University of Ghana http://ugspace.ug.edu.gh t- DEPARTMENT OF HEALTH POLICY, PLANNINC AND MANAGEMENT SCHOOL OF PUBLIC HEALTH COLLEGE OF HEALTH SCIENCES UNIVERSITY OF GHANA, LEGON COST-EFFECTIVENESS ANALYSIS OF HOME MANAGEMENT OF FEVERS IN CHILDREN UNDER-FIVE YEARS IN THE DANGME WEST DISTRICT OF GHANA BY JUSTICE NONVIGNON (Student Number 10319531) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN FULFILLMENT OF THE REQUIREMENT FOR THE AWARD OF DOCTOR OF PHILOSOPHY (PHD) PUBLIC HEALTH DEGREE DECEMBER 2011 SC~OOL OF PUBLIC HEALTH LIBRARY ~_39 LEGON University of Ghana http://ugspace.ug.edu.gh DECLARATION I declare that this thesis is my own original research work conducted under supervision, except where specific references have been made and duly acknowledged, and that this work has not been submitted in part or in whole to any other institution for the award of a similar or any other degree. CANDIDATE Signature: ----------~~ Justice Nonvignon SUPERVISORS NJ, JJ).tL~V~ Signature: --. -- - .. --------- ~ Dr. Moses Aikins (Primary Supervisor) Professor John o. Gyapong (Secondary Supervisor) 11 University of Ghana http://ugspace.ug.edu.gh ABSTRACT Background Home management of childhood illnesses has been promoted as a way of managing childhood illnesses within communities. This approach can be life-saving, especially in rural areas of developing countries where geographical access to essential health services is limited. Many sub-Saharan African countries have adopted this approach and developed policies on its implementation. Ghana is implementing this policy, targeting malaria, pneumonia and diarrhea. However, there is little empirical evidence on the cost-effectiveness of these interventions. The objective of this study was, therefore, to assess the cost-effectiveness of two strategies of home management of fevers in children under-five in Ghana using antimalarials only or in combination with antibiotics. Methods Data were collected as part of a cluster randomised controlled trial with a step- wedged design. Approximately 12,000 children aged two months to 59 months in the Dangme West District in southern Ghana were covered. The costs and cost- effectiveness of the two interventions were assessed compared with the "control period" before the clusters were randomized as part of the step-wedged design. In addition, a three-year budget of the home-based care (HBC) currently being implemented nationally in Ghana were estimated by adjusting parameters from the trial. Costs were analyzed from societal perspective. Under-five anaemia cases averted, deaths averted and disability-adjusted life years (DALYs) averted were used as effectiveness measures. III University of Ghana http://ugspace.ug.edu.gh Results The total financial costs of the interventions were US$99,690.32 (antimalarial only) and US$138,321.21 (antimalarial plus antibiotic). The total economic costs for the interventions were US$204,394.72 (antimalarial only) and US$260,93 1.49 (antimalarial plus antibiotic). Recurrent costs constituted 89% and 90% of the total direct costs of antimalarials only and antimalarial plus antibiotic respectively. The number of eligible fever cases treated were 5,818 (antimalarial only) and 6,601 (antimalarial plus antibiotic). The anaemia cases averted were 1,361 (antimalarial only) and 1,147 (antimalarial plus antibiotics) and the deaths averted were 79.1 and 79.9 for antimalarial only and antimalarial plus antibiotic respectively. The DALYs averted were 2,264.8 (antimalarial only) and 2,284.6 (antimalarial plus antibiotic). The results show that cost per anaemia case averted was US$150.18 and US$227.49 for antimalarial only and antimalarial plus antibiotic respectively. The cost per death averted was US$2,585.58 for antimalarial only and US$3,272.20 for antimalarial plus antibiotic. Cost per DALY averted were US$90.25 (antimalarial only) and US$114.21 (antimalarial plus antibiotic). Further, the home management interventions together led to societal savings of about US$165,083. The results also show that caregivers of children under-five incur intangible costs such as pain and loss of leisure that are difficult to quantify. Additionally, the budgets to cover 50% of the target population under HBC for 2012 were also estimated to be about US$1.6m illion for antimalarial only and about US$3.4million for antimalarial plus antibiotic. Conclusion Home management of fevers is a cost-effective way of reducing morbidity and mortality in children under five years in Ghana. IV University of Ghana http://ugspace.ug.edu.gh DEDICATION This thesis is dedicated to my wife and son - Angela and Jaison. Angela, your prayers, patience, love and never-ending encouragement have paid-off with the completion of this work. The time and financial resources you and Jaison had to sacrifice to allow me work on this thesis have also paid-off. I love you both dearly. v University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENT I thank the God of my life for His sufficient Grace that enabled me complete this work. The research component of my studies was funded by the UNICEF/UNDP/Worid Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR) and the doctoral training component was funded by the Knowledge Enriches Programme of the Netherlands Organization for the Advancement of Tropical Research (WOTRO/NWO) as part of the Malaria- Pneumonia project. I appreciate thesis support from the African Doctoral Dissertation Research Fellowship - a collaborative programme of International Development Research Centre and African Population and Health Research Centre. I am highly indebted to Dr. Moses Aikins and Professor John Gyapong, my supervisors for their sacrifices and guidance. Mrs. Mercy Abbey, Dr. Amanua Chinbuah and Dr. Margaret Gyapong, (Director of the Dodowa Health Research Centre) and other staff of the centre, especially the Malaria-Pneumonia group are also appreciated. The following individuals and groups are also acknowledged for their support in diverse ways: Prof. Fred Binka and Prof. Richard Adanu, past and current Dean of the School of Public Health; Ms. Ruby Nsakie; Mr. Kojo Aniah; Dr. Christine A. Clerk; Mr. Samuel Bosomprah; Ms. Najat Dauda; other staff of the School of Public Health.; my 21 Nonvignon siblings; colleagues of the Bethlehem Methodist Youth Fellowship, Pokuase; Professor Emeritus Piet Kager (University of Amsterdam, The Netherlands); Mr. Jeroen van Spijk (formerly of University of Amsterdam); Dr. Catherine Goodman, Dr. Rob Baltussen; Dr. Lesong Conteh; staff, scientific advisors and Fellows of ADDRF 2010 cohort; Professor and Mrs. Oduro- Afriyie. God bless you all. vi University of Ghana http://ugspace.ug.edu.gh Table of Contents DECLARATION ii ABSTRACT iii DEDICATION v ACKNOWLEDGEMENT vi Table of Contents vii List of Figures xii List of Tables xiii List of Abbreviations xv CHAPTER ONE · · 1 1.0 INTRODUCTION 1 1.1 Background 1 1.2 Statement of the problem 3 1.3 Objectives 4 1.3.1 General objective 4 1.3.2 Specific objectives 4 1.4 Study significance 5 1.5 Conceptual framework 5 CHAPTER TWO 10 2.0 LITERATURE RE,V1EW 10 2.1 The epidemiology of malaria and pneumonia in Ghana 10 2.2 The health system of Ghana 12 2.2.1 An overview 12 2.2.2 Policies governing the control of childhood illness 13 vii University of Ghana http://ugspace.ug.edu.gh 2.2.3 Challenges and opportunities 16 2.2.4 Home-based care and health systems strengthening 16 2.3 Types of economic evaluation 18 2.4 Cost-effectiveness analysis: background, definition and foundations 19 2.5 Empirical studies 22 2.5.1 Costing of malaria/pneumonia interventions 22 2.5.2 Cost-effectiveness of preventive malaria/pneumonia interventions 26 2.5.3 Cost-effectiveness of malaria/pneumonia treatment interventions 30 2.5.4 Cost-effectiveness of malaria/pneumonia home management interventions. 35 2.5.5 Cost-effectiveness of other childhood interventions 39 2.6 Summary 39 CHAPTER THREE 41 3.0 METHODOLOGY 41 3.1 The malaria-pneumonia home management trial.. 41 3.1.1 Background and objectives 41 3.1.2 Trial site/area 41 3.1.3 Target population 42 3.1.4 Trial components 42 3.1.5 Trial design and interventions 43 3.1.6 Data collection and trial phases 45 3.1.7 Community-based agents 46 3.1.8 Health facilities 47 3 .2 Cost-effectiveness study 47 3.2.1 Study population 47 3.2.2 Sampling method and sample size for indirect cost 47 VIII University of Ghana http://ugspace.ug.edu.gh 3.2.3 Data collection and data sources 48 3.2.4 Data collection techniques and tools 49 3.2.5 Quality control 50 3.2.6 Data analysis 50 3.2.6.1 Estimating direct programme costs 50 3.2.6.2 Estimating indirect costs 52 3.2.6.3 Estimating total costs 53 3.2.6.4 Savings 53 3.2.6.5 Intangible cost 54 3.2.6.6 Estimating programme effectiveness 55 3.2.6.7 Cost-effectiveness Ratios (CERs): 58 3.2.6.8 Estimating budget impact of Rome-based Care 61 3.2.6.9 Sensitivity analyses 62 3.2.7 Ethical considerations of the study 62 3.2.8 Reliability and validity of results 64 3.2.9 Study limitations 64 CHAPTER FOUR 66 4.0 RESULTS 66 4.1 Home management trial costs 66 4.1.1 Total financial cost of Malaria-only Arm (AAQ) 66 4.1.2 Total financial cost of Malaria-Pneumonia Arm (AAQ+AMX) 66 4.1.2 Total economic cost of AAQ Arm 72 4.1.3 Total economic cost of Malaria-Pneumonia (AAQ+AMX) Arm 75 4.2 Cost savings of the study interventions 79 4.3 Intangible costs 79 IX University of Ghana http://ugspace.ug.edu.gh 4.4 Programme effectiveness 82 4.4.1 Estimated mortality rates of the study interventions 82 4.4.2 Estimated number of cases treated and anaemia cases 83 4.4.3 Estimated expected number of deaths 83 4.4.3 Estimated Disability-adjusted life years 83 4.5 Cost-effectiveness of trial interventions 84 Note: With the exception of cost per death averted and cost per DALY averted, other ..................... parameters are the same for both standardized and unstandardized effects. 85 4.6 Sensitivity analyses of costs and cost-effectiveness estimates 85 4.6.1 One-way sensitivity analysis of cost estimates 85 4.6.2 Multi-way sensitivity analysis of cost estimates 86 4.6.3 One-way sensitivity analysis of cost-effectiveness estimates 87 4.6.4 Multi-way sensitivity analysis of cost-effectiveness estimates 88 4.7 Estimated budget of Home-Based Care programme 89 4.7.1 Estimated budget for antimalarials only HBC 90 4.7.2 Estimated budget for antimalarials plus antibiotics HBC 90 CHAPTER FIVE 93 5.0 DISCUSSION 93 5.1 The home management trial., 93 5.1.1 Costs of implementing the home management interventions 93 5. I.1. I Financial costs 94 5.1.1.2 Economic costs 95 5.1.2 Cost-effectiveness of home management interventions 98 5.1.2.1 Usefulness of cost-effectiveness ratios 102 5.1.3 Cost savings of home management interventions 103 x University of Ghana http://ugspace.ug.edu.gh 5.2 Implications of results for home-based care approach 104 CHAPTER SIX 108 6.0 CONCLUSION AND RECOMMENDA nONS 108 6.1 Conclusion 108 6.2 Recommendations 109 6.2.1 Government of Ghana/Ministry of Health/Ghana Health Service 109 6.2.2 Decentralized local government structures and local communities 109 6.2.2 Research community 110 REFERENCES III APPENDIX I: Baseline Survey Form 126 APPENDIX II: Time Tracking Tool 134 APPENDIX III: Sick Child Form 135 APPENDIX IV: Morbidity survey forms 4 and 5 137 APPENDIX V: Projected under-five population for HBC districts, 2011-2013 144 APPENDIX VI: Number of communities and eligible fever cases 147 xi University of Ghana http://ugspace.ug.edu.gh List of Figures Figure 1: Types of costs and consequences of health services and programmes 7 Figure 2: Components of the home management trial 43 Figure 3: Diagrammatic presentation of cost-effectiveness analysis 60 Figure 4: Financial recurrent and capital costs of AAQ Arm 67 Figure 5: Financial recurrent and capital costs of AAQ+AMX Arm 67 Figure 6: Total economic cost of the AAQ study Arm 74 Figure 7: Direct economic costs of study Arms by activity category 78 Figure 8: Total economic cost of the AAQ+AMX study Arm 78 Figure 9: Responses to Likert items by study arm 81 Figure 10: One-way sensitivity analysis of cost-effectiveness estimates 88 Figure 11: Multi-way sensitivity analysis of cost-effectiveness estimates 89 XII University of Ghana http://ugspace.ug.edu.gh List of Tables Table 1: Empirical studies on costs of malaria/pneumonia interventions 25 Table 2: Empirical studies on cost-effectiveness of preventive malaria/pneumonia interventions 28 Table 3: Empirical studies on cost-effectiveness of malaria/pneumonia treatment interventions 32 Table 4: Cost-effectiveness studies targeting both preventive and treatment malaria/pneumonia interventions 34 Table 5: Empirical studies on cost-effectiveness of home management of malaria/pneumonia '" 37 Table 6: Empirical studies on cost-effectiveness of other childhood illness interventions 38 Table 7: Home management trial design 45 Table 8: Parameters used in sensitivity analyses 62 Table 9: Financial cost of the AAQ study Arm by activity 69 Table 10: Financial cost of AAQ+AMX study Arm by activity 70 Table 11: Summary of financial costs for AAQ and AAQ+AMX study arms 71 Table 12: Direct economic cost of AAQ Arm by activity 73 Table 13: Indirect cost of the home management trial 74 Table 14: Total economic cost of the home management trial. 74 Table 15: Direct economic cost of AAQ+AMX Arm by activity 76 Table 16: Summary of economic costs for AAQ and AAQ+AMX study arms 77 Table 17: Cost savings of study interventions 79 Table 18: Summary statistics of responses to Likert items 80 Table 19: Distribution of responses (%) by study arm 82 XIII University of Ghana http://ugspace.ug.edu.gh Table 21: Estimated standardized effectiveness indicators of study interventions 84 Table 23: Cost-effectiveness ratios for study arms 85 Table 24: One-way sensitivity analysis of cost estimates 86 Table 25: Multi-way sensitivity analysis of cost estimates 87 Table 26: One-way sensitivity analysis of cost-effectiveness estimates 87 Table 27: Multi-way sensitivity analysis of cost-effectiveness estimates 89 Table 28: Estimated budget of Rome-based Care - AAQ Arm 92 Table 29: Estimated budget of Rome-based Care - AAQ+AMX Arm 92 xiv University of Ghana http://ugspace.ug.edu.gh List of Abbreviations AAQ Artesunate Amodiaquine AMX Amoxicillin ARI Acute Respiratory Infection CBA Community-Based Agent CEA Cost-Effectiveness Analysis CER Cost-Effectiveness Ratio CHAG Christian Health Association of Ghana CHPS Community-based Health Planning and Services Consumer Price Index CUA Cost-Utility Analysis DALY Disability-Adjusted Life Year DDT Dichlorodiphenyltrichloroethane DHRC Dodowa Health Research Centre DMHIS District Mutual Health Insurance Scheme DYLL Discounted Years of Life Lost GHS Ghana Health Service HBC Home-Based Care HMF Home Management of Fever HMIS Health Management Information System HMM Home Management of Malaria HMT Home Management Trial ICER Incremental Cost-Effectiveness Ratio IMCI Integrated Management of Childhood Illnesses IPT Intermittent Preventive Treatment xv University of Ghana http://ugspace.ug.edu.gh IPTc Intermittent Preventive Treatment in children IPTi Intermittent Preventive Treatment in infants IPTp Intermittent Preventive Treatment in pregnancy IRS Indoor Residual Spraying ITN Insecticide- Treated Net MDG Millennium Development Goal MOH Ministry of Health NHIS National Health Insurance Scheme NMCP National Malaria Control Programme QALY Quality-Adjusted Life Year RDT Rapid Diagnostic Test SDHT Sub-District Health Teams SMR Standardized Mortality Ratio SP Sulfadoxine Pyrimethamine TBA Traditional Birth Attendant WHO World Health Organization YLD Years of Life lived with Disability YLL Years of Life Lost xvi University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE 1.0 INTRODUCTION 1.1 Background Malaria and pneumonia are leading causes of under-five deaths in sub-Saharan Africa, causing 22% and 21% of under-five deaths respectively (Black et aI., 2003). Studies have shown that there are overlaps in the symptoms and clinical features of both diseases (O'Dempsey et ai., 1993; English et ai., 1996; Kallander et aI., 2004). Both diseases are characterized by fever - especially in children under-five years who constitute the most-at-risk population (Okiro & Snow, 2010) - but also have other symptoms such as cough, fast breathing and signs of severe illness such as difficult breathing, inability to eat, and convulsions. These symptoms make it difficult to distinguish between the two diseases (which are the major causes of childhood fevers) in the absence of laboratory investigations. Consequently, fevers - including malaria and pneumonia - in African children have been treated presumptively because many health facilities in the region lack diagnostics (Onwujekwe et ai., 