acceptability was measured as willingness to accept and require higher coverage and longer treatment durations comply with the duration of ivermectin treatment. A Likert (Colebunders et al., 2018). The contribution of community scale was used to measure attitudes and perceptions. Attitude knowledge, attitudes and perceptions towards onchocerciasis was measured by asking their opinions and feelings on and its influence on participation and acceptability of CDTI onchocerciasis transmission, discrimination and distribution in Ulanga have not been fully investigated. This study of ivermectin drugs by CDDs. Perception was measured by examined the influence of knowledge, attitudes and asking their experience and beliefs on the way ivermectin is perceptions on community participation and acceptability of distributed in the community, the benefits of ivermectin the CDTI program in Ulanga district, Tanzania. treatment, responsibility of women as community drug distributors, and role of government in running the METHODS onchocerciasis control program. The questionnaire was pre- Study design and setting tested among 42 community members (10% of the 422 A mixed methods community-based study involving calculated sample size) and the feedback obtained was used quantitative and qualitative approaches was conducted in to improve the instrument. Ulanga district from June to July 2018. The Ulanga district, For the qualitative component, focus group discussion and one of the six districts in Morogoro region, has 40 rivers, of interview guides were used to explore participants’ which 34 rivers flow throughout the year. Some of the knowledge, attitudes and perceptions toward onchocerciasis perennial rivers such as Luli, Lukande, Mbalu, Mzelezi, and the CDTI program. Mzingizi and Ruaha constitute breeding sites for S.damnosum s.I. (vector) that transmits onchocerciasis. Data analysis The quantitative data were entered, cleaned, coded and Study participants double-entered into a computer. Analysis was done using the The study population consisted of community members statistical package for social sciences (SPSS) version 22 above 15 years of age, who had been residents of their (IBM Corp., Armonk, NY, USA). Socio-demographic respective villages for more than 12 years and who voluntary characteristics, knowledge, attitudes and perceptions were agreed to participate. summarized in frequency tables with proportions and their 95% confidence intervals (CI). 29 Research Article East African Journal of Applied Health Monitoring and Evaluation
3 To measure knowledge, a scale was developed from eight multiple-choice questions that carried a total of 12 correct Table 1: Demographic characteristics of study participants responses. A correct response was scored 1 and an incorrect (N=422) response was scored 0. Scores were added to make one scale that ranged from 0-12 points. Cut-off points were developed Variables N (%) from the mean knowledge score and categorized as follows: Sex 80 - 100% (10-12 points) indicated high knowledge level; 50% Males 211 (50.0) – 79% (6-9 points) indicated moderate knowledge; and less than 50% (0-5) indicated low knowledge. Females 211 (50.0) Attitudes and perceptions were measured with a five-point Age (years) Likert scale. For attitudes, there were 7 items each scored < 29 100 (23.7) from 1 to 5. These were subjected to reliability analysis and Cronbach’s alpha value of 0.638 was obtained. This was 30 - 49 212 (50.2) followed by running a sum score, and the attitude scale was 50 - 69 83 (16.1) formed with a range from 7 to 35 points. The mean score was Above > 69 27 (6.4) used to classify attitudes as negative or positive. Negative Education level attitudes ranged from 8 to 23 points, while positive attitudes ranged from 24 to 35 points. For perceptions, there were 15 No formal education 54 (12.8) items each scored from 1 to 5. Reliability analysis for all 15 Pre primary education 42 (10.0) items resulted in Cronbach’s alpha value of 0.714. Primary education 239 (56.7) Accordingly, all 15 items were added and a perception scale was formed, which ranged from 15 to 75. The mean score Secondary education 70 (16.6) was calculated so as to classify perceptions as negative or Post-secondary education 17 (0.4) positive. Negative perceptions ranged from 15 to 45 points, Occupation while positive perceptions ranged from 46 to 75 points. Peasants and livestock The qualitative data from IDIs and FGDs were transcribed 299 (70.9) and then organized into segments of text; analysis and keepers interpretations were done by clustering similar and related Traders 74 (17.5) topics together to form major emerging themes. Employed 33 (7.8)
