SCHOOL OF PUBLIC HEALTH COLLEGE OF HEALTH SCIENCES UNIVERSITY OF GHANA FACTORS INFLUENCING COMPLIANCE OF PRESCRIBERS WITH MALARIA TEST- BASED CASE MANAGEMENT POLICY IN EFFUTU MUNICIPALITY. BY ALEXANDER ASAMOAH (10179774) A THESIS SUBMITTED TO THE SCHOOL OF PUBLIC HEALTH, UNIVERSITY OF GHANA, IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE AWARD OF THE MASTER OF PHILOSOPHY DEGREE IN APPLIED EPIDEMIOLOGY AND DISEASE CONTROL University of Ghana http://ugspace.ug.edu.gh ii JUNE, 2014. University of Ghana http://ugspace.ug.edu.gh i DECLARATION I, Alexander Asamoah, declare that except for other people’s investigations which have been duly acknowledged, this thesis is the result of my own original research undertaken under supervision and that it has neither in whole nor in part been presented for another degree in this university or elsewhere. Author: ALEXANDER ASAMOAH ……………………………………………. Date............................................................... MPhil Resident, School of Public Health, College of Health Sciences, University of Ghana, Legon. Academic Supervisor: DR. SAMUEL OKO SACKEY ..…………………………………………… Date................................................................ Department of Epidemiology and Disease Control, School of Public Health, College of Health Sciences, University of Ghana, Legon University of Ghana http://ugspace.ug.edu.gh ii DEDICATION Dedicated to my wife Mavis Asamoah and son Ellis-Roi Kofi Asamoah. University of Ghana http://ugspace.ug.edu.gh iii ACKNOWLEDGEMENT My utmost gratitude goes to the Jehovah God Almighty for His protection and strength to bring me this far through this work. I am very grateful to my supervisor Dr. Samuel Sackey for his guidance and directions that helped in shaping this project. I also express my sincere gratitude to Mr Anthony Dongdem for his mentoring activities, the entire faculty of the School of Public Health, and the faculty of the Ghana Field Epidemiology and Laboratory Training Programme, Legon for their comments and useful suggestions. My deepest gratitude goes to Dr. Amoussou, the Effutu Municipal Director of Health Services and the entire management and staff of all the health facilities who granted me permission to conduct the study in their facilities. I am also thankful to Messrs. Ama Annobil and Mrs. Diana Awo Zanu, as well as all those who helped in the data collection. I wish to thank the prescribers, patients and caregivers who consented to participate in this study. Notably, I am grateful to the President’s Malaria Initiative for funding this research to a successful end. University of Ghana http://ugspace.ug.edu.gh iv ABSTRACT Introduction: Malaria remains a major public health preventable and treatable mosquito- borne ailment. A test-based case management of malaria and targeted use of Artemisinin- based Combination Therapy (ACT) for treatment has proven to reduce over-diagnosis and overtreatment and therefore recommended as the main control strategy. But compliance by prescribers is still low. Most districts still manage malaria presumptively with treatment of negative test results with ACT. This study was to determine factors that influence the compliance of prescribers with the test-based malaria case management policy in Effutu Municipal. Methods: A cross sectional study was conducted to extract both qualitative and quantitative data from health facility records and prescriber interviews as well as assess prescribers' malaria management of patients. Univariate analyses of categorical variables were expressed as frequencies and proportions. Bivariate analysis was used to show associations between selected independent variables and patient testing as well as treating patients according to test results. Results: Of 175 patients and 25 prescribers assessed for compliance, 125 (71.4%) and 13 (52%) were females respectively. Prescribers complied with the policy for 15 (8.6%) patients suspected of uncomplicated malaria. Factors identified to influence testing included patient age 13 - 45 years OR=1.26(95%CI =0.50-3.20), and measured temperature of ≥37.5oC 2.40(0.66-8.76), patient NHIS status 3.54(0.44-27.99), prescriber age ≤35 years 1.52(0.68- 3.42), prescriber female sex 1.74(0.81-3.73), prescriber cadre as physician assistant 2.08(0.79-5.44) and years of experience <6 years 1.71(0.69-4.23), health facility factors such as mission/religious operating authority 5.08(1.67-15.45) and having a functional laboratory University of Ghana http://ugspace.ug.edu.gh v or five microscopists. Factors identified to influence treating according to test results included patient age >45 years 1.50(0.17-13.22), and measured temperature of 37.5oC or more 1.23(0.15-9.97), prescriber age ≤35 years 2.15(0.45-10.29), prescriber male sex 2.04(0.51-8.23), prescriber cadre as medical officer and years of experience < 6years 2.17(0.28-25.87), health facility factors such as lower health facility types 6.00(1.33-27.05), government operating authority 8.40(1.27-55.40) having a functional laboratory 1.56(0.13- 18.95) and five microscopists 1.22(0.14-10.48). Conclusion: The prescriber compliance with the malaria test-based case management policy in the Effutu Municipal at patient level was low. From this study, prescribers at Mission/Religious operating health facility significantly tested more patients before treatment than those in private hospitals. However, prescribers at government operating health facilities and lower health facility types significantly treated patients according to test results than those in private hospital and hospital facilities respectively. Key words: Malaria, prescribers, compliance, test-based management University of Ghana http://ugspace.ug.edu.gh vi TABLE OF CONTENTS DECLARATION ........................................................................................................................ i DEDICATION ...........................................................................................................................ii ACKNOWLEDGEMENT ........................................................................................................iii ABSTRACT.............................................................................................................................. iv TABLE OF CONTENTS.......................................................................................................... vi LIST OF TABLES ..................................................................................................................... x LIST OF FIGURES .................................................................................................................xii LIST OF ABBREVIATIONS .................................................................................................xiii CHAPTER ONE ........................................................................................................................ 1 1.0 INTRODUCTION ............................................................................................................... 1 1.1 Background .......................................................................................................................... 1 1.2 Problem Statement ............................................................................................................... 4 1.3 Conceptual Framework ........................................................................................................ 6 1.3.1 Explanation of Conceptual Framework......................................................................... 6 1.4 Justification .......................................................................................................................... 7 1.5 General Objective ................................................................................................................ 7 1.6 Specific Objectives .............................................................................................................. 7 CHAPTER TWO ....................................................................................................................... 8 2.0 LITERATURE REVIEW .................................................................................................... 8 University of Ghana http://ugspace.ug.edu.gh vii 2.1 Malaria Case Management and Burden ............................................................................... 8 2.2 Prescriber compliance with malaria test-based management recommendation .................. 9 2.3 Willingness of caregivers to accept policy and impact of compliance .............................. 14 2.4 Factors influencing prescribers' compliance with test-based policy .................................. 15 2.5 Prescriber's adherence to Guidelines ................................................................................. 17 CHAPTER THREE ................................................................................................................. 20 3.0 METHOD........................................................................................................................... 20 3.1 Study Design ...................................................................................................................... 20 3.2 Study Location ................................................................................................................... 20 3.2.1 Demography ................................................................................................................ 20 3.2.2 Environmental Factors ................................................................................................ 21 3.2.3 Health Services............................................................................................................ 21 3.3 Variables ............................................................................................................................ 22 3.4 Sampling ............................................................................................................................ 23 3.4.1 Study Population ......................................................................................................... 23 3.4.2 Sampling Size.............................................................................................................. 23 3.4.3 Sampling method......................................................................................................... 24 3.4.3.1 Health Facilities .................................................................................................... 24 3.4.3.2 Prescribers and Patients ........................................................................................ 24 3.4.5 Data Collection Techniques and tools......................................................................... 26 3.4.6 Ethical Clearance......................................................................................................... 27 University of Ghana http://ugspace.ug.edu.gh viii 3.4.7 Training of Interviewers .............................................................................................. 28 3.4.8 Pre-testing and review of data collection tools ........................................................... 28 3.5 Data Collection .................................................................................................................. 28 3.6 Data Quality Control .......................................................................................................... 30 3.7 Data Processing and Analysis ............................................................................................ 30 3.7.1 Data Processing ........................................................................................................... 