2010). In Ghana, malaria and acute respiratory infections (ARIs) - principally pneumonia - are the leading causes of fever in children under five years, often ranking first and second on the list of the top ten communicable diseases (Ghana Health Service, 2008; MOHINMCP, 2010). World Health Organization (WHO) estimates show that in 2006, about 7.3 million malaria cases occurred in Ghana, resulting in about 25,000 deaths (World Health Organization., 2008). Malaria accounts for over 40% of out- patient attendance and 25% of under-five deaths in Ghana (Ghana Health Service, 2008; MOHINMCP, 2010) whilst Pneumonia accounts for about 22% of hospital University of Ghana http://ugspace.ug.edu.gh admissions in tertiary health facilities in the country (Dakubo & Commey, 1996) and is responsible for 20% of under-five deaths (MOHINMCP, 2010). Current control strategies for both malaria and pneumonia in Ghana include early diagnosis and prompt treatment, particularly among those at risk of death and severe complications. The recommended first line treatments are artesunate - amodiaquine for malaria and amoxicillin for pneumonia. Until recently, the official interventions for both diseases were geared towards the formal health sector, though rural populations have little or no access to health facilities (Salisu & Prinz, 2009). Consequently, many caregivers of children seek treatment from outside the formal health sector (mostly untrained providers) (Ahorlu et aI., 1997; Nonvignon et al., 2010). Treatment in health facilities is often sought as a last resort especially in rural areas where geographic, financial and cultural access is a problem (Dun yo et aI., 2000). In developing countries, home management is commonly used to treat a number of illnesses, including malaria and acute respiratory infections (Goodman et al., 2006). Studies in Africa have shown that given adequate information and supervision, fevers can be treated promptly and appropriately in communities (Kidane & Morrow, 2000; Sirima et al., 2003). These studies showed that with adequate information and improved access to user-friendly pre-packed antimalarials and antibiotics, caregivers could recognise and seek appropriate treatment for most fever cases occurring at home. As a result, home management of malaria (HMM) has been endorsed by WHO as a sure way of managing tropical illnesses like malaria in developing countries where a large portion of the population has little geographical access to health 2 University of Ghana http://ugspace.ug.edu.gh facilities (World Health Organization, 2004) and its implementation has started in several countries in sub-Saharan Africa - 18 countries by 2008 (Ajayi et aI., 2008). The National Malaria Control Programme (NMCP) of the Ghana Health Service (GHS) is currently rolling out home management of malaria, acute respiratory infections and diarrhoea (called home-based care - HBC) as part of control of childhood diseases. 1.2 Statement of the problem A major concern with scaling up the implementation of health care programmes in developing countries has been the costs associated with such programmes (Hanson et al., 2004). The health sector faces difficult trade-offs regarding the allocation of scarce human, financial and other resources across a spectrum of diseases that constitute a burden on the sector. Therefore, in most cases, the clinical effectiveness of an intervention is not the only factor which influences the implementation or scaling up of desirable health interventions; costs of the given intervention and its cost-effectiveness compared to other existing or new interventions tend to playa role in the decision regarding which intervention is selected. Even though the NMCP has begun implementing HBC as part of efforts to reduce child mortality, the evidence regarding the costs and cost-effectiveness of HMM in Ghana is limited. Consequently, the conduct of a cost-effectiveness analysis is necessary in that, apart from providing information on the costs of implementing the community-based interventions, it will also provide information regarding the resources that could be saved by adopting the most cost-effective intervention compared to other alternatives. In this regard, the analysis of costs and cost- effectiveness serves as important tools that aid decision-making to scale up the most 3 University of Ghana http://ugspace.ug.edu.gh effective health care intervention taking into account the budget constraints of the policymaker. 1.3 Objectives 1.3.1 General objective The general objective of the study is to estimate the costs and cost-effectiveness of home management of fevers (HMFs) in children under-five years in the Dangme West District of Ghana. 1.3.2 Specific objectives The specific objectives of the study are to 1. Calculate the costs associated with HMFs in children under age five (i.e. direct and indirect costs to the health sector, caregivers and community); 2. Estimate the cost savings of HMFs to society (i.e. savings in expenditure accruing to caregiver and society, and caregiver savings in work time; 3. Describe the intangible costs associated with under-five fevers; 4. Calculate the effects in terms of number of under-five fever cases treated, anaemia cases averted, deaths averted and disability-adjusted life years (DALYs) averted; S. Calculate the cost-effectiveness of HMFs (i.e. cost per anaemia case averted, cost per death averted and cost per DALY averted); 6. Analyze the policy implications of the cost estimates for home-based care in Ghana. 4 University of Ghana http://ugspace.ug.edu.gh 1.4 Study significance First, the NMCP of the GHS has started implementing the HBC strategy. However, little information is available in the literature regarding the cost implications and cost- effectiveness of HMFs in Ghana. Consequently, findings of the study will prove useful for policy purposes. Additionally, since the home management trial, which the current study was part of, was undertaken under the Research and Development Division of the GHS, the findings of the study will likely stimulate discussions and further research that could feed into policies regarding the implementation and scale- up of the HBC strategy. Second, this study is also likely to influence health policy in other countries, especially in sub-Saharan Africa, who have embraced the idea of HMM. Finally, this study is the third in Africa to analyze the cost-effectiveness of HMM. However, it is the first to assess the cost-effectiveness of adding antibiotics to antimalarials during HMM. Therefore, the study will add to existing literature on cost and cost- effectiveness analysis of HMFs and other illnesses in particular given the scanty literature on such studies. 1.5 Conceptual framework Economic evaluation is based on the principle of opportunity cost; every resource used has a corresponding opportunity cost (i.e. the benefit that would have been derived had the resource been used in satisfying an alternative need). Economic evaluations also handle issues of efficiency i.e. achieving a goal at least (opportunity) cost and maximizing benefit (Donaldson et aI., 2002). Drummond et al.(l987) provide a conceptual framework for economic evaluation of health care programmes which is the guiding framework upon which Drummond et al. (2005) was developed. The 5 University of Ghana http://ugspace.ug.edu.gh framework presents the relevant lines of cost and proposes practical ways of measuring such costs. Due to the relevance and practicality of the framework, this study adapts a similar framework. Figure 1 shows the conceptual framework used for this study. The framework has two broad components showing costs and consequences of health care programmes. The cost component identifies three categories of costs. The first two categories identify the costs of organising the home management programme and the third category identifies the costs associated with standard health care options (e.g. health facilities, drug store and other over-the-counter options). The first category consists of direct costs to the health sector of organizing and operating the HMFs programme. These costs include those incurred on training and monitoring of the community-based agents (CBAs), which come in the form of recurrent (e.g. supplies, personnel, vehicle operation and maintenance, building space) and capital costs (e.g. equipment, vehicles). Such costs are identified by listing the items (i.e. "ingredients) used up in the programme and involve variable costs and fixed costs. The second category comprises costs incurred by caregivers of children under-five years under the home management programme, which include out-of-pocket expenses and inputs that caregivers and family/community make into treatment. This component includes the time spent by community-based agents in treating children who have fever. There are also indirect costs incurred by caregivers. These indirect costs include the caregiver's time lost from work, which constitutes production losses to the caregiver. Psychic costs such as pain, anxiety, and suffering are intangible costs usually borne by patients. However, for the purpose of this study, the psychic costs of caregivers rather than the patients are measured since the patients are children below five years of age 6 University of Ghana http://ugspace.ug.edu.gh who could hardly express themselves, thus, may not be able to provide responses to questions posed. Psychic costs are not included in cost-effectiveness analysis. Therefore, though the current study attempts to measure such costs, they are not included in estimating the cost-effectiveness of the interventions. Figure 1: Types of costs and consequences of health services and programmes COSTS CONSEQUENCES Cost of the home management 1 I. Changes in physical, social, andprogramme emotional fUlioning (effects)I. Costs to the health sector of organising and operating the HMFs programme ( e.g. training and II. Changes in effects Tm"I ...,. ,c"'Tm"I;to.n...n....g.9".10 ............ +" c....o....m......m......u...~.n....ny;hr -o'h.asecrI 0 ;t"I3,....+Vj_ lJ ""'-'" agents - CBAs, facilitators and J costs a. relating to under- }. supervisors' time and allowances, five morbidity and deaths DIrect supplies, equipment, capital) averted, DALY s effects II. Costs to caregiver of under-five and community of HMFs b. relating to - Caregiver saVings} out-of-pocket expenses Direct in expenditure caregiver's input into Direct } cost Health sectortreatment savings in effects expenditure CBA's time caregiver's time Indirect} savings in lost Indirect (productivity losses) costs work time } effects psychic/ intangible costs Cost of other options including health facility-based option III. Costs of other treatment options for fevers in under-fives out-of-pocket expenses (e.g. treatment, travel, Direct drugs) cost caregiver's input into } treatment caregiver's time Indirect (producti vity losses) } cost Source: Adapted from Drummond et al. (1987, 2005). 7 University of Ghana http://ugspace.ug.edu.gh The third category of costs presented in Figure 1 is a modification to the original framework proposed by Drummond et al. (1987) using a similar costing approach by Kirigia (1998). This category presents costs of using other treatment options (including health facility-based treatment, drug stores, etc.) for fever in children under-five years of age and such costs include out-of-pocket expenses and other caregiver inputs. For the purpose of the current study, the costs from this category will be the basis for calculating the savings made by the HMFs interventions. There are several perspectives from which costs can be measured. These perspectives include patient perspective, provider perspective, donor/funder perspective and societal perspective. The identification of a costing perspective is important in that it provides the focus of costing and determines which costs need to be included in the study (Drummond et al., 1987), as some items may constitute cost from one perspective but not cost to another perspective. A typical example is travel cost, which is cost from patient's point of view but not cost from the provider's point of view. The broadest perspective is the societal perspective which includes all other perspectives. This study will measure costs from the societal perspective, because the home management programmes brings about significant costs to caregivers and the community as well as the health sector and the exclusion of anyone of these costs implies that not all costs associated with the programme are measured. Figure 1 also presents the consequences of health programmes. The first category under consequences deals with therapeutic outcomes of the given health care programme. These outcomes are effects which relate to changes in the physical, social or emotional functioning of individuals and can be measured objectively. For the 8 University of Ghana http://ugspace.ug.edu.gh purpose of this study, the therapeutic effects may lead to changes in effects measured in natural units (second category) - number of under-five fever cases appropriately treated, deaths averted, and DALY s averted. The changes in effects have direct and indirect components. Direct effects include changes (reduction) in under-five morbidity and number of deaths, and DALY s as well as caregiver and health sector savings in expenditure. Indirect effects relate to caregiver's savings in work time lost due to a child's illness. 9 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO 2.0 LITERATURE REVIEW This chapter presents a review of literature on malaria and pneumonia. The chapter begins with a review of the epidemiology of malaria and pneumonia in Ghana, which is followed by a review of the policies governing the control of childhood illnesses and the challenges faced by the health system and opportunities. Further, the home- based care approach to the management of childhood illnesses is discussed. The definition of and background to cost-effectiveness analysis and some theories underpinning the concept of cost-effectiveness are then presented. The chapter concludes with a critical review of empirical cost and cost-effectiveness studies. 2.1 The epidemiology of malaria and pneumonia in Ghana Malaria is a mosquito-borne protozoal disease which was first treated in Peru in the 1600s using the bark of the Cinchona tree (Cahill, 2004). Plasmodium, which causes malaria in humans is transmitted to humans by the Anopheles mosquito, which picks up the plasmodium parasite from an animal with parasitaemia. Contact with the Anopheles mosquito could lead to infection with plasmodium falciparum, plasmodium vivax, plasmodium ovale or plasmodium malariae. It is plasmodium falciparum, however, which causes majority of the clinical cases and mortality (Bozdech et al., 2003; Cahill, 2004). In Ghana, malaria is caused by three main parasite species i.e. plasmodium falciparum, plasmodium malariae and plasmodium ova/e. However, about 90% of malaria infections in the country are caused by plasmodiumfalciparum (Ministry of Health, 2009). The entire population of the country is at risk of the disease and transmission occurs throughout the year with some seasonal variations, especially during the rainy season between April and July (Ministry of Health, 2009). 10 University of Ghana http://ugspace.ug.edu.gh Malaria is a major cause of morbidity in Ghana. Since 2001 - though exhibiting a downward trend - the disease has consistently accounted for more than 30% of all outpatient cases with a high of 45.3% in 2003 and a low of 31.7% in 2008 (National Malaria Control Program, 2010). Children under-five years of age and pregnant women are most vulnerable to malaria. In 2009, the disease was responsible for 48.9% of under-five outpatient cases and 11.5% of outpatient attendance for pregnant women (National Malaria Control Programme, 2009). Malaria also remains a major cause of admissions, accounting for 35.9% of all admissions in public health facilities in the country. Again, the disease accounts for 48.8% of under-five admissions and 14.9% of admission of pregnant women (National Malaria Control Programme, 2010). The most important cause of mortality in Ghana especially in children under-five years of age, malaria caused 16.8% of all deaths in public health facilities in 2009. The disease has been responsible for over 20% of deaths in children under-five years of age annually since the year 2000 with a high of 34.6% in 2008. Malaria was also responsible for 9.4% of maternal deaths in 2009 (MOH, 2009; NMCP, 2010). Pneumonia is an inflammatory condition that affects the lung and is usually caused by bacterial, viral or fungi infection (Leach, 2009). The disease is normally associated with fever, respiratory symptoms, and difficult breathing. In Ghana, pneumococcus meningitis and invasive pneumococcal disease are prevalent (Holliman et al., 2007). Pneumonia is responsible for 20% of under-five deaths annually in Ghana with the majority of the deaths occurring in infants (Ghana News Agency, 2010). 