Semi-Quantitative Evaluation of Access and Coverage (SQUEAC) Kiyawa LGA CMAM Program Jigawa State, Northern Nigeria June-July 2014

Transcription

1 Semi-Quantitative Evaluation of Access and Coverage (SQUEAC) Kiyawa LGA CMAM Program Jigawa State, Northern Nigeria June-July 2014 Joseph Njau, Ifeanyi Maduanusi, Chika Obinwa, Francis Ogum, Zulai Abdulmalik, and Janet Adeoye ACF International 1

2 ACKNOWLEDGEMENTS Kiyawa LGA SQUEAC 1 has been implemented through generous support of the Children Investment Fund Foundation (CIFF). The ACF Nigeria Coverage Assessment Team are grateful to the following for their valuable contributions towards the successful completion of the Kiyawa Local Government Area (LGA) SQUEAC assessment. Firstly, Gunduma Health System Board (GHSB) Jigawa State authorized the implementation of the SQUEAC assessment in Kiyawa LGA. Usman Tahir (Director General GHSB), and Kabir Ibrahim (Director Primary Health Care GHSB) were very supportive throughout the exercise, acting as the link between the State and the CMAM 2 Coverage assessment team. Aisha Aminu Zango, the State Nutrition Officer (SNO), Husaini Ado of National Population Commission (NPopC), Ibrahim Haruna Head of Department (HoD) Kiyawa LGA Water, Sanitation and Hygiene (WASH) Program and Suleiman Muhammad the Nutrition Focal Person (NFP) of Kiyawa LGA are acknowledged for providing the coverage assessment team with client/beneficiary records, anecdotal program information and the LGA maps. Gloria Njoku, the ACF s Head of Base, at Dutse office is appreciated for introducing the team to the various stakeholders and giving logisitical support to the SQUEAC investigation team. Peter Magoh the ACF security manager, and Abubakar Kawu (program support officer) conducted a very helpful security and logistics assessment prior to the implementation of the SQUEAC assessment. Abubakar played a key role in logistics and administrative support during the SQUEAC implementation period. Joseph Njau (ACF CMAM Coverage Program Manager) provided technical support in the implementation of the assessment while Sophie Woodhead (of Coverage Monitoring Network team - CMN) gave useful insights in compilation and validation of this SQUEAC report. Ifeanyi Maduanusi (ACF Coverage Deputy Program Manager) led the coverage assessment team in training enumerators, supervision of the assessment and writing of the SQUEAC report. The ACF CMAM Program Coverage Officers (Chika Obinwa, Zulai Abdul Malik, Janet Adeoye and Francis Ogum) led the enumerator teams and were instrumental to ensure the quality of daily SQUEAC study activities. The effort of the enumerators - the foot soldiers in collecting information during the study is acknowledged. Different stakeholders including health workers, caregivers, traditional and religious leaders, traditional birth attendants and other interviewees who, despite their busy schedule gave very useful information regarding the CMAM Program in Kiyawa LGA are highly appreciated. Coverage Assessment Team ACF International 1 Semi Quantitative Evaluation of Access and Coverage 2 Community-based Management of Acute Malnutrition 2

3 Table of contents 1. EXECUTIVE SUMMARY INTRODUCTION OBJECTIVES METHODOLOGY DESCRIPTION OF FIELD ACTIVITIES RESULTS AND FINDINGS STAGE 1: ROUTINE MONITORING AND PERFORMANCE DATA-IDENTIFYING POTENTIAL AREAS OF LOW AND HIGH COVERAGE ROUTINE MONITORING (OTP CARDS) AND PERFORMANCE DATA PROGRAM EXITS (DISCHARGE OUTCOMES) ADMISSION TRENDS MUAC AT ADMISSION LENGTH OF STAY (LOS) FROM ADMISSION TO RECOVERY NUMBER OF VISITS BEFORE DEFAULT DEFAULTERS AND ALL EXITS BY LOCATION TIME TO TRAVEL TO CMAM (OTP) HF PLOT/DISTANCE FROM TREATMENT CENTRE CONCLUSION OF THE ROUTINE MONITORING (OTP CARDS) ANALYSIS STAGE 1: QUALITATIVE DATA-INVESTIGATION OF FACTORS AFFECTING PROGRAM AND COVERAGE QUALITATIVE SAMPLING FRAMEWORK QUALITATIVE INFORMATION AWARENESS AND PERCEPTION OF THE PROGRAM, COMMUNITY MOBILIZATION, AND PEER-TO-PEER REFERRALS IN COMMUNITIES COMMUNITY VOLUNTEER (CV) ACTIVITY AND HIGH DEFAULT RATE HEALTH SEEKING BEHAVIOUR IN COMMUNITIES STOCK-OUT OF DATA TOOLS AND ROUTINE DRUGS LEADING TO CHARGES FOR OTP CARDS AND ROUTINE DRUGS STOCK-OUT AND MISUSE OF RUTF, AND OCCASIONAL CLOSURE OF OTP SITES LARGE TURN-OUT OF BENEFICIARIES, LONG WAITING TIME, AND UNAVAILABILITY OF SHADES, MATS AND BENCHES HEALTH WORKERS ACTIVITIES, NON-ADHERENCE TO CMAM GUIDELINES, AND TRAINING DATA TRIANGULATION CONCEPT MAP STAGE 2: SMALL AREA SURVEY AND SMALL STUDY SMALL AREA SURVEY CASE DEFINITION RESULT OF SMALL AREA SURVEY SMALL STUDIES SAMPLING METHODOLOGY CASE DEFINITION RESULT OF QUANTITATIVE SMALL STUDY SMALL STUDY ON DEFAULTERS RESULTS OF DESCRIPTIVE SMALL STUDY CONCLUSION OF SMALL AREA SURVEY AND SMALL STUDY DEVELOPING THE PRIOR HISTOGRAM OF BELIEF CONCEPT MAP UN-WEIGHTED BARRIERS AND BOOSTERS WEIGHTED BARRIERS AND BOOSTERS TRIANGULATION OF PRIOR

4 6.4.6 BAYES PRIOR PLOT AND SHAPE PARAMETERS STAGE 3: WIDE AREA (LIKELIHOOD) SURVEY CALCULATION OF SAMPLE SIZE AND NUMBER OF VILLAGES TO BE VISITED FOR LIKELIHOOD SURVEY QUANTITATIVE SAMPLING FRAMEWORK CASE FINDING METHOD AND CASE DEFINITION QUALITATIVE DATA FRAMEWORK RESULTS OF THE WIDE ARE SURVEY POSTERIOR/COVERAGE ESTIMATE DISCUSSIONS RECOMMENDATIONS ANNEXURE ANNEX1: SCHEDULE (DETAILED) OF IMPLEMENTED ACTIVITIES IN KIYAWA SQUEAC ANNEX2: PARAMETERS USED IN PRIOR BUILDING AND SAMPLE SIZE CALCULATION ANNEX3: CONCEPT MAPS-TEAM A AND B ANNEX 3: ACTIVE AND ADAPTIVE CASE FINDING PROCEDURE ANNEX6: SUMMARY OF THE SMALL STUDY FINDINGS-KIYAWA LGA List of figures FIGURE 1: MAP OF NIGERIA (TOP RIGHT) SHOWING JIGAWA STATE AND JIGAWA STATE MAP SHOWING KIYAWA LGA (INDICATED BY ARROW). MAPS CAN BE DOWNLOADED AT STATE_10024.HTML#.U_CQ7SIG-P FIGURE 2: MAP OF KIYAWA LGA SHOWING THE LOCATION OF THE 5 CMAM HFS (HFS)... 9 FIGURE 3: EXIT TRENDS FOR KIYAWA LGA CMAM PROGRAM-JANUARY 2013 TO APRIL FIGURE 4: ADMISSION TREND OF KIYAWA LGA CMAM PROGRAM AND THE SEASONAL CALENDAR OF EVENTS FIGURE 5: ADMISSION MUAC FOR KIYAWA LGA CMAM PROGRAM FIGURE 6: LENGTH OF STAY (LOS) FROM ADMISSION TO RECOVERY FIGURE 7: PROPORTION OF EXIT MUACS -RECOVERED FIGURE 8: PLOT OF NUMBER OF VISITS BEFORE DEFAULT FIGURE 9: DEFAULTERS MUAC ON EXIT FIGURE 10: PROPORTION OF EXIT MUACS AT POINT OF DEFAULT FIGURE 11: DEFAULTERS VS EXITS-KIYAWA AND OTHER LGAS FIGURE 12: TIME TO TRAVEL TO CMAM SITES -CASES IN PROGRAM AND DEFAULTERS FIGURE 13: TIME TO TRAVEL TO CMAM SITES FOR EACH OF THE 5 CMAM HEALTH FACILITIES FIGURE 14: REASONS FOR NOT ATTENDING THE CMAM PROGRAM-SMALL AREA SURVEY FIGURE 15: ANALYSIS OF REASONS FOR DEFAULT AND CONDITION OF CASES FOUND IN THE STUDY IN STAGE FIGURE 16: ILLUSTRATION OF TRIANGULATION OF PRIOR FIGURE 17: BAYESSQUEAC BETA-PRIOR DISTRIBUTION PLOT SHOWING THE SHAPE PARAMETERS AND THE SUGGESTED SAMPLE SIZE 43 FIGURE 18: KIYAWA MAP DIVIDED INTO QUADRANTS FOR SPATIAL SAMPLING OF VILLAGES FIGURE 19: BARRIERS TO PROGRAM ACCESS AND UPTAKE-WIDE AREA SURVEY (WAS) FIGURE 20: DISTRIBUTION OF QUADRANTS ACCORDING TO COARSE COVERAGE ESTIMATES FIGURE 21: BAYES PLOT SHOWING PRIOR, LIKELIHOOD AND POSTERIOR (CONJUGATE ANALYSIS) List of tables TABLE 1: PARAMETERS USED IN ANALYSES OF LIKELIHOOD SURVEY TABLE 2: SUMMARY OF OUTCOME DATA EXTRACTED FROM BENEFICIARY/OTP CARDS AT A GLANCE TABLE 3: SOURCES AND METHODS USED TO GET INFORMATION IN THE BBQ TOOL FOR KIYAWA LGA TABLE 4: BARRIERS, BOOSTERS & QUESTIONS FINDINGS AND SOURCES OF INFORMATION TABLE 5: SIMPLIFIED LOT QUALITY ASSURANCE CLASSIFICATION OF SMALL AREA SURVEY RESULTS TABLE 6: SIMPLIFIED LOT QUALITY ASSURANCE CLASSIFICATION OF SMALL STUDY RESULTS TABLE 7: CMAM HFS AND VILLAGES SELECTED FOR DEFAULTER STUDY