Ethical considerations Unemployed 16 (3.8) Ethical clearance was obtained from the Muhimbili Duration of residence (years) University of Health and Allied Sciences (MUHAS) Ethical < 29 133 (31.5) Review Board, while permission to conduct the study was obtained from the administrative authorities of Morogoro 30 - 49 186 (44.0) region and Ulanga district. Informed verbal consent for 50 - 69 79 (18.8) participation in the study was obtained from each participant. > 69 24 (5.7)
RESULTS Respondents knew that the following were symptoms of Socio-demographic characteristics of the study onchocerciasis: skin itching (86.3%), followed by skin participants rashes (72.3%), blindness (71.8%), skin depigmentation A total of 422 respondents were interviewed in 422 (58.1%) and skin lesions (43.8%). A majority of participants households with a sex ratio of 1:1 with half (50.2%) being (96.4%) were aware of modern treatment; however only 11.8% 30-49 years old (Table 1). More than two thirds (70.9%) of knew ivermectin as the drug of treatment and prevention. participants depended on small-scale agriculture and Respondents (44%) who had lived in Ulanga for 30-49 years livestock keeping as sources of income while less than half were more likely to report the correct type of treatment (44%) were residents for 30-49 years. compared to those (31.5%) who had lived in Ulanga for 29
or fewer years. From the total respondents, 207 (49.2%) had Knowledge of onchocerciasis low level of knowledge; 150 (35.5%) had moderate level of Of 422 respondents interviewed, 397 (94.1%) had ever heard knowledge; and 65 (15.4%) had high level of knowledge about onchocerciasis (Table 2). Only 78 (19.6%) correctly regarding onchocerciasis and ivermectin treatment. knew filarial worm as the causative agent; however, 247 Focus group discussions provided more depth about (62.2%) thought the black fly to be the causative agent. Of knowledge of the disease. Most participants were aware of respondents, 119 (28.2%) knew onchocerciasis is transmitted the disease and knew its local name, sign and symptoms, but from one person to another, and 217 (51.4%) reported that they did not clearly define the disease and cause of it. One of onchocerciasis is not transmitted from one person to another. the participants said: However, 301 (71.3%) correctly knew that the black fly bite was the mode of transmission; the rest (28.7%) reported “…The local term for onchocerciasis is wankoseka, when mosquito bites, contact with infected persons, and aerosols. 30 Research Article East African Journal of Applied Health Monitoring and Evaluation
4 you say wankoseka even children can understand, the only Yes 303(71.8) thing I know about wankoseka is a disease of itching No 47(11.1) caused by vifuna, the moment vifuna bite you will suffer severe itching.” (Female, Uponera, 40-49 years) Don’t know 72(17.1) Skin rashes Table 2: Knowledge of the disease among community Yes 305(72.3) members (N=422) No 42(10) Indicative questions on Don’t know 75(17.8) N (%) knowledge Type of treatment used Heard about onchocerciasis Modern 407(96.4) Yes 397(94.1) Traditional 15(3.6) No 25(5.9) If Modern, which drug is used Causative agent of Ivermectin 86(20.4) onchocerciasis Albendazole 73(17.3) Filarial worm 78(19.6) Don’t know 299(70.9) Black fly 247(62.2) Onchocerciasis can be prevented Mosquito 37(9.3) Yes 253(60) Witchcraft 2(0.5) No 83(19.7) Don’t know 33(8.3) Don’t know 86(20.4) Transmission from person to person Acceptability of community-directed treatment with ivermectin Yes 119(28.2) A majority of respondents (83.4%) agreed that they would No 217(51.4) take ivermectin, while 16.6% didn’t agree (Table 3). Among Don’t know 86(20.4) those who would not agree to take ivermectin, the reasons were fear of side effects (52.9%), mistrust of the method of Mode of transmission of dose calculation (17.1%), and because they were not sick onchocerciasis (30%). Most (74.4%) were willing to comply with CDTI for Black fly bite 301(71.3) the duration of 12 to 15 years in order to eliminate Contact with infected person 10(2.4) onchocerciasis. Itching (49.1%) and body swelling (16.7%) were mentioned as side effects of taking ivermectin. Mosquito bite 42(10) During focus groups, participants were concerned that Through breath 6(1.4) there