11 University of Ghana http://ugspace.ug.edu.gh 2.2 The health system of Ghana 2.2.1 An overview The health sector is organised around the Ministry of Health, which has the overall oversight of the health system and is responsible for policy formulation, coordination, resource mobilization, monitoring and evaluation. The Ghana Health Service was created by Act 525 of Parliament (Government of Ghana, 1996) and is responsible for service delivery and implementation of policies (MOH, 2009). The teaching hospitals are autonomous institutions accountable directly to the Ministry of Health and have their own boards. The autonomy of the teaching hospitals is also enshrined in Act 525 (1996). The Ghana Health Service has regional, district and sub-district directorates which are responsible for service delivery at the respective levels and accountable to the headquarters. The health care system in Ghana follows a three-tier pattern; it has primary, secondary and tertiary levels. The primary level comprises community level services provided by community-based volunteers and community health clinics also known as community-based health planning and services (CHPS) compounds. Sub-district health teams (SDHTs) oversee the provision of health care services at the community level. District hospitals are part of the secondary level and serve as referral facilities to primary level providers and also refer to regional hospitals which are in the tertiary level. The tertiary level has teaching hospitals which serve as the highest referral points (MOHINMCP, 2010). There are private providers and non-governmental organizations, including the Christian Health Association of Ghana (CHAG) who together provide over 40% of health care services in the country (WHO Regional 12 University of Ghana http://ugspace.ug.edu.gh Office for Africa, 2009). There exists traditional health services providers who form an essential part of health service delivery in the country (Senah et aI., 2001). The National Health Insurance Scheme (NHIS) was introduced in 2005 as a measure to reduce the financial burden of health care on individuals, making health care financially accessible (WHO Regional Office for Africa, 2009). The National Health Insurance Act 650 (Government of Ghana, 2003), which was passed into law in 2004 provided that scheme be administered through District Mutual Health Insurance Schemes (DMHIS) and private scheme. The scheme is mainly tax-funded (WHO Regional Office for Africa, 2009) and also covers free maternal health care which was introduced in 2008. The health care system of Ghana faces public health challenges posed by mainly infectious diseases, including malaria, pneumonia, diarrhoea and a host of non- communicable diseases. A major public health concern relates to child survival. A number of disease conditions that are treatable are responsible for the relatively high rate of under-five mortality rate of 80 per 1000 live births. These include malaria (25%), pneumonia (20%), diarrhoea (17%) and other neonatal conditions (27%). Additionally, estimates show that more than 50% of all deaths are associated with malnutrition (MOHINMCP, 20 10). 2.2.2 Policies governing the control of childhood illness Malaria, pneumonia, diarrhoea, and malnutrition or vitamin A deficiency are the main causes of under-five mortality in Ghana. Until recently, management of these conditions follow a "programme centred approach" where these conditions are managed separately by their respective programmes. However, the current child health policy of the country focuses on a "child centre approach" with all programmes 13 University of Ghana http://ugspace.ug.edu.gh following a single integrated plan in the management of these illnesses. The child health policy aims at improving "population coverage of effective child health interventions" and the main goal is "to reduce child mortality to 40 per 1000 live births by 2015" (Ministry of Health, 2007). The policy revolves around the following guiding principles: • The continuum of care for the mother and the child which targets all life stages of the child from pregnancy through childhood; • The continuum of care for the health system which ensures the delivery of child health interventions at all levels including the home, community and health facilities; • A regular review of the minimum essential package of medicines targeting child health as part of the essential medicines list of the Ministry of Health; • Collaboration between all stakeholders in the implementation of child health activities' • Contribution of child health activities to achieving health-related Millennium Development Goals; • Child health programmes will be informed by conventions and treaties to protect children, to which Ghana is a signatory. The Ministry of Health, Ghana Health Service and other partners started the implementation of community-based interventions in 2011 as part of a home-based approach to the management of childhood illnesses including malaria, acute 14 University of Ghana http://ugspace.ug.edu.gh respiratory infections and diarrhoea (MOHlNMCP, 2010). The approach focuses on the following areas: • Correct recognition of signs and symptoms of malaria, ARJ and diarrhoea • Provision of correct and timely management for simple/ uncomplicated malaria, ARI and diarrhoea in accordance with national protocols. • Recognition of the danger signs of severe illness • Prompt referral of cases beyond the ability of CBAs/caregivers to the appropriate level • Provision of adequate support services and care: • Use of antimalarial suppository at the community level for severe malaria cases before referral to the hospital/clinic. • Behaviour change communication on prevention, recognition, early care seeking and full compliance with treatment for malaria, ARJ and diarrhoea • Promotion, distribution and re-treatment of insecticide-treated nets (fTNs) • Promotion of intermittent preventive treatment for pregnant women (IPTp) • Identification and referral of pregnant women due for rPT • Promotion (community mobilisation and sensitization) of indoor residual spraying (IRS) in targeted districts • Promotion of good hygiene including washing of hands with soap and water • Promotion of appropriate clothing when the weather changes 15 University of Ghana http://ugspace.ug.edu.gh As part of the HBC, local community members are selected, trained and equipped to act as CBAs. These CBAs diagnose and provide drugs for uncomplicated cases of malaria, pneumonia and diarrhoea and refer complicated cases to nearest health facilities. The CBAs are selected in consultation with opinion leaders and community members. 2.2.3 Challenges and opportunities The health sector in Ghana faces health systems challenges which fight against efforts at ensuring that all people have access to good quality health care services, thereby improving the overall health of the people and the wealth of the nation. Important challenges include inadequate health infrastructure which implies that many people, especially in rural areas have very little physical access to health care services. There is also inadequate health human resource in the country, with the doctor/population ratio of 13,683 (Ghana Health Service, 2008). The weak health. management information system (HMIS) also raises concerns about the quality of data and non-use of data from the HMIS. Additionally, there are challenges with funding of the health sector which impedes efforts to scale up priority interventions (MOH, 2009). 2.2.4 Home-based care and health systems strengthening The World Health Organization describes a health system as one that "consists of all organizations, people and actions whose primary intent is to promote, restore or maintain health" (de Savigny & Adam, 2009). The goals of a health system are "improving health and health equity in ways that are responsive, financially fair, and make the best, or most efficient, use of available resources" (de Savigny & Adam, 2009). Health systems, especially in developing countries, are generally weak, thereby making efforts at improving interventions to the people in greatest need unfruitful. 16 University of Ghana http://ugspace.ug.edu.gh This weakness poses a challenge to efforts toward the attainment of the health-related Millennium Development Goals targets of reducing child mortality, improving maternal health and combating HIV/AIDS, tuberculosis and malaria. In order for health systems to deliver the services to those in the greatest need, the WHO has developed a framework which has six health systems building blocks: service delivery; governance and leadership; financing; human resources; information; and medical products, vaccines, technologies (de Savigny & Adam, 2009). The interactions and relationships among the building blocks serve to strengthen the system to better provide essential services. Therefore, any intervention that seeks to strengthen the health systems lead to an interaction between two or more building blocks. Home-based care interventions are primarily service delivery interventions. As such, health care services are provided within communities and by local community members (under the supervision of District and sub-District Health Administration) to ensure continuity of service provision. However, this generates an interaction between service delivery and other health systems building blocks as they influence each other. The HBC programmes benefit from the active participation of decentralized health structures, working with opinion leaders and community members, which ensure community participation. In terms of human resources, staff from the decentralized health structures and health facilities benefit from capacity building programmes and serve as supervisors to CBAs. 17 University of Ghana http://ugspace.ug.edu.gh Financing of the HBC is usually done by the health sector and stakeholders as drugs are mostly provided without charge. Data are collected by the CBA and supervisors and these data are useful for planning purposes by District Health Administration. Given the nature of HBC interventions, it is evident that they are a platform that strengthens the health system. 2.3 Types of economic evaluation Economic evaluation refers to the analysis of the costs (inputs) and consequences (outputs) of alternative courses of action (Drummond et al., 2005). Economic evaluation is a comparative analysis because it compares the costs and consequences of alternative progmmmes and enables some judgement to be made regarding which of the alternative programmes is worth doing. The key economic concepts of scarcity and choice are underlying concepts in economic evaluation; resources are scarce (i.e. not available in the quantities that would have allowed society to satisfy all its wants) and as a result, choices are needed and economic evaluation serves as one of the tools for aiding choices. There are four basic types of economic evaluation (Hagberg, 2007). These types are cost-effectiveness analysis (CEA), cost-utility analysis (CUA), cost-benefit analysis and cost-minimization analysis. The first three types of evaluation examine both the costs and consequences of alternative programmes. However, these techniques differ in the way they measure outcome (Drummond et al., 1987). Cost-effectiveness analysis measures the costs of alternative programmes in monetary units and the consequences in natural units (e.g. number of cases averted, lives saved). In this case, the consequences of the programme under consideration are difficult to measure in 18 University of Ghana http://ugspace.ug.edu.gh monetary terms. However, cost-benefit analysis quantifies and measures both costs and consequences in monetary units. Cost-utility analysis measures costs in monetary units and consequences in utility (i.e. preferences or satisfaction for a set of health outcomes). An example of utility measure is quality-adjusted life year (QAL V). The CUA technique is sometimes regarded as a type of CEA as it also does not quantify consequences in monetary terms and both seem to address a similar issue (McGuire et aI., 2000; Drummond et aI., 2005). Cost-minimization analysis (sometimes called cost analysis) assumes that the alternative programmes being evaluated achieve similar outcomes or same degree of outcomes. Thus, only costs are evaluated and the programme that achieves the outcome at least cost is preferred. Drummond et al. (2005) refer to this type of evaluation as partial economic evaluation since only costs are considered. 2.4 Cost-effectiveness analysis: background, definition and foundations Cost-effectiveness analysis is an essential tool for comparing the costs of alternative health interventions to the expected health gains or outcomes (Musgrove & Fox- Rushby, 2006). The technique was developed in the 1960s and first applied by the military and has since been used extensively in evaluating health care programmes. Cost-effectiveness analysis is applied where different health interventions are expected to produce similar health outcomes (each achieving the outcome at different degrees) at different costs. The cost-effectiveness ratio (CER), is a ratio of total costs to total health benefits of a given health programme. The CER is sometimes called average cost-effectiveness ratio. Incremental cost-effectiveness ratio (leER) is calculated as the net costs (i.e. the difference in costs of two interventions) divided by the net benefits/effects (i.e. the difference in benefits of two interventions). The result 19 University of Ghana http://ugspace.ug.edu.gh of such analysis is stated in terms of cost per unit of outcome (e.g. cost per life-year saved). The technique can be used either to evaluate a series of mutually exclusive alternative interventions or to evaluate a single intervention (Musgrove & Fox- Rushby, 2006). In each case, a comparison is made between the intervention or interventions under consideration and the existing programme (sometimes called the "do-nothing" option). The decision rule is that the intervention that produces the least CER is adjudged most cost-effective. Cost-effectiveness analysis in particular - and economic evaluation in general - are linked to the concept of economic efficiency (McGuire et aI., 2000; McGuire, 2001). Economic efficiency comes in various forms. Technical efficiency is concerned with achieving maximum output level from a given input level. Defined under health, technical efficiency refers to effectiveness, where the highest possible health outcome is achieved given the level of resource input. Productive efficiency is concerned with achieving a given level of health outcome at minimum cost. A third form of efficiency - allocative efficiency - concerns obtaining the highest value (i.e. in terms of utility) from the output, given the cost input. Cost-effectiveness analysis measures technical and productive efficiency (McGuire et aI., 2000; Donaldson et al., 2002). Nevertheless, CEA can also be linked to a form of economic efficiency known as Pareto efficiency - named after an Italian economist Vilfredo Pareto who first used the concept. Pareto efficiency or Pareto optimality is a welfare economic condition which is achieved if an allocation of the (maximum) outputs is made among persons to the extent that a re-allocation or further allocation will tend to make some better and at least one person worse off (Brownstein, 1980; McGuire et aI., 2000; McGuire, 2001). 