30 3.7.2 Data Analysis .............................................................................................................. 31 CHAPTER FOUR.................................................................................................................... 32 4.0 RESULTS .......................................................................................................................... 32 4.1 Characteristics of Study Population ................................................................................... 32 4.1.1 Health facility characteristics ...................................................................................... 32 4.1.2 Prescriber Characteristics ............................................................................................ 35 4.1.3 Patient Characteristics ................................................................................................. 37 4.1.3.1 Patient Occupation ................................................................................................ 38 4.1.3.2 Patients' preference to test-based management of malaria ................................... 39 4.2 Assessment of the level of compliance by prescribers ...................................................... 41 4.2.1 Prescriber prescription practice ................................................................................... 43 4.3 Factors associated with testing suspected malaria cases before treatment. ...................... 44 4.3.1 Patient Factors ............................................................................................................. 44 4.3.2 Prescriber Factors ........................................................................................................ 47 4.3.3 Health Facility Factors ................................................................................................ 49 University of Ghana http://ugspace.ug.edu.gh ix 4.4 Factors associated with treating suspected malaria cases according to test results. .......... 51 4.4.1 Patient Factors ............................................................................................................. 51 4.4.2 Prescriber Factors ........................................................................................................ 53 4.4.3 Health Facility Factors ................................................................................................ 56 CHAPTER FIVE ..................................................................................................................... 58 5.0 DISCUSSION .................................................................................................................... 58 CHAPTER SIX ........................................................................................................................ 68 6.0 CONCLUSION AND RECOMMENDATIONS .............................................................. 68 REFERENCES ........................................................................................................................ 70 APPENDICES ......................................................................................................................... 76 APPENDIX 1: CONSENT FOR PARTICIPANTS ............................................................ 76 APPENDIX 2: DATA COLLECTION TOOLS .................................................................. 88 University of Ghana http://ugspace.ug.edu.gh x LIST OF TABLES Table Page Table 1. Types of health facilities by sub municipal, Effutu municipal, 2014........................32 Table 2. Types of health facilities by operating authorities, Effutu municipal, 2014 ........... ..33 Table 3. Facility level indicators for malaria test-based case management, Effutu municipal, 2014........................................................................................................................34 Table 4. Characteristics of prescribers by sub-municipal, Effutu Municipal, 2014. ........... ..36 Table 5. Responses of Prescribers by Cadre, Effutu Municipal, 2014.....................................37 Table 6. Distribution of patients interviewed by age, sex and by type of health facility, Effutu Municipal, 2014...............................................................................................38 Table 7: Performance of Prescriber towards compliance with malaria test-based case management policy...................................................................................................................42 Table 8: Reasons for prescriber preferences to anti-malarials other than AA by sex, Effutu Municipality, 2014........................................................................................................44 Table 9: Patient factors associated with being tested before prescribed treatment, Effutu Municipal, 2014............................................................................................................46 Table 10. Perceived factors influencing testing for malaria parasites before treatment by prescribers, Effutu Municipal, 2014....................................................................47 Table 11: Prescriber factors associated with testing of patients seen before prescribing treatment, Effutu Municipal, 2014...........................................................................................48 Table 12: Health Facility factors positively associated with testing before prescribing treatment, Effutu Municipal, 2014...........................................................................................50 Table 13: Patient factors positively associated with treating cases according to test results, Effutu Municipal, 2014............................................................................................................52 University of Ghana http://ugspace.ug.edu.gh xi Table Page Table 14. Perceived factors influencing treatment of uncomplicated malaria according to test results by prescribers, Effutu Municipal, 2014.............................................53 Table 15: Prescriber factors positively associated with treating cases according to test results, Effutu Municipal, 2014................................................................................................55 Table 16: Health facility factors associated with treating patients according to test results, Effutu Municipal, 2014...............................................................................................57 University of Ghana http://ugspace.ug.edu.gh xii LIST OF FIGURES Figure Page Figure 1. Conceptual framework: factors influencing prescriber's compliance.........................6 Figure 2 Proportion of patients tested and treated accordingly and prescribers by health facility, Effutu Municipality, 2014...............................................................................35 Figure 3. Distribution of patient interviewed by occupation, Effutu Municipal, 2014............39 Figure 4 Proportion of patients' preference to comply with malaria test-based case management policy, Effutu Municipal, 2014...........................................................................40 Figure 5: Proportion of patient's preference to comply with malaria test-based case management policy by sex, Effutu Municipal, 2014. ..............................................................41 Figure 6. Proportion of suspected malaria cases tested and appropriately prescribed ACT by health facilities, Effutu Municipal, 2014....................................................................43 University of Ghana http://ugspace.ug.edu.gh xiii LIST OF ABBREVIATIONS AA Artesunate-Amodiaquine ACT Artemisinin-based Combination Therapy AL Artemeter Lumefantrine CHPS Community-based Health and Planning Services CHAG Christian Health Association of Ghana DHIMS District Health Information Management System IMCI Integrated Management of Childhood Illness IPT Intermittent Preventive Treatment IRS Indoor Residual Spraying ITNs Insecticides Treated Nets LQAS Lot's Quality Assurance Sampling NHIA National Health Insurance Authority NHIS National Health Insurance Scheme NMCP National Malaria Control Programme OPD Outpatient Department OTSS Outreach Training Supportive Supervision PMI Presidents Malaria Initiative RDT Rapid Diagnostic Test WHO World Health Organisation University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE 1.0 INTRODUCTION 1.1 Background Malaria is a preventable and treatable mosquito-borne disease, whose main victims are children under five years of age and pregnant women mostly in Africa (WHO, 2012b). It is an ailment of major public health concern and has been estimated to cause over one million deaths worldwide with 90% occurring in sub-Saharan Africa. The disease is also directly responsible for 20% (one in five) of childhood deaths in Africa, and indirectly contributes to illness and deaths from respiratory infections, diarrhoeal diseases and malnutrition. In Ghana, malaria used to be hyper-endemic and it is estimated to account for 40% of all out patient attendance and over 25% of under-five mortality. To control malaria, the strategy in Ghana used to be case management based on prompt recognition and presumptive treatment as its main focus, using artemisinin based combinational therapy (ACT) Artesunate Amodiaquine as the first line drug (Ghana Ministry of Health, 2009). For this reason numerous interventions such as Intermittent Preventive Treatment (IPT) for pregnant women, use of Insecticide Treated Nets (ITNs), Indoor Residual Spraying (IRS) for malaria prevention as well as the use of Artemisinin based combinational therapy (ACTs) for treatment, malaria diagnosis and pharmaceutical management and drug quality for malaria case management as reported in President's Malaria Initiative document, (FY2012) have been implemented and therefore expecting malaria cases to decrease. However, higher numbers of malaria cases are still being recorded. The National Malaria Control Programme (NMCP) in its annual report (2010) confirms the increasing cases of malaria from about 3.1 million cases in 2007 to about 3.8 million in 2010. The proportion of malaria to total OPD cases also increased from 32.5% in 2009 to 34% in 2010 as reported in the Ghana Health Service annual report for year 2010. University of Ghana http://ugspace.ug.edu.gh 2 Meanwhile, some researchers have discovered over-diagnosis and overtreatment of malaria as one of the main reasons for the higher cases (Nanyingi, 2008; Okebe et al, 2010 and Abdelgader et al, 2012) despite the provisions of laboratory kits and the recommendation of test-based management of malaria by WHO. The concern for possible resistance of the Plasmodium parasites to the artemisinin based combination therapy due to overtreatment has been raised by some researchers and WHO has reported resistance in the countries Cambodia, Myanmar, Thailand and Vietnam. In Ghana, over-diagnosis is further affirmed by the National Malaria Control Programme (NMCP) in its 2010 annual report stating that "cases were predominantly presumptively diagnosed and most febrile illnesses were wrongly captured as malaria and therefore reinforced the need to confirm all cases by laboratory tests through microscopy or rapid diagnostic tests (RDTs)". Nonetheless, the cost of using Artemisinin-based Combinational Therapies (ACT) in treating malaria is very expensive compared to that of Chloroquine used in the past and the global fund for AIDS, TB and malaria is spending millions to fund the use of the ACTs across Africa. As such, over-diagnosis of malaria coupled with its attendant overtreatment of cases greatly increases the economic burden to control the disease. Following the evidence of over-diagnosis and overtreatment with the high cost implication and possible resistance