were misconceptions about the use of height Don’t know 63(14.9) measurements as a method of dose calculation; most participants did not trust this method. Signs and symptoms of onchocerciasis “…Height measurement is used as method of dose Itching calculation, but I do not trust this method, and I think it is the reason of the side effects people get; because how can Yes 364(86.3) a person with 60kg given the dose similar to that of a No 14(3.3) person with 80 kg with excuse that they have same height. Don’t know 44(10.4) I think other measurements should also be taken such as Lesion weight before giving the drugs.” (Female, Uponera, 30-39 years) Yes 185(43.8) No 114(27) Attitudes towards onchocerciasis and ivermectin Don’t know 123(29.1) treatment Among the 422 participants, 343 (79.1%) believed that Skin depigmentation onchocerciasis is a bad disease, and 338 (80.1%) believed Yes 245(58.1) that it’s very sad when you see a person with onchocerciasis No 63(14.9) (Table 4). Similarly, 257 (60.9%) agreed that affected individuals worry about consequences of onchocerciasis. Don’t know 114(27) More than half of the respondents (53.3%) were against the Blindness idea of stigmatization of the affected people. Itching was
31 Research Article East African Journal of Applied Health Monitoring and Evaluation
5 identified as the main side effect after taking ivermectin are at risk of contracting the disease; while 64 (15.2%) (71.1%). From the total respondents, 227 (53.8%) had a agreed or strongly agreed that onchocerciasis is caused by positive attitude towards onchocerciasis control and witchcraft. More than one third (39.1%) of the respondents ivermectin treatment while the rest (46.2%) had a negative agreed that ivermectin treatment will eliminate attitude towards onchocerciasis control and ivermectin onchocerciasis; while 195 (46.1%) held the opinion that treatment. ivermectin has other health benefits [with 106 (25.1%) agreeing and 89 (21%) strongly agreeing]. Table 3: Acceptability of Community Directed Treatment Regarding the operations of the CDTI program, 107 (25.4 %) were not sure if CDDs are capable of handling side with Ivermectin Program among Community Members in effects. However, 160 (37.9%) agreed CDTI is efficient in Ulanga District (N=422) fighting and preventing onchocerciasis. Furthermore, a Indicative statement on majority [308 (71.8%)] agreed that onchocerciasis control N (%) acceptability should be run by the government. From the total respondents, 236 (56%) of respondents had a positive perception towards Would take ivermectin as onchocerciasis control and ivermectin treatment while the distributed by CDDs rest [186 (44%)] had a negative perception towards Yes 352(83.4) onchocerciasis control and ivermectin treatment. Perceptions towards onchocerciasis were also supported No 70(16.6) with qualitative information from focus groups. Reasons for not taking (n=70) Onchocerciasis-related discrimination and stigmas among Fear of side effects 37(52.9) infected people were not occurring in the community, according to participants, as they are treating each other Don’t trust the dose 12(17.1) equally. One of the participants said: Not sick 21(30) Willingness to comply with 12 to “…How can my fellows discriminate me because of 15 yearly wankoseka, I was infected with wankoseka but no one treated me badly, I continued to share alcohol with others” rounds of CDTI (Male, Uponera, 40-49 years) Yes 314(74.4) No 30(7.1) Participants were also probed to see if there were local superstitions about the cause of onchocerciasis. All of them Not sure 78(18.5) denied that this was the case. For example: Experiences with ivermectin side
effects “…There is no relationship between superstition and onchocerciasis; it’s a disease like other diseases that Yes 269(63.7) requires hospital treatment.” (Male, Uponera, 20-29 years) No 96(22.7) Don’t remember 57(13.5) Knowledge, attitudes and perceptions of community drug Ivermectin side effects distributors (CDDs) During in-depth interviews, no CDDs knew the correct experienced causative agent of onchocerciasis. Four of them responded Itching 132(49.1) that Vifuna (black flies) were the causative agent and one of Dizziness 18(6.7) them said that mosquitoes were the causative agent. Most CDDs were only aware of black fly bite as the mode of Headache 22(8.2) transmission and did not clearly know how transmission Vomiting 15(5.6) occurs. All CDDs were able to mention the symptoms of Diarrhea 