Ethiopian Population Based National TB Prevalence Survey Research Protocol

Federal Ministry of Health Ethiopia Ethiopian Population Based National TB Prevalence Survey Research Protocol Amendment Request July, 2010 Addis Ababa 0

Contents 1. Introduction... 4 2. Aims an d Objectives... 6 3. Study Design... 7 4. Sampling methods... 11 5. Screening strategy... 1414 6. Case definition...18 7. Organization... 20 8. Training... 24 9. Survey operation... 2525 10. Data management and analysis...34 11. KAP study... 3938 12. Quality assurance... 3938 13. Reporting and dissemination of results 39 14. Budget (in USD)... 40 15. Ethical Consideration... 4944 16. Timeline... 51 17. References... 54 18. Annexes... Error! Bookmark not defined.55 Annex 1: Census Form (household register)... Error! Bookmark not defined.55 Annex 2. ID and invitation card... Error! Bookmark not defined.56 Annex 3 a: Individual Survey Form... Error! Bookmark not defined.57 Annex 3 b: TB history interview... 59 Annex 4: Field x-ray log book... 60 Annex 5: Central x-ray log book... 61 Annex 6: Laboratory Register for TB suspects... 62 Annex 7: TB cases register... 63 Annex 8: Central x-ray reading and sputum results... 64 Annex 9: Cluster report... Error! Bookmark not defined.65 Annex 11. Questionnaire for KAP study on TB... 75 Annex 12. Survey Consent Forms and Information sheets... 76 1

Abbreviations AFB CXR CSA DOTS EHNRI FMOH HIV ISN IUATLD LJ MDG MDR TB MTB NTP PI PPS SC SCT TAG TB WHO ZN SOPs Acid fast bacilli Chest x-ray Central Statistics Authority Directly Observed Therapy, Short-course Ethiopian Health and Nutrition Research Institute) Federal Ministry of Health Human immunodeficiency virus Individual Survey Number International Union Against TB and Lung Disease Lowenstein-Jensen Millennium Development Goals Multi-drug resistant tuberculosis Mycobacterium tuberculosis National TB Programme Principal Investigator population proportion to size Steering Committee Survey Coordinating Team Technical Advisory Group Tuberculosis World Health Organization Ziehl Neelsen Standard operating procedures 2

Summary of the proposed study This proposal is to conduct a nationwide TB prevalence survey to establish the burden of TB disease, estimate the proportion of cases detected and treated by the National Tuberculosis Programme and identify causes of low case detection in Ethiopia. This study is being conducted by a research team from the Ethiopian Health and Nutrition Research Institute (EHNRI), the Federal Ministry of Health (FMOH), in collaboration with local partners and technical assistance by the World Health Organization (WHO) and funding through the Global Fund, USAID/TBCAP and the WHO. The budget is approximately 2.8 million U.S. dollars. The survey will be held using a representative sample of approximately 46,514 adults in 85 population clusters, stratifying by rural, urban and pastoralist areas. The field operation is estimated to take approximately 8 months to complete. Cases of TB disease will be detected using a TB symptom screen and a chest X-ray for all eligible persons aged 15 years and above. Those with TB symptoms and/or an abnormal chest x-ray will submit two sputum specimens that will be examined microscopically for acid-fast bacilli and cultured for Mycobacterium tuberculosis in the National Reference Laboratory (EHNRI) located in Addis Ababa. Individuals detected with TB will be referrred for appropriate case management. 3

1. Introduction Tuberculosis (TB) is a chronic infectious disease caused by Mycobacterium tuberculosis (MTB). Almost one-third of the world population (about 2 billion people) is infected with M. tuberculosis and during the past decade there has been a resurgence of tuberculosis. Currently, TB is the leading cause of mortality among infectious diseases worldwide but 95 percent of TB cases and 98 percent of deaths due to TB occur in developing countries (Blanc and Uplekar, 2003). TB is one of the diseases covered by the sixth Millennium Development Goal (MDG), with the target of halting and reversing the prevalence of TB by 2015 (United Nations, 2010). The DOTS (Directly Observed Therapy, Short course) strategy is the backbone of global tuberculosis control, with the aim of finding 70 percent of new smear positive cases of tuberculosis and curing 85 percent of them (WHO, 2009). TB prevalence is one of the key indicators of the MDGs and the Global Stop TB Plan. TB prevalence surveys are one of the most effective tools to monitor the impact of the TB program, especially in countries where the incidence and prevalence are high. In many high TB incidence countries, routine surveillance data based on TB case notification are limited because not all persons diagnosed and on treatment for TB may be counted (e.g., underreporting in private and some public health facilities).as a result, an unknown proportion of TB cases remain undiagnosed and untreated. Thus, the true epidemiology of the disease remains unknown. In addition to monitoring program impact, data on TB incidence and prevalence are also essential to establish and re-evaluate the national TB program s objectives, strategies and interventions. Prevalence surveys have already been conducted in many of the top 22 high TB burden countries; few countries have conducted more than one. In Ethiopia there has been no national TB prevalence survey conducted to date. TB situation in Ethiopia Ethiopia is located in the northeastern part of Africa, also known as the Horn of Africa, In 2007, 4