factors against the highly effective ACTs, WHO recommended a test-based diagnosis and management of malaria across all age groups (WHO, 2010). The policy requires that, malaria tests be conducted for suspected cases and also treat cases according to test results by treating only cases with positive test results for uncomplicated malaria. Ghana through the NMCP and Ghana Health Service has adopted this policy to test all suspected of malaria before treatment (WHO, 2012; MOH, 2009; NMCP, 2010) and therefore expects prescribers, patients and other stakeholders to adhere to it to University of Ghana http://ugspace.ug.edu.gh 3 achieve the desired outcome. Testing could be done in the laboratory using microscopy or Rapid Diagnostic Tests kits (RDT). Recent studies conducted in Sudan and Uganda by Abdelgader et al, (2012) and Nanyingi, (2008) respectively showed a low adherence to the recommended test-based malaria case management policy. The 2010 NMCP annual report also stated that only about 31% of cases suspected to be malaria were tested in the year under review. Also, the current urban malaria study (2013) revealed that, the proportion of children living in Accra and Kumasi (which are urban areas in Ghana) and tested positive for malaria after reporting fever in the last two weeks was less than 7%. The study therefore recommended that health care professionals should be educated to perform malaria tests on all suspected cases, and also current practices for diagnosis and treatment of fever should be modified to reflect the low prevalence of malaria in most city neighborhoods to help avoid over-diagnosis and overtreatment. The lower rates recorded for malaria cases that are tested notwithstanding the proportion of the cases tested that are treated according to test results shows a lower adherence to the policy of testing all suspected cases before treatment and according to test results. There is therefore a need to investigate the factors that necessitate a suspected malaria case to be tested and also treated according to test results. Meanwhile numerous researches in adherence studies have mostly focused on characteristics and factors of patients that influence adherence of a clinical guideline, the provider they encounter and the organizational setting of the facility they visit (Feldman et al, 1997; Mckinlay et al, 1998 and Mckinlay et al, 2002) with less consideration of the prescriber's characteristics and attitudes involved. Studies by Mckinlay et al (2007) have shown that the adherence of prescribers in the case of Physicians with guidelines varies with different types of “patient” and with the length of clinical experience at a lower rate of less than 20% following another third of a set of guidelines. Also, a research on perceived barriers University of Ghana http://ugspace.ug.edu.gh 4 affecting prescribers adherence to clinical guidelines in Holland with a group of Dutch general practitioners showed that a lack of agreement with guidelines, lack of evidence as well as lack of knowledge relating to guidelines by some prescriber together with other barriers (Lugtenberg et al, 2009) work together to influence medical decisions as against the signs and symptoms of the problem itself. These are among few studies that have focussed on a set of guidelines considering the variety of barriers that should be addressed to improve guideline adherence (Cabana et al, 1999). In addition, guideline studies often focus on barriers regarding the guideline as a whole, rather than on barriers operating at the level of the individual recommendations within the guidelines (Boivin A, Legare and Gagnon 2008; Kasje, Denig and Haaijer-Ruskamp 2002; Smith, Walker and Gilhooly 2004; Cranney, Warren, Barton, Gardner and Walley, 2001). As different recommendations within the same guideline can have different barriers, it might be more useful to focus on barriers of individual recommendations to optimize the strategies needed for implementation of guidelines in practice. 1.2 Problem Statement As the over-diagnosis and over treatment of malaria persists, compliance of prescribers with the policy of universal laboratory diagnosis has been found to be low though varied compliance among prescribers in various settings to this policy guideline exists. In Ghana, the urban malaria study showed that only less than 7% of urban malaria cases were tested as opposed to the malaria case management policy. Analysis of DHIMS data by Bonku (2012) showed that proportion of confirmed cases was still low. Also, analysis of DHIMS data for the Effutu Municipal area showed that, only 15% of all uncomplicated malaria cases were tested to be positive. For cases 5 years and above, the confirmed OPD uncomplicated malaria cases were 8.4%, 3.2%, 18.7% and 14.8% for the period 2009, 2010, 2011 and 2012 University of Ghana http://ugspace.ug.edu.gh 5 respectively. That of 2013 (July) was 20.7%. Also for 2013 half year, total uncomplicated malaria cases confirmed was 22%. In the first quarter of 2013, only 4609 (23%) out of 20123 cases of OPD uncomplicated malaria were tested with 2008 (43.6%) such cases testing positive. However, 14018 (70%) of the cases were treated with ACTs aside cases treated with other anti-malarials. Meanwhile in an effort to implement the policy successfully, the NMCP report in 2010 indicates that education training for prescribers has been done and diagnostic services have been improved through the supply of microscopes with the support of the Global Fund and the United States- President Malaria Initiative (US–PMI). In addition, capacity building among laboratory staff has also been enhanced. The 2012 PMI malaria operational plan report indicates that, over 7015 clinicians have been trained to accurately treat malaria with RDT's as well as reaching about 240 clinical laboratories through their Outreach Training & Supportive Supervision (OTSS) in the year 2011. Moreover, in the 2012 funding year, PMI 7 proposes to procure additional laboratory equipment, as necessary, and continue to build capacity for microscopy and RDT use. With all these progress made, compliance rate to the policy is still low and therefore medical decisions by some prescriber to comply with the new policy might be influenced by factors other than what is stated in the policy guidelines as well as the training and logistics provided. The factors prescribers seek or barriers to overcome to readily request a laboratory test to confirm suspected malaria cases and also wait to treat according to the results as key guideline components in the malaria test-based case management policy, is yet to be thoroughly studied. Therefore this study aims to determine the factors that influence the compliance of prescribers with the test-based malaria case management policy. University of Ghana http://ugspace.ug.edu.gh 6 1.3 Conceptual Framework Figure 1. Conceptual framework: factors influencing prescriber's compliance 1.3.1 Explanation of Conceptual Framework Several factors could lead to influence prescriber's medical decision. However, the factors that would be studied for its influence on prescribers' medical decision to comply with the malaria test policy or not have been classified as Health facility factors, Patient Factors, External Policy Factor  NHIS Policy Prescriber Factors 1. Age 2. Sex 3. Professional Category 4. Professional Rank 5. Years of service 6. Awareness of test policy 7. Training about policy Patient Factors 1. Age 2. Sex 3. Occupation 4. Patients Symptoms 5. NHIS status Health Facility Factors 1. Facility type 2. Accreditation to NHIS 3. Availability of functional lab 4. Number of qualified microscopists 5. Availability of logistics 6. Availability of Treatment guidelines Prescriber's compliance University of Ghana http://ugspace.ug.edu.gh 7 Prescriber factors and External policy factor. These factors are presumed to influence the prescribers' compliance with the test-based policy either positively or negatively. 1.4 Justification According to the Ghana malaria control strategic plan for 2008 – 2015, the control programme hopes to reduce malaria burden by 75% by 2015 and also to have all uncomplicated cases follow the test treat track (T3) policy adopted to help attain their goal. Part of these strategies is to improve early diagnosis and effective management of malaria in all health facilities. Therefore these factors when identified would help make recommendations to the National Malaria Control Programme to strengthen, improve and optimize control strategies needed to improve compliance of prescribers with the test-based malaria policy towards the attainment of their goal. 1.5 General Objective To determine factors influencing compliance of prescribers with malaria test-based case management policy. 1.6 Specific Objectives 1. To determine the level of compliance among prescribers with the test-based policy 2. To identify factors that influence prescribers' decision to test cases 3. To determine factors that influence prescribers' decision to treat cases according to test results. University of Ghana http://ugspace.ug.edu.gh 8 CHAPTER TWO 2.0 LITERATURE REVIEW 2.1 Malaria Case Management and Burden Malaria case management has over the past been presumptive where patients with fever or history of fever are quickly diagnosed of malaria and treated with Chloroquine which was inexpensive. Following the report on Chloroquine resistance to malaria parasites, more expensive but highly effective ACTs were recommended. Considering the cost of malaria treatment and the presumptive requirement for malaria diagnoses, the economic burden involved in managing malaria therefore became unnecessarily huge. Therefore researchers significantly proved over-diagnosis using presumptive diagnosis of malaria and showed the possibility of unnecessary overtreatment and ACT drug resistance to malaria parasites. Subsequently, a test-based management of malaria was recommended to ensure effective use of ACTs for patients who test positive to malaria. Most countries have adopted this recommendation but are experiencing non compliance among the prescribers in spite of numerous interventions made. Malaria is one of the world's major public health concerns, contributing to 219 million clinical cases of malaria in 2010 and an estimated 660,000 deaths. India has the highest malaria burden in South East Asia accounting for an estimated 24 million cases per year followed by Indonesia and Myanmar. However about 90% of all malaria deaths occur in Africa which is also the most affected continent with Nigeria, Democratic Republic of the Congo, United Republic of Tanzania, Uganda, Mozambique and Cote d’Ivoire accounting for 47% of malaria cases (WHO, 2012a) For centuries, malaria has impaired productivity, economic growth, child development and learning, and health status on a large scale. The disease also takes a high toll on households and health care systems. As per the estimates of the World Health University of Ghana http://ugspace.ug.edu.gh 9 Organization, malaria reduces GDP growth by approximately one percentage point per year (WHO, 2008). In Ghana, malaria used to be hyper-endemic and it is estimated to account for 40% of all out patient attendance and over 25% of under-five mortality (Ghana Ministry of Health, 2009). 