12(4.5) onchocerciasis, the most common symptoms being severe itching, skin depigmentation, skin lesions and blindness. Few Body swelling 45(16.7) CDDs were able to mention the correct drug for treatment for Body pain 25(9.3) onchocerciasis. The following is one of the statements given by CDDs: Perceptions towards onchocerciasis and ivermectin treatment “….How can I memorize the name of the drug, I only know Out of 422 respondents, 321 (76.1%) agreed or strongly how they look like and how to distribute them.” (Male, agreed that onchocerciasis is a common health problem in Isongo, 51 years) their community (Table 5). A majority [308 (73%) and 315 (74.1%) respectively] agreed that onchocerciasis is a CDDs believed that some of community members have dangerous disease and a curable disease. Approximately a positive attitudes and perceptions towards ivermectin as quarter (24.2%) of respondents strongly disagreed that they being very useful in treating those infected, and at the same 32 Research Article East African Journal of Applied Health Monitoring and Evaluation
6 time preventing those uninfected from getting onchocerciasis, in the community and transmission has decreased while others had negative attitudes and perception toward compared to before. Now when distribution of ivermectin ivermectin due to side effects, mistrust of method of dose is late people are coming to ask me what is happening. Or calculation, fear of becoming impotent or being uninfected. why are they not getting the medication?” (Female, Isongo, 49 years) “…Community members know the importance of using ivermectin, it has helped to treat a lot of infected people
Table 4: Attitude towards Onchocerciasis and Ivermectin treatment (N=422) Attitudes statements on disease Strongly Agree Not Disagree Strongly and treatment agree sure disagree No (%) No (%) No (%) No (%) No (%) Onchocerciasis is a bad disease 133(31.5) 201(47.6) 41(9.7) 26(6.2) 21(5) It’s very sad when you see a person with the diseases 120(28.4) 218(51.7) 52(12.3) 17(4) 15(3.6) A person with the disease worries about its consequences 83(19.7) 174(41.2) 83(19.7) 30(7.1) 52(12.3) A person with the disease avoids contacts with other community member 22(5.2) 66(15.6) 109(25.8) 86(20.4) 139(32.9) Sometimes Ivermectin causes itching 102(24.2) 198(46.9) 35(8.3) 48(11.4) 39(9.2) Ivermectin is preferably given by female CDDs 111(26.3) 126(29.9) 91(21.6) 50(11.8) 44(10.4)
CCDs are happy to help patients without payment 68(16.1) 83(19.7) 39(9.2) 74(17.5) 158(37.4)
Side effects were mentioned as a barrier of community hardening, vomiting, dizziness and diarrhea especially participation in the CDTI program. All CDDs said ivermectin when it’s the first time to take ivermectin, but when has several side effects, but that they differ from person to ivermectin is taken for several years these side effects person. The following were side effects mentioned: tends to subside.” (Female, Isongo, 49 years)
“…The main side effects are swelling of the body, skin
Table 5: Perceptions towards Onchocerciasis and Ivermectin treatment (N=422) Perceptions statements on disease Strongly Agree Not Disagree Strongly and treatment agree No (%) sure No (%) disagree No (%) No (%) No (%) Disease is common 124(29.4) 197(46.7) 54(12.8) 26(6.2) 21(5) Disease is dangerous 134(31.8) 174(41.2) 77(12.8) 29(6.9) 8(1.9) Disease is curable 144(34.1) 171(40.5) 58(13.7) 27(6.4) 22(5.2) Risk of contracting disease 97(23) 89(21.1) 97(23) 37(8.8) 102(24.2) Disease associated with witchcraft 23(5.5) 41(9.7) 65(15.4) 76(18) 217(51.4) CDTI will eliminate disease 71(16.8) 165(39.1) 101(23.9) 55(13) 30(7.1) CDTI has other health benefits 89(21) 106(25.1) 144(34.1) 63(14.9) 20(4.7) CDDs households visits increase ivermectin uptake 134(31.8) 146(34.6) 79(18.7) 40(9.5) 23(5.5) CDDs can manage side effects 79(18.7) 122(28.9) 107(25.4) 63(14.9) 51(12.1) CDDs should get incentives 107(25.4) 139(32.9) 83(19.7) 47(11.1) 46(10.9) Female CDDs are more persuasive 78(18.5) 118(28) 120(28.4) 51(12.1) 55(13) CDTI is the most effective disease control tool 81(19.2) 160(37.9) 112(26.5) 24(5.7) 45(10.7) Persons with disease should be isolated 35(8.3) 41(9.7) 101(23.9) 77(18.2) 168(39.8) Disease has socio-economic impact 119(28.2) 148(35.1) 90(21.3) 36(8.5) 29(6.9) Disease control should best be run by government 160(37.9) 143(33.9) 47(11.1) 23(5.5) 49(11.6)
All community drug distributors interviewed agreed that they were given education on how to manage side effects.