the total population was estimated to be 74 million (Central Statistical Authority, 2007). Administratively the country is divided into nine Regional States and two City Administrative Councils (Addis Ababa and Dire Dawa). Each Regional State is further divided into Zones, then Woredas and Kebeles. Zones are the second administrative level after Region, and they are composed of Woredas (similar to districts), which are the third administrative level. Finally, Kebeles, which are within Woredas, are the lowest administrative level in Ethiopia. Kebeles consist of blocks or household groups (number and size of blocks vary), which are known as Gotes (or Gares) and have an average population size of 5,000 persons. In urban centers, the administrative structure is different, because City Councils don t have Zones and Woredas. Instead, Addis Ababa is composed by 10 sub-cities, each one composed of Kebeles. Dire Dawa is directly divided into Kebeles, At present, in Ethiopia there are a total of 94 Zones, 810 Woredas, and 15,022 Kebeles. According to the World Health Organization (WHO) Annual Global TB Report (2009), Ethiopia ranked seventh in the world for TB burden, with an estimated annual TB incidence (all forms) of 378 new cases per 100,000 persons and 163 new smear positive cases per 100,000 persons per year (WHO, 2009). The estimated prevalence of all forms of TB is 579 per 100,000 population and 286 per 100,000 for smear positive TB. The national TB control program has achieved 100 percent geographical and 92 percent health facility DOTS coverage. Despite the extensive expansion of DOTS service in the country, the program performance indicators, in particular the case detection rate for smear positive TB, remain unsatisfactorily low. For instance, even though the treatment success rate was 84 percent in 2007/2008, the case detection rate was 34.5 percent. Rationale Between 2007 and 2009, the WHO annual estimate of smear positive TB incidence in Ethiopia has increased from 152 to 163 per 100,000 persons. The current WHO TB estimates for Ethiopia are extrapolated from an exercise undertaken in 1997, using the assumptions of a 5

50 percent case detection rate, 48 percent DOTS coverage and trends based on high HIV prevalence countries in the region. It is questionable whether these outdated assumptions take into consideration crucial factors, such as the extensive DOTS service expansion and the recent establishment and expansion of Health Extension Workers in TB prevention and control activities in the community. Furthermore, the current adult HIV prevalence in Ethiopia is 2.3 percent, which is much lower than the prevalence of HIV infection in other sub-saharan Africa countries. Thus, these factors might have affected the estimation of the true TB prevalence and incidence and the distribution of the disease in Ethiopia. However, the quality of existing program and surveillance information on the ground concerning the TB burden and trends is not sufficient to decipher this. In order to clarify the TB epidemiological situation in Ethiopia and to strengthen the TB program using an evidence-based approach, the Ministry of Health is planning to conduct a population based national TB prevalence survey in 2010, with financial support from Global Fund TB round six grant, in collaboration with in-country partners working on TB, and technical support from WHO. 2. Aims and objectives Primary aim The primary aim of the national prevalence survey is to estimate the prevalence of pulmonary TB in Ethiopia in 2010, as a basis for evaluation of current performance in case detection and as a baseline measurement for subsequent surveys in the future. To estimate the prevalence in the general population, the prevalence among adults aged 15 year or more will be measured. Primary objectives 1. To determine the prevalence of smear positive TB 2. To determine the prevalence of culture positive TB 3. To determine the prevalence of symptoms suggestive of TB 4. To determine the prevalence of radiological abnormalities suggestive of TB 6