2.2 Prescriber compliance with malaria test-based management recommendation Effective case management practice using universal parasitological testing of malaria has become the target of most control strategies and interventions in Africa. This owes to the fact that presumptive diagnosis and treatment in the era of Chloroquine administration was less effective and would be more costly in the era of Artemisinin-based combinational therapy and hence recommend a laboratory diagnosis and treatment of malaria. Several studies therefore have extensively confirmed over-diagnosis and hence overtreatment in most settings due to presumptive treatment using varied methods (Nanyingi, 2008; Okebe et al, 2010, Chandler, Whitty and Ansah, 2010 and Abdelgader et al, 2012). In Gambia, a cross- sectional survey was carried out in two urban primary health facilities by Okebe et al (2010) during and outside the malaria transmission season and the results showed that only 33.2% of patients enrolled mostly children (0-15yrs) were tested during the wet season (p=0.003). Also, more children under five years were tested than older children (p = 0.022) which also showed that a positive test result was 4.4 times more likely in the older children (p = 0.010) than in the under fives. As such the claim that children under five years are more susceptible to malaria and therefore often diagnosed as such were found to be most significantly misdiagnosed and wrongly treated compared to the other age groups in both seasons. Meanwhile only 4.7% (10/215) and 12.5% (37/297) of patients tested positive for malaria parasites in the dry and wet seasons respectively. As much as most of the cases that tested negative were treated for malaria, the research findings showed that extrapolation of the University of Ghana http://ugspace.ug.edu.gh 10 proportion of cases unnecessarily treated for malaria suggested that 7,636 (95% CI: 7,586 - 7,686) of the 8,410 may have received anti-malarials for illnesses other than malaria. The results further showed that, in both seasons, the negative predictive value of a primary health facility slide was above 97%. This study therefore provides evidence for over-diagnosis and overtreatment of malaria and further non compliance with the recommended test-based malaria management policy by prescribers. In Ghana a randomised control open label clinical trial conducted by Ansah et al, (2009) in four districts also confirmed over-diagnosis and overtreatment of malaria. In this study the researchers randomly assigned patients suspected of uncomplicated malar ia into either rapid diagnostic tests in one arm or microscopy or clinical diagnosis in another arm. The results showed that 51.6% of the 1400 patients who tested negative were treated for malaria in the rapid diagnostic test arm in a microscopy setting compared with 55.0% of the 1389 patients in the microscopy arm (adjusted odds ratio 0.87, 95% CI 0.71 to 1.1; P=0.16). Also in the clinical setting, 53.9% of the 1072 patients in the rapid diagnostic test arm who tested negative were treated for malaria compared with 90.1% of the 1090 patients with negative slides in the clinical diagnosis arm (odds ratio 0.12, 95% CI 0.04 to 0.38; P=0.001). Here presumptive treatment of malaria is clearly shown to greatly misdiagnose and over treat malaria. This further portrays a non compliance of prescribers with the universal parasitological confirmation and management of malaria. In Uganda, a study considering compliance of prescribers with ACT prescriptions by Sears et al, (2013) as part of a sentinel site malaria surveillance programme collected and analysed data on patient visits in six health facility over a two-year period after providing training in malaria test-based case management practice. In this study laboratory confirmed malaria cases were analysed in two groups. Patients visit with uncomplicated malaria diagnosis were considered as ACT candidates and those with other forms of malaria who may University of Ghana http://ugspace.ug.edu.gh 11 not have been considered as ACT candidates. Out of 46,265 patients classified as ACT candidates, 94.5% were correctly prescribed ACT. Artemether- lumefantrine prescriptions formed 97.3% of ACT prescribed. Meanwhile ACT candidates who did not receive Artemether- lumefantrine were prescribed artemisinin-naphthoquine (AN, n = 870, 2.0%), dihydroartemisinin-piperaquine (DP, n = 326, 0.7%), and a single patient was prescribed artesunate plus sulphadoxine pyrimethamine (SP). However the findings that young children and woman of childbearing age had higher odds of failure to receive an ACT prescription could suggest severe forms of malaria confirming reports that show higher susceptibility of malaria in those groups (WHO, 2012b). Though the study revealed significant differences across the sites of study in the proportion of patients who were not prescribed ACT, higher rates of compliance with prescribe ACT achieved cannot be overemphasized. Also, the fact that the researchers provided training for prescribers before the studies to achieve these results makes it imperative that higher rates of compliance could be duly expected in settings where training have been provided. In Malawi, Chinkhumba et al (2010) in a cross sectional studies also assessed prescribers compliance with malaria RDT test results after providing a cadres-specific training in malaria RDT test-based case management for health care workers. Their results showed that, prescribers accurately treated positive cases (98%) with anti-malarials but however felt reluctant to refrain from treating 58% of patients who tested negative to RDT with anti-malarials contrary to the training provided. The study however did not indicate the type of anti-malarials prescribed by the prescriber to the patients. In Sudan, a cross sectional cluster sample survey conducted by Abdelgader et al, (2012) in public health facilities also showed that, only 46% of the 1,643 consultations for febrile outpatients were parasitologically tested and 35% of the febrile patients were however both tested and treated according to test results. Among patients who tested positive, 64% University of Ghana http://ugspace.ug.edu.gh 12 were treated with ACT while 24% were treated with artemether monotherapy. Nevertheless, only 17% of those who tested negative were treated for malaria. The study however indicated important gaps in the availability of ACTs, diagnostic capacities and coverage with malaria case management activities as well as a variable readiness among health facilities and health workers to successfully implement the recommended malaria policy. Meanwhile, a follow-up study conducted in Sierra Leone by Gerstl, Dunkley, Mukhtar, Baker and Maikere in 2010 showed that prescribers' positive adherence to ACT treatment is not enough though necessary. This is because their study to measure patients adherence to ACT therapy showed that, of 118 patients, 22.9% did not take one or more tablets prescribed (certainly non-adherent), 28.8% took incorrect doses (probably non- adherent) with 48.3% completing doses correctly (probably adherent). This therefore tells that adherence to anti-malarial ACT therapy should be emphasized for both prescribers and patients. In a similar study conducted by Lawford et al, (2011) in Kenya, the results however showed 64% (588/918) of the 918 patients treated with AL to be probably adherent, 31.7% (291) considered definitely non-adherent and 4.3% (39) considered as probably non-adherent. Significant factors found to predict patients' adherence was strongest for patient knowledge of the ACT dosing regimen (OR= 1.76; 95% CI = 1.32-2.35) before patient age (OR = 1.65; 95% CI = 1.02-1.85). In Nigeria, Meremikwu et al, 2007 analysed uncomplicated malaria patients records and compared prescribing practice between public and private health facilities in Cross River State. The results of the audit showed that prescribers in the public facilities were more likely to document patients' history or physical examinations than their colleague in the private facility. However, prescribers of both facility-types exhibited similar practice and prescribing behaviour where very few of the WHO guidelines were followed. University of Ghana http://ugspace.ug.edu.gh 13 Uzochukwu et al, (2010) in a study conducted in South Eastern Nigeria compared RDT and ACT availability and use between 74 public and private health facilities found that only 31.1% of health facilities used RDTs to diagnose malaria. Public health facilities compared to private ones were found to use RDTs the more with RDT use in urban areas also higher than rural areas. However RDT healthcare provider awareness and health facility availability were 61.1% and 53.3% respectively. Factors accounting for non use of RDT were their unreliability and costs as reported in a different setting by Nanyingi, (2008), issues of supply as well as prescribers' ignorance and preference for other methods of diagnosis. In India a convenience survey showed varied physicians malaria case management practice with reference to method of diagnosis and overall treatment, among state and public or private health facilities. The private sector of healthcare for the treatment of uncomplicated malaria showed the strongest predictor (OR 8.0, 95%CI: 3.8, 17) of artemisinin monotherapy prescription. Also more private sector physicians used RDTs exclusively and prescribed more artemisinins alone and other ACTs compared to those in the public sector (Mishra et al, 2011). A cost benefit analysis from Burkina Faso showed that, it was more expensive to presumptively manage adult malaria in both dry and rainy seasons. Also, in the dry season, the test-based strategy for treating malaria was better for both children and adults than the presumptive strategy. Management of adult malaria with the test-based system was more economical in the dry season than in the rainy season. (Bissofi et al, 2011). Meanwhile, a prospective observational study in India about the cost analysis of prescription patterns of anti-malarials revealed that prescribers have least concern about the cost of the therapy involved in treating malaria (Shantveer Halchar et al, 2012) and hence a need for continued awareness pertaining to rational use of drugs should be ensured as they are also encouraged to adhere to treatment guidelines. University of Ghana http://ugspace.ug.edu.gh 14 2.3 Willingness of caregivers to accept policy and impact of compliance In Ghana, Baiden et al, (2012) conducted a cross sectional study among caregivers in a rural dwelling to determine factors that might influence caregiver's acceptability of RDT based malaria case management and their concern about the denial of ACT treatment on testing to RDT. The results showed that 98% of the caregivers preferred the test-based malaria management over the usual presumptive treatment and were also willing to be denied ACT treatment when their children test negative for malaria (OR 0.57, 95%CI 0.33–0.98). However, caregivers who had valid (adjusted O.R. 1.30, 95% CI 1.07–1.61) or expired (adjusted O.R. 1.38, 95% CI 1.12–1.73) health insurance were reluctant to be denied ACT to their RDT-negative children compared to caregivers who never had a health insurance cover. Meanwhile caregivers' acceptance could be enhanced by their engagement in the procedures of the test while a negative attitude of prescribers could undermine caregivers' acceptance. A similar study using focus group discussions and in-depth interviews conducted by Ezeoke et al, (2012) in urban and rural areas of South Eastern Nigeria showed both public and private health service providers and community members' willingness to test for malaria. They agreed it would distinguish malaria from other illnesses with similar symptoms and hence enable the provision of an appropriate treatment. However, the reliability of negative tests was a major challenge as they believe sometimes conflict with clinical symptoms. The cost of tests and lack of testing facilities were other concerns for acceptability of test-directed treatment as against symptoms-directed treatment of malaria as reported by Uzochukwu et al, (2010) with similar method and setting as well as Nanyingi (2008) in Uganda. The researchers henceforth believe a behavioural change of both providers and patients towards malaria tests would be needed to promote test-directed treatment in patient management. Chuma, Okungu and Melyneux, (2010) in an earlier study in Kenya used multiple data collection methods and found that the interaction of multiple factors related to affordability, University of Ghana http://ugspace.ug.edu.gh 15 acceptability and availability of test diagnostics influenced the test-based management of uncomplicated malaria. The interacting factors regarding acceptability were also found to be prescriber-patient relationship, patient expectations as well as perception of treatment effectiveness as also reported by Baiden et al, (2012) in Ghana, distrust in the quality of care and poor adherence to treatment guidelines. 