33 Research Article East African Journal of Applied Health Monitoring and Evaluation
7 One of the CDDs said: Attitudes and perception of community members towards onchocerciasis and CDTI “…If a person complains of itching or swelling of body Attitudes and perceptions towards onchocerciasis and the after taking medication, I know these are side effects of CDTI intervention have major implications for the uptake medication hence I administer aspirin/panadol to subside and therefore coverage of an intervention, which also affects the side effects.” (Male, Uponera, 70 years) the success of the control of the disease (Colebunders et al., 2018). The present findings show that almost half of study DISCUSSION participants had negative attitudes, which implies that Knowledge of community members on onchocerciasis coverage may not be optimal because those with negative and CDTI attitudes may be non-compliant to the intervention. The Inadequate level of knowledge on onchocerciasis affects findings are consistent with the studies done in DRC, Nigeria coverage and community participation in CDTI programs, and Cameroon (Makenga et al., 2017; Brieger et al., 2012; conceivably leading to a high prevalence in those areas Okeibunor et al., 2011). (Okwara et al., 2017). Despite onchocerciasis being a major health problem in Ulanga, less than a fifth of the study Attitudes and perception of community drug distributors participants correctly knew the causative agent; the In Ulanga, CDDs held positive attitudes and perceptions remaining participants held the misconceptions that the towards use of ivermectin for onchocerciasis control. It was disease is caused by black flies and mosquitoes as well as observed that CDDs and community members believed that witchcraft. Similarly, more than a quarter of participants had ivermectin is the best drug for treatment and preventing misconceptions that transmission is by mosquitoes bites, onchocerciasis because of the obvious benefits. However, contact with infected person, and aerosols. This observation there were some negative attitudes and perceptions due to is consistent with studies conducted in Nigeria and side effects and doubt of the method of dose calculation and Northwestern Ethiopia (Adeoye, Ashaye and Onakpoya, fear of impotency or being uninfected. 2010; Weldegebreal et al., 2014; Afolabi et al., 2016). In one study, respondents were knowledgeable on the LIMITATIONS manifestations of onchocerciasis, possibly reflecting their The major limitations of this study were accuracy of recall long experiences with these symptoms (Zouré et al., 2014). and personal bias from the respondents. Participants were asked questions that required them to recall information such Acceptability of CDTI as side effects experienced as a result of using ivermectin Most (83%) respondents were willing to take ivermectin and medication; this might have led to under- or over-reporting about three quarters were willing to comply with treatment of the information. Furthermore, this cross-sectional design for 12 to 15 years in order to eliminate the disease. These cannot determine if the measured outcomes had an impact on findings have positive implications for the success of the continued transmission of onchocerciasis. CDTI program because acceptability and compliance to ivermectin treatment is crucial in controlling onchocerciasis Public health implications in endemic areas, and it has proven to be successful in In Ulanga district, a majority of community members were reducing transmission, morbidity and elimination of the aware of the presence of onchocerciasis in the community. disease (Kim et al., 2015). Fear of side effects, mistrust of However, there was a borderline level of knowledge on the method of dose calculation and being