Secondary objectives 1. To measure the prevalence of cervical lymphadenitis among study participants; 2. To assess the knowledge, attitude, and practice of the population concerning TB 3. To assess health seeking behavior among participants with TB symptoms 3. Study Design A national cross-sectional multistage cluster survey will be carried out in 2010. The prevalence survey is designed to provide a prevalence estimate for the population for the entire country. It incorporates a representative sample of the national population so that conclusions can be drawn about the country as a whole. In order to increase the precision of the study and the accuracy of results, the sampling will be divided in three strata (urban, rural and pastoralist, as further explained at page12). Multistage cluster sampling will be performed. The selection of Woredas will be performed first utilizing probability proportion to size (PPS), followed by sampling of Kebeles (basic sample unit) by PPS. Within a selected Kebele, Gotes/Gares (household groups) will be selected to reach the required calculated sample size. The survey will be conducted among adults (aged 15 years and older). All eligible and consenting study participants will be screened for TB symptoms during one-on-one interviews and for radiologic evidence of TB using a chest X-ray (CXR). Screened persons with current TB symptoms (cough for 2 weeks or more) and/or abnormal CXR results will be asked to submit two sputum samples (spot and morning samples) for bacteriological confirmation. All sputum samples will be sent to the National TB Reference Laboratory (Ethiopian Health and Nutrition Research Institute) and examined with microscopy for acid fast bacilli. Culture will be performed on all morning specimens. Prevalent TB cases will be identified based on results of bacteriological examination (smear and culture) according to the case definitions in this protocol as recommended by the WHO task force on TB prevalence surveys. 7

Inclusion and Exclusion criteria Exclusion criteria for sampling frame: Of the 810 Woredas within the country, 38 will be excluded from the sampling frame from the beginning because it is not feasible to conduct the survey in these areas due to security and logistical reasons. Even though the population living in these 38 Woredas account for only the 3 percent of the total population of Ethiopia, excluding them may bias the results of the survey, which will be clearly stated in the final survey report and when conclusions are drawn from the study. Ultimately, the sampling frame will include 772 Woredas, which account for the 97 percent (77,174,699 persons) of the total population of the country. The population living in each Woreda has been derived during the last national census (2007). Exclusion criteria during Woreda/Kebele sampling: Once sampling is completed and if an entire selected Woreda is determined to not be accessible or if the survey is unexpectedly found to be not feasible in that Woreda (e.g due to natural disasters, epidemics, etc) another Woreda from the same Zone will be selected to replace the first one. Similarly, if the selected Kebele is not accessible or if the survey is not feasible for major unexpected reasons another Kebele from the same Woreda will be selected to replace the first one. Exclusion criteria in the selected Kebele: To increase the feasibility of data collection, and to avoid bias as a result of inclusion of special high risk groups, the following settings, which may fall into the selected clusters, will be excluded from the study: - Military compound 8

- Diplomatic compounds - Confined setting: Jail/prison, refugee camps - Hospitals - Schools and universities and dormitories - Orphanages - Monasteries - Homeless persons Individual eligibility criteria for participation in the study: The individual eligibility criteria for study participation take into consideration a person s age, residence in the cluster area and, in case of temporary visitors, length of their stay in the cluster area. Because bacteriological positivity is rare and collection of sputum samples is difficult in persons below 15 years of age, the exclusion of this age group will greatly reduce the workload while simultaneously having only a small effect on the prevalence estimate for the total population. A census will be carried out to provide updated and accurate information on the resident population in the selected survey areas, specifically focusing on migration and membership within each household. For this reason, living in the survey area for at least two weeks prior to the day of the census has been chosen as eligibility criterion. Individual inclusion criteria: Age > 15 years Residents who have stayed at least one night in a household during the 14 days prior to the census day Visitors who have stayed in a household for at least the past 14 days prior to the census day Individual exclusion eligibility criteria: Age < 15 years 9

Residents who have been away during entire past 14 days from a household Visitors who arrived and stayed in the household less than 14 days prior to the census. The inclusion/exclusion criteria for participation - involvement in screening take into consideration the eligibility criteria and the provision of informed consent. Individuals can be included in the survey screening only if they have been fully informed of the objectives and procedures of the study and have freely given informed consent. Individuals who are unable or unwilling to provide informed consent will not be included in the survey screening. The consent procedure will be carefully described and approved by an ethics review committee. Individuals below 18 years need the informed consent to be provided by a family member or a guardian 18 years or older. The individual inclusion criteria for study participation include the following: Eligible individuals, based on study criteria Individual informed consent provided (or one family member/guardian 18 years or older for persons below 18 years or disabled persons who are unable to provide written consent by themselves). Persons aged 15-17 who live independently, without any parental or guardian supervision will be asked to provide consent for themselves. The individual exclusion criteria for study participation include: Persons who refuse to provide consent Persons who are unable to provide consent or whose guardians cannot provide consent (e.g. severe physical or mental disability impairing ability to provide informed consent, or age<18 years without consent from an older guardian or parent) Individuals who are eligible for the study but refuse to provide consent will be counted as absentees. Additional study components 10