2.4 Factors influencing prescribers' compliance with test-based policy A study by Nanyingi (2008) in Uganda identified some factors among health workers and laboratory staff preventing their adherence to laboratory diagnoses in malaria case management. Shortage of laboratory staff and training of health workers were the main considerable hindrances towards the adherence. Although health workers admitted the effective and economical role of laboratory diagnosis in malaria control, they were also concerned about possible long waiting time, unreliable results, and finally doubted the ability of a single drop of blood to reveal any possible malaria parasite in the blood. Also treatment of malaria mostly preceded laboratory diagnosis as a special situation in under fives who are more susceptible to malaria. But the question of how committed health workers would adhere to routine laboratory diagnosis and management of malaria in patients five years and above was not addressed. However health workers said routine laboratory diagnosis would only be considered in children and adults with confusing diagnosis. These study findings were observed in a similar study in Nigeria by Uzochukwu et al, (2011) where health workers perceived RDTs to be the most effective among microscopy and clinical diagnosis of malaria. Meanwhile, the health workers prescribed ACTs in 74% of RDT-negative results. The researchers therefore recommended in-depth studies to determine why such health worker behaviour occurs. University of Ghana http://ugspace.ug.edu.gh 16 In Kenya, a cross sectional survey was conducted to assess health facility and health worker readiness to implement their anti-malarial drug policy. The assessment also included factors that influence Arthemether lumefantrine prescription for the treatment of uncomplicated malaria in children under five years of age. The results showed variable readiness of health facility and health workers towards the implementation with 89% of 193 facilities with stocked AL, 55% of 227 health workers had access to guidelines, 46% received in-service training on AL and only 1% of facilities had AL wall charts. Factors found to be associated with better prescribing practice were higher cadre of health workers, in-service training pertaining to the use of AL, positive malaria test , main complaint of fever and high temperature. Though the researchers concluded that changes in clinical practice might take a longer time than anticipated, adherence to new guidelines would be improved by provision of successful interventions which are scaled up to increase coverage ( Zurovac, Njogu, Akhwale, Hamer, and Snow, 2008). In Ghana, Dodoo et al, (2009) in a cohort-event monitoring study conducted showed that, though only 3.2% of 2,831 uncomplicated malaria diagnoses were laboratory confirmed with the highest proportion in the 5-12 years age group, predictors of ACT first line therapy prescription by physicians were laboratory-confirmed diagnosis (adjusted OR 9.7 [5.2–18.2]), age 5 years and above, and attending a government facility as against a private facility which is contrary to the report of Misra et al, (2011) in India. Also patients above 12 years we re less likely to be co-prescribed antibiotics than patients under 5 years though analgesics and antibiotics were the most co-prescribed per patient. Factors that may influence prescribers' adherence to the implementation of the malaria test-based and effective use of ACT were found to be patients' age, diagnostic confirmation (Zurovac, Njogu, Akhwale, Hamer, and Snow, 2008) and concurrent conditions suspected by prescriber during consultations. University of Ghana http://ugspace.ug.edu.gh 17 Meanwhile, a cross sectional survey conducted in Tanzania by Mubi et al, (2013) in government primary healthcare facilities showed that, malaria confirmation for only 30% of patients that were tested (31/105) among 168 fever patients seen at health facilities with available diagnostics. Anti-malarial prescriptions were given to all positive-test patients, 14% of negative-test patients and 28% of presumptively diagnosed patients for malaria. Patients with negative results compared to those with positive results and no-tested patients compared to those tested were more likely to be prescribed antibiotics. Factors that influenced adherence to test results were RDT stock outs, staff shortage and health worker perceptions as reported by Nanyingi (2008) in a different setting. Selemani et al, (2013) in a repeated cross sectional health facility survey in rural Tanzania also assessed health worker factors associated with correct prescription for uncomplicated malaria. Their findings showed that health workers with three or more years of working experience had significantly higher odds of prescribing correctly (aOR 2.9; 95%CI 1.2-7.1; p =0.019) than others and health worker cadre specifically Clinical officers (aOR 2.2; 95% CI 1.1-4.5; p = 0.037), and nurse aide or lower cadre (aOR 3.1; 95% CI 1.3- 7.1; p = 0.009) were also more likely to prescribe ACT correctly than medical officers. Meanwhile, their study revealed that training on ACT use, supervision visits, and availability of job aids were not significantly associated with prescribing correctly whilst health worker age less than or equal to 35 years could not determine any association to prescribing correctly 2.5 Prescriber's adherence to Guidelines Cabana et al, (1999) identified that, clinical guidelines had limited effect in changing the behaviour of physicians who were then the main prescribers. They therefore stated a problem that showed how little knowledge exists about the process and factors involved in changing physician practices pertaining to their response to guidelines. The researchers University of Ghana http://ugspace.ug.edu.gh 18 therefore systematically reviewed literature to identify barriers to physician adherence to clinical practice and developed a framework for improvement. The results of their search found that 293 questions about barriers included seven categories of barriers after grouping them into common themes. They found that, these barriers affected the physician in three main ways. These are physician's knowledge (which involved lack of knowledge or lack of familiarity of guideline), physician's attitude (involving lack of agreement, lack o f self- efficacy, lack of outcome expectancy, or the inertia of previous practice) and physician's behaviour also involving external barriers. The external factors could be described as patient, guideline recommendation as well environmental factors. Though these findings could not be generalized since the barriers varied in different settings, it still offered a rational approach toward improving guideline adherence and a framework for future research. Mckinlay, Lin, Freund and Moskowitz (2002) in an experimental study of an unexpected influence of a physician attributes in medical decision showed that, patients' attributes and physician attributes independently did not influence physicians' actions. Meanwhile younger physicians ordered more test for a polymyalgia rheumatica (PMR) diagnosis and for younger-older patients than older physicians who equivalently diagnose the condition in elderly patients and also order few tests. More so, younger physicians' diagnosed depression in more male patients as older physicians diagnosed more females for depression. This study showed that, indeed age of prescriber together with age and sex of patients has an impact on physicians' actions and therefore both physician and patient attributes interactively rather influence medical decisions of prescribers. In 2009, Lugtenberg, Schaick, Westert and Burgers analysed barriers among Dutch general practitioners to determine why physicians don't adhere to guideline recommendations during their practice. This qualitative study conducted six focused group discussions using 30 General Practitioners with an average of seven per session. Factors that prevented physicians University of Ghana http://ugspace.ug.edu.gh 19 from complying with key recommendations in clinical guidelines for 56 key recommendations were discussed separately for various groups of physicians and sessions involved. The results of the study showed that, the barriers or factors varied greatly within guidelines with the most perceived barriers being lack of agreement with the recommendations because of lack of applicability or evidence (68%), environmental factors such as organisational constraints (52%), lack of knowledge regarding the guideline recommendations (46%), and guideline factors such as unclear or ambiguous guideline recommendations (43%). Among the three levels of barriers developed into a framework by Cabana et al (1999) the researchers found that barriers related to the attitude of physicians occurred for 91% of the key recommendations followed by prescriber behaviour factors (82%) and lastly knowledge- related barriers of prescribers also perceived for 46% of the key recommendations. Though they concluded that the barriers varied greatly among recommendations, it is however imperative that issues of guideline recommendations bordering on awareness or familiarity, agreement, prescriber self-efficacy, expectation of the outcome, inertia of previous practice or sufficient motivation, patient factors, guideline recommendation factors and environmental factors together or independently influence a prescriber's ability to comply with recommendations of clinical guidelines. These should therefore be highly considered during implementation of strategies focussing on key recommendations of clinical guidelines to improve compliance in their practice. University of Ghana http://ugspace.ug.edu.gh 20 CHAPTER THREE 3.0 METHOD 3.1 Study Design A cross sectional health facility survey was conducted in all four sub municipals of the municipality and obtained both qualitative and quantitative data from prescribers, patients, and health facilities to assess factors that influence prescribers' compliance with malaria test-based case management policy in the municipality. 