uninfected were causative agent, transmission, treatment and prevention. reasons given for non-compliance, and this can affect the Furthermore, nearly half of the population held negative participation of people in CDTI programs and hence slow attitudes and perceptions towards the CDTI program. interruption of onchocerciasis transmission in this area. Inadequate level of knowledge coupled with negative Similar observations were made in Morogoro, Tanzania attitudes and perceptions in the community mean that (York et al 2015). coverage may not be optimal. Those with negative perceptions may be non-compliant to ivermectin treatment Knowledge of community drug distributors and therefore affect total coverage of the CDTI program, CDDs play an important role in engaging the community to leading to persistent transmission. participate in the control program; they can either influence In order to address this problem, public health education positively or negatively the success of such a program should be reinforced in the community so as to improve the (Krentel et al., 2017). In Ulanga, all of the CDDs interviewed level of knowledge regarding onchocerciasis and ivermectin were familiar with onchocerciasis, though the majority of treatment as well as changing the negative perceptions and them had inadequate knowledge. This is possibly due to attitudes towards CDTI. Health education should focus on inadequate training in the biology of the disease, as the the causative agent, transmission, signs and symptoms, training mostly covered drug distribution and management ivermectin treatment and prevention in order to ensure better of side effects (York et al., 2015). Poor knowledge of CDDs understanding about onchocerciasis control. affects participation of community members in the control program, as community members may not trust them (York ACKNOWLEDGEMENTS et al., 2015). The authors are grateful to District Administrative Secretary and Ulanga District Medical Officer who permitted this 34 Research Article East African Journal of Applied Health Monitoring and Evaluation
8 study to be carried out in their jurisdiction areas. Sincere Maegga, B.T.A., Kalinga, A.K., Kabula, B., Post, R.J. thanks should also go the study participants and all research and Krueger, A. (2011). Investigations into the isolation of assistants for their cooperation and support during field work. the Tukuyu focus of onchocerciasis (Tanzania) from S. damnosum s.l. vector re-invasion. Acta Tropica, 117(2), CONFLICTS OF INTEREST pp.86–96. There are no conflicts of interest. Makenga Bof, J.-C., Mpunga, D., Soa, E. N., Ntumba, F., Bakajika, D., Murdoch, M. E., & Coppieters, Y. (2017). AUTHORS’ CONTRIBUTIONS Onchocerciasis in the Democratic Republic of Congo: VM designed the study, did data collection, analysis and Survey of knowledge, attitude and perception in Bandundu interpretation and drafted the manuscript. DT and DK province. Journal of Infection and Public Health, 10(5), participated in study design and critically revised the 600–607. https://doi.org/10.1016/j.jiph.2017.01.017 manuscript. Mwaiko, G. L., Mtoi, R. S. and Mkufya, A.R. (1990). Onchocerciasis prevalence in Tanzania. Cent Afr J Med, REFERENCES 36(4), pp.94–6. Adeoye, A.O., Ashaye, A.O., Onakpoya, O.H. (2010). Mweya, C.N. (2007). Onchocerciasis situation in the Perception and attitude of people toward onchocerciasis Tukuyu focus of southwest tanzania after ten years of (river blindness) in South Western Nigeria. Middle East ivermectin mass treatment. Volume 9, Issue 3, p.Pages 174- African Journal of Ophthalmology, 17(4), 310–314. 179. Afolabi, O., Okaka, C., Oke, I., and Oniya, M. (2016). Neglected Tropical Diseases Control Program, (2016). Knowledge, Attitude and Perception of Onchocerciasis and Neglected Tropical Disease Control Program. 