5 TABLE 8: WEIGHTED BARRIERS AND BOOSTERS OF KIYAWA LGA CMAM PROGRAM TABLE 9: PARAMETERS FOR SAMPLE SIZE CALCULATION FOR LIKELIHOOD SURVEY TABLE 10: BARRIERS TO PROGRAM ACCESS AND UPTAKE-WAS TABLE 11: RESULTS OF THE LIKELIHOOD (WIDE AREA) SURVEY TABLE 12: DISAGGREGATED SAM CASES PER QUADRANT AND COARSE ESTIMATE DURING THE WIDE AREA SURVEY TABLE 13: FRAMEWORK OF ACTION POINTS TO ADDRESS BARRIERS OF KIYAWA CMAM PROGRAM

6 ABBREVIATIONS ACF CIFF CMAM CV FMOH GHSB IEC LGA NFP INGO OTP PHC RUTF SAM SLEAC SNO SMART SMOH SQUEAC VI Action Contre La Faim Action Against Hunger ACF International Children Investment Fund Foundation Community-based Management of Acute Malnutrition Community Volunteer Federal Ministry of Health Gunduma Health System Board Information Education and Communication Local Government Area Nutrition Focal Person International Non-governmental Organization Outpatient Therapeutic Program Primary Health Care Ready to Use Therapeutic Food Severe Acute Malnutrition Simplified Lot quality assurance sampling Evaluation of Access and Coverage State Nutrition Officer Standardized Monitoring Assessment of Relief and Transitions State Ministry of Health Semi Quantitative Evaluation of Access and Coverage Valid International 6

7 1. Executive summary The Community-based Management of Acute Malnutrition (CMAM) in Kiyawa LGA of Jigawa State commenced in October CMAM is implemented in the five CMAM sites 3 by the Jigawa State s Gunduma Health System Board (GHSB), Kiyawa Local Government Area (LGA) and primarily supported by UNICEF D Field Office, Bauchi. Since inception of the CMAM program a SQUEAC 4 assessment had not been done in Kiyawa LGA. However, a recent SLEAC assessment 56 by Valid International (VI) unveiled a low classification 7 of coverage with 7 children out of total of 47 SAM cases found in Kiyawa LGA being in the program. This gave a coarse estimate of 14.8% which give a picture of program coverage. Kiyawa LGA was chosen for a SQUEAC assessment to identify positive and negative factors affecting CMAM program coverage, build capacity of SMoH and LGA staff, and to proffer recommendations to improve the CMAM coverage. Routine program data and client/beneficiary/out Patient Therapeutic Program (OTP) records from May 2013 to May 2014 were extracted and analyzed. Five health Facilities (HFs) which offer the CMAM services 8 and 10 villages 9 which form part of the catchment population of these HFs in Kiyawa LGA were visited to obtain additional qualitative information about the CMAM program using different sources 10 and using different methods 11. All the information was triangulated by the various sources and methods until no new information was forthcoming and then analyzed into negative factors (barriers) and positive factors (boosters) which affect program coverage. The key barriers to program access and coverage identified included; high number of defaulters, lack of motivation of community volunteers to conduct case-finding and defaulter tracing, non-adherence to CMAM national guideline/protocol, generalized stock-out routine drugs, stock-out of data tools for recording beneficiary information, consumption of RUTF by adults, and siblings of SAM children in communities, weak mechanism for delivery of RUTF to HFs. The key boosters include; large turnout of beneficiaries 12, good awareness of the program in communities, good opinion of the program in communities, willingness of beneficiaries to stay in the program despite challenges 13, peer-to-peer referrals by caregivers, good working relationship between Health workers (HWs) and CVs. 3 All the five CMAM sites Katuka, Garko, Maje, Kwanda and Katanga 4 Semi Quantitative Evaluation of Access and Coverage 5 D Field Office is located in Bauchi, Bauchi State. 6 Chrissy B., Bina S., Safari B., Ernest G., Lio F. & Moussa S.; Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Survey of Community-based Management of Acute Malnutrition program; Northern States of Nigeria-(Sokoto, Kebbi, Zamfara, Kano, Katsina, Gombe, Jigawa, Bauchi, Adamawa, Yobe, Borno). Valid International. February The SLEAC used a three-class classifier with 20% and 50% as the thresholds; <= 20% is low; >20% to <=50% is moderate coverage while >50% 8 All the five CMAM sites Katuka, Garko, Maje, Kwanda and Katanga were visited. 9 Villages visited include; Nafara, Dangoli, Tsirma, Jama a Dawa, Kakarawa, Gorumo, Yelwa, and Danfasa. 10 Care-givers, Health Workers (HWs), Community Volunteers (CVs), community leaders, religious leaders, majalisa, teachers, traditional healers, traditional birth-attendants (TBAs), and women group, etc. 11 Semi-structured interview, in-depth interview, observations and informal group discussions 12 This is a potential for enhancing peer to peer referral and less dependence on CVs 13 For instance there are caregivers who sleep-over at CMAM facilities till they are attended. 7

8 Homogenous coverage across the wards in Kiyawa LGA was confirmed. Program failures caused by placing some costs (hidden charges) on OTP cards and routine drugs to be borne by beneficiaries was also confirmed; defaulters were studied and reasons for defaulting were also analyzed. After conjugate analysis of the prior and likelihood survey, a point coverage estimate of 48.5% (41.3% %. CI; 95%) 14 was reported. Thus the coverage was significantly higher than that unveiled in the recent SLEAC results for Kiyawa LGA due to reasons explained in detail in the report. Recommendations: key recommendations include provision of data tools for CMAM sites, provision of routine drugs, removal of the hidden charges on beneficiaries to ensure free CMAM services, strengthening of the delivery mechanism of RUTF to HFs, supportive supervision of HWs at CMAM HFs to enhance adherence to CMAM protocol and establishment of innovative ways to improve motivation of CVs. 2. Introduction Jigawa State is located in the Northwest of Nigeria. It is bordered by States of Kano and Katsina in the West, Bauchi in the South, Yobe in the East and shares nn international border with Niger Republic in the North.. The State has 27 Local Government Areas (LGAs) 15. According to the 2006 census, the State has a total population of 4,348,649 million inhabitants 16. The population growth of the state is estimated at 3.5 % with about 48 % of the population falling under the age of fifteen. Out of the total population about 2.9 million are considered to be productive adults. Eighty per cent (80%) of the population is found in the rural areas and is made up of mostly Hausa, Fulani and Manga (a Kanuri dialect). The pattern of human settlement is nucleated, with defined population centres 17. Kiyawa LGA is about 30 km away from Dutse (Jigawa State Capital). It is bordered by Dutse, Birnin Kudu, Buji, Kafin Hausa, and Jahun LGAs, as well as Bauchi State. The people of Kiyawa are mostly Hausa and Fulani. The major occupation is farming; while the religion is predominantly Islam. Kiyawa LGA is under the traditional leadership of the Emir of Dutse, with two district heads in Shuwarin and Kiyawa districts in charge of the traditional affairs of the LGA. At the commencement of implementation of Community-based Management of Acute Malnutrition (CMAM) in Kiyawa LGA in October 2011, 5 CMAM sites were established 18. The program is supported by UNICEF D Field Office, which provides technical support. The health workers 19 run the services at the CMAM Health Facilities (HFs) on a weekly basis. A Nutrition Focal Person (NFP) for the LGA supervises the five CMAM sites and reports to the State 14 Results are expressed with a credible interval of 95%. 15 Auyo, Babura, Biriniwa, Birnin Kudu, Buji, Dutse, Gagarawa, Garki, Gumel, Guri, Gwaram, Gwiwa, Hadejia, Jahun, Kafin Hausa, Kaugama, Kazaure, Kiri Kasama, Kiyawa, Maigatari, Malam Madori, Miga, Ringim, and Roni 16 National Population Commission UNICEF Nigeria uses a model of 5 CMAM sites in each LGA for treatment of malnourished children. This is presently, being reviewed for possible scale-up. 19 HWs are employees of the Gunduma Health System Board -GHSB) 8

9 Nutrition Officer (SNO) who coordinates all nutrition activities in the State. The SNO who reports to the State Director of Primary Health Care, also serves as the linkage to the UNICEF Nutrition Specialists in UNICEF regional Office. Figure 1: Map of Nigeria (top right) showing Jigawa state and Jigawa state map showing Kiyawa LGA (indicated by arrow). Maps can be downloaded at Figure 2: Map of Kiyawa LGA showing the Location of the 5 CMAM HFs (HFs) 9