3.2 Study Location 3.2.1 Demography The study was conducted in the Effutu Municipality. The Effutu Municipal area is in the Central Region as one of the 20 districts, municipalities and metropolis in the region. It is situated between latitudes 50 20' N and longitudes 00 25 W and 00 37 W on the eastern part of the Central Region Ghana. It is winded and sandwiched by Greater Accra Region and the districts - Ga Rural, Agona and Gomoa. It is bordered to the north by Agona Municipal, north-east by the West Akim Municipal, to the south by the Gulf of Guinea, to the east by Gomoa East District and Ga West Municipal, and on the west by the Gomoa West District. The Municipal covers an area of 417.3 square kilometers (163 sq miles) with a population of 68,597 according to Population and Housing Census of 2010 giving a population density of 164 persons per square kilometer. The municipal population forms 3.1% of the regional population and consists of 32,795 males and 35,802 females giving a sex ratio of 1:1.1 (Ghana Statistical Service, 2012). The 2013 projected population for the municipality was 75,176 (male-35940 and female-39239) and that of children under five, five years and above and women in their reproductive age are 15,035, 60,141 and 17,290 University of Ghana http://ugspace.ug.edu.gh 21 respectively. The Effutu Municipal is sub divided into four sub municipals namely Ansaful, Essuekyir-Gyahadze, Winneba East and Winneba West. 3.2.2 Environmental Factors The Effutu Municipal lies within the west semi-equatorial region along the Coast of the Gulf Of Guinea and it experiences two rainfall patterns: the major rainy season starts from April and ends in July and minor rainy season from September to November. The municipal also experiences two wind systems - the South-West Monsoon winds (Rainfall wind) and North-East Trade Winds (Harmattan Wind). The annual rainfall figures of the municipal are quite low ranging from 400mm to 500mm along the coast but are higher in the hinterlands with the mean annual rainfall ranging between 500mm and 700mm. The mean annual minimum and maximum temperature are 220C and 280C respectively. The vegetation of the municipality is made up of semi-deciduous forest (forming 70% of the municipality) and coastal savannah grassland. 3.2.3 Health Services The municipal has thirteen health facilities including four functional Community- based Health and Planning Services (CHPS) zones, one health centre, one clinic, five hospitals of the district level category including a specialist hospital, one public health specialty unit and one maternity home. The municipal and Specialist hospital serve as the referral point for all the other health facilities. The health service in the municipality is provided by both private (3) and public facilities (made of Government (8), quasi- government (1) and CHAG (1)). Pharmacy Shops and licensed chemical sellers play a supplementary role as well traditional medical practitioners predominated by Traditional Birth Attendants commonly called TBAs. The levels of health service delivery are as follows; University of Ghana http://ugspace.ug.edu.gh 22  Hospital  Health centres / clinics  Reproductive and Child Health centres  Community clinics (CHPS compounds). Malaria continues to top the list of top ten causes of OPD attendance in the district, contributing 13,084 and 37,397 cases in 2011 and 2012 respectively. 3.3 Variables The main dependent variable was the prescriber's compliance with the test-based policy. The varied independent variables of interest were as follows: 1. Patient variables  Age of patient  Sex of patient  Occupation  Patient symptoms  NHIS status 2. Prescriber Variables  Age of prescriber  Sex of prescriber  Professional Category/ Cadre of health worker  Rank  Years of service  Awareness of test policy  Training about policy University of Ghana http://ugspace.ug.edu.gh 23 3. Health Facility  Type of facility  Facility accreditation to NHIS  Availability of functional laboratory  Number of qualified microscopists  Availability of logistics for testing  Availability of guidelines and protocol for malaria treatment 3.4 Sampling 3.4.1 Study Population The study population included health facilities, prescribers involved in outpatient malaria case management and patients diagnosed of uncomplicated malaria by health facilities in the Effutu municipality. 3.4.2 Sampling Size All four sub-districts in the Effutu Municipality were selected for the study. All the 13 health facilities in the four sub-districts were selected for the study but nine out of the thirteen health facilities participated in the study while one clinic declined consent. Two CHPS zones could not give treatment but only involved in Child welfare clinics and home visits and therefore did not diagnose and treat uncomplicated malaria. The other health facility was a Specialty Public health unit which attends to skin diseases and hence did not diagnose and treat uncomplicated malaria. Twenty-five (25) prescribers performing general outpatient consultation encountered on the days of visit within the health facilities were included in the study. University of Ghana http://ugspace.ug.edu.gh 24 A sample of seven patients who were diagnosed of uncomplicated malaria and within the inclusion criteria were selected as they are diagnosed for each prescriber. A total of 175 patients were therefore included in the study. 3.4.3 Sampling method The Effutu municipality was conveniently selected and all four sub-municipals were included in the study. The sub-municipals were Ansaful, Essuekyir-Gyahadze, Winneba East and Winneba West. Each sub municipal had at least one public health facility. 3.4.3.1 Health Facilities All the categories of health facilities in the sub-municipals were included in the study. These included the municipal hospital, health centre, CHPS compounds, a clinic/maternity home. This was to ensure that all facilities providing different levels of care and with different case loads and various cadres of prescribers contributing to total cases in the municipality were fairly represented and obtained in their maximum numbers. 3.4.3.2 Prescribers and Patients At each health facility, the primary sampling unit for the compliance assessment part of the study was a prescriber-patient diagnoses and treatment for uncomplicated malaria. Prescribers who performed general outpatient consultations were eligible for inclusion. Eligible prescribers who consented to participate in the study were there fore included. Any patient five years and above diagnosed of uncomplicated malaria by a prescriber and has not been diagnosed of malaria in the same facility two weeks before the study was eligible for inclusion. The first seven sampling units that met the inclusion criteria were selected at each health facility. This was based on the assumption that the order in which patients arrived at a University of Ghana http://ugspace.ug.edu.gh 25 health facility was not associated with prescribers' compliance with a medical policy. The selection of seven diagnoses and treatment for the compliance assessment was based on the Lots Quality Assurance Sampling (LQAS) technique which is based on binomial probabilities. The LQAS assumes that prescribers should comply with medical policies correctly at least 95 percent of the time. Using the binomial equation, at least six diagnoses and treatments per prescriber are needed to determine this compliance level. We assessed seven outpatient uncomplicated malaria diagnoses and treatments per prescriber. In cases where more than one child presented by a caretaker were diagnosed and treated for uncomplicated malaria, one child was selected at random for inclusion. The other children were excluded from the study and the next patient who met the inclusion criteria was sampled. Patients less than five years old, malaria diagnosed in-patients, patients diagnosed and treated for malaria within two weeks in the same facility and pregnant women attending the facility for antenatal care were excluded from the study. Patients less than five years were excluded to avoid conflicts of under five childhood diseases management and the malaria test-based management policy to ensure better findings. In-patients were also excluded because laboratory investigation could occur more likely as part o f the case admission management and not as a result of test-based management policy compliance. Patients previously diagnosed and treated for malaria within two weeks before the study were excluded to ensure the findings better represent initial malaria test-based case management practice. Pregnant women attending the health facility for antenatal care were also excluded to avoid excuse of low birth weight and maternal mortality reduction goal. Patients who met the inclusion criteria were selected as they came to collect drugs prescribed from the dispensary. All the prescribers performing outpatients' consultations together with malaria University of Ghana http://ugspace.ug.edu.gh 26 diagnoses and treatments were included in the study for assessment of their compliance with test-based policy and interview. 3.4.5 Data Collection Techniques and tools Three methods were used to collect the data. These included 1. Assessment of diagnoses and treatments with a patient interview questionnaire. A structured questionnaire was used to interview patients for their preference of the policy and to capture the procedures used by the prescriber to diagnose and treat them for uncomplicated malaria. 2. Assessment of the health facilities using a Health Facility Audit Questionnaire: General issues relevant to malaria test-based case management such as staffing, availability of functional laboratory, logistics and equipment as well as standard guidelines were assessed. 3. Interview of Prescribers using a Prescriber Interview Questionnaire: Prescribers who performed general outpatient consultations with malaria diagnoses and treatments were interviewed with the questionnaire to assess their characteristics, knowledge and awareness of policy as well as training on the use of RDTs and work experience. University of Ghana http://ugspace.ug.edu.gh 27 3.4.6 Ethical Clearance Ethical approval was sought from the Ghana Health Service (GHS) ethical review committee. Permission was also sought from the Effutu Municipal Health Directorate and the Facility Heads or In-Charges of the health facilities involved in the study. Informed consent was obtained from prescribers, patients and their caregivers and confidentiality assured before the study. They were duly informed about the purpose, procedures, risks and benefits of participating in the study. Study participants who could not read had the consent form read and explained to them before an impartia l witness. Study participants who agreed to participate in the study were required to sign or thumbprint (left thumb) the consent form to show their consent. Assents for children participants were obtained from their caregivers on their behalf. Confidentiality of information was ensured before and after the study and used only for the purpose indicated in the study. The information was collected privately and securely stored without names or any traceable identity to the participants and in a file accessible only to the research team. A code linked to a particular name was kept confidential. The assessment of compliance and the interviewing of the participants involved were done solely by trained professionals including the principal investigator. Only health facilities, prescribers and patients who agreed were assessed and interviewed. There was no risk to be experienced by the participants for participating in the study. The participants were however informed of possible minor discomfort for being interviewed or in answering of some questions which they could choose not to answer. Voluntary participation of the study was explained to participants and informed of their liberty to withdraw consent and discontinue participation at any time without attracting any penalty. Participants were not coerced to partake in the study and there was no direct benefit or University of Ghana http://ugspace.ug.edu.gh 28 monetary compensation for participation besides that for the Municipal Health Management Team and the Ghana Health Service for planning health delivery services. 3.4.7 Training of Interviewers The field workers to assist in the study were trained three days prior to the start of data collection. They were trained to understand the purpose of the study, data collection techniques and tools and their appropriate interpretation. Simulated exercises were repeated to ensure an increased agreement and consistency among the field workers and between them and the trainer. 