5th Annual Ivermectin Treatment in Idogun Community, Ondo State, Joint Planning Meeting. Ministry of Health, Community, Nigeria. British Journal of Medicine & Medical Research, Development, Gender, Elderly and Children. 13(4), 1-7. Neglected Tropical Diseases Control Program, (2017). Brieger, W.R., Okeibunor, J.C., Abiose, A.O., NTDSCP, 2016/2017 Annual Program Report. Ministry of Ndyomugyenyi, R., Wanji, S., Elhassan, E. and Amazigo, U. Health, Community, Development, Gender, Elderly and V. (2012). Characteristics of persons who complied with and Children. failed to comply with annual ivermectin treatment. Tropical Okeibunor, J. C., Amuyunzu-Nyamongo, M., Onyeneho, Medicine & International Health, 17(7), pp.920–930. N. G., Tchounkeu, Y. F. L., Manianga, C., Kabali, A. T., and CDC, (2017) CDC-Neglected Tropical Diseases -The Leak, S. (2011). Where would I be without ivermectin? Burden of Onchocerciasis. Available at: Capturing the benefits of community-directed treatment with https://www.cdc.gov/globalhealth/ntd/diseases/oncho_burde ivermectin in Africa. n.html [Accessed March 12, 2018]. Okwara, K., Iwunze J.I., Njoku F.U., Nwachukwu, M. and Colebunders, R., Basáñez, M.-G., Siling, K., Post, R. J., Heanacho J.N. (2017) Knowledge and perception on Rotsaert, A., Mmbando, B., Hopkins, A. (2018). From river onchocerciasis and benefits of ivermectin in parts of imo blindness control to elimination: bridge over troubled water. river basin following long term treatment *Amaechi A. Infectious Diseases of Poverty, 7(1), 21. Global Research Journal of Science, 5(1), pp.21–31. Crump, A., Morel, C.M. and Omura, S. (2012). The Weldegebreal, F., Medhin, G., Weldegebriel, Z. and onchocerciasis chronicle: From the beginning to the end? Legesse, M. (2016) Knowledge, attitude and practice of Trends in Parasitology, 28(7), pp.280–288. community drug distributors’ about onchocerciasis and Endale, A., Erko, B., Weldegebreal, F. and Legesse, M. community directed treatment with ivermectin in Quara (2015). Predictors of compliance with community-directed district, North Western Ethiopia. BMC research notes, 9, treatment with ivermectin for onchocerciasis control in Kabo p.206. area, southwestern Ethiopia. Parasites & Vectors, 8(1), p.99. World Health Organization, (2017). Onchocerciasis fact Fleming, F.M., Matovu, F., Hansen, K.S. and Webster, sheet. Geneva:WHO. J.P. (2016). A mixed methods approach to evaluating Avaiableat:http://www.who.int/mediacentre/factsheet/fs374/ community drug distributors performance in the control of en. [Accessed March13, 2018]. neglected tropical diseases. Parasites & Vectors, 9(1), p.345. York, K.J., Kabole, I., Mrisho, M., Berry, D.M. and Kim, Y.E., Remme, J.H.F., Steinmann, P., Stolk, W.A., Schmidt, E. (2015) Factors Affecting Community Roungou, J.B. and Tediosi, F. (2015) Control, elimination, Participation in the CDTI Program in Morogoro, Tanzania. and eradication of river blindness: scenarios, timelines, and Journal of Nursing Scholarship, 47(1), pp.96–104. ivermectin treatment needs in Africa. PLoS neglected Zouré, H.G.M., Noma, M., Tekle, A.H, Amazigo, U.V., tropical diseases, 9(4), p.e0003664. Diggle, P.J., Giorgi, E.and Remme, J.H.F. (2014). The Krentel, A., Gyapong, M., Mallya, S., Boadu, N.Y.A., geographic distribution of onchocerciasis in the 20 Nyamongo,M., Stephens, M. and McFarland, D.A. (2017). participating countries of the African Programme for Review of the factors influencing the motivation of Onchocerciasis Control: (2) pre-control endemicity levels community drug distributors towards the control and and estimated number infected. Parasites & vectors, 7, p.326. elimination of neglected tropical diseases (NTDs) J. F. Friedman, ed. PLOS Neglected Tropical Diseases, 11(12), p.e0006065. 35