10 SQUEAC has not been carried out since the commencement of CMAM program in Kiyawa to determine access and coverage of the program. However, a recent SLEAC study conducted late 2013 by Valid International reported a coarse estimate of 14.89% (where 7 SAM cases were covered by the CMAM program out of 47 SAM cases found in the LGA during the assessment). Therefore, Kiyawa LGA CMAM coverage was classified as low and was one of the LGAs (Kiyawa) identified for an in-depth investigation through a SQUEAC assessment. The purpose was to gather additional information on positive and negative factors affecting access and coverage of CMAM program. The SQUEAC assessment was undertaken in Kiyawa LGA in the month of July Objectives The SQUEAC investigation was guided by the following specific objectives; 1. To investigate the barriers and boosters to program and coverage in Kiyawa LGA. 2. To evaluate the spatial pattern of program coverage in Kiyawa LGA. 3. To estimate overall program coverage in Kiyawa LGA. 4. To make relevant recommendations in order to reform and to improve the CMAM program, 5. Build the capacity of SMoH and LGA staff to conduct a SQUEAC assessment. 4. Methodology The SQUEAC methodology was adapted to suit the SQUEAC investigation of Kiyawa LGA CMAM program. The methodology used are explained in detail as follows; Stage 1 Quantitative data: Routine program data, and data from client information recorded in the OTP cards were extracted and analyzed into various plots; admission trends; exit trends; time to travel to site; Length of stay from admission to cure; MUAC at admission; number of visits before default; MUAC at default; defaulters by location. The derived plots indicated some factors that affect the program such as existence of active case finding, met need, far distance and defaulting tendencies. Qualitative data Barriers, boosters and questions: Qualitative information was collected from different sources, using different methods. The information was analyzed into a barriers, boosters and questions (BBQ) to capture the positive and negative factors that affect the program and its coverage. The data was triangulated to adduce evidence. Additional information was 10

11 collected to confirm the evidence gathered before, until no further information was forthcoming around a certain theme/topic in a process referred as sampling to redundancy. This information was further analyzed into weighted and un-weighted barriers and boosters which were scored according to the perceived weight each barrier or booster had on the program coverage (either negatively or positively). The process of weighing barriers and boosters is discussed in detail in separate section (making the prior section). Concept map: the factors that affect the program were analyzed into a concept maps that showed the relationship between them. The concept maps are annexed in section 9.3 annex 3. Stage 2 data Based on the information and evidence gathered in stage 1, a small area survey and two small studies were conducted to investigate the spatial pattern of coverage and factors that may be affecting coverage. Small Area Survey The small area survey data was analyzed using simplified Lot Quality Assurance Sampling (LQAS) technique to test a hypothesis. This was done by examining the number of Severe Acute Malnutrition (SAM) cases found (n) and the SAM cases covered (c) in the program. The threshold value (d) was used to determine if the coverage was classified as satisfactory or not. Value (p) was used to denote a minimum standard used as a measure of high, moderate or low coverage 20. The recent SLEAC assessment coverage classification (with a coarse estimate of 14.98%) indicated that the coverage in Kiyawa LGA was low, that is below 20% 21. Given this coverage classification in Kiyawa LGA, consideration could have been given to use the 2 standard 3 class classifier to set the values of p1 and p2 as the lower and higher limit of classification of coverage to test spatial coverage hypothesis in the current SQUEAC assessment. However, in Kiyawa SQUEAC assessment the standard (p) was adapted as 50% 22 for analyses of small area survey data using simplified LQAS. This was because it was highly likely that the coverage could have considerably improved at the time of this SQUEAC investigation compared to the period when SLEAC assessment was implemented. The low coverage unveiled by SLEAC assessment could be attributed to long duration of stock-out of RUTF at the time 23. In the current SQUEAC investigation, the information gathered in stage 1 gave a picture of a program that was likely to have a high coverage. Therefore, using the 20 SPHERE standards has recommended minimum coverage for Therapeutic programs in rural, urban, and camp settlements. These thresholds are 50%, 70% and 90% coverage for TFP program run in the contexts of rural, urban and camp areas respectively. 21 The two standard three class classifier classifies coverage as follows: Low coverage-20% and below; Moderate coveragegreater than 20% up to 50%; high coverage-above 50% 22 Previously conducted SLEAC coarse estimate was not considered to set the value of p in Kiyawa LGA. Therefore, the sphere standard of coverage in a rural setting <= 50% for low coverage; and >50% for high coverage was used. 23 The Wide Area Survey in SLEAC assessment considers the children who are in the program (covered) against the total number of SAM cases found during the survey. Large number of Current (SAM) cases were not covered during SLEAC due to lack of RUTF stocks in HFs at the time; which may have made many beneficiaries stay home till RUTF was available at the HFs 11

12 SPHERE standard of coverage mentioned above is sufficiently sensitive to show disparities in spatial coverage in Kiyawa as the context has changed. Therefore: The value of (p) used that was used was 50%. The formula for deriving (d) is shown below: = = = 2 If the number of covered cases exceeded value (d), then the coverage was classified as being satisfactory. However, if the number of covered cases found did not exceed value (d) then the coverage was classified as being unsatisfactory. The combination of the (n) and (d) was used as the sampling plan. The results of the small area survey were shown in table 5 and reasons for coverage failure obtained from the small area survey were plotted in figure 6 in results section of this report. Small Study Two small studies were conducted; that is, a quantitative and a descriptive small study respectively. The small study investigated the number of beneficiaries that were charged for CMAM routine drugs and OTP registration cards and therefore, establish whether this was major cause of coverage failure. The data obtained was analyzed quantitatively using simplified LQAS and utilized the threshold (p) of 50% (as use in small area survey). This way the classification thereof, established whether charging beneficiaries for the mentioned service contributed to coverage failure. In this study, the (n) parameter was number of caregivers with SAM cases that were interviewed. A descriptive small study of defaulters was done to understand the reasons for defaulting and results presented on table 6 and illustrated in Figure 14.The results of the small study are explained in the results section of this report. Stage 3 data The prior: The tools that have been used revealed a rich set information about coverage. It also identified barriers to access and care, as well as spatial coverage of the program. Therefore the prior of the program was estimated through use of the following tools 24 : Belief histogram Weighted barriers and boosters Un-weighted barriers and boosters Calculation of the total positive and total negative factors illustrated in the concept map. The prior was established in a beta prior distribution with prior shaping parameters and plotted on BayesSQUEAC calculator 25. The beta prior distribution 26 expresses the findings of 24 Each listed process is discussed in detail in the body of the report. The recent SLEAC coarse estimate was not used incorporated while calculating the prior. This was because it was observed that it could possibly bias the result since CMAM sites were closed down at the time coinciding with the SLEAC survey in Kiyawa LGA in Downloaded free from 26 As illustrated in the example in figure 21 in results section of this report. 12

13 stage 1 and 2 in similar ways to the likelihood survey s (binomial distribution) as described in a sections below. The BayesSQUEAC calculator also suggested a sample size at 10% precision. The likelihood survey yielded data that was analyzed to give program coverage. The data was organized into the parameters shown in the table 1 below. The binomial distribution of the likelihood results are shown in figure 21 in results section in this report. Table 1: Parameters used in analyses of likelihood survey Parameters Values Current cases in the program (x) Current SAM cases not in the program (y) Total current SAM cases (x+y=n) Point coverage 27. CI 95% = ( ) 100 ( + ) The program coverage (posterior). The process of combining the prior and the likelihood to arrive at the posterior (also referred as conjugate analysis 28 ) was used in arriving at the program coverage in this SQUEAC investigation. This meant that the prior information about coverage (i.e. the findings from the analysis of routine programs data; the intelligent collection of qualitative data; and the findings of smallarea surveys, and small studies) collected using the bayesian technique 29 was helpful to provide information about overall coverage of the program (expressed in a beta prior distribution). As such, all the relevant information that was collected in stage 1 and 2 were used together with the survey data that were collected in stage 3 of the SQUEAC investigation to give an overall picture about the program and to unveil the headline coverage. A conjugate analysis which is used to provide the program coverage (as described below) requires that the prior and the likelihood are expressed in similar ways. The conjugate analysis combined the beta distributed prior with a binomial distributed likelihood to produce a beta distributed posterior (see figure XX). Met need is calculated as: = ( ) 27 Point coverage gives overall accurate measure of this program 28 A conjugate analysis requires that the prior and the likelihood are expressed in similar ways. 29 Bayesian methods allowed findings from work done prior to a survey to be combined with data from the survey. In this case survey data are treated as just another source of information and are used to update the prior information The main advantage in using the Bayesian approach in this as well as in all SQUEAC investigation are:1) Smaller survey sample sizes are required compared to larger population based dummy surveys2) It provides a framework for thinking about SQUEAC data that has been collated and analyzed in stage 1 and stage 2 13

14 5. Description of field activities SQUEAC planning, training and implementation A letter from the Federal Ministry of Health (FMOH) notifying the State Ministry of Health (SMOH) of the intended SQUEAC assessment of Kiyawa CMAM program was delivered through the SNO. The ACF Coverage Assessment Team arrived in Dutse 30 on 14 th May 2014 to commence the SQUEAC investigation. A letter requesting approval to conduct the assessment was sent to Jigawa State Health Research Ethics Committee through Gunduma Health System Board (GHSB) before the daily SQUEAC activities could begin. The Coverage Assessment Team was granted access to visit the CMAM HFs and to extract data from beneficiary/client records as at 24 th June Advertisement for enumerators and recruitment was completed, with most of the enumerators being indigenous members of the Kiyawa LGA. A total of 6,231 Outpatient Therapeutic Program (OTP) card information of beneficiaries from 5 CMAM sites who have exited the program from May 2013 to May 2014 were extracted into a computer database. Two days theoretical training for successful enumerators, SMOH, LGA and National Population Commission (NPopC) staff was done in 2 days (30 th June and 1 st July 2014). Qualitative information was gathered by visiting all the five CMAM sites and 10 villages (2 villages per OTP site). Due to the fasting period (Ramadan), intermittent breaks were given to the enumerators during the SQUEAC assessment so that they could rest. The small area survey and two small studies were conducted on 12 th and 13 th July while the likelihood survey was conducted in the period 15 th and 18 th July Dissemination workshop of the SQUEAC findings was done at the State level and involved GHSB, Kiyawa LGA, as well as the traditional ruling council of Kiyawa (conducted on 4 th of August 2014). A joint session of all the stakeholders helped make recommendations to improve the Kiyawa CMAM program. A detailed list of daily activities during the Kiyawa SQUEAC assessment is contained in annex 1 of this report. 6. Results and findings This section summarizes the results of stage 1, 2 and 3. The various data were organized as follows: 6.1. Stage 1: Routine monitoring and performance data-identifying potential areas of low and high coverage. In stage 1 the routine data, anecdotal program information and performance data were analyzed to study the effectiveness of the CMAM program, the outreach activities and potential areas of low and high coverage, time taken to travel to site among others. This analysis formed the basis of identifying locations that would be suitable to collect the 30 Jigawa State capital where the coverage team resided during Kiyawa SQUEAC. Dutse is about 30 kilometers from Kiyawa. 14