3.4.8 Pre-testing and review of data collection tools The data collection tools were pre-tested at the Swedru Government hospital in the Agona West district which has similar settings as that in the Effutu municipality and reflected the local conditions. The pre-test provided the clear and understandable questions required for the appropriate response and the necessary modifications made based on the pre-test. 3.5 Data Collection We conducted the survey from December 25, 2013 to January 31st 2014. All the thirteen health facilities in the various sub municipals within the Municipality were visited. However, one Clinic, University Clinic located in the Winneba west sub municipal declined to participate in the study. Three other health facilities, the Public Health Specialty Unit, Ansaful CHPS zone and the Essuekyir CHPS zone located in the Winneba East, Ansaful and Essuekyir-Gyahadze sub municipals respectively were excluded from the study. The Public health specialty unit attended to skin diseases only whereas the Ansaful CHPS zone and the Essuekyir CHPS zone were not operating in full capacity. They were not to diagnose nor give University of Ghana http://ugspace.ug.edu.gh 29 medications but only limited to carryout Child Welfare Clinics and home visits. They therefore referred suspected cases to the nearest fully operating facilities as at the time of the study. The team of trained professionals visited the remaining nine facilities in the mornings of selected days of the week. The team introduced its members to the health facility in- charges and unit heads involved and sought permission to begin the study. A written permission was sought from the facility in-charges. The written permission was approved and a representative from the facility introduced the team members to the various departmental heads and staff involved in the study. All the 25 prescribers who diagnosed and treat uncomplicated malaria in the various health facilities were eligible for their compliance with the test-based policy to be assessed. Their consent was sought and enrolled in the study for the assessment. Patients diagnosed of uncomplicated malaria by enrolled prescribers were recruited at the dispensary based on the inclusion criteria and their willingness to participate in the study. Patients and Caregivers who consented to participate in the study were recruited on arrival to the dispensary. They were given identification numbers and interviewed as they wait for their turn to take their medication or after they receive their medication before they leave the health facility. The independent trained professional team assessed and recorded the diagnoses and treatment given by the prescriber without interfering with the case management practice of the prescriber. This procedure was repeated for six other patients per prescr iber enrolled into the study. The prescribers were interviewed after collecting information from all corresponding seven patients and facility in-charges at their free time or after the entire consultations. Health facility in-charges and unit heads who could provide adequate information on key areas relevant to the study were interviewed at their offices of work after obtaining patient information. This was done to assess the health facility for facilities and logistics pertinent to malaria test-based case management using a Facility Audit Questionnaire. University of Ghana http://ugspace.ug.edu.gh 30 3.6 Data Quality Control Field workers who were health personnel and experienced in the health sector were used to ensure quality data collection. They were trained and evaluated to ascertain their ability to provide accurate assessments and document the true responses of the study participants. The principal investigator supervised field workers during the data collection and cross checked from other health records of participants in the facility for correctness. The principal investigator ensured double entry of data into Epi-Info software version 7.1.2.0 by independent data entry personnel. All discrepancies were resolved by reference to original data collection tools. Data validation programs introduced during the development of the data entry template was used to ensure validation of data during the data entry process. 3.7 Data Processing and Analysis 3.7.1 Data Processing Double entry of data was done by two data entry personnel independently into Epi- Info software version 7.1.2.0. Data cleaning and verification was done to ensure good quality and accurate data. For each variable entered, the frequencies were run to identify number of missing variables and incorrectly entered data. Analyses were done on variables on which data was entered. When necessary, missing data was excluded from analyses whilst correction of incorrectly entered data was done by listing the variables with the "allow update" function enabled and the necessary corrections made. The data was properly stored and backed up on an external hard drive. University of Ghana http://ugspace.ug.edu.gh 31 3.7.2 Data Analysis Univariate and bivariate analysis were performed and used as descriptive and inferential statistics respectively. Odds ratios were used to determine associations among variables. Confidence intervals were used to determine statistical significance. Compliance of prescribers was determined by assessing prescriber uncomplicated malaria case management for seven patients. A prescriber was said to have complied to malaria policy if at least six of the patients seen were tested to confirm malaria diagnosis and treated according to the test results. Also a patient was said to be tested if a prescriber requested a malaria test to be done by RDT, microscopy or both on suspicion of uncomplicated malaria and the results obtained and returned to the prescriber before any treatment is prescribed. A patient was also said to be treated according to test result if only positive malaria test results were prescribed with anti-malarial drugs. Univariate analyses of categorical data were expressed as frequencies, proportions and percentages. Cross tabulations and frequencies were used to analyse data descriptively. Appropriate measures of central tendency and dispersion were employed. Bivariate analyses were done using unadjusted odds ratio and their corresponding 95% confidence interval to assess the association between some patients, prescriber, health facility independent variables and testing before treatment as well as treating according to test results. The unadjusted odds ratios were calculated for each predictor variable. Multivariate analysis was not done because most of the associations obtained from the unadjusted odds ratio were statistically insignificant. University of Ghana http://ugspace.ug.edu.gh 32 CHAPTER FOUR 4.0 RESULTS 4.1 Characteristics of Study Population 4.1.1 Health facility characteristics Of the nine health facilities, five (55.6%) were hospitals, one (11.1%) clinic/ maternity home, one (11.1%) health centre and two (22.2%) CHPS compounds. One (11.1%) of the health facilities was located in the Nsuekyir-Gyangyanadze sub-municipal and two (22.2%) in the Ansaful sub-municipal. The Winneba East and Winneba West sub municipals had three (33.3%) facilities each participating in the study. None of the sub municipals had all the different categories of health facilities participating in the study ( table 1). Table 1. Types of health facilities by sub municipal, Effutu municipal, 2014. Sub Municipal Hospital Clinic Health Centre CHPS Maternity Home TOTAL Ansaful 2 0 0 0 0 2 Nsuekyir- Gyangyanadze 0 0 0 1 0 1 Winneba East 2 0 0 1 0 3 Winneba West 1 0 1 0 1 3 TOTAL 5(55.6%) 0 1(11.1%) 2(22.2%) 1(11.1%) 9(100) Five (55.6%) of the health facilities were government facilities. The mission/ religious and private-for-profit operating authorities had one (11.1%) and three (33.3%) facilities respectively (table 2). University of Ghana http://ugspace.ug.edu.gh 33 Table 2. Types of health facilities by operating authorities, Effutu municipal, 2014. Hospital Clinic Health Centre CHPS Maternity Home TOTAL Operating Authority Government 2 0 1 2 0 5 Quasi- government 0 0 0 0 0 0 Mission / Religious 1 0 0 0 0 1 Private for profit 2 0 0 0 1 3 TOTAL 5 0 1 2 1 9 All the nine health facilities accepted NHIS from patients. Meanwhile, one (11.1%) CHPS compound, Zongo CHPS was not independently accredited by NHIS to use its facility name for NHIS claims and therefore operated its NHIS under another health facility, Winneba Health Centre. About 33% (3/9) of the health facilities in the municipality (Winneba Municipal Hospital, Klimovic Memorial Hospital and Otoo Memorial Hospital) accept other heath insurances. From the health assessment, all the nine health facilities had the capacity to test for malaria. Seven (77.8%) health facilities had functional laboratories. Two (22.2%) facilities, Gyangyanadze CHPS and Zongo CHPS compounds did not have functional laboratories but used malaria RDTs to test for malaria parasites and had RDTs available within the last six and three months. Of the facilities with functional laboratories, there was an average of three (3) microscopists per health facility. All the twenty four (24) microscopists had undergone training in malaria parasite microscopy. All the nine facilities had on average three (3) staff University of Ghana http://ugspace.ug.edu.gh 34 trained in the use of the RDT to diagnose malaria and also had the capacity to test for malaria parasites. Seven (77.8%) of the facilities confirm malaria using microscopy, eight (88.8%) of the facilities confirmed malaria with RDT whilst 66.7% confirm malaria using both microscopy and RDT. One (11.1%) facility confirms malaria with microscopy only and two (22.2%) by RDT only. All the seven (77.8%) facilities which confirmed malaria by microscopy also used Giemsa stain reagent to prepare blood films for malaria parasite detection. The facilities also had distilled water available for dilution of the Giemsa stock. Microscopists in seven of the facilities have been trained in malaria parasite microscopy. Seven of the facilities have had staff trained in malaria case management within the last six months. All nine health facilities had staff trained in the use of RDT to diagnose malaria. The 2010 standard treatment guidelines and 2010 anti-malarial drug policy was available in seven and five health facilities respectively (Table 3). Table 3. Facility level indicators for malaria test-based case management, Effutu municipal, 2014. Indicator Number of Facilities Percentage functional lab f functional microscope staining reagent in last 3 months staining reagent in last 6 months Availability of Microscopist Training of microscopists in malaria parasite microscopy RDTs in last 3 months RDTs in last 6 months 2010 standard Treatment guidelines 2010 Anti-malarial drug policy Training of staff in the use of RDT to diagnose malaria Training of staff in case management within last six months 7 7 7 7 7 7 7 6 7 5 9 7 77.8 77.8 77.8 77.8 77.8 77.8 77.8 66.7 77.8 55.6 100 77.8 n=9 Only one (11.1%) health facility, Otoo Memorial Hospital had no RDTs available within the last six months from the study. However, all nine health facilities had RDTs University of Ghana http://ugspace.ug.edu.gh 35 available within the last three months from the study. Of the facilities with functional laboratories, a functional microscope was available for malaria parasite microscopy. Otoo Memorial Hospital was the health facility with most prescribers (26.9%) that participated in the study followed by Winneba Health Centre (23.1%) which also tested the most of patients tested (57.1%). The facility that treated the most (100%) of cases tested according to test result was Gyangyanadze CHP zone as shown in Fig 2 below. Figure 2: Proportion of prescribers, patients tested and treated according to test results by health facility, Effutu Municipality, 2014. 