20 University of Ghana http://ugspace.ug.edu.gh Cost-effectiveness analysis can inform and guide policymakers to allocate health resources based on the desire to responds to a population's greatest health needs. Garber and Phelps (1997) argue that decisions based on CEA can have favourable welfare economic properties if both costs and outcomes are measured properly. In this regard, CEA can be used to determine optimal allocation of health resources (Garber & Phelps, 1997). Under certain conditions, market forces - in competitive markets - lead to technical, allocative, productive and Pareto efficiency. However, if allowed to operate on it own, the health care market is unable to achieve such efficiency due to market failures. This failure of health care markets implies that regulation is required to make these markets attain efficiency. Criticisms of CEA have mainly been methodological issues (McGuire et aI., 2000; Donaldson et aI., 2002). However, questions have been raised about the relevance of CEA as a decision making criterion, as CEA tends to reduce the whole issue of whether or not to implement a health intervention over others into a "simple ratio" of costs and effectiveness (Diamond & Kaul, 2009). Further, there is the criticism that the use of thresholds in interpreting CERs represents a "...gross oversimplification of a highly complex process" (Diamond & Kaul, 2009). In spite of these criticisms, CEA is generally regarded as a relevant scientific approach that illuminates the understanding of policymakers on the worth of new interventions for the purposes of medical decision making and public policy (Weintraub & Cohen, 2009). Cost-effectiveness analysis ensures that an intervention that yields the best cost-effect combination is recommended and if implemented, some resources could be freed for use by other competing health needs. It must be noted that in CEA, a critical 21 University of Ghana http://ugspace.ug.edu.gh assumption is that the goal of the decision maker is to achieve economic efficiency. With this in mind, any decisions based on the results of CEA is consistent with economic efficiency (McGuire et al., 2000). The current study uses CEA due to the fact that the health outcomes being considered cannot be measured in monetary terms. 2.5 Empirical studies Different categories of empirical studies that analyzed the costs and cost-effectiveness of health interventions were reviewed. The review covers preventive and treatment interventions as well as home management interventions of malaria and pneumonia. Studies on the cost-effectiveness of other childhood diseases were also reviewed. 2.5.1 Costing of malaria/pneumonia interventions Table 1 presents studies that analyse the cost of malaria and pneumonia interventions. The review comprises seven studies covering the period 2005 and 2009. Most of the studies reviewed were in Africa - three in Ghana (Asante et al., 2005; Grabowsky et al., 2005; Akazili et al., 2007), two in Kenya (Guyatt et al., 2002; Ayieko et al., 2009) and one in Tanzania (Adam et al., 2005). One study (Yeung et al., 2008) was conducted in Cambodia (Asia). The interventions evaluated include hospital services, rapid diagnostic test (RDTs), blister packaging and increasing access to diagnostics to remote communities, distribution of ITNs and IRS. Five of the studies analyzed the costs of malaria interventions with the remaining one (Ayieko et al., 2009) analysing costs of both malaria and pneumonia interventions. One study (Adam et al., 2005) assessed the cost of Integrated Management of Childhood Illnesses (IMCl) as against routine care, with the IMCI approach covering malaria, pneumonia, diarrhoea, measles and malnutrition. Two of the six studies (Grabowsky et al., 2005; Yeung et al., 2008) analyzed the financial cost of the interventions whilst the remaining five 22 University of Ghana http://ugspace.ug.edu.gh analyzed economic costs. Three of the studies analyzed costs from provider's perspective and one from household perspective. Guyatt et al. (2002) and Adam et al (2002) are the only studies that analyzed cost from societal perspective. From Table I, malaria treatment costs US$6 per episode and between US$9 and about US$16 per household in Ghana whereas costs of hospital treatment of malaria in Kenya are estimated to range between US$45 and US$96. Treatment costs of pneumonia in Kenya are comparatively higher, ranging from $43 to $177 per case depending on the category of the hospital. In Cambodia, strategies targeting malaria prevention and treatment cost between US$7 and US$21. The costs of managing childhood illnesses under the IMCI approach were reported to be lower (US$11.19) than under routine services in health facilities (US$16.09) in Tanzania. The studies presented in Table 1 provide estimates of costs to households, the health sector and society in the treatment and prevention of malaria and pneumonia in developing countries. These estimates are useful for planning purposes. However, the use of provider perspective, which is relatively easier to undertake, excludes costs from patients or household perspective. Two studies used societal perspective which includes cost from all perspectives. Furthermore, one of the studies analyzed financial costs, which do not represent the real or opportunity cost of such interventions though financial cost estimates are useful for planning. Again, even though malaria and pneumonia both present similar symptoms and are difficult to distinguish in the absence of diagnostics, only two out of the six studies analyze hospital cost of the two disease conditions. 23 University of Ghana http://ugspace.ug.edu.gh Finally, it is difficult to compare the cost estimates from these studies across countries because of varying costing methodology and differences in year in which the cost data were collected. 24 University of Ghana http://ugspace.ug.edu.gh Table 1: Empirical studies on costs of malaria/pneumonia interventions SIN Reference, country Interventions Disease/target pop. Costing Type of cost Cost (US$) perspecti ve 1 Ayieko et al. (2009) Hospital cost for malaria, Malaria, Provider Economic $43.23 - 95.58 Kenya, pneumonia, other pneumonia (malaria) childhood illnesses Children >5 $43.36-177.14 , - (pneu.) 2 Malaria, Provider $7.48 - 21.12 Yeung et al. (2008) Rapid diagnostic test; General population Cambodia Blister-packaging locally; Strategies to increase Not stated access to diagnosis and treatment to remote communities o' 3 Akazili et al. (2007) - Malaria, Household Economic $6.39 per case Ghana General population $9.39 per household Graboswsky et al (2005), Distributing ITNs during Malaria, Provider Financial $3.74 4 Ghana measles vaccination Children >5 5 Asante et al. (2005), Malaria, - Economic $15.79 per case Ghana - General (household) $1.9 per case (health facilities) 6 Adam et al. (2005) Integrated Management of Multiple* Societal Economic $11.19 (AMCl) Tanzania Childhood Illnesses versus $16.09 (routine) routine care 7 Guyatt et al. (2002), IITN and IRS Malaria, Provider & Economic $0.88/person protected Kenya General population societal (IRS) $2.34/person protected (ITN) Note: *Included malaria, pneumonia, diarrhoea, measles and malnutrition 25 University of Ghana http://ugspace.ug.edu.gh 2.5.2 Cost-effectiveness of preventive malaria/pneumonia interventions Eleven studies on cost-effectiveness of preventive malaria/pneumonia interventions were reviewed and are presented in Table 2. These studies were conducted between 1999 and 2009. Seven of the studies were conducted in sub-Saharan Africa, three in Asia and one study (Sinha et ai., 2007) was a multi-country study. The interventions evaluated include vaccination campaigns, distribution and use of ITNs, IRS and intermittent preventive treatments (IPTs). Furthermore, the studies range from controlled trials to programme evaluations. Five of the studies were on malaria in children, three on malaria in mixed population (i.e. children and adults), and one (Butraporn et ai., 1999) on migrant population. There were two studies on pneumonia - one in under-fives (Sinha et aI., 2007) and the other in the elderly (Cai et al., 2006). The studies use different outcome measures, such as number of cases treated, episodes averted, deaths averted, life-years saved and DAL Ys averted. Each of the studies provides insights into the cost-effectiveness of various health interventions that can be used to prevent malaria and pneumonia, thereby informing the decision to scale-up the most effective interventions. The review shows that cost- effective interventions exist to prevent malaria/pneumonia; a case could be averted with as low as US$0.6, a DALY averted with US$3.7 and death prevented with US$100. The maximum amounts required for averting one case, saving one DALY and preventing one person from dying due to malaria or pneumonia are US$87, US$100 and US$1,950 respectively. However, out of the 11 studies, two were interventions targeting pneumonia with the remaining nine on malaria, which shows that cost-effectiveness literature - especially 26 University of Ghana http://ugspace.ug.edu.gh in Africa - focuses more on malaria, even though pneumonia kills many more people. None of the cost-effectiveness studies reviewed examined interventions that deal with the two diseases together. Furthermore, six studies analyzed the cost-effectiveness of interventions aimed at preventing malaria/pneumonia in children with four of these focussing on under-fives. Literature on the cost-effectiveness of interventions targeted at children under-five years are important given that the age group suffers most from the brunt of the diseases and many interventions target this age group. Further, the studies reviewed used different outcome measures and this makes comparability of results difficult. Disability-adjusted life-year is a measure which makes different studies comparable. However, not all the studies use DALY. Additionally, some of the studies analyzed costs from only provider's perspective which does not take into account costs to patients or society. Some studies also analyzed cost from only patient perspective whilst others did not explicitly state from which perspective costing was done. Since these studies seek to inform policy on the choice of interventions to improve societal health, societal perspective of costing would have been a better approach. Finally, comparing the cost-effective estimates reported by these studies is difficult given the varied methodology and cost year. 27 University of Ghana http://ugspace.ug.edu.gh Table 2: Empirical studies on cost-effectiveness of preventive malaria/pneumonia interventions SIN Reference, country Interventions Disease/target pop. Study type Outcome measure C E estimate 1 Hutton et al. (2009) Intermittent Preventive Malaria, Trial Episodes averted US$1.6-4.7 Mozambique & Treatment in infants Children Deaths averted US$l 00.2-30 1.I Tanzania (IPTi) DAL Ys averted US$3.7-11.2 2 Temperly et al. School-based delivery of Malaria, Controlled trial, costs Number treated US$29.S4 (200S) IPT School children US$150 - 11 Butraporn et al. lambdacyhalothrin- Malaria Trial, patient costs Cases averted US$0.59 (1999) treated nets; Migrant pop. US$0.74 Thailand spraying with DDT 29 University of Ghana http://ugspace.ug.edu.gh 2.5.3 Cost-effectiveness of malaria/pneumonia treatment interventions Table 3 presents 12 empirical studies on malaria/pneumonia treatment interventions. Ten of the studies were conducted in sub-Saharan Africa, one in Asia and the remaining one in both continents (Tan-Torres et al., 2005). The studies cover the period 2001 and 2009. The interventions evaluated include diagnostics and case management using different drugs, mostly combination therapies. The study designs range from observational studies, randomised controlled trials and multi-centre trials. Two studies evaluated interventions targeting under-fives, seven targeted mixed population and two target pregnant women (Goodman et ai., 2001a; Wolfe et al., 2001). Ten of the studies evaluate malaria interventions, one on pneumonia, measles and diarrhoea and one on fevers (Shillcutt et al., 2008). Further, the outcome measures used include cases diagnosed, cases treated, deaths averted, DALY s averted and discounted years of life lost (DYLL) averted. Like the studies reviewed earlier, these studies provide useful information on cost-effective interventions in treating malaria and pneumonia. The review indicates that one DALY could be averted with less than US$25 using treatment interventions. However, about two-thirds of the studies did not use outcome measures (i.e. DALY) that make comparability of cost-effectiveness estimates easy. Besides, majority of the studies focus on malaria only. Furthermore, only two of the 11 studies focus on interventions targeting under-fives who bear the greatest burden of malaria and pneumonia. Wolfe et al. (2001) and Goodman et al. (2001a) evaluated interventions targeting pregnant women. 30 University of Ghana http://ugspace.ug.edu.gh Four of the studies reviewed evaluated both preventive and treatment interventions and are presented in Table 4. The studies cover 1999 to 2008. Three of these studies were on Africa and one on Brazil. All four studies use DALY as outcome measure, and this makes the results comparable. However, one study was conducted on malaria and pneumonia interventions whilst three were conducted on malaria. Additionally, one study was conducted on under-fives and pregnant women. From these studies, one DALY could be averted with between US$l and US$466. 31 University of Ghana http://ugspace.ug.edu.gh Table 3: Empirical studies on cost-effectiveness of malaria/pneumonia treatment interventions SIN Reference, country Interventions Diseaseitarg_et_pop. Study type Outcome measure CE estimate 1 Lubell et a1.(2009) Treatment using Quinine; Malaria, Multi-centre trial, Deaths averted $140 Bangladesh, Indonesia Treatment using Artesunate mixed population costs from provider and Myanmar perspective 2 Chanda et a!. (2009) Clinical diagnosis; Malaria, Observational study, Cases diagnosed $6.5 - 17.1 Zambia Microscopy; mixed population provider costs correctly Rapid Diagnostic Tests .") Uzochukwu et al.(2009) Presumptive treatment; Malaria, ICER using decision Deaths averted $221 Nigeria Rapid Diagnostic Tests; mixed population analysis (ICER) Microscopy 4 Shillcutt et al. (2008) Presumptive treatment; Fevers, Decision tree DAL Y averted NiA SSA Microscopy; mixed population analysis Rapid Diagnostic Test 5 Chanda et a!. (2007) Artemesinin combination Malaria, Average CEA using cases successfully $8.57-10.65 Zambia therapies (ACTs); mixed population data from health treated; Sulphadoxine- system proportion of pyrimethamine cases proceeding to severe malaria 6 Lubell et al. (2007) Rapid diagnostic test; Malaria, patients correctly N/A Tanzania Microscopy mixed population treated 7 Wiseman et al. (2006) Amodiaquine+Sulphadoxine Malaria, Randomised trial Resource savings $0.53-10.01 Tanzania -pyrimethamine; under-fives and malaria case Amodiaquine+ Artes unate; averted Artemether-Lumefantrine 8 Tan- Torres et al. (2005) Case management of Pneumonia Cost-effectiveness Number of NiA SE Asia and SSA pneumonia; Measles analysis pneumonia, Oral rehydration therapy; Diarrhoea, measles and 32 University of Ghana http://ugspace.ug.edu.gh SIN Reference, country Interventions Disease/target pO_E: Study type Outcome measure CE estimate Supplementation/ under fives diarrhoea cases fortification of staple food averted; with vitamin A or zinc; Number of deaths Provision of supplementary due to pneumonia, food with counselling on measles and nutrition; diarrhoea averted Immunization against measles. 9 Coleman et al.(2004) ACT versus SP Malaria Decision analysis DALYs saved N/A SSA 10 Goodman et al. (200 Ia) Weekly chloroquine Malaria, CEA using Discounted years $1.13-2.14 chemoprophyl. Pregnant women standardized of life lost Monthly sulphadoxine- analytical (DYLL) $1.30-2.31 pyrimethamine framework 11 Goodman et al. (200 Ib) Changing firstline drug from Malaria, Model using DALYs averted >$25 SSA CQ to SP mixed population hypothetical cohort 12 Wolfe et al. (2001) Presumptive treatments with Malaria, Decision analysis Low-birth weight Kenya sulphadoxine- pregnant women prevented; pyrimethamine; DALYs saved $10-23 Febrile case management 33 University of Ghana http://ugspace.ug.edu.gh Table 4: Cost-effectiveness studies targeting both preventive and treatment malaria/pneumonia interventions SIN Reference, country Interventions Disease/target pop. Study type Outcome CE estimate measure 1 Hansen and Chapman Indoor Residual Spraying; Malaria, pneumonia, DALYs saved $35-466 (2008) Treatment as an outpatient; Others Zimbabwe Treatment as an inpatient; Other interventions targeting various diseases 2 Morel et al. (2005) Insecticide treated bed nets; Malaria Model using DAL Y averted 1nt.