15 qualitative data that would provide more information regarding factors affecting the CMAM program. The data extracted from OTP records of the five HFs in Kiyawa LGA are summarized in Table 2 below. Table 2: Summary of outcome data extracted from beneficiary/otp cards at a glance. CMAM HF Exit records Dead Defaulter Nonrecovered Recovered Transferred to SC Missing data Garko Katanga Katuka Kwanda Maje Total Proportion 100% 0.03% 48.53% 2.42% 43.04% 0.13% 5.84% The findings of the quantitative and qualitative data analysis is described in the following sections Routine monitoring (OTP cards) and performance data Program exits (discharge outcomes) The exit trend of the performance data was plotted and smoothened in spans of 3 months median and average (M3A3) respectively. The plots are presented in figure 3 below. 15

16 Figure 3: Exit trends for Kiyawa LGA CMAM program-january 2013 to April 2014 From the exit trends above, the Kiyawa CMAM program is observed not to be effective. The recovery rate is below SPHERE standard 31 of 75% throughout the period under review. November and December 2013 had zero recovery rate with no client discharged as recovered. During this time the CMAM sites were not administering services due to stock-out of RUTF. On the other hand, defaulter rate was observed to be high, above the recommended 15% standard throughout same period. The peak of defaulting was in the month of November 2013, when all the children in the program had become defaulters as the CMAM services were not operational. The summary of extracted routine data tabulated in Table 1 above showed that about 48% of all exits during the review period were defaulters. Compared to the SPHERE standards only the death rate was low Admission trends The figure 4 below shows the admission trend of Kiyawa LGA CMAM program in relation to the seasonal calendar of events. 31 For a rural TFP program SPHERE standards and in line with Nigerian CMAM guidelines the recommended performance rates are as follows: recovery rate >=75%; defaulter rate of <15%; death rate of <10%. A good program should have non recovery rates of below 5%. 16

17 Figure 4: Admission trend of Kiyawa LGA CMAM program and the seasonal calendar of events The peak in admission in the month of May 2013 may have been as a result of high measles incidence in this period 32 (no evidence could be obtained from the LGA health records 33 ) and commencement of rainy season with increased diarrhea 34 episodes contributing to incident SAM cases presented for admission into the CMAM program. The decline in admissions from May to June 2013 coincided with commencement of weeding of farmlands which is mainly done by women. The slight increase in admissions from June to July 2013 may be attributed to peak of hunger and food prices as the household food reserve had depleted through both consumption and use of seeds for planting. As the rains and flooding increases the admissions reduce steadily from July to September; this discourages some caregivers from attending the CMAM site. At this period, the female labor demand (weeding 35 ) increased; at the same time 32 Measles is associated with malnutrition. 33 The seasonal calendar was developed with key information of the HOD Health and the Agriculture Department of Kiyawa LGA, however, this was based no verifiable record could be produced. 34 Diarrhea diseases contribute burden of disease and are associated with high malnutrition 35 Women are highly engaged in weeding of farm lands, as a result, they are likely not able to attend the CMAM services due to the conflicting priority of weeding farms lands. 17

18 the rainy season and flooding subsided. The admission was seen to have increased significantly in the month of October. The increase in admission could not be sustained due to stock-out of RUTF in November and December During the period of RUTF stock out, the five CMAM sites were closed, and zero admission was recorded for the two months. As the CMAM sites reopened in January 2014, admissions were witnessed as defaulters return to seek treatment. The month of January also coincides with the female labour demand in processing of harvested crops which usually occurs from December to January, and as such some women are occupied and could not access CMAM services. More so, availability of foods in households in January 2014 may have contributed to reduced incident SAM cases. At the end of processing activities (which engages women labour), caregivers had time to attend the program, thus, admissions may have rose steadily from February to May Besides this, other reasons that can be attributed to this increase in admissions include gradual depletion of household food reserves, commencement of hot season which is associated with increased cases of measles, and diarrhea which are major factors contributing to increase in incidence of malnutrition MUAC at admission Figure 5: Admission MUAC for Kiyawa LGA CMAM program The analyses of the extracted data showed a median MUAC on admission to be 105mm. There were very small number of children admitted with MUAC above the admission criteria. The plot also indicated that a relatively large group of SAM cases (below 110mm) were not identified early and referred accordingly. This is a pointer to a probable weak active case finding activity by CVs in the LGA. Digit preference and heaping at digits ending with 9 and 4 was noticeable, as such, erroneous MUAC measurement or lack of verification of MUAC during admission was suspected. 18

19 Length of stay (LoS) from admission to recovery Figure 6: Length of Stay (LoS) from admission to recovery The median length of stay (LoS) of 6 weeks shown above was plotted from all the discharged recovered children in the program during the period of May 2013 to May 2014 with exit MUAC of 125mm and above. The satisfactory LoS result was treated with caution based on two major reasons. Firstly, the CMAM program performance indicators were observed to be less than satisfactory due to low recovery rates and high defaulter rates did not meet the recommended SPHERE standards of >=75% and less than 15%, respectively. (See section ). Similarly, table 1 above showed that about 48% of all exits during the review period were defaulters. A program that has low performance is likely to have Long LoS and high defaulter rate. Secondly, further analysis of MUACs on exit of those reported as recovered showed that a significant proportion of the children discharged as recovered did not attain the exit MUAC of 125mm or more (illustrated in figure 7 below). 30% of discharged recovered children are suspected to have been discharged erroneously with MUAC of less than 125mm, with significant number being current cases of SAM. Therefore, the quality of the monitoring data and service delivery was suspected to have errors and therefore, the short length of stay of 6 weeks showed above may not likely reflect the true picture of the Kiyawa LGA CMAM program A good CMAM program has a short length of stay (LoS). It could be that SAM cases were detected early and admitted into the program. The treatment episode is shorter for SAM cases who are detected early and have no medical complications. This also, keeps the cost of SAM treatment low. 19

20 Figure 7: Proportion of exit MUACs -recovered Number of visits before default Figure 8: Plot of number of visits before default The median number of weeks SAM children stay in the program before default was found to be 4 weeks. Significant number of beneficiaries were noticed to withdraw early in the program. There is a likelihood that a large proportion of the defaulters in the community are likely to be current SAM cases 37. This is illustrated in further analysis of MUAC at default for all the defaulters. The results are shown in figure 9 and 10 below. 37 Defaulters that have had below 4 visits are likely to be current cases of SAM. See SQUEAC/SLEAC technical manual on interpretation of defaulters at < 4 weeks and those at >= 4 weeks. 20

21 Figure 9: Defaulters MUAC on exit Figure 10: Proportion of exit MUACs at point of default Defaulters and all exits by Location The analysis of extracted client information in the period May 2013 to May 2014 showed that about 85% of defaulters (2,459 out of 3,024 recorded) were mainly visiting Katanga and Katuka health facilities 38 in Kiyawa LGA. (See Table 1 above). The defaulters from other LGAs near Kiyawa are also illustrated in figure 11 below. 38 Katanga and Katuka health facilities had the highest number of admissions from the client information and routine data 21

22 Figure 11: Defaulters vs exits-kiyawa and other LGAs The data was analyzed further for defaulters in Katanga and Katuka who exited in the last quarter (March to May 2014) of the review period. The analyses showed that more than half of the defaulters were from outside Kiyawa LGA (the catchment area) where there are no CVs. However, significant number of defaulters were also coming from Kiyawa LGA despite the presence of community volunteers. Therefore, it was suspected that that community volunteers are not following-up with home visits for absentees. The LGAs where defaulters were coming from were shown in Figure 11 above Time to travel to CMAM (OTP) HF plot/distance from treatment centre Time to travel to site by caregivers were plotted as illustrated below. However, it was noted that most beneficiary/otp cards had missing or erroneous information on time to travel to CMAM HF. Other sources and methods were used to get additional information on time to travel to CMAM HF. Key informants in each of the health facility were used to give additional information and also to verify the existing information on time to travel from different villages to the CMAM HFs. The reconciled data was analyzed and plots of time it took the beneficiaries to travel to each of the CMAM HF was presented below. 22

23 Figure 12: Time to travel to CMAM sites -cases in program and defaulters Figure 13: Time to travel to CMAM Sites for each of the 5 CMAM health facilities Median time to travel for all the CMAM health facilities by clients was found to be 1 hour 30 minutes. Significant number of the beneficiaries were observed to be travelling from villages 23