0 20 40 60 80 100 120 P e r c e n ta g e s Health Facility Proportion of Prescribers Proportion of Patients Test proportion treated accordingly proportion 4.1.2 Prescriber Characteristics All the 25 prescribers consented to participate in the study. Of the 25 prescribers, seven (28%) were medical officers, ten (40%) were physician assistants and the rest, various categories of nurses. Thirteen (52%) out of the 25 prescribers were females. The medical University of Ghana http://ugspace.ug.edu.gh 36 officers and physician assistants comprised of 57% and 40% females respectively. The youngest age was 24 years whilst the oldest was 76 years with a median age of 29 years (inter-quartile range 27.5 - 44.0). Seven (28%) of the prescribers did not have any training in malaria test-based case management and of the twenty (80%) who were aware of the malaria test-based case management policy, sixteen (64%) had training in malaria test-based case management policy. Five (20%) of the prescribers feel they do not need any additional training to manage malaria better and 3 (12%) were unwilling to comply with the malaria test-based case management policy as shown in the table below: Table 4. Characteristics of prescribers by sub-municipal, Effutu Municipal, 2014. Prescriber Characteristics Sub Municipal Overall Winneba East n (%) Winneba West n (%) Ansaful n (%) Nsuekyir Gyangyanadze n (%) n (%) Sex Female 4 (57.1) 5 (38.5) 3 (75) 1 (100) 13 (52) Male 3 (42.9) 8 (61.5) 1 (250 0 (0) 12 (48) Trained Ever 5 (71.4) 10 (76.9) 1 (25) 1 (100) 17 (68) Never 2 (28.6) 2 (15.4) 3 (75) 0 (0) 7 (28) Missing 0 (0) 1 (7.7) 0 (0) 0 (0) 1 (4) Awareness of Policy Yes 4 (57.1) 12 (92.3) 3 (75) 1 (100) 20 (80) No 3 (42.9) 0 (0) 1 (25) 0 (0) 4 (16) Missing 0 (0) 1 (7.7) 0 (0) 0 (0) 1 (4) Willingness to comply Yes 7 (100) 10 (76.9) 4 (100) 1 (100) 22 (88) No 0 (0) 3 (23.1) 0 (0) 0 (0) 3 (12) Need Additional Training Yes 6 (85.7) 10 (76.9) 3 (75) 1 (100) 20 (80) No 1 (14.3) 3 (23.1) 1 (25) 0 (0) 5 (20) TOTAL 7 (100) 13 (100) 4 (100) 1 (100) 25(100) University of Ghana http://ugspace.ug.edu.gh 37 Eighteen (72%) of the prescribers said they routinely request malaria test for patients five years and above suspected of uncomplicated malaria before prescribing treatment while seventeen (68%) said they routinely treat malaria cases according to test results in patients five years and above (Table 5). Table 5. Responses of Prescribers by Cadre, Effutu Municipal, 2014. Prescriber Responses Cadre of Prescriber Overall Medical Doctors n (%) Physician Assistants n (%) Nurses & Midwives n (%) Community Health Nurses n (%) n (%) Sex Female 4 (57.1) 4 (40) 3 (60) 2 (66.7) 13 (52) Male 3 (42.9) 6 (60) 2 (40) 1 (33.3) 12 (48) Routinely test cases Yes 4 (57.1) 8 (80) 4 (80) 2 (66.7) 18 (72) No 3 (42.9) 2 (20) 1 (20) 1 (33.3) 7 (28) Routinely prescribe according to test results Yes 5 (71.4) 8 (80) 3 (60) 1 (33.3) 17 (68) No 2 (28.6) 2 (20) 2 (40) 2 (66.7) 8 (32) TOTAL 7 (100) 10 (100) 5 (100) 3 (100) 25(100) 4.1.3 Patient Characteristics Out of the 175 patients interviewed, 125 (71.4%) were females and 41 (23.4%) were aged twelve years and below. The median age of the patients was 28 years (inter-quartile range 14.5 - 48.0). Most 81 (46.3%) adult patients were the age group 13 - 45 years old. The records for paediatric patients from 5 - 12 years old were 41 (23.4%). There were two (1.1%) patient records (one each from the hospital and health centre category) for which age of patient was not provided. The age and sex distribution of the patients interviewed by the type of health facility in which they were seen are shown in table 6. University of Ghana http://ugspace.ug.edu.gh 38 Table 6. Distribution of patients interviewed by age, sex and by type of health facility, Effutu Municipal, 2014. Patient Characteristics Type of Health Facility Overall Hospital n (%) Health Centre n (%) Maternity Home n (%) CHPS n (%) n (%) Sex Female 82 (73.2) 22 (62.9) 7 (100) 14 (66.7) 125 (71.4) Male 30 (26.8) 13 (37.1) 0 (0) 7 (33.3) 50 (28.6) Age 5 -12 23 (20.5) 9 (25.7) 1 (14.3) 8 (38.1) 41 (23.4) 13 - 45 55 (49.1) 13 (37.1) 4 (57.1) 9 (42.9) 81 (46.3) 46 - 60 13 (11.6) 9 (25.7) 1 (14.3) 2 (9.5) 25 (14.3) >60 20 (17.9) 3 (8.6) 1 (14.3) 2 (9.5 ) 26 (14.9) Missing 1 (0.9) 1 (2.9) 0 (0) 0 (0) 2 (1.1) TOTAL 112 (100) 35 (100) 7 (100) 21 (100) 175 (100) 4.1.3.1 Patient Occupation Majority 62(35.8%) of the patients interviewed were students and pupils followed by traders 31(17.9%). Patients who were on retirement were 13 (7.5%) and very few 4 (2.3%) were temporarily unemployed. There were two (1.1%) patients whose interview could not capture their occupation. These were excluded from the analysis in Fig 3. University of Ghana http://ugspace.ug.edu.gh 39 Figure 3: Distribution of patients interviewed by occupation, Effutu Municipal, 2014. 0 5 10 15 20 25 30 35 40 P e r c e n ta g e Occupation 4.1.3.2 Patients' preference to test-based management of malaria More patients, ninety one (52.3%) preferred the prescribers' decision (As prescriber decides) to manage them upon diagnosis of malaria whiles 67 (38%) preferred (Yes) to be tested and treated according to test results. One of the patient's preferences could not be captured and was therefore not used in the analysis. The figure 4 shows the proportion of patients' preference to comply with malaria test-based case management policy. University of Ghana http://ugspace.ug.edu.gh 40 Figure 4: Proportion of patient's preference to comply with malaria test-based case management policy, Effutu Municipal, 2014. 0 10 20 30 40 50 60 Yes No As prescriber decides P e r c e n ta g e Patients preference More males 42% (21/50) preferred (Yes) to be tested and treated according to test results upon malaria diagnosis compared to females 36.8%(46/125). Of the patients who preferred the prescriber's decision (As prescriber decides) in managing them for uncomplicated malaria, males 54% (27/50) dominated. On the other hand, more females did not (11.3%) preferred(No) to be tested and treated according to test results than males as shown in Fig 5. University of Ghana http://ugspace.ug.edu.gh 41 Figure 5: Proportion of patient's preference to comply with malaria test-based case management policy by sex, Effutu Municipal, 2014. 0 10 20 30 40 50 60 yes No As prescriber decides P e r c e n ta g e Patient's preference Male Female 4.2 Assessment of the level of compliance by prescribers Of the sampled seven uncomplicated malaria cases diagnosed per prescriber, the highest number tested by a prescriber before prescribing appropriate treatment was 6 (85.7%). The tested cases that were mostly treated according to test results were 15 (42.9%). None of the twenty five (25) prescribers appropriately tested and treated at least six of the sampled cases to attain compliance. The compliance level of the prescribers with the malaria test-based policy was therefore zero (Table 7). University of Ghana http://ugspace.ug.edu.gh 42 Table 7: Performance of Prescriber towards compliance with malaria test-based case management policy, Effutu Municipality, 2014. Health Facility Prescribers Code Malaria Cases n (%) Cases tested n (%) Cases tested positive n (%) Cases treated according to test results n (%) Prescriber compliance (Yes-- ≥6/7 No--- <6/7) Winneba Health Centre A 7 (100) 1 (14.3) 1 (100) 1 (100) No B 7 (100) 2 (28.6) 2 (100) 2 (100) No C 7 (100) 3 (42.9) 3 (100) 3 (100) No D 7 (100) 2 (28.6) 2 (100) 2 (100) No E 7 (100) 0 (0) 0 (0) 0 (0) No Municipal Hospital A 7 (100) 2 (28.6) 1 (50) 1 (50) No B 7 (100) 0 (0) 0 (0) 0 (0) No Zongo CHPS A 7 (100) 1 (14.3) 0 (0) 0 (0) No B 7 (100) 1 (14.3) 0 (0) 0 (0) No Otoo Memorial Hospital A 7 (100) 1 (14.3) 0 (0) 0 (0) No B 7 (100) 1 (14.3) 0 (0) 0 (0) No C 7 (100) 1 (14.3) 0 (0) 0 (0) No D 7 (100) 1 (14.3) 0 (0) 0 (0) No E 7 (100) 0 (0) 0 (0) 0 (0) No F 7 (100) 0 (0) 0 (0) 0 (0) No G 7 (100) 2 (28.6) 1 (50) 1 (50) No Klimovic Memorial Hospital A 7 (100) 0 (0) 0 (0) 0 (0) No B 7 (100) 1(14.3) 1 (100) 1 (100) No Coast For Christ Baptist Hospital A 7 (100) 6 (85.7) 0 (0) 0 (0) No B 7 (100) 2 (28.6) 1 (50) 1 (50) No C 7 (100) 1 (14.3) 0 (0) 0 (0) No Trauma and Specialist Hospital A 7 (100) 3 (42.9) 2 (66.7) 2 (66.7) No B 7 (100) 1 (14.3) 0 (0) 0 (0) No Gyangyanadze CHPS A 7 (100) 1 (14.3) 1 (100) 1 (100) No Bethel Maternity Home A 7 (100) 2 (28.6) 0 (0) 0 (0) No TOTAL 175(100) 35(20) 15(42.9) 15(42.9) 25 University of Ghana http://ugspace.ug.edu.gh 43 4.2.1 Prescriber prescription practice Of the 175 patients sampled for the compliance assessment, only 15 (8.5%) were tested and treated appropriately with anti-malarials. Presumptive treatment was given to majority 140 (80%) of the patients and twenty (11.4%) patients who tested negative were also given anti-malarials regardless of the test results. Of six (66.7%) facilities which appropriately prescribed ACTs for malaria treatment, Artesunate Amodiaquine AA was less prescribed being prescribed only in two (33.3%) facilities as shown in Fig 6 below. Figure 6: Proportion of suspected malaria cases tested and appropriately prescribed ACT by health facilities, Effutu Municipal, 2014. 0 5 10 15 20 25 30 35 40 45 P e r c e n ta g e s Health Facilities Patient Tested Tested Positive Appropriately prescribed ACT Appropriately prescribed AA The responses of prescribers concerning preference for AA for treatment of malaria showed that eight (32%) rarely prescribe AA for malaria treatment whilst 15 (60%) sometimes did. Moreover, eight (32%) never prescribed AA for malaria treatment. Most 17 University of Ghana http://ugspace.ug.edu.gh 44 (68%) prescribers prescribed other anti-malarials than AA for fear of adverse reactions. Four (16%) prescribers mentioned patient preference as a reason for prescribing other anti- malarials than AA (Table 8). Table 8: Reasons for prescriber preferences to anti-malarials other than AA by sex, Effutu Municipality, 2014. Reasons why prescribers sometimes prescribe other anti-malarials than AA Sex of Prescribers Total Male n (%) Female n (%) n (%) Fear of Adverse Reaction Prescriber Personal Choice Stock out of drug Lack of confidence in drug Patient's Preference Clinical decision 9 (36) 0 (0) 5 (20) 4 (16) 3 (12) 0 (0) 8 (32) 1 (4) 3 (12) 1(4) 1 (4) 1 (4) 17 (68) 1 (4) 8 (32) 5(20) 4 (16) 1 (4) n = 25 4.3 Factors associated with testing suspected malaria cases before treatment. 4.3.1 Patient Factors At bivariate analysis, patient factors associated with being tested for malaria parasites before treatment were: patient aged 13 - 45 years (OR=1.26, 95%CI = 0.50 - 3.20), when compared with ages 5-12 years, measured temperature of 37.5 o C or more (OR = 2.40, 95%CI = 0.66- 8.76) and patient with NHIS status (OR = 3.54, 95%CI = 0.44 - 27.99) as shown in table 9. Meanwhile, the odds of being tested before treatment were the same in both male and female patients (OR = 1.00, 95%CI = 0.44 - 2.27). Patient occupation as students had higher odds of testing than government workers (OR= 1.26, 95%CI= 0.32 - 5.07). Patient symptoms that University of Ghana http://ugspace.ug.edu.gh 45 were associated with being tested before treatment included Loss of appetite (OR = 1.26 95%CI = 0.44 - 3.63) when compared with patients who had fever. All these analysis were positively associated with being tested before treatment of malaria. However, all these differences were statistically insignificant (Table 9). University of Ghana http://ugspace.ug.edu.gh 46 Table 9: Patient factors associated with being tested before prescribed treatment, Effutu Municipal, 2014. Variable Patients seen by Prescribers N(%) Conforming Bivariate Unadjusted OR 95%CI p-value