$9-151 Sub-Saharan Africa Indoor residual spraying; Mixed population retrospecti ve Case management with CQ; costs Case management with SP; Case management with non ACT; Case management with ACT; 1PT with SP in pregnancy 3 Akhavan et al. (1999) Case treatment; vector control Malaria, Programme Lives saved $2672 Brazil Mixed pop. evaluation DALYs saved $69 4 Goodman et a1. (1999) Childhood-related preventive Malaria, Model using DALYs averted $1-85 Sub-Saharan Africa strategies; under-fives and hypothetical Pregnancy-related preventive pregnant women population strategies; based on Case management strategies model life table 34 University of Ghana http://ugspace.ug.edu.gh 2.5.4 Cost-effectiveness of malaria/pneumonia home management interventions Table 5 presents a summary of studies on the cost-effectiveness on home management interventions. In all, five studies are presented - four of the studies on Africa (i.e. one each in Zambia, Uganda, Kenya and Ghana) and one being a multi-country analysis. The studies cover the period 2006 and 2011. All five studies evaluated malaria interventions with four targeting children and one on all age groups. The study designs included randomised placebo-controlled trial, programme evaluation and Markov modelling. Four of the studies analyzed costs from provider perspective and two used societal perspective. With the exception of Conteh et al. (2010) and Chanda et al. (2011) who use cases averted and cases diagnosed and treated respectively, the other three studies used DALY s as outcome measure. Goodman et a!. (2006) evaluated a training programme for shopkeepers and communities in rural Kenya aimed at encouraging appropriate use of over-the-counter antimalarial drugs for childhood fevers. Lubell et al. (2010) used a Markov model to evaluate the cost-effectiveness of a programme that treated febrile children presumptively with pre-packed antimalarials provided by trained community members in Uganda. Similarly, Chanda et al. (2011) evaluated cost-effectiveness of community distribution of pre-packed antimalarials using community health workers. Conteh et al. (2010) evaluated the cost-effectiveness of using community volunteers to deliver different regimen of intermittent preventive treatments in children (IPTc) in rural Ghana. Finally, Tozan et ai. (2010) assessed the cost-effectiveness of using pre- referral rectal artesunate to treat children suspected to have severe malaria in areas with little access to health facilities. This was done alongside providing advice to caregivers to comply with referrals. 35 University of Ghana http://ugspace.ug.edu.gh However, only two of the studies reviewed examined community distribution of pre- packed ACTs, one using community drug distributors (Lubell et al., 2010) and the other using community health workers (Chanda et al., 2011). To date, these are the only studies on cost-effectiveness of home treatment of malaria using pre-packed ACTs. This review shows that there is paucity of studies on cost-effectiveness of HMM approach recommended by the WHO (WHO, 2004) and adopted by eighteen African countries by 2008 (Ajayi et al., 2008) The use of DALY s as outcome measure makes the cost-effectiveness estimates comparable. Thus, home management programmes could be used to avert one DALY with as low as US$3.85-18.38 (Goodman et aI., 2006). None of the studies targeted fevers due to malaria and pneumonia, which the current study does. Besides, the only study evaluating home management interventions in Ghana deals with malaria prevention (Conteh et al., 2010). The use of provider perspective by Chanda et al. (2011), Tozan et al. (2010), and Goodman et al. (2006) in costing home management interventions raises concerns regarding the omission of key costs. This is because caregivers incur direct and indirect costs as part of home (or community) interventions. The use of provider perspective ignores these important costs and is likely to present costs that do not truly represent society. Besides, Conteh et al. (2010) did not seem to cost the time of community-based volunteers through whom the interventions were delivered. This leads to the omission of such an important cost type from the analysis. 36 University of Ghana http://ugspace.ug.edu.gh Table 5: Empirical stu d'res on cost-e ffectiveness 0fh orne management 0f mrnalaarni a/tm/ eumoma SIN Reference, Interventions Disease/target Study type Costing Outcome CE estimate country, evaluated pop. perspective measure cost year I Chanda et al. Home management Malaria, - Provider Cases $4.22 (2011), Zambia using pre-packed Mixed population appropriately 2009 ACTs & RDT diagnosed and treated 2 Tozan et al. I dose of rectal Malaria, - Provider DALYs Int.$77-1173 (2010), Africa artesunate Under-five averted 3 Lubell et al. Home management Malaria, Markov Societal DALY - (20 I0), using pre-packed Under-five modelling averted Uganda, 2007 ACTs 4 Conteh et al. rPTe using Malaria, Randomised Societal Case averted $67.77 -211.80 (2010), Ghana Amodiaquine Under-five placebo- 2005 Artesunate or SP controlled trial 5 Goodman et al. Educational Malaria, Programme Provider Death $105.92-505.42 (2006), Kenya programme for Children 59 months). The secondary target population included caregivers of all children and community members who served as CBAs. Though this secondary population did not receive treatment, they played key roles in implementing the interventions at the community level. 3.1.4 Trial components The trial comprised three different evaluations as shown in Figure 2 - social science; epidemiology; and economic evaluation. 42 University of Ghana http://ugspace.ug.edu.gh The social science component developed and implemented appropriate health communication messages for home management of fevers. The activities of this component were done by assessing health promotion channels through which households could easily access and make use of the interventions, developing appropriate materials and implementing and evaluating the effectiveness of these strategies. The epidemiology component developed the design and implemented the community-based interventions that aimed at improving early recognition and appropriate treatment of fevers in children under-five years of age and assessed its impact on morbidity and mortality. This study is the economic evaluation component which uses CEA to evaluate the interventions. Figure 2: Components of the home management trial Home management trial I I " Epidemiology Social Science, I' Economic Evaluation Evaluation L Ev~luation j I I I ': ,,:.', ,r" Impact Assessment Morbidity & 'bri- Mortality Analysis of cornmunrcarfonstrategies I' '"-~8~~'~i]~' 3.1.5 Trial design and interventions The HMT was a controlled trial with the interventions introduced in a step-wedge manner (Chinbuah et aI., In preparation). It involved a cluster randomised controlled design with two intervention arms: 43 University of Ghana http://ugspace.ug.edu.gh Antimalarial-only arm (AAQ): Febrile children in these clusters received artesunate- amodiaquine if diagnosed by CBAs as exhibiting signs of uncomplicated fever. First dose treatment was observed by the CBA. Antimalarial plus antibiotic arm (AAQ+AMX): Febrile children in these clusters received artesunate-amodiaquine plus amoxicillin if diagnosed by CBAs as exhibiting signs of uncomplicated fever. Again, first dose treatment was observed by the CBA. The step-wedged design offered the opportunity to study a "control group" which comprised all clusters before they were randomised into the two treatment groups. The 376 communities in the district were grouped into 115 clusters each comprising approximately 100 children under-five years. The cluster design was chosen for the following reasons: First, the design was chosen to prevent treatment cluster contamination. Through the social science component of the trial, caregivers were educated to recognise the signs and symptoms of fever. If randomisation was done on an individual basis, caregivers in the control group would be exposed to the intervention which could likely to change their behaviour. Further, the cluster design enhances subject adherence through interaction between caregivers and community members (i.e. social pressures, expected behaviour). In the first year of the trial, no intervention was introduced; all study communities theoretically served as control. In the second year, CBAs were selected and trained after which two-thirds of the clusters were randomised to receive the interventions (AAQ or AAQ+AMX) while the remaining one-third served as the control. In the 44 University of Ghana http://ugspace.ug.edu.gh third year, the control clusters were randomised to receive AAQ or AAQ+AMX as shown in Table 7. Table 7: Horne management trra. I d esi.gn Study Year Intervention Arm Controlled group Arm I Arm 2 1 Baseline Baseline Baseline 2 ACT Only ACT+AMX No intervention '1 J ACT Only ACT+AMX ACT Only or ACT+AMX 4 Devolution to health system 3.1.6 Data collection and trial phases During the first year, baseline data on morbidity (haemoglobin, parasite density counts and anthropometric measurements) were collected through a household survey. Subsequently, two rounds of morbidity surveys were conducted in each year, one before and the other after the major rainy seasons each preceded by sensitization of communities. The surveys were conducted to collect data on morbidity, behaviour change and mortality. During each survey, ten clusters were covered (three randomly selected from each arm and one chosen at random through a ballot). All children within the selected clusters constituted the sample for the survey. Caregivers of these children were also asked questions related to treatment-seeking, cost of treatment and travel and waiting times (i.e. for those who visited health facilities). The implementation of the trial was structured into four phases: Formative phase: Activities undertaken during this phase included sensitization of health workers, and community entry and selection of CBAs. Operational phase: Activities undertaken included training of CBAs and field supervisors, and monitoring ofCBAs. 45 University of Ghana http://ugspace.ug.edu.gh Intervention phase: Activities undertaken included implementation durbars and distribution of drugs, monitoring and supervision of CBAs, and implementation of education programmes. Post-intervention phase: The main activities undertaken here were closing durbars. The trial activities were grouped into three main categories - training, community mobilization and monitoring and supervision. Training programmes included initial and refresher trainings for CBAs and field supervisors. The purpose of the training was to ensure continuous update of skills of CBAs and field supervisors. Community mobilization included community sensitization and selection of CBAs, information, education and communication activities. The third category was monitoring and supervision of CBAs by field supervisors. 3.1.7 Community-based agents Community-based agents are sometimes referred to in literature as Community Health Volunteers, Village Health Volunteers, Lay Health Workers, or Community Drug Distributors (Haines et aI., 2007). These CBAs were community members whose services were used to implement the interventions. The CBAs were trained by health workers and project staff to diagnose uncomplicated fever in children under-five years of age and administer appropriate treatment at the community level while referring complicated cases to the nearest health facilities. Opinion leaders and members of communities concerned played a key role in identifying and selecting community members for training as CBAs. The selection was done during sensitization programmes. The project team, together with health 46 University of Ghana http://ugspace.ug.edu.gh care workers at the sub-district level, met with opinion leaders and community members and explained the objectives of the HMT to them, after which community members were asked to nominate people who had the requisite skills to be trained as volunteers. Community members came up with some criteria - such as residence, availability and dedication - which aided them in their nominations. By the end of the trial, 660 CBAs had been selected and trained. 3.1.8 Health facilities The district and sub-district health facilities played key roles in sensitizing communities and educating them on the importance of the trial. Staff from the health facilities served as facilitators during training of CBAs. Health facilities also recognised referrals from CBAs, attended to such referred cases and recorded details into books provided by the trial secretariat. 3.2 Cost-effectiveness study 3.2.1 Study population The primary and secondary population used in the study (Le. children under-five, caregivers and CBAs) has been described under section 3.1.3. 3.2.2 Sampling method and sample size for indirect cost For the purpose of recording the travel time of caregivers and the time of CBAs that went into treatment and supervision (which were used to calculate the indirect cost of the study arms), four-day training activity (one day in each sub-district) was held for a sample of CBAs. The sample size was calculated using Epilnfo software. The following parameters were used in calculating the sample size: the population of active CBAs as at January 2009 (i.e. 603); expected frequency of 50% (by default) 47 University of Ghana http://ugspace.ug.edu.gh and the worse expected of 40%. A total of 91 CBAs were selected and trained. The CBAs were randomly selected from the two arms using population proportional to size. 3.2.3 Data collection and data sources Dodowa Health Research Centre (DHRC): The HMT was organised and implemented by DHRC. Trial cost data were obtained from financial and logistics records of the Centre. Travel time for CBAs and supervisors attending trainings and other activities were also obtained from the trial records. In addition, the useful life of capital items was collected from DHRC accounts office and the store of the District Health Administration (DHA). Caregivers: Data on sources of treatment utilised by caregivers was collected from the survey records and used in estimating the proportion of under-five fever cases seeking care from various treatment sources. These proportions were used in estimating the savings of the HMT. In addition, data on three categories of cost (i.e. direct, indirect and intangible costs) incurred by caregivers were collected from the survey records and CBA records. The direct and indirect costs, together with effectiveness data, were used to estimate the CERs while the intangible costs were described. Community-based agents: Data on quantity of drugs used, treatment records of children, caregiver time and time of CBAs that went into treatment and supervision were obtained from CBA records. 48 University of Ghana http://ugspace.ug.edu.gh WHO-CHOICE database: Unit cost estimates for treatment at health facilities were obtained from WHO-CHOICEl (Choosing Interventions that are Cost-Effective). The WHO-CHOICE database has costs of all components of treatment - except drugs and diagnostics - updated for 2007 cost estimates. Costs of drugs and diagnostics were obtained from the baseline survey data. These cost estimates were used to estimate the cost saving of the HMFs approach. 3.2.4 Data collection techniques and tools Dodowa Health Research Centre: Data on the trial costs were collected from receipts, invoices and other financial records. Caregivers: Data were collected from caregivers through interviews during the baseline surveys, using structured questionnaires (see Appendix I). Data on caregiver travel time were also collected through interviews using time-tracking forms (see Appendix II). Community-based agents: The CBAs were provided with sick child forms on which they recorded treatment statistics such as demographics of sick child, child's condition and drugs given to the child (see Appendix III). Further, CBAs were given forms to record the time they and caregivers spent on treatment of a child and the time they spent with field supervisors during bi-weekly supervision visits (see Appendix II). All forms were submitted weekly or monthly to field supervisors who monitored activities of CBAs. I WHO-CHOICE has estimates of treatment costs for various illnesses for different levels of health facilities for 192 countries, including Ghana. 49 University of Ghana http://ugspace.ug.edu.gh 3.2.5 Quality control Pre-data collection stage: Selection and training: Field workers were selected through written examinations and interviews and were trained on data collection techniques and tools. The trainings were facilitated by the trial staff and others with considerable experience in research. Pre-test of questionnaires: Questionnaires and forms were pre-tested after the trainings and then reviewed to ensure relevance, appropriate wording and smooth flow of questions. Pre-tests were done each time before actual field work begun and ensured minimization of errors from the field. Data collection stage: During data collection, field workers were monitored by supervisors to ensure compliance. Supervisors led the field workers to edit completed questionnaires and forms daily. The editing ensured that errors and inconsistencies were dealt with. Data entry: After each data collection exercise, data were entered into Epidata using double entry and logical checks. Verifications were then run for the double entry to check and correct errors committed by data entry clerks. The trial had a data manager who ensured data were entered and verified properly. 