24 outside Kiyawa LGA, so also, are defaulters (see section ). Most of the villages outside Kiyawa LGA where beneficiaries were coming are mainly distant communities as recorded on the beneficiary records. It was also observed that some beneficiaries use motor bikes (Achaba) as the medium of transport, while significant number travel to the HF while walking. Plots for individual CMAM facilities indicated similarities in time to travel to site of clients come as shown in figure 13 above Conclusion of the routine monitoring (OTP cards) analysis The analyses of information obtained from beneficiary/otp cards gave indications of the potential factors that were possibly affecting coverage in Kiyawa LGA. However, the information above needed to be investigated and understood further using the qualitative methodology. Major points noted for further investigation include; 1. Adherence to national CMAM guideline during admission and discharge 2. Charges incurred by caregivers while accessing CMAM services 3. High default rate 4. Distance as a factor affecting access 6.2. Stage 1: Qualitative data-investigation of factors affecting program and coverage Qualitative sampling Framework Qualitative information about the CMAM program in Kiyawa LGA was sought and obtained from different stakeholders (sources) so as to compliment the information obtained from the quantitative data analyses of the extracted beneficiary/client records. Different methods were used while collecting such information from stakeholders. The information collected were analyzed, while sources and methods used were triangulated as evidence arose. All the 5 CMAM HFs were visited. In each of the HFs, at least four caregivers (who were randomly chosen), HWs, and CVs were interviewed. Motorcyclists were also asked questions, as well as husbands of caregivers of beneficiaries in-program. Ten communities were visited to collect additional information about the program from community members. These communities that were visited were purposively selected based on the following key reasons; 1) distances from CMAM site, 2) pattern of admission and defaulting. At least, a community leader, religious leader, teacher, provision shop seller, patent medicine vendor, majalisa (community age-group social gathering in tea places and shades), and women group were interviewed in each of the communities visited, using different methods to obtain information about the CMAM programme in Kiyawa LGA. Information obtained were identified and analyzed in respect of the effect it had on the program (that is as a booster or a barrier to the program). Each piece of evidence was considered and if it needed further verification, then a different source with/or different method was used to seek further information which was compared with previously gathered information. The questions arose, aided in collecting more information from the same or different sources using appropriate methods to obtain more clarity on the piece of information before it was finalized as a barrier 24

25 or booster. In this way, using the different methods and sources repeatedly reached a point where no new information could be obtained 39. This is how sampling to redundancy was achieved Qualitative information The qualitative information obtained from the field are summarized and explained under thematic headings as follows Awareness and Perception of the Program, Community Mobilization, and Peer-to-peer Referrals in Communities All respondents interviewed in different communities reported that they were aware of the CMAM program in Kiyawa LGA. Nevertheless, there was evidence of poor community mobilization in all the communities visited. Large number of respondents in the communities reported that nobody had sensitized or told them about the CMAM program. However, it was established that community mobilization was done when the Kiyawa LGA CMAM program commenced in October The present number of beneficiaries accessing CMAM services were mainly referred by their peers (peer-to-peer referral by caregivers) in communities. This was noticed to thrive mostly on the good opinion of the Kiyawa CMAM program. Community members said that they appreciated the CMAM program in the Kiyawa LGA. Notably, none of the respondents reported negative opinion of CMAM program. The respondents interviewed in both HFs and communities were able to relate the program with good outcome in terms of recovery of many malnourished children that had accessed the CMAM program. Likewise, communities are largely aware of the existence of the program which has largely been as a result of the good opinion about the program that is shared within the community. In the period the CMAM program has been running in Kiyawa LGA (for about three years), information about what the program does in treatment of malnutrition had diffused to communities in the LGA, with caregivers contributing immensely as information or news carriers. The good opinion and awareness which the Kiyawa CMAM program enjoy in communities within Kiyawa LGA was so strong that health workers in CMAM HFs are used to resolve non-compliant 40 cases during immunization campaigns. Non-compliant community members on recognizing the health workers from CMAM health facilities allow their children to receive immunization vaccines Community Volunteer (CV) Activity and High Default Rate Each of the five CMAM HFs had 25 CVs assigned to them and documented. However, only a handful of the CVs were reported to be working. The referral activity of CVs could not be tracked because referrals slips were not being used. However, as the SQUEAC investigation 39 The process of collection of the information is iterative and as such pieces of information are investigated repeatedly until no new information is forthcoming. 40 Caregivers who hitherto would not allow their children to receive polio immunization were reported to have accepted their children to be immunized once they see a health worker from a CMAM HF due to the good opinion they have about the CMAM program. 25

26 was going on, referral slips were being distributed to the CMAM HFs for the first time by the Nutrition Focal Person (NFP). Therefore, the SQUEAC investigation team could not verify referral by CVs from the records. Though the CVs interviewed in all the CMAM sites maintained that they refer SAM cases to the CMAM HFs, only two out of all caregivers 41 interviewed in all the five CMAM sites reported that they were referred by a CV. On each OTP day, the CVs come to the CMAM HFs to help the health workers in taking anthropometric measurements and some other sundry activities. Some of the measurements taken by CVs were noticed to be incorrect, and are not usually validated by health workers. Most CVs reported that they are not able to carry out tracing of absentees and defaulter nor do screening for malnutrition within communities properly due to lack of means of transport to travel to distant areas. Therefore, the very high number of defaulters observed in quantitative analyses of data could be linked to poor CV activities in terms of follow-up and tracing of absentees/defaulters. Another factor reported by health workers and program staff is that most beneficiaries come from neighboring LGAs 42 and States 43 (see the section on defaulter vs. exits by location). The CVs had been trained only once since the inception of the program in Kiyawa LGA Health Seeking Behaviour in Communities Responses on the health seeking behavior in communities on treatment of malnutrition varied. Some respondents interviewed in different communities and HFs reported that malnourished children are first taken to traditional healers and patent-medicine vendors before coming to the CMAM HFs for treatment 44. A local herb (Tsimi) was reported as being used in communities for treating malnutrition. On the other hand, many respondents reported that SAM children are taken to the health facility for treatment 45. Therefore, the health seeking behavior is presently mixed, and many caregivers, sometimes, visit elsewhere first before going to the CMAM HFs for treatment. This could be the possible explanation for the low median admission MUAC of 105mm which indicates that most admissions were admitted late into the CMAM program Stock-out of Data Tools and Routine Drugs leading to charges for OTP cards and Routine Drugs During the extraction of client information from the OTP cards, it was observed that pieces of papers were used as registration cards. These pieces were of varying sizes and colours, and are written upon in different formats, with most of the sensitive beneficiary/client information missing. Further inquiries revealed that there has been stock-out of data tools 41 A caregiver each in Maje and Katuka reported being referred by a CV. Others reported peer-to-peer referrals, and passive referrals by health workers from non-cmam HFs. 42 Dutse, Jahun, Birnin Kudu, and Miga LGAs 43 Bauchi and Kano 44 Two caregivers in Katuka CMAM HF, two CVs in Katuka HF, a provision shop seller in Nafara (3km from CMAM HF), a religious leader from Dangoli (about 15km from CMAM HF); a caregiver from Garko CMAM HF, Traditional healer of Tsirma (3km from Garko CMAM HF), a TBA, religious leader, majalisa and provision shop seller in Jama ar Dawa reported that traditional herbs are used to treat malnutrition because they are easily accessible in their communities. A Community volunteer from Maje CMAM HF reported that traditional medicine is preferred for treating malnutrition in communities. Traditional healer in Yelwa community (about 6 km) from Kwanda also reported that herbs are preferred for treating malnutrition in communities. 45 This was reported by the rest of the interviewees in the various communities and HFs. 26

27 leading to HWs resorting to using different types of papers to record client information. Some HWs reported that they used to contribute money among themselves to photocopy beneficiary/otp cards for clients 46. Stock-out of drugs and data tools was reported by HWs in all the CMAM HFs. However, the HWs did not have a clear reason for charging beneficiaries varying charges for the OTP cards and routine drugs (especially, Amoxicillin and ACT). Some caregivers-in-program 47 reported that they pay for OTP registration cards, and are also charged if their ration card got torn or missing. 48 This information was also, collaborated by that given by provision shop sellers, Achaba 49 riders, husbands of care-givers in-program, and four HWs 50. A CV in Katanga intimated that sometimes they lend money to care-givers to buy drugs so that the care-givers would not be sent home without RUTF. This is because RUTF is never given to OTP beneficiaries who are unable to buy the routine drug (especially, amoxicillin) on admission, which is a directive given to health workers Stock-out and Misuse of RUTF, and occasional closure of OTP sites Stock-out of RUTF was reported for the month of October to December 2013 (see section Admission Trend). During the period mentioned, all the CMAM services were stopped as there was no RUTF available to be given to the beneficiaries. The period of cessation of CMAM services coincided with the time when Valid International conducted SLEAC survey in Kiyawa. The result of the SLEAC survey reported that Kiyawa LGA had only 7 SAM cases covered by the CMAM program out of 47 SAM cases found 52. At the time of the SQUEAC investigation, CMAM health facilities were reported to be closed. This was at the time when the HWs were having a refresher training on CMAM at the LGA secretariat. On the other hand, there were discrepancies in supply chain management of RUTF as was also observed in the field. In Katuka CMAM HF, some caregivers were observed going home without weekly RUTF ration. HWs on ground explained that the supply they received had finished in e process of giving few RUTF to a large number of beneficiaries. However, it was noted that RUTF was being misused by HWs and CVs. Two CVs in Katanga CMAM HF reported that they are given 3 satchets of RUTF weekly to motivate them so that they can lend a hand at the HF. They were quoted reporting The RUTFs given to us were meant as an incentive to motivate us. A community leader from Jama ar Dawa community under Garko CMAM HF had reported that a HW from Katuka CMAM HF usually bring RUTF to share to them in their community. However, the HW did not agree that this ever happened. Consumption of RUTF by adults and siblings of SAM children in-program were also reported 46 The health worker in-charge of Katanga, Katuka and Garko CMAM HFs. 47 All 8 caregivers interviewed in Katuka CMAM HF, and the one interviewed in Maje CMAM HF. 48 The amount reported varied from NGN 200 to NGN 50 for new cards and for replacing thorn or missing cards. 49 Achaba refers to motor cycle as a medium of transport. The Achaba riders usually carry caregivers from their communities to the HFs in OTP days and wait for them as their children get treatment before ferrying them back to the community. 50 Two health workers each in Maje and Katanga CMAM HF said that sometimes care-givers are asked to pay N100 for drugs. 51 This was also collaborated by the response of the SNO. 52 Chrissy B., Bina S., Safari B., Ernest G., Lio F. & Moussa S.; Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Survey of Community-based Management of Acute Malnutrition program; Northern States of Nigeria-(Sokoto, Kebbi, Zamfara, Kano, Katsina, Gombe, Jigawa, Bauchi, Adamawa, Yobe, Borno). Valid International. February

28 and observed 53, respectively. Many children and adults in communities reported that they have tasted the RUTF. Further inquiries showed that those children were never admitted SAM cases in CMAM HFs Large Turn-out of Beneficiaries, Long waiting Time, and Unavailability of Shades, Mats and Benches During the visits to the CMAM HFs, very large number of OTP beneficiaries was observed in Katuka and Katanga HFs. Garko and Kwanda HFs were also observed to have significant number of OTP beneficiaries, however, Maje CMAM HF had significantly few number of OTP beneficiaries. The large turnout of clients in CMAM HFs indicated accessibility, and willingness of caregivers to utilize the CMAM services. The lack of capacity to control the crowd at the HF resulted in long queues and waiting time for OTP beneficiaries. Additionally, lack of shades for the beneficiaries in all the CMAM HFs was observed. Beneficiaries gladly make use of Dogo Yaro tree as shelters, and sometimes have to queue under the sun, as observed by the SQUEAC investigators. There is also, lack of mats and benches for caregivers at the CMAM HFs. Most caregivers were also, seen sitting on the ground as they waited for their turn to be attended to by the HW. In Katuka and Garko HFs, it was observed that some local merchants display portable mats for caregivers to buy and use at the CMAM HF. Importantly, it was also, reported that some caregivers sleep-over at HFs of Katuka, Katanga and Maje HFs prior to the OTP day so that they would gain first entry and be attended to on time Health Workers Activities, Non-adherence to CMAM guidelines, and Training Health workers reported that sometimes they are under pressure by caregivers who insist that their children should be admitted into the CMAM program. A CV serving Katuka HF, reported that sometimes HWs may bow to the pressure so as to get the goodwill of the caregivers. Further enquiries to explain the erroneous discharge of clients by HWs as observed on the routine data revealed that HWs priority was to discharge clients as soon as the client has been in the program for up to 8 weeks without necessarily placing emphasis on improvement (in terms of weight or MUAC), or whether the beneficiary had recovered or not 54. This was possibly the reason for the erroneous discharge of clients with MUAC less than 125mm (as identified in the quantitative analyses of data). Furthermore, this situation may have largely contributed to the short length of stay from admission to cure (see the section on length of stay). Health workers were noticed not to be validating anthropometric measurements taken by CVs at the CMAM HF in Katuka. A health worker in Kwanda CMAM HF was observed assigning MUAC and weight arbitrarily without taking measurements. This confirms suspicions raised in section A caregiver was observed giving RUTF to a healthy sibling that accompanied her to Katuka CMAM HF. 54 The health workers were directed to follow this line of action since there are very large number of clients; those who were not recovered after eight weeks were to be discharged and referred to inpatient care. Nevertheless, this was observed to conflict with the guideline which directed that such action can be taken after a SAM child had stayed 12 weeks in the program (13 visits) and is yet to be recovered. 28

29 The large turnout of beneficiaries may have been felt as a burden by the health workers. Some of the health workers complained that they are not being paid any additional money for the additional job they are doing compared to their counterparts in non-cmam HFs. Health workers in CMAM HFs clamor for incentives so as to boost their morale. This was also pointed out by the NFP, who reported that he has received such complaints often from health workers in all the CMAM HFs. On the other hand, some caregivers reported that some health workers dispose bad attitude towards them, specifically, in Katuka and Katanga. Thus, terms such as friends of health workers and the rich were used by the respondents to describe those given preferential treatment, specifically, in Katuka CMAM HF. Some caregivers in Katanga and Katuka reported that threats of discharge from the program was been used to keep them complying with the health workers. However, some caregivers reported that they are treated well by the health workers. Passive referrals by health workers from non-cmam HFs in Kiyawa LGA was reported in all the CMAM sites by some of the caregivers interviewed during the health facility visits. Some caregivers interviewed in Katuka, and Kwanda said they were referred by health workers in and neighboring State (Bauchi) to the CMAM HFs in Kiyawa LGA. This was also confirmed by the health facility in-charge of Kwanda CMAM HF Data triangulation Information in the SQUEAC investigation was obtained when SQUEAC tools 55 were used on diverse sources using diverse methods. The triangulation process was done to ensure conformity of the evidence accumulated before adoption as a negative or positive factor that affects the program coverage. The information that was obtained was analyzed into barriers and boosters and the relationship between them drawn to give a clearer picture of the program coverage in concept maps. (See the activities that ensured in the investigation in annex 1 and concept map in annex 3). The barriers and boosters identified, and the sources and methods used to obtain such information during the investigation are tabulated in in the tables 3 and 4 below. Table 3: Sources and methods used to get information in the BBQ tool for Kiyawa LGA. Codes Source Method Codes 1 Client record Extraction A 2 Carers Semi Structured Interview B 3 Health facility Observation D 4 Majalisa Informal Group Discussion E 5 SNO In-depth interview C 6 Health worker Semi Structured Interview /In-depth B,C Interview 7 NFP In-depth interview C 8 CVs Semi Structured Interview /In-depth B,C Interview 9 Religious leader Semi Structured Interview B 55 Tools include the simple structure, semi structured questionnaires; observation checklists; illustrations in form of pictures, words or phrases; various forms to fill out extracted data from routine data etc. 29

30 10 Community leader Semi Structured Interview B 11 Provision shop Semi Structured Interview B 12 Patent medicine dealer Semi Structured Interview B 13 TBA Semi Structured Interview B 14 Traditional healer Semi Structured Interview B 15 Teacher Semi Structured Interview B Table 4: barriers, boosters & questions findings and sources of information BOOSTERS SOURCES BARRIERS SOURCE 1 Peer-to-peer referral 2B, 6B, 6C Poor Health seeking behavior 2B, 8B, 9B, 11B, 12B, 14B, 4E 2 Passive referrals by HWs in non- CMAM health facilities; Health workers from non-cmam facilities support in weekly services to beneficiaries in CMAM sites 2B, 6B, 6C Stock-out of Data tools (admission and ration cards) resulting in use of piece of paper as cards; Caregivers are charged money for OTP cards and for replacement of lost/thorn OTP cards in Katuka, Katanga, Maje, sites 1A, 2B, 3D, 6B, 8B, 6C 3 Good health seeking behavior 2B, 4E, 6B, 6C, 8B, 9B, 10B, 11B, 12B, 13B, 14B, 15B Generalized stock-out of routine drugs(amoxicillin) resulting in caregivers paying for Amoxicillin and ACT 2B, 6B, 8B, 11B, 6C, 8C 4 Large turnout of beneficiaries accessing CMAM services 3D, 6B, 6C Poor attitude of health workers; preferential treatment given to the rich/friends of health workers 2B, 3D, 5C 5 Willingness of caregivers to sleep over at OTP sites in order to access CMAM services 3D, 6C, 8C, 10B Over-burden of health workers due to very large number of beneficiaries accessing the OTP services; long waiting time at CMAM sites 3D, 6B 6 Good opinion about the CMAM program in communities 2B, 4E, 6B, 6C, 8B, 8C, 9B, 10B, 11B, 12B, 13B, 14B, 5B Lack of shades for beneficiaries in all the OTP sites; no mats and seats for OTP beneficiaries 2B, 3D, 6B, 7C, 8C, 7 Health workers are trained thrice on CMAM since inception (once yearly) 6B, 6C, 7C Wrong measurement of weight and MUAC by CV who are used for taking anthropometric measurement. HW assign MUAC arbitrarily. 3D, 8C 30

31 8 Good collaboration of health workers and CVs and good attitude of some health workers towards caregivers 6B, 8B Sharing and consumption of RUTF among healthy siblings and children beyond the age of five years; consumption of RUTF by adults; Caregivers does not understand how the program works 3D, 4E, 9B, 11B, 13B, 14B, 15B, 9 Referrals by some community CVs 2B, 6B, 8B Community volunteers are not motivated; conduct poor community mobilization and sensitization, very poor active case finding and defaulter tracing. Community volunteers clamor for incentives 6B, 8B, 6C, 7C, 8C 10 Good awareness of the program in communities 4E, 9B, 10B, 11B, 13B,14B, 15B Faulty supply chain management from LGA to CMAM sites leading to stock-out on OTP days 2B, 3D, 6B, 7C 11 Selection of CMAM site for intervention by SURE-P of Federal Government 3D Non-adherence to CMAM protocols a. Non-compliance with discharge criteria (discharge with MUAC <12.5cm b. Arbitrary assigning of MUAC measurements, evident by erratic MUAC movements on client cards 1A, 6B, 8B 12 7C, 8C High number of defaulters 1A, 6B, 6C, 5C, 7C Concept map The coverage team was split into two Teams that is, A and B. Each team drew a concept map illustrating the positive and negative relationships existing between positive and negative factors unveiled from the from the field visits interview and observations. Epigram softwareversion was used to draw the concept maps presented in the in annex Stage 2: Small Area Survey and Small Study. At the completion of stage 1, the major factors that may be affecting coverage in Kiyawa LGA based on the results of analysis of information gathered were identified. Four factors identified include; 1. Location and accessibility of CMAM Sites, 56 Epigram software was developed by Mark Myatt and is available on 31