3.2.6 Data analysis 3.2.6.1 Estimating direct programme costs Costing was done from the societal perspective, which considers costs from the perspective of households (transport, food and related expenses) and the health system. The study estimated both financial and economic costs. Financial costs were 50 University of Ghana http://ugspace.ug.edu.gh the actual expenses made (i.e. salaries, supplies, rent) whilst economic costs captured actual expenses plus donations and the time of volunteers. Whereas financial costs were only analyzed during the cost analysis, economic costs were used in the cost- effectiveness analysis. Three categories of costs were considered in the analysis; direct costs, indirect costs and intangible costs. Direct costs are the direct expenses incurred on the programme activities and have capital and recurrent components. Capital costs: These costs were incurred on items with useful life of more than one year (i.e. vehicles and equipment). The cost information was based on the entire trial period, from the pre-intervention phase to the end of the intervention phase. Capital costs were annualized using equivalent annual cost framework, with the useful life of items and the recommended discount rate of 3% (Gold et al., 1996). Discounting refers to the practice of reducing the "future" costs (or benefits) into the present show that future costs (or benefits) may not carry the same weight as present costs (or benefits). Capital items have useful life of more than one year. Therefore, in order to calculate the equivaient annual cost of such items, the original costs must be discounted. The useful life of capital items were estimated on the basis of discussion with the Accountant and the Stores Manager at the DHRC and DHA. The useful life estimates were: five years for vehicles; three years for motorcycles and helmets; two years for television sets, flip chart stands, generator, digital camera and DVD players; and one and half years for VCDs and extension cables. Annual capital costs were calculated for each activity. Recurrent costs: Recurrent costs included cost of items with useful life of less than one year. The items were medical supplies (e.g. drugs), stationery supplies, allowances and salaries of staff involved in undertaking activities and implementing 51 University of Ghana http://ugspace.ug.edu.gh the interventions at the community level, and cost of vehicle operation and maintenance. Building spaces were considered as recurrent and market rental costs were used for them. The time of trial personnel and health workers involved in training CBAs and other activities were valued using their gross monthly salaries (inclusive of social security) and added to the recurrent cost of the respective activity. Total direct cost: Total direct programme cost was calculated as a summation of capital cost and recurrent cost for each year. These costs were divided into the two study arms using the proportion of CBAs and the number of communities under each arm. However, cost of drugs which were specific to particular study arm was allocated to the respective study arm. 3.2.6.2 Estimating indirect costs Indirect costs were productivity losses. For the trial, this included the time of CBAs who served as the centre around which the activities were implemented but were not remunerated for their services. Caregivers also gave up some productive hours in order to accompany their sick children to the CBAs. Indirect cost was analyzed using the human capital approach. This approach estimates the value of productive time lost to illness by multiplying the time lost with the proportion of income that would have accrued (Asante et aI., 2005). Data on the time of caregivers and CBAs were collected by CBAs. This study used the average agricultural labour wage per day because the predominant economic activity in the study district is agriculture. Admittedly, the use of the local agricultural wage rate may have led to an over-estimation of costs since the approach assumes that all individuals are employed and allocates income to individuals who are unemployed. 52 University of Ghana http://ugspace.ug.edu.gh However, the unavailability of unemployment data for the sector limits the options available. Besides, some of the caregivers were engaged in non-market activities (e.g. housewives). Therefore, though these people may be classified as unemployed, an estimation of their indirect (or real) cost is important. 3.2.6.3 Estimating total costs The method for allocating the total costs to the two study arms depended on the activity which generated the costs. For instance, for training programmes, the proportion of CBAs belonging to each arm relative to the total number of CBAs was used to allocate the total costs. For community mobilization activities, the proportion of communities in each arm relative to the total number of communities was used. However, for cost items that were specific to an arm (e.g. cost of antibiotics drugs for AAQ+AMX), the entire cost was allocated to the respective arm. Total Financial costs: The total financial costs were calculated as the recurrent cost plus the capital cost. Total economic costs: The total economic costs were calculated as direct costs plus indirect costs. Costs in the local currency (Ghana Cedi) for each year were adjusted to 2009 (final year) costs using the Consumer Price Indices (CPI) for health goods for the respective years. The CPIs were obtained from the Ghana Statistical Service2. The calculation of total cost was done in Microsoft Excel spreadsheet. 3.2.6.4 Savings In the Dangme West District, about 75% of all under-five fever cases used home- based remedies and over-the-counter providers (Nonvignon et aI., 2010) as the first 2 CPIs were obtained from various quarterly bulletins of "Consumer Price Index" published by the Ghana Statistical Services. 53 University of Ghana http://ugspace.ug.edu.gh point of call with 25% seeking care from health facilities. The proportion of under- five fever cases seeking treatment from various health care options and unit cost of these services were obtained from the baseline survey and WHO-CHOICE database. Savings on treatment: The savings of the HMF were estimated as the product of the proportion of cases under standard care (i.e. seeking care from each health care option), the total number of cases treated (under the HMF) and the unit cost of services under the standard care options. There were savings on drugs (provided without charge under the HMT) and diagnostics (there were no diagnostics under the HMT, hence, no cost on these), transport cost, indirect cost (i.e. waiting and travel time valued using the indirect cost estimation approach discussed earlier). Savings on funerals: Savings on funeral expenses were also calculated as the average cost of funeral of under-five death multiplied by the total number of deaths averted by the HMF programme. Direct funeral savings were on coffin, cloths for the corpse, and refreshment for mourners. The number of work days lost by caregivers as a result of the funeral constituted indirect savings. The cost savings were also caJcu lated in Microsoft Excel spreadsheet. 3.2.6.5 Intangible cost Intangible cost constitutes the pain, suffering and anxiety incurred by caregivers who had to attend to their children during episodes of fever. Such costs are difficult to quantify and are, therefore, not usually included in costs and cost-effectiveness analyses. However, since intangible costs could be substantial, especially in developing countries, the study collected data on these using questionnaires with a five-point Likert Scale (see Appendix IV). The scale ranged from" 1" representing "strongly agree" to "5" representing "strongly disagree". The answers between the 54 University of Ghana http://ugspace.ug.edu.gh two points are "agree", "neither agree nor disagree", and "disagree" in that order. Four items were analyzed. Item 1 assessed whether a child's illness was a source of worry to the caregiver. Items 2 and 3 sought to elicit the views of caregivers on whether or not the performance of their normal daily tasks, including leisure activities, are affected by the illness of their children. Item four assessed whether the iilnesses of under-fives pose economic difficulties to caregivers at home due to extra expenses they have to make. The distributions of responses to these four items were presented using cross tabulations. Further, composite scores' were generated in STAT A (StataCorp, 2007) for items 1, 2 and 4 based on standardized values of zero mean and standard deviation equals one. The index generated was then used to compare if there were significant differences between AAQ and AAQ+AMX. The results of the intangible cost data were not included in the cost-effectiveness analysis. 3.2.6.6 Estimating programme effectiveness The effectiveness indicators used in the study were the number of deaths averted and disability-adjusted life years (DALYs) averted. These indicators were estimated as follows. Cases treated: The number of eligible fever cases treated for each study arm was obtained from the sick child forms. In instances where CBAs treated fever in people other than the target group (i.e. two month to 59 months) or treated cases that were ] A two-sample t-test (with equal variances) was calculated in STAT A with command "ttest intangible, byf studyarrn)" with the view to test differences between mean scores of AAQ and AAQ+AMX. diIT=mean(AAQ) -mean(AAQ+AMX); 1-10: diff= 0, 55 University of Ghana http://ugspace.ug.edu.gh not fever, these were regarded as ineligible and were not counted as part of the eligible cases treated. Anaemia cases averted: The number of anaemia cases were collected during biannual morbidity surveys and prevalence was assessed by determining the proportion of surveyed children with haemoglobin levels lower than llmg/dJ. Deaths averted: The number of deaths from each arm and the control were obtained from the mortality records of the project. However, since the population of under- fives vary across the intervention arms and the control as a result of the trial design, standardized mortality ratios (SMRs) were calculated to make the deaths comparable across the control and the intervention arms. The indirect standardization method (Kirkwood & Sterne, 2005) was used to calculate SMRs. The method is used to calculate the expected number of deaths using mortality rates from a standard population and contrasting this expected value to the number of deaths observed in the specific index population (i.e. the control in this case). The SMR is the ratio of the total observed number of deaths in the index population to the expected deaths (Kirkwood and Sterne, 2005 pp. 268-269). Therefore, assuming that AAQ and AAQ+AMX had the same population as the control, the expected number of deaths in these arms were calculated and used for the cost-effectiveness analysis. For each study arm, deaths averted were calculated as the difference between number of deaths in the control and the expected number of deaths in the respective intervention arm. Disability-adjusted life years averted: Disability-adjusted life years are the sum of years of life lost (YLL) and years of life lived with disability (YLD) due to a condition. However, discounted YLL (DYLL) which is the mortality component of 56 University of Ghana http://ugspace.ug.edu.gh DALYs presents the majority of malaria (Goodman et al., 2006) and presumably pneumonia disease burdens. An assessment of the impact of an intervention on DALY s requires the calculation of DALY s with and without the intervention (Rushby & Hanson, 2001). The following formula by Murray and Lopez (1994) as used by Rushby and Hanson (2001) was used to calculate DYLL in a Microsoft Excel Spreadsheet: Where r = discount rate K = age weighting modulation factor C = age weighting correction constant f3 = parameter from the age weighting function a = age of death L = life expectancy at age Q. The following parameters were used; average age of death of 2 years (from mortality records of the project); average life expectancy at age 2 of 6S.4 years (Ghana Statistical Service, 2011b); no age weighting (Goodman et al., 2006) and discount rate of 3 percent (Murray & Lopez, 1996a; 1996b). This study calculated the average DYLL and multiplied the result by the deaths in the control and each intervention arm to obtain the total DALYs for AAQ and AAQ+AMX. Then, the difference between DALYs of each arm and the control was the DALYs averted by each arm. 57 University of Ghana http://ugspace.ug.edu.gh 3.2.6.7 Cost-effectiveness Ratios (CERs): Cost Effectiveness Ratios for the intervention arms were calculated as follows: CER, Total cos t,= ------'--- (2) Effec t , i = 1,2 Where CER; = cost-effectiveness ratio; i = Intervention arms of the trial (i.e. AAQ or AAQ+AMX) The CERs are average CERs and not incremental CERs. Each of the effectiveness measures was used alternatively as the denominator to calculate the CER. Two sets of CERs were calculated, one using the standardized effects and the other using the unstandardized effects. The calculation of CERs was done in Microsoft Excel. Figure 3 diagrammatically summarizes the flow of the analysis. Fevers have two effects; morbidity and mortality, which bring about direct costs, indirect costs and intangible costs. Direct costs consist of programme costs (capital and recurrent costs). community costs (mainly out-of-pocket expenses), and health facility costs (i.e. costs incurred by caregivers at facilities). Indirect costs consist of mainly community costs (i.e. caregiver's time lost from work due illness of child) while intangible costs are psychic costs associated with the illness (e.g. pain and suffering, and loss of leisure time). There are also costs incurred by the health facility in treating the sick child and these are made up of mainly recurrent costs (e.g. personnel time and drugs). The first two categories (i.e. direct and indirect costs) were added to give total intervention cost while the third category (i.e. intangible costs) was analyzed and 58 University of Ghana http://ugspace.ug.edu.gh described. The unit costs and proportion of cases that would have been treated at standard care then used to calculate the savings. The total intervention costs and the effectiveness indicators were used to calculate the CERs. 59 reg-elre J: Ocsg-nlmmaCrcUnpirveseernsraitryto on f oGr'choasrn-earr e hcntvtpen:/e/susgsapnaarcyser.sug.edu.gh · Programme costsCapital e.g. vehicles, · equipment t--- ,.. Recurrent e.g. supplies,personnel, building space, vehicle operation &maintenance Community costs ,.. Direct costs· Caregiver out-of-pocket ·· Implementation costsDirect~ I-- expenses (e.g. treatment, Community costs ~costs I--drug, diagnostics, · Health facility costs Total rl transport)L_. interventionMorbidity I cost Communitv cost I-- ~ Indirect costs Fever · Caregiver time lost from · Community costs V t-- work i.e. number of days Indirect spent away from work for - f---+ cost f--+ · treatment- CSA time treating sick hildL.j ·Intangible costs'--- ~· Psychic costsMortality Psychic cost·· DescriptionIntangible r--. Analysis using Likert I- ~ cost " J Savings · Health facility costs L ] Recurrent e.g, supplies, personnel, building space, CERs vehicle operation &maintenance l Effectiveness indicators·· Deaths avertedDALYs averted 60 University of Ghana http://ugspace.ug.edu.gh 3.2.6.8 Estimating budget impact of Home-based Care Currently, the NMCP is iruplementing a horne-based care approach, targeting malaria, pneumonia and diarrhoea. Elements of the HBC are similar to the trial conducted in Dangme West District (e.g. use of CBAs). The implementation of the HBC programme started in 34 districts in 2011 and is expected to be scaled to cover additional 36 districts and 26 districts in 2012 and 20 l3 respectively. Within each district, the implementation started in selected communities and the coverage will be scaled up gradually until the entire district is covered. The financial cost estimates from the trial were adjusted to cost of training and supervising CBAs and providing them with medicines to work under the HBC programme. The following parameters were used: 200 children per CBA; average fever episodes per child per year of 3; market price of medicines. The under-five population figures for 2011 - 2013 for the districts to be covered by HBC were obtained from Ghana Statistical Service (Ghana Statistical Service, 2011 a) (see Appendix V). , Using the above parameters, a three-year budget for scaling up HBC according to NMCP roll out