32 2. Accessibility of services in terms of fees paid for OTP cards and routine drugs by beneficiaries, 3. Ability of the beneficiaries to stay in the program to recovery and factors affecting it (in terms of LoS and Defaulting) 4. Quality of the CMAM services offered in terms of adherence to the CMAM protocol by HWs. To investigate the heterogeneity of the program probable coverage and factors that lead to program failure, hypothesis were formulated around two of the above factors (that is Point 1 and 2). Two small studies were conducted to study each of point 3 and 4 above. Each of the factors were studied in detail and reported below Small Area Survey Hypothesis 1: Coverage is homogeneous in all areas, and is classified as above 50% 57 in wards where CMAM sites are located, as well as in wards where CMAM sites are not located. To test this hypothesis a total of four communities which are all distant from existing CMAM HFs, and are located in Wards without CMAM HFs were selected. On the other hand, four communities which are not distant from CMAM HFs and are in Wards where CMAM HFs are located were selected. Small area survey was used in each of the selected communities in order to test the formulated hypothesis Sampling Methodology The villages were selected purposively based on the characteristics used in setting the hypothesis. Active and adaptive case-finding 58 was conducted in the communities during the small area survey Case Definition Severe Acute Malnutrition (SAM case) was defined as Children (6-59) months, with MUAC <115mm and or bilateral pitting oedema. SAM case covered: Refers to a SAM case identified as defined above and is currently enrolled in a CMAM site or Stabilization Centre (SC). The status is verified when beneficiary shows the investigator the RUTF packets and/or ration Card. SAM case not covered: Refers to a SAM case who is not currently enrolled in a CMAM program or the SC. The case is also confirmed as not in the program when the beneficiary shows the investigator the RUTF packets and/or ration Card. 57 This is based on the 2 standard 3 classification of coverage as less than 20% as low, 20% - 50% as moderate, and greater than 50% as high for rural areas. 58 See the procedure of the active and adaptive case finding annexed to this report 32

33 Recovering case: A child (6-59) months old, with MUAC above 115mm and is enrolled in a CMAM program at the time of the investigation. This case is verified when beneficiary shows the investigator the RUTF packets and/or ration Card Result of Small Area Survey The results of small area survey are presented in the table 5 below Table 5: Simplified Lot Quality Assurance classification of small area survey results Coverage Wards Location Total SAM ( ) Decision rule = Covered(C) Not covered (NC) Recovering case (RC) Hypothesis basis Wards without CMAM sites Wards with CMAM sites Balago Kalkutan 5 = 10 = Malkiba Shuwarin Gurduba Fake Chuwina Katanga Katanga 10 = 11.5 = Garko Ali Sabon Gari Maje Tesher Gamji Katuka Chikin Gari Figure 14: Reasons for not attending the CMAM program-small area survey Interpretation of the results LQAS classification technique was applied and the results are as follows: Wards without CMAM HFs 33

34 The number of SAM cases found that were covered was 11. The Decision rule calculated above = 10 Since 11 is greater than 10, the coverage in the Wards without community volunteers was classified as above 50%. Wards where CMAM HFs are located The number of SAM cases found that were covered was 15. The decision rule calculated was 11. Since 15 is greater than 11, the coverage in Wards where CMAM HFs are located, is therefore classified as high (above 50%) Small Studies Two small studies were conducted. A quantitative small study was done to verify hypothesis 2, based on accessibility of CMAM services in terms of hidden fees paid by OTP beneficiaries. While a descriptive small study was done to shed light on defaulting in Kiyawa LGA CMAM program. Hypothesis 2: CMAM services in Kiyawa LGA has hidden charges incurred by more than 50% of the beneficiaries Sampling Methodology Two CMAM HFs (Katuka and Katanga) where more than half of all the program beneficiaries in Kiyawa LGA are accessing CMAM services were purposively chosen. On an OTP day, twenty caregivers in-program in each of these CMAM HFs were selected randomly by balloting. The selected caregivers were asked questions bordering on the fees they pay for beneficiary/otp registration cards and routine drugs Case Definition Care-giver with a child (6-59) months currently enrolled in the CMAM program in Kiyawa LGA. Incurred hidden charges: Refers to care-giver as defined above who paid hidden fees for OTP registration cards and drugs Did not incur hidden charges: Refers to care-giver as defined above who did not pay hidden fees for OTP registration cards and drugs Result of Quantitative Small Study The results of the small study are presented in the table 6 below: 34

35 Table 6: Simplified Lot Quality Assurance Classification of small study results CMAM HFs Care-givers inprogram Interviewed ( ) = Experienced hidden charges Did not experience hidden charges Katuka Katanga Interpretation of the results LQAS classification technique was applied and the results are as follows: Hypothesis Two: Katuka CMAM HF The number of caregivers out of the total sampled who paid hidden charges for OTP card and drugs were 18. The Decision rule calculated above was 10 Since 18 is greater than 10, the number of beneficiaries who paid hidden charges to access CMAM services was classified to be more than 50%. Katanga CMAM HF The number of caregivers who paid hidden charges for OTP card and drugs were 16. The Decision rule calculated = 10 Since 16 is greater than 10, the number of beneficiaries who paid hidden charges to access CMAM services are classified as greater than 50% Small Study on Defaulters A descriptive small study was conducted to understand the major reasons for defaulting in Kiyawa LGA CMAM program. From the client records extracted (see table 1 and session ), it was observed that about half of the exits of all the clients were defaulters. About two-third of the defaulters were from Katanga and Katuka CMAM HFs. Furthermore, about half of these defaulters from Katanga and Katuka CMAM HFs are from Kiyawa LGA, while the rest are from neighboring LGAs. Therefore, Katanga and Katuka HFs were purposively selected to study defaulting. Two villages under Katuka CMAM HFs and a village under Katanga HF where majority of the defaulters were observed to be coming from were selected. 35

36 Table 7: CMAM HFs and villages selected for defaulter study CMAM HFs Villages Number of Defaulters in the Database Number traced Katuka Nafara 9 2 Katuka 10 0 Katanga Kawari Results of Descriptive Small Study Nafara community had nine defaulters on the database of client records. However, during the defaulter tracing, only two out of the nine defaulters were traceable. The community leader of Nafara explained that the remaining names (both that of the caregivers and children, as well as their family name) are not from their community. The community leader pointed further that those beneficiaries from Nafara Kudu community in Birnin kudu LGA, usually report that they come from Nafara of kiyawa LGA while accessing the CMAM services. The second community, Katuka very close to the CMAM site was also visited to trace 10 defaulters obtained from the client records. A CV that works in the Katuka CMAM health facility was chosen as the village Guide. When the names of caregivers and children were mentioned, the Guide confirmed that none of the names were from Katuka. Another CV from Katuka community was recruited as a Guide for the defaulter tracing, however, she also said they have no such names in their community. Finally, a third CV was chosen as a the village guide to trace the defaulters, and he confirmed that beneficiaries from other communities mostly outside Kiyawa LGA uses Katuka as their address on their OTP cards to avoid being rejected since they are not under Kiyawa LGA. However, in Kawari village under Katanga CMAM HF, twenty one defaulters were traced and interviewed out of all the twenty-two defaulters obtained from the client database. The responses of all the defaulters traced and interviewed during the defaulter study are summarized in annexed 6 of this report. The analysis of the reasons of default are analyzed in figure 15 below. 36

37 Figure 15: Analysis of reasons for default and condition of cases found in the study in stage Conclusion of Small Area Survey and Small Study Hypothesis 1: The result of the LQAS comparison of the Wards with CMAM HF and the Wards without CMAM HF pointed out possible homogeneity of coverage across Kiyawa LGA. The hypothesis tested positive for probable homogenous coverage in Kiyawa LGA across political wards. Therefore, the hypothesis that coverage is probably homogenous in all Wards despite location of CMAM site was accepted. Hypothesis two Forty 59 (40) caregivers were randomly sampled from two CMAM HFs and privately asked if they incurred hidden charges. Initially, caregivers were reluctant to respond because they were threatened by the health workers that they would be expelled from the program. Therefore no reliable information was obtained from these caregivers. In the face of this development the SQUEAC investigators adopted another method of enquiry and interviewed caregivers who had received the RUTF ration from the CMAM HFs, and were on their way back home. Thirty-four caregivers confirmed individually that they all paid NGN200 (about USD1.25) for beneficiary/otp cards on their first visit. Additionally, NGN Caregivers in each of the facilities of Katanga and Katuka. 37

38 was also paid by each of the caregivers to buy a bottle of Amoxicillin syrup (routine drugs). The caregivers reported that health workers in the CMAM HFs usually send any caregiver that is not ready to pay any of these charges home. The team also observed a caregivers paying money to a health worker for Amoxicillin syrup at Katuka CMAM HF. Additional information from two Achaba riders, and three commercial motorists who were waiting to convey caregivers back home, a gateman, and a husband of caregiver-in-program in Katuka and the village head of Chikin Gari in Katuka confirmed that the OTP cards are sold for (200) and torn/misplaced ones replaced with (100). More so, the in-charge of Katuka CMAM HF finally confirmed that caregivers buy routine drugs as a measure so that the HF can be able to purchase the same for them. The in charge noted that the SAM children would deteriorate often because of lack of routine drugs at the HF. This meant that the routine drugs were arranged to be delivered by the NFP and then given to the caregivers at a cost on OTP days. The In-charge also confirmed that there was stock-out of beneficiary/otp cards for many months leading the health workers to make photocopies with their money but could not confirm that they collect money from the caregivers as payment for the cards. While the small study was going on, it was observed that at a particular point, the CMAM services stopped, and caregivers sent home without RUTF ration for the week. Investigation by the SQUEAC team revealed that the RUTF supply for the CMAM HF was exhausted. Therefore, caregivers numbering more than two hundred (including those referred by the SQUEAC investigation Team during the Small Area Survey active and adaptive case finding) went home without RUTF. Caregivers and health workers alike confirmed that it was not the first time RUTF supply was exhausted and caregivers sent home without RUTF ration. Therefore, the small study was concluded by accepting the hypothesis that more than half of the CMAM beneficiaries incur hidden charges while accessing the CMAM program in Kiyawa LGA. On defaulting, it could be seen that many of the defaulters who claim to be from Kiyawa LGA give wrong address. Therefore, most of the defaulters in the Kiyawa LGA CMAM program can be said to be from outside Kiyawa LGA (the catchment area). 6.4 Developing the prior Histogram of Belief. A histogram of beliefs was constructed with teams discussing and reaching consensus on what could be the most likely coverage in Kiyawa LGA CMAM program. The belief of the coverage team on what the program coverage could be was obtained, and minimum and maximum coverage was also identified by individuals within the team. Individual team members were told to write on a paper what they believed should be the coverage in Kiyawa LGA. This was used to construct a histogram of belief of the coverage team. The central tendency of the histogram on the coverage of beliefs was set at 50% with a minimum 30 % and maximum 70%. Prior 1: Histogram of belief = 50% 38

39 6.4.2 Concept Map The SQUEAC investigation team was split into Team A and B, and each team drew a concept map based on the barriers and boosters obtained in the field. A minimum of 1 score was given to a factor that bore minimum impact of coverage while a maximum score of 4 was given to a facto that had a maximum impact on coverage. Team A concept map has a total of 20 barriers, and 12 boosters while team B had 19 barriers and 12 boosters. To ascertain the contribution of positive and negative factors to the program coverage the following calculations were done: Prior calculated from concept map Team A: =20 4=80 =12 4=48 = {(0+48)+(100 80)} 2 =34 Team A prior estimation=34% Prior calculated from concept map B: =19 4=76 =12 4=48 = {(0+48)+(100 76)} 2 Team B prior estimation = 36% Average prior calculated from concept map of Team A and Team B =(34+36)/2=35 =36 Prior 2: Prior developed by use of concept map =35% Un-weighted barriers and boosters The barriers and boosters were consolidated and refined, and reduced to 12 barriers and 11 boosters in number. Using the largest number 60, the maximum score was scaled so that neither the sum of positive scores nor the sum of the negative scores can exceed 100%. Thus: Therefore; 60 The largest between the list of barriers and boosters is the barriers (12 in number). This is the number used to estimate the value that can be assigned to the score indicating maximum impact. Here this value was assigned to each of the barriers and boosters assuming that each has an equal impact on program coverage. 39

40 = = 8.33 =8 To calculate the contribution of barriers and boosters to coverage: =12 8=96 =11 8=88 = {(0+88) + (100 96)}/2=46 Prior 3: Prior estimated from un-weighted barriers and boosters = 46% Weighted barriers and boosters. The coverage team discussed and weighted each of the barriers and boosters with regards to their perception on the contribution of each barrier or booster to the coverage of the CMAM program in Kiyawa LGA. To reach on a score, team members discussed extensively before finally agreeing upon a score. The highest possible score assigned to a barrier or booster was 10 while 1 was the lowest score. The table below shows that the weight assigned by the teams and average/final score given to the barriers and boosters. Table 8: Weighted barriers and boosters of Kiyawa LGA CMAM Program No BOOSTERS Average Score BARRIERS Average Score 1 Peer-to-peer referral 10 Poor Health seeking behavior 2 2 Passive referrals by HWs in non- CMAM health facilities; Health workers from non-cmam facilities support in weekly services to beneficiaries in CMAM sites 6 Stock-out of Data tools (admission and ration cards) resulting in use of piece of paper as cards; Caregivers are charged money for OTP cards and for replacement of lost/thorn OTP cards in Katuka, Katanga, Maje, sites 8 3 Good health seeking behavior 6 Generalized stock-out of routine drugs(amoxicillin) resulting in caregivers paying for Amoxicillin and ACT 6 4 Large turnout of beneficiaries accessing CMAM services 10 Poor attitude of health workers; preferential treatment given to the rich/friends of health workers 3 40

41 5 Willingness of caregivers to sleep over at OTP sites in order to access CMAM services 5 Over-burden of health workers due to very large number of beneficiaries accessing the OTP services; long waiting time at CMAM sites 3 6 Good opinion about the CMAM program in communities 10 Lack of shades for beneficiaries in all the OTP sites; no mats and seats for OTP beneficiaries 3 7 Health workers are trained thrice on CMAM since inception (once yearly) 3 Wrong measurement of weight and MUAC by CV who are used for taking anthropometric measurement. HW assign MUAC arbitrarily. 1 8 Good collaboration of health workers and CVs and good attitude of some health workers towards caregivers 6 Sharing and consumption of RUTF among healthy siblings and children beyond the age of five years; consumption of RUTF by adults; Caregivers does not understand how the program works 3 9 Referrals by some community CVs 3 Community volunteers are not motivated; conduct poor community mobilization and sensitization, very poor active case finding and defaulter tracing. Community volunteers clamor for incentives 7 10 Good awareness of the program in communities 10 Faulty supply chain management from LGA to CMAM sites leading to stockout on OTP days 3 11 Selection of CMAM site for intervention by SURE-P of Federal Government 4 Non-adherence to CMAM protocols c. Non-compliance with discharge criteria (discharge with MUAC <12.5cm d. Arbitrary assigning of MUAC measurements, evident by erratic MUAC movements on client cards 7 12 High number of defaulters 7 Total 73 Total 53 Therefore: = (73+(100 53) 2 =60 41

42 Prior 4: Prior estimate from weighted barriers and boosters = 60% Triangulation of Prior Prior estimate 61 was then calculated by triangulation of all the prior estimates obtained from various methods. It is illustrated by figure 16 below. = ( )/4=47.75 Prior mode = 47.75% Figure 16: Illustration of triangulation of prior Bayes Prior Plot and Shape Parameters The teams were able to estimate (or give an informed guess of) the lowest and highest possible coverage 62 for Kiyawa LGA CMAM program. The SQUEAC investigation team believed that coverage of CMAM in Kiyawa LGA could not be lower than 30%, and not higher than 70%. Then alpha (26.8) and beta prior (28.2) shaping parameters were calculated using the BayesSQUEAC calculator 63. This also resulted into a Beta-prior distribution plot using the SQUEAC calculator software, version 3.01 as shown in the figure 17 below. The Bayes plot of the prior also suggested a sample size that was adopted in the likelihood survey described below. 61 The average of the coverages is a credible value of the mode of the prior. It is also referred as the mode of the probability density of the coverage. 62 These are also referred as minimum probable value and maximum probable value for coverage. 63 The BayesSQUEAC calculator can be downloaded free from 42

43 Figure 17: BayesSQUEAC beta-prior distribution plot showing the shape parameters and the suggested sample size 6.5 Stage 3: Wide area (likelihood) survey After developing the prior mode using information collected in stage 1 and 2, the likelihood survey was built into the Kiyawa SQUEAC investigation to add to the existing information (analyzed in stage 1 & 2). This was done so as to provide a headline coverage of the program. The procedures of implementing the wide area survey are described below Calculation of Sample Size and Number of Villages to be visited for likelihood survey The number of representative sample of SAM cases was calculated using the BayesSQUEAC calculator to be 53 SAM cases at 10% precision (results are expressed at CI; 95%). The number of villages that was needed to be visited to obtain a minimum of 53 SAM children aged 6-59 months was calculated using the formula below: Nvillages = (N SAM cases) (N(median population size all ages) x percentage of under fives in the population x prevalence )

44 Table 9: Parameters for sample size calculation for likelihood survey Parameters Value 1 SAM cases 53 2 N(Median population size of all ages) Percentage of under-fives in the 18% population 64 4 SAM prevalence % The median population was preferred to average population in calculating the number of villages to be visited. The median population size for all the communities/settlements was calculated to be 400 and was used to calculate the number of villages to be visited in the likelihood survey. Therefore the sample calculation is given as: Nvillages = 53 ( x 0.013) = 56.6 =57 57 villages had to be visited so as to get a minimum of 53 SAM cases Quantitative sampling framework Spatial sampling of villages was done by dividing the Kiyawa LGA maps into quadrants. The quadrants were numbered accordingly. Only quadrants having up to half or more of its area covering Kiyawa LGA map was selected. Furthermore, diagonals of the quadrants were drawn so as to identify the center of each quadrants with a red dot. A total of 19 quadrants were selected as shown in the figure 18 below. Then the number of villages (n) to be visited in each quadrant was calculated as follows; to be visited in each quadrant = = 3 =3 The distribution of the quadrant on the Kiyawa map are illustrated in the figure below: 64 Source: National Bureau of Statistics 65 Severe Acute Malnutrition results of Mid Upper Arm Circumference (MUAC) for the National Nutrition and Health Survey. May

45 Figure 18: Kiyawa map divided into quadrants for spatial sampling of villages Case Finding Method and Case Definition Active and adaptive case-finding method was used during the wide area survey. The case definition was a child: Aged (6-59) months With a MUAC of less than 11.5 cm, and or With bilateral pitting oedema Qualitative data Framework During the likelihood/wide area survey, each SAM case that was identified and was not in the CMAM program 66 was regarded as non-covered case. Therefore, a questionnaire was administered to the non-covered caregiver so as to collect information on possible reasons for the SAM child not being in the program. The analysis of these reasons or barriers to access and uptake is illustrated in figure 19 and table 9 below 66 As verified by show of RUTF or ration card by beneficiary 45

46 Figure 19: Barriers to program access and uptake-wide area survey (WAS) Table 10: Barriers to program access and uptake-was Reasons for non-attendance (barriers) Frequency Lack of knowledge or have wrong information 21 Caregiver does not know child is malnourished 14 The carer didn t know that her child can be readmitted 3 No knowledge of the program 1 Co wives convinced her that her child is healthy 1 A CV had told caregiver that her child was not eligible. 1 Caregiver thought it was necessary to be enrolled at the hospital first 1 Limited access to services 7 No transport fare 3 Far distance 3 Caregiver can t afford money paid for card and drugs 1 Rejection 21 The child has been rejected before as not eligible 5 Caregiver said child has been discharged CMAM HF 12 Caregiver thought her child would be rejected 2 The caregiver said the child was rejected after defaulting 1 Rejected by health worker after child relapsed 1 Caregiver priorities prevents attendance 6 Caregiver said she was too busy to attend 3 The mother is sick 3 Others Husband refused caregiver to attend 3 Child vomits or stools after eating RUTF 3 Caregiver prefers patent medicine dealer 1 Caregiver was not given RUTF when she attended the CMAM HF 1 46