plan was estimated (in 2011 Ghana Cedis). The estimated budget excludes other operational expenditures of the proposed programme such as administrative costs. Consumer Price Indices for 2011 were used for the calculation with an assumption of fairly constant CPIs over 2011 to 2013. The budget was estimated under two different scenarios. The first scenario assumed that under the HBC, antimalarials only are used whilst the second scenario assumed that antimalarials plus antibiotics are used. 61 University of Ghana http://ugspace.ug.edu.gh 3.2.6.9 Sensitivity analyses First, one-way sensitivity analyses were performed using the following parameters: lower (2%) and higher (5%) discount rates; useful life of vehicle of 8 years instead of 5 years used in calculating capital cost; national daily minimum wage rate instead of the local daily wage rate in estimating indirect cost; alternative drugs (Arthemeter Lumefantrine for and Co-trimoxazole) which are also recommended by the Ministry of Health instead of Artesunate Amodiaquine; During the trial, CBAs offered their services as volunteers who were not remunerated. Aan assumption that CBAs were given allowances equal to their indirect costs was used as part of the sensitivity analysis. Further, an assumption of higher attrition rates among CBAs was also used. A further assumption used was that mortality rates were 10% lower and between 1% and 20% higher than the observed mortality rates. Finally, multi-way sensitivity analysis was also performed, using four scenarios as shown in Table 8. Table 8: Parameters used in sensitivity analyses Scenario Parameters used 3% discount; 10% mortality reduction; national wage rate.; AL & AL+CT 2 5% discount; 15% mortality increase; national wage rate.; AL & AL+CT 3 50% CBA attrition, CBAs paid allowance, other parameters same as in base case scenario 4 5% discount; 15% mortality increase; AL&AL+CT; national wage; 50% CBA attrition + CBA allowance 3.2.7 Ethical considerations of the study Approval: Approval was sought from CBAs to provide services without remuneration. This was done before they were recruited for the study. Additionally, the District Health Management Team and health facilities in the study district were informed of the study and approved of their participation in the training of CBAs and to attend to cases referred by CBAs. 62 University of Ghana http://ugspace.ug.edu.gh Selection of subjects: Communities in the study area were grouped into clusters, which were randomised to receive different treatments. All children in the selected communities were included in the study. In order to prevent the effect of pressure from participants, all children in the same community received treatment with the same drug and all communities in the same cluster received treatment using the same drug. Risk/benefits of the study: Apart from possible side-effects of drugs used, no risks were associated with this study. Direct benefits to the subjects included free health service in terms of free drugs for cases treated by CBAs. As benefit to the health sector, findings would inform health policymakers in their decision to scale up HBC. Privacy!confidentiality: In order to ensure that the privacy of subjects were protected, data collected from subjects were safeguarded and used only for the purpose for which they were collected. The confidentiality of participants was assured. Voluntary participation and compensation: Caregivers were informed that participation in the study was voluntary. Furthermore, no compensation was to be made for participation. Data storage and usage: Data were collected through surveys. The data collected were stored in a database created for the trial and used for research purposes only. No unauthorized persons were granted access to the data. Informed consent and assent: Consent for data collection was sought from caregivers. First, objectives of the project and the possible impacts and side effects of the drugs were explained to the caregivers. Before each survey, caregivers were made to sign Of 63 University of Ghana http://ugspace.ug.edu.gh thumb print consent forms to indicate assent. Each caregiver was then given a copy of the signed consent form to keep. Conflicts ofinterest. There were no conflicts of interest. Funding for the study: The study was funded by the UNICEF/UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases and the Knowledge Enriches Programme of the Netherlands Organization for the Advancement of Tropical Research. Ethical clearance: Ethical clearance was granted by the Ethical Review Committees of the Ghana Health Service and the World Health Organization. 3.2.8 Reliability and validity of results The implementation of the study covered an entire district. Consequently, conclusions drawn from the results of this study provide reliable and valid estimates and may be generalized to other districts with similar setting. However, the findings of such a study may not be generalized to urban districts since health and other conditions in urban areas differ from those in rural areas. 3.2.9 Study limitations The study was bound by the following limitations: 1. The study sought measure the relevant costs and effects associated with the home management of fevers in children under five years. it must be noted that not all costs were measured; only relevant ones were measured. Besides, any measurement errors in the costs and effects automatically feed into the cost- effectiveness ratios. 64 University of Ghana http://ugspace.ug.edu.gh 2. In estimating the three-year budget of the HBC strategy, the unit costs derived from the trial were applied, which may have led to an over or under-estimation of the budget. This point is made because the trial made use of more CBAs and more frequent supervision that is actually the case under routine conditions. 3. Finally, some of the assumptions used in estimating costs, effects and cost- effectiveness of the HBC may be more theoretical, implying that real-life situations may differ slightly. For instance, it was assumed that mortality rates were not different across the districts. However, in reality, mortality rates differ across districts in Ghana. 65 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR 4.0 RESULTS This chapter presents the (financial and economic) costs and the effects as well as the CERs of the HMT interventions. The sensitivity analyses of the costs and cost- effectiveness estimates are also presented in the chapter. Additionally, the estimated CER and budget of the HBC are presented in the chapter. 4.1 Home management trial costs 4.1.1 Total financial cost of Malaria-only Arm (AAQ) The financial cost represents only direct costs. The total financial cost of AAQ was GH:ou not give all ;the drugs as prescribed? '.-.- ".. r ~ 88. NA 22 How far is your home from the place you obtained the Not far Orthodox medicine'rWe could make this "how long did "Far' '., take from your home to the place you obtained the Very Far Orthodox medicine (in hour/minutes)? 88,NA .. In that case they only have to write hour or minutes and we determine if it is far or near. 23 -If you took transport from your home to the place you obtained the Orthodox medicine, how much did you GHO ......................... spend on transport? 24 Section On Cough (Ask this section only if child had 1. Yes 2. No (if cough in the last 2 weeks.) no skip to 21) 88. NA 25 When (name) had cough did you discuss with anybody? 1. Yes 2. No (if no skip to 21) 88. NA 26 If Yes, who did you discuss with? I. Mother 2. Father 3 Grandma 4. Grandpa 5.Aunt 6. Uncle 7. Sibling 8. Other specify ........ 129 University of Ghana http://ugspace.ug.edu.gh 88. NA 27 1. Gave tepid Sponging What did you do at home when you noticed 2. Gave orthodox medicine at that (Mention name of child under five) had home cough? 3. Gave herbal treatment at home 4. Gave an enema 5. Did nothing, go to chemical seller to buy drugs 6.Did nothing, go to Health facility 7. Did nothing, go to CBA for Rectal Artesunate 8. Did nothing/wait a while 9. Other (specify) . 88. NA 28 What drug did you give your child when he/she Amoxicillin . had cough? ....... 1 Septrin . ..2 Other . antibiotic ..: ~ 3 .. Chloroquine : , ; .. : : A Amodiaquine- Artesunate 5 Other anti- ~. -:,- malarial. 6;."'[ Paracetamol/painkiller .;: 7 "'\ . Multivitamin 8 Iron drops 9 Herbal medications 10 Other (specify) 11 DK 99 Not Applicable 88 29 How soon after the time you noticed your child Same day had cough did you give this drug? Next day Other(specify) . 88. NA 30 How many days after the onset of cough did 1. Same day you seek treatment outside your home for the 2.Next day child? 3. Other specify) . Did not seek treatment outside 88.NA 31 Who took the decision? 1. Self 2. Spouse 3.Both spouses 130 University of Ghana http://ugspace.ug.edu.gh 4. Grandmother /Grandfather of child 5.other (specify) 88. NA If orthodox medicine was given, where was the drug obtained? Drugs at home From neighbour/friend at home CBA Health facility Chemical seller ShoplPharmacy Other (specify) ................. 88. NA 32 Did you give all the drugs as prescribed? 1. Yes 2. No 88. NA 33 If no, how many tablets remained? Number ................................... 88 Not Applicable 34 How much of the syrup remained? 1. Almost full, 2. Half empty 3. Almost empty 88. NA 35 Why did you not give all the drugs as prescribed? 88. NA 36 How far is your home from the place you obtained the Not Far Orthodox medicine? Far , .' ~ 3. Very Far. o· 88.NA .' 37 Section on "FastlDifficult Breathing'{Ask' this section .. ,LYes 2.:No (if no skip to 34) only if child had fast/difficult breathinz) .: ..88. NA 38 When (name) had fast/difficult breathing did you 1. Yes 2. No (if no skip to 34) discuss with anybody? 88. NA 39 If Yes, who did you discuss with? 1. Mother , - 2. Father., .., -- .... _ . ., 3 Grandmao. 4. Grandpa 5.Aunt 6. Uncle 7. Sibling - 8. Other specify ........88. NA 40 1. Gave tepid Sponging What did you do at home when you noticed 2. Gave orthodox medicine at that (Mention name of child under five) had home fast/difficult breathing? 3. Gave herbal treatment at home 4. Gave an enema S. Did nothing, go to chemical seller to buy drugs 6.Did nothing, go to Health facility 7. Did nothing, go to CBA for Rectal Artesunate 8. Did nothing/wait a while 9. Other (specify) .................... 131 University of Ghana http://ugspace.ug.edu.gh 88. NA 41 (If orthodox medicine) What drug did you give Amoxicillin . your child when he/she has fast/difficult ... 1 breathing? Septrin . .2. Other antibiotic 3 Chloroquine . .4 Amodiaquine- Artesunate 5 Other anti- malarial 6. Paracetamol/painkiller ............ 7 Multivitamin 8 Iron drops 9 Herbal medications 10 Other (specify).............. . .11 DK 99 Not Applicable 88 42 How soon after the time you noticed your child Same day - .had fast/difficult breathing did you "give this. .Next day drug? .. Other(specify) , .. :. ~ . 88. NA, .. ' 43 How many days after the onset of fast/difficult 1. Same day breathing did you seek treatment outside your 2. Next day home for the child? 3. Other (specify) . 4. Did not seek treatment _.' outside 88. NA 44 Who took the decision? 1. Mother 2. Father 3 Grandma 4. Grandpa 5. Aunt 6. Uncle 7. Sibling 8. Otherspecify ........ 88. NA If orthodox medicine was given, where was the drug Drugs at home obtained? 2. From neighbour/friend at home CBA Health facility Chemical seller ShoplPharmacy 7. Other (specify) . 88.NA Did you give all the drugs as prescribed? 1. Yes 2. No 88. NA If no, how many tablets remained? Number 132 ~PAJD SCHOOL OF PUBLICHEALTH LIBRARY ~.':'d LEGON University of Ghana http://ugspace.ug.edu.gh · .................................. 88 NA How much of the syrup remained? Almost full Halfempty Almost empty 88. NA Why did you not give all the drugs as prescribed? 88. NA How far is your home from the place you obtained the Not Far Orthodox medicine? Far Very Far 88. NA ------- .•...~- • -~ •• ""'t. 133 University of Ghana http://ugspace.ug.edu.gh APPENDIX II: Time Tracking Tool TREATMENT ARM (PLEASE TICK) COMMUNITY: SUB DISTRlCT: FS CODE SECTION A (TOTAL TREATMENT/CAREGIVER TRA VEL TIME) SN DATE CHILD ID TREATMENT TIME CAREGIVER IN- REMARKS STARTING ENDING TRAVEL TIME TIME TIME HOUR/MINUTES 1 2 3 4 5 -- ,SECTIO~B (TOTAL TlMESPij;NT WITH FIELD SUPER VISOR ,UmNG MONITORING - VISITJ, '_-, -'·.f- MEETINGS DATE' STARTING TIME ENDING TIME 1ST MEETING WITH FS 2NU MEETING WITH FS -_ -;"-0;" ,- _, . -. CCB DATE __ ~ __ 134 University of Ghana http://ugspace.ug.edu.gh APPENDIX III: Sick Child Form MANAGEMENT OF THE SICK CHILD AGE 2 MONTHS UP TO 59 MONTHS Child's ID I CBAName .. Community . Sub district. ... Name of child .. Date ofBirth: .... Age . Date . First visitD Follow lip visit0 Date Sex of child: Male ... . Female . Name of mother .. "If child is less than 2 months old or 5 year s or more, refer immediately. Ask caregiver: What are the child's problems? Observe child and Circle Yes or No against each question or problem Yes Docs child have fever? HOT BODY Hedola I.FNO FEVEn No REFER How many days has the child had fever'! ... Days Yes (FYES REFER Has child been treated for fever within the last month? No CHECK FOR DANGER SIGNS Danger sign present? Yes Is child able to drink or breastfeed? 1'10 [FNO REFF;R Yes Docs the child vomit everything?' IFYES REFER No Yes Has child had convulsion? Is child convulsing now? 11<' YES REFER No Is the child hard to awaken.qr unconscious? Yes· IF YES REFER No Yes Diles the child have cough? No If yes, for how many days has child been coughing? ............ Days Yes IJiYES REFER Does the child have difficult breathing? No If yes, for how many days has child had difficult breathing? Count the breaths of the child for one minute and record. REFER Child not yet I year old (2 months II months) Ifbreathing rate is50 or more refer Ifbreathing rate is Child more than I year old (12 months-59 months) REFER40 or more Yes IF YES, REFER Does child have Chest in-drawing? No Yes IFYES, REFER Does child have Stridor (noisy breathing)? No 135 University of Ghana http://ugspace.ug.edu.gh Docs the child have Diarrhoea? Yes No If yes, for how long has child had Diarrhoea? Yes If YES REFER Is there blood in the stools" No Yes IFYES REFER Does the chi ld have sunken eyes? No Yes IFYES REfER Does skin pinch return slowly? No Yes IFYES REFER Has the chi Id lost a lot of weight (become very thin?) No Yes IFYES REFER Arc the child's palms pale or white" No rFYES REFER Docs the child have swollen feet? Yes No Treatment Chart According to Age of Child Artesunate Amodiaquine (sachet) Give the sick child 2 sachets a day for 3 days Age , 2 months up to l lrnonths Orange Pack ~.,.,,\ :~L~~ Pink Pack ·1 TreatrTI:en(C;:hartAccording.to~e ,of Child . Amoxicillin . the sick child one tablet in (he and one in the evening for J days '-----1 ~ 2 months up to II months CfIE,,, _ - . ._--_ ~ __ .. ~ Silver Pack ~ ~'" ,., Iyear up to 3 years g Gold Pack • 3 years up to 5 years ..,. .,.. -------_........_.---_ .. .................. _ .. _ ..... ._ _ -_ ..Pink Pack ~ ~ If this is a follow up visit Yes (after 2days) Ask and check Is the child well? child's condition No IFNO REFER Action Taken 1. Drugs given and sent home. 2. Drugs given and referred. 3. No Drugs given, referred trnrnedtatetv 136 University of Ghana http://ugspace.ug.edu.gh \, ., APPENDIX IV: l\'Iorbiditysurvey forms 4 and 5 HOME AND COMMUNITY MANAGEMENT OF MALARIA AND PNEUMONIA IN CHILDREN UNDER-FIVE: A CLUSTER RANDOMISED CONTROLLED TRIAL IN SOUTHERN GHANA -TWO-WEEK RECALL OF FEVER SURVEY FORM 4: HEALTH SEEKING CARE MODULE Field Worker COULdl_---L.~_,_J1 Survey Serial No: , , T0 be admm..istere d to Caregrvers or every Chi1ld under 5 woh has been 1ill IiII the past two weeks WHO SOUGHT CARE OUTSIDE THEMSELVES. Name of Caregiver , CAREGIVER .ID Narne of Child .. ,CIllLD ID WHERE DID CAREGNER SEEK CARE FROM? 23. What about-your child's illness caused you to seek care from so Always take my child to the clinic else outside your own home remedy/self treatment ' available (one response) 2. child was not getting better (probe for stage of the illness) 3. Child was getting worse 4. Emergency situation set in 5. Other , . ......................... University of Ghana http://ugs, pace.ug.edu.gh 25 You said ..... was the very first symptom, (Ref Form 3Q3.)How many days after you noticed ..... (Name first symptom) 1. Sam,e .day (within 24 hours= DAY 0) 4. Forth Day (72- 96 hours Day did you seek care outside your house? (assist mother to count) 3) (Use the day of the week to calculate the appropriate answer) 2. Next day (after 24 hours < 48 = DAY 1) 5.More than 3 days < than 1wk (Day , I 4-7 ) . , 3. Third Day ( 48-72 hours = Day 2) 6. More than 7 days " 26. Did you ...,{·.,i'r., 27. Were you asked 29. If you did not pay for this service, this service? to pay for this' did you pay for why did you oat pay? For those who sought care at a health 1. Yes 2. No service? ' this service in facility: 88.NA 99DK 1Yes 2.No Cedis? 1. Exempted 2.m member. 88.NA 99.DK 88.NA 99.DK 88. NA 99.DK - .' ,~/ University of Ghana http://ugspace.ug.edu.gh 31 How much did you pay for transportation? L Cost of transport ............ , 2~Walked 88.NA 32 How many hours/minutes did you spend from home to the facility to seek service at this place SECTION G2: SPECIAL FOOD COST AND PRODUCTIVITY LOSSES OF,CARE GIVER G33 How many days was the child ill? (DAYS) " G34 Did you or anyone prepare any special food (different from. the usual household meal) for the child on the day(s) he/she was ill? 1. Y~,s 2. No 88. NA G3S How much did the ingredients for Total cost preparing the special food cost (in Day 1 of illness Day 2 of illness Other illness day(s) cash and/or in kind in nearest Cedis)? G36 How long did it take to prepare and Day 1 of illness Day 2 of illness Other illness day(s) Total Minutes/hours administer the food? G37 Who normally pays for treatment when any of your children or : your child falls ill? L'Mother 2. Father 3. Other (specify) G38 Did you or someone else in the family prepare a traditional or any - other treatment for the child? 1. yes 2. No (If no skip to Question G41) G39 How much did the ingredients used in the preparation' .of the treatment cost in total? (in cash and in kind to nearest Cedisj; 88. NA G40 How long did it take to prepare the traditional medicine and administer (i.e., during the days when this was done)? (Minutes/hours) 88.NA G41 Who went with the child to the health facility? 1.Mother 2. Father 3. Grandma 4. Grandpa 5. Aunt (Multiple responses allowed) 6. Uncle 7. Sibling 8. Other (specify) 88. NA G42 What would the person have been doing if not to have taken the 1. Work 2. Household chores. 3. Leisure 4. Social activities child for treatment? S.Other (specify) 88.NA G43 Did anyone take your place in this work/activity when y~)Utook the child for treatment? 1. Yes 2. No 88. NA G44 If Yes who worked for you? , 1. Mother 2, Father 3. Grandma 4. Grandpa 5. Aunt 6. Uncle 7. Sibling 8. Other (specify) 88. NA G4S How long did the person work for? (Days/Hours) G46 Did you have to pay this person in kind or cash for the work he/she 1. Yes 2. No 88. NA did for you? G47 If you paid the person who worked for you in cash, how much was it? ~~,~A. G48 If payment was in kind, what is the estimated.monetary va\ue of it? ~ R.1l NA J University of Ghana http://ugspace.ug.edu.gh G49 Because you were spending time looking after the sick child, do l.Yes 2.No 88.NA you think your work has suffered? G50 If Yes, what will be the effect? G51 Normally how many hours do you work in a day? . (Hours) 88. NA G52 Normally how many days do you work in a week? t, . (Days) 88. NA G53 When (name) has fever, how many days do you lose from work? I i (Days) 88.NA Now, I'm going to read a few statements related (name's) illness, After each please tell me the extent to which you agree or disagree with the statement _1_D../!_tions1: . Stronf(/jl_ Agree 2. Agree 3. Somewhat Agree 4.};)isagree 5.· Strongly Disagree) G54 My child's sickness is always a worry to me 1. Strongly Agree 2. Agree 3. Somewhat Agree 4. Disagree 5. Strongly Disagree G55 When my child is well, I am not able to perform all the tasks I need to do at work 1. Strongly Agree 2. Agree 3. Somewhat Agree 4. Disagree 5. Strongly Disagree G56 When my child is sick, I am not able to perform all. the tasks I 1. Str,OnglyAgree 2. Agree 3. Somewhat Agree need to do at work .' 4. Disagree 5. Strongly Disagree G57 Life at home is not normal when my child is well 1. Strongly Agree 2. Agree .. 3. Somewhat Agree 4. Disagree . 5. Strongly Disagree G58 Life at home is difficult when my child is sick 1, Strongly Agree 2. Agree 3. Somewhat Agree 4. Disagree 5. Strongly Disagree G59 My family and I are able to observe our leisure times whether my child is sick or well 1. Strongly Agree 2. Agree 3. Somewhat Agree 4. Disagree 5, Strongly Disagree University of Ghana http://ugspace.ug.edu.gh HOME AND COMMUNITY MANAGEMENT OF MALARIA AND PNEUMONIA IN CHILDREN UNDER-FIVE: A CLUSTER RANDOMISED CONTROLLED TRIAL IN SOUTHERN GHANA -TWO- WEEK RECALL OF FEVER SURVEY FORM 5: REFERRAL MODULE Survey Serial No: Child under 5 who has been ill inthe past two I Date of Visit SERVICES 41. Did you 42. Were you asked 43. How much 44. If you did not pay for this service, receive this to pay for this did you pay for why did you not pay? For those who sought care at a service? service? ' this service in health facility: 1. Yes 2. No 1 Yes 2. No,; . Cedis? 1. Exempted 2.HI member. 88. NA . 99DK 88. NA 99.DK 88. NA 99.DK 88. NA 99. DK \4\ University of Ghana http://ugspace.ug.edu.gh 1. Cost of transport . 2. Walked: 88.NA How many hours/minutes did you spend from home to the' to seek service at this ..I SECTION G: SPECIAL FOOD COST AND PRODUCTIVITY LOSSES OF CARE GIVER 50. When the child was referred was he/she admitted the point of referral LYes 2.No 88. NA 51. How many d~s was the child admitted? iDAYS) 52. Did you or anyone prepare any special food when the child was 1. Yes 2. No 88. NA admitted (different from the usual household meal) ? 53. How much did the Day I of admission Day 2 of admission Other admission day(s) Total cost ingredients for preparing the special " food cost (in cash and/or in kind in I , nearest Cedis)? 54. How long did it take Day 1 of admission Day 2 of admission Other admission day(s) Total to prepare and Minutes/hours administer the food? "- 142 University of Ghana http://ugspace.ug.edu.gh 55. Who normally pays for treatment when any of l.Mother 2. Father 3. Other (specify) _your children or your child falls ill? 56. Who went with the child to the referral point? l.Mother 2..Father 3. Grandma 4. Grandpa 5. Aunt (Multiple responses allowed) 57. 6. Uncle· 7. Bibling 8. Other/specify)What would you have been doing if you were 88. NA not taking the child for treatment? l.Work 2.JHousehold chores, 3. Leisure 4. Social activities5. Other (specifyl 58. Did anyone take your 88.NAplace in this 1. Yes 2. No 88. NA work/activity when you took the child for , treatment? . 59. If Yes who worked for you? 1.Mother 2, Father 3. Grandma 4. Grandpa 5. Aunt 60. 6. Uncle 7. Sibling 8. Other (specify) 88. NAHow long did the person who worked for? 61. (Days/Bours)Didyou have to pay this person in kind or cash 1. Yes 2. No 88. NA for the work he/she did forj'_ou? 62. If you paid the person who worked for you in cash, how much was it? 88.NA 63. If payment was in kind, what is the estimated , monetary value of it? 88.NA 64. Because you were spending time looking after 1. Yes 2. No 88. NA the sick child, do you think your work has suffered? 65. If Yes, what will be the effect? Thank you very much CCB l43 University of Ghana http://ugspace.ug.edu.gh ..I APPENDIX V: Projected under-five population for HBC districts, 2011-2013 2011 2012 0-11 12-23 201324-59 0-11 12;-23 24-59 District months 0-11 12-23months 24-59months Total U5 months months months Total U5 months months months Total U5 Asunafo South 2,575 2,575 7,726 12,877 2,602 2,602 7,807 13,012 2,629 2,629 7,886 13,144 Atebubu-Amanten 2,463 2,463 7,390 12,317 2,489 2,489 7,468 12,446 2,515 2,515 7,544 12,573 Jaman North 3,245 3,245 9,736 16,226, 3,279 3,279 ~837 16,396 .. 3,312 3,312 9,937 16,562 Jaman South 2,593 2,593 7,779 12,965 2,620 2,620' 7,860 13,101 2,647 2,647 7,940 ... 13,234 Sene ,3,079 3,079 9,238 15,39;7 3,112 3,112 9,335 15,558 3,143 3,143 9,429 15,716 Tain 3,302 3,302 9,907 16,512 3,337 3,337 10,011 16,685 3,371 3,371 10,113 16,855 Amansie West 4,460 4,460 13,380 22,300 ,4,544 4,544 13,631 22,719 4,627 4,627 13,882 23,137 *Sekyere Central 3,931 3,931 11,792 19,654 3,986 3;986 11,959 19,932 4,060 4,060 12,179 20,299 *S~ere Afram Plains 3,931 3,931 ,_11,792 19,654 3,986 3;986 11,959 19,932 4,060 4,060 12,179 20,299 Amansie Central 3,182 3,182 9,546 15,9lO' 3,242 3',242 9,725 16,209 3,301 3,30] 9,904 16,507 *Bosome Freho 3,931 3,931 11,792 19,654 3,986 3,9816 i1,959 19,932 4,060 , 4,060 12,179 20,299 Ahafo Ano South 4,784 4,784 14,352 23,920 4,874 4,874 14,622 24,370 4,964 4,964 14,891 24,818 Elura-Seko 3,225 3,225 9,676 16,127 3,286 3,286 9,858 16,430 3,346 3,346 10,039 16,732 \ 44 University of Ghana http://u,gs'.pace.ug.edu.gh ., I Nkwanta North 4,482 4,482 13,445 22,408' 4,488 4;488 13,464 22,440 4,491 4,491 13,474 22,456 "Nkwanta South- - - - I- - ' .- *Biakoye - - - 2,375 2,375 7,124 11,874 ' ::2,378 2,378 7,135 11,891 2,380 South Dayi 2,380 7,140 11,900 1,136 1,136 3,409 5,682 ,1,138 1,138 3,414 5,690 1,139 Adaklu Anyigbe 1,139 3,416 5,694 1,507 1,507 4,520 7,533 1,509 r ,509 4,526 7,544 1,510 1,510 4,530 7,549 Ga West 12,863 12,863 38,590 64,31 i' 13,220 q~220 39,659 66,099 13,579 13,579 40,736 67,893 AQa South - ,- - " - - - - - . ~.. . - Dangme West 2,853 2,853 8,559 14,266 2,932 2:932 8,797 14,661 3,012 3,012 9,035 15,059 *Akyemansa 4,065 4,065 12,194 20,323 4,068 4,068 12,203 20,338 4,076 4,076 12,229 20,381 Akwapim South 3,670 3,670 11,010 18,3S0 3,673 3,673 11,018 18,364 3,681 3,681 11,042 18,403 Akwapim North 3,270 3,270 9,809 16,348 3,272 3,272 9,816 16,360 3,279 3,279 9,837 16,395 Atiwa 2,691 2,691 8,0721 13,454 2,693 2;693' 8,078 13,464 2,699 2,699 8,096 13,493 Kwaebibirem 5,412 5,412 16,237 27,062 5,416 ,5,416 16,249 27,082 5,428 5,428 16,284 27,140 Kwahu North 4,108 4,108 12,325 20,542 4,111 4,111 12,334 20,557 4,120 4,120 12,360 20,601 I *Kwahu East 4,065 4,065 12,194 20,323 4,068 4,068 '12,203 20,338 4,076 4,076 12,229 20,381 , , Suhum Kraboa Coal tar 5,236 5,236 15,708 26,181 .5,240 5,240 15,720 26,200 5,251 5,251 15,754 26,256 145 University of Ghana http://ugspace.ug.edu.gh I Ahanta West ,3,925 3,925 11,775 19,625 3,991 3,991 11,973 19,955 4,057 4,057 12,171 20,284 Axim 5,824 5,824 17,472 29,120 5,730 5,730 17,189 28,648 5,824 5,824 17,472 29,120 Sefwi Wiawso 6,166 6,166 18,499 30',832 6,270 6,270 18,810 31,350 6,373 6,373 19,120 31,867 Was sa Amenfi West 5,674 5,674 17,022 28,371 5,770 5,770 17,309 28,848 5,865 5,865 17,594 29,324 \ *Prestea Huni Valley 5,396 5,396 16,187 26,978 5,440 5,440 16,320 27,200 5,530 5,530 16,589 27,649 Total 129,420 129,420 388,259 647,099 130,750 130,750 392,249 653,749 132,404 132,404 397,213 662,021 Average U5 pop. 3,806 3,806 11,419 19,032 3,846 3,~46 11,537 19,228 3,894 3,894 11,683 19,471 Year 1 (34 districts) 129,420 129,420 388,259 647,099 , Year 2 (36 additional .. districts) , I 269,191 269,191 807,572 1,345,954 Year 3 (26 additional I districts) 296,904 373,847 373,847 1,121,542 1,869,236 Notes: *Average population for other districts in the same region used as these districts were created after the projections were made "Projected population of these districts are part of another district from which, they were created .J Based on information from districts, the proportion ofU5 out of total pop isLO% (4% for 0-11; 4% for 12-23; 12% for 24-59 groups). These proportions were applied to the total U5 population to generate population for the age group \46 University of Ghana http://ugspace.ug.edu.gh APPENDIX VI: Number of communities and eligible fever cases AAQArm AAQ+AMXArm Item 2007 2008 2009 Total 2007 2008 2009 Total Total number of CBAs 179 298 234 711 204 362 286 852 Number of CBAs who treated at least one fever case 142 224 175 541 165 278 228 671 Number of communities 85 163 163 411 103 202 202 507 Number of eligible cases treated 1,455 2,285 2,078 5,818 1,526 2,808 2,267 6,601 " , -. - 147 University of Ghana http://ugspace.ug.edu.gh APPENDIX VII: Publications and conference presentations A. A conference paper presented at the ih World Congress of the International Health Economics Association held in Beijing, China, 15th12th - July, 2009. Title: Treatment choices for fevers in children under-five years in a rural Ghanaian district Presenter: Justice Nonvignon Abstract: Few health care demand studies examine the choice of treatment services for childhood illnesses. Besides, in their analysis, most of the existing studies compare alternative treatment options to a single option, usually self-medication. This study aims at examining the factors that influence the choices that caregivers of children under five years make regarding treatment of fevers due to malaria and pneumonia. This study uses 2006 household socio- economic survey and Health and Demographic Surveillance data covering caregivers of 529 children under-five years in the Dangme West District and applies a multinomial probit (MNP) technique to model the choice of treatment services for fevers in under-fives in rural Ghana. Four care options are considered: self-medication, over-the-counter providers, public providers and private providers. I , I The findings indicate that longer travel, waiting and treatment times encourage people to I prefer.self-medication and informal providers to public and private providers. Caregivers with .. health insurance coverage: also prefer-care frompubLicprovfders compared to informal 01' private providers. Caregivers with higher incomes prefer public and private providers 'over self-medication while higher treatment charges and longer times at public facilities encourage caregivers to resort to private providers. Also, caregivers of female under-fives prefer self- care while caregivers of male under-fives prefer public providers to self-care, implying gender disparity in the choice of treatment. The results of this study imply.that efforts at .. curbing under-five morbidity due to malaria and.pneumonia 'need t6 take into account care-: . seeking behaviour of caregivers of under-fives as well as implementation of strategies. Keywords: Choice of health care, malaria and pneumonia; under-five illnesses, multinomial probit model Co-authors: Moses Aikins, Margaret A. Chinbuah, Mercy Abbey, Margaret Gyapong, Bertha Garshong, Saviour Fia and John Gyapong A University of Ghana http://ugspace.ug.edu.gh B. A conference paper presented at the 8th World Congress of the International Health Economics Association held in Toronto, Canada, 13thloth - July, 2011. Title Opportunity cost of community health volunteers in the implementation of community management of under-five fevers in a rural setting in Ghana Presenter: Justice Nonvignon Abstract Background Accelerating progress towards attainment of Millennium Development GOoal Four of reducing child mortality requires taking child health services closer to the homes and within communities. Consequently, community-based health interventions that seek to substantially reduce - if not eliminate - the lack of access to health services in remote communities in developing countries are promoted by various stakeholders, including the World Health Organization. Given the shortage of trained health workers in developing countries, such community-based interventions have largely made use of community health volunteers who are community members trained specifically to help implement interventions within the community. Usually, community health volunteers are not remunerated even though they play a key role in the success of community health programmes. Some studies on costing of community-based health interventions in developing countries tend to exclude the opportunity cost to community health volunteers who are involved in the implementation of such, interventions even though these also constitute economic costs of implemeriting the interventions. This study estimates the opportunity costs. to community health volunteers duririg implementation of a 'prpgramme on community .management of fevers in children . under-five years in a rural district in southern Ghana. .' .. -- .' .- - Method The community management programme was implemented throughout the rural district over a period of four years with community health volunteers used implementing the interventions for three years (i.e, year two to yearfourjvDataori the total time of community health ·1 volunteers (383 volunteers in year. two, 660 in year three and 520 in year four) that went into the programme implementation (including trainings, treatment of children and time spent with supervisors) were collected prospectively. Opportunity costs were estimated using the average wage for unskilled workers in the rural district. Microsoft Excel was used to calculate opportunity costs. Results The preliminary results show that on average, a total of 145.73 hours of work were spent per volunteer over the three-year period. Calculated separately for each programme year using unskilled labour rate in the district and converted into 2009 Ghana cedis, the preliminary results show that on average, a community health volunteer time cost GH