A KAP survey will be conducted among 10 percent of the study participants in order to capture information on general knowledge, attitudes and practices about TB. Individuals who have a positive symptom screening (defined as cough for 2 weeks or more), will be interviewed about their health seeking behavior, in order to clarify factors which affect service utilization, to better understand the utilization of the government and private health system, and possible reasons for not seeking treatment. Furthermore, study participants with a positive TB history in the last two years, will be interviewed to explore in details the classification of TB, the previous/current TB treatment, outcome, and responsible health facility. All TB cases identified during the survey will be offered HIV counseling and testing, as per national policy, after linking them with the routine health system. 4. Sampling methods Sample size determination: For the sampling frame, data from Central Statistics Authority (CSA) derived from the last national census conducted in 2007 were used. According to census data, the projected population size during the proposed study period (2010) is estimated to be 79,731,054. The required sample size for the prevalence survey is estimated to be 46,514. The sample size was calculated with the assumption that the WHO smear positive TB prevalence estimation for 2007 (286 per 100,000) might be an over estimation. So, we adjusted our prevalence estimate to be 200 per 100,000 to be more conservative in order to get an adequate sample size. To calculate the required sample we used the following formula: n= Z 2 p (1-p)/d 2 Where n- sample size p- estimated prevalence of smear positive TB. z- standard normal value at 95% confidence interval=1.96 d- relative precision We consider a design effect of 1.5, participation rate of 85% and relative precision of 20%. The proportion of adult greater than 15 is 55% (data extrapolated from last national census, 11

2007). n=1.96 2 x0.002(1-0.002)x1.5x0.55 = 46,514 (0.2x0.002) 2 x0.85 In order to select 46,514 individuals, 85 clusters/kebeles will be selected (548 persons/cluster). Stratification In order to increase the efficiency of sampling and precision of final results, stratification in sampling has been applied as follows. The general population of Ethiopia is generally divided in three sub-groups, according to living area and conditions: 1. Population living in urban area 2. Population living in rural area 3. Pastoralist or nomadic population In this study, the definition of urban population is for people who live in towns with a minimum population size of 2,000, and which is registered by Ethiopian Central Statistics Authority (CSA) as urban. The urban population represents the 16 percent of the total population. The rural population is defined as people whose main source livelihood is agriculture (other than pastoralist); the population living in this area represents the 76 percent of the total population. The pastoralist population is defined as people whose main source of livelihood is livestock, with which they move seasonally in search of fresh grazing land and water; pastoralists represent the 8 percent of the total population. Hence, this study will be conducted with three strata: rural, urban and pastoralist. Sample unit and sample size According to the experiences of prevalence surveys in Asia, field activities (census, screening 12

and examinations) for a cluster with 520-600 adults are comfortably completed in 5 or 6 days. In order to complete one cluster within one week, the size of a cluster unit will be approximately 550 eligible persons (> 15 years old), with accepted ranges of 520-600 persons per cluster. To select 46,514 people, this means a total of 85 clusters will be selected from the three strata. The number of clusters per strata will be allocated according to the population proportion of each stratum: urban, rural and pastoralists. For each stratum, Woredas will be initially selected by probability proportion to size (PPS) sampling. Then, in each selected Woreda, a list of Kebeles with the corresponding population will be provided by the Woreda authorities. Kebeles will be selected by PPS as a basic sampling unit. Precise mapping will be done in selected Kebeles. In case the selected Kebele is smaller than the minimum cluster size (520), another Kebele should be added. This will be done by going to the adjacent Kebele (Note: According to the census data, we do not anticipate this to be the case). In most cases the selected Kebele is bigger than the maximum cluster size (600), thus it will be necessary to conduct the survey in only a limited part of the Kebele. Within each selected Kebele, there are household groups, or blocks. A serial number will be assigned to each block and one block will be randomly chosen for the initial survey area. Once that is finished, from there, the team will continue to the next consecutive block to the north, and then moving in a clockwise direction until the desired sample size has been achieved. All persons meeting the eligibility criteria on the first day of the survey (census day) will be included as eligible persons. The number of clusters selected from each stratum is proportional to the population size living in that area: 1. Rural population: 63 clusters 2. Urban population: 14 clusters 3. Pastoralist population: 8 clusters 13

Within the pastoralist areas, the Somali region has a unique administrative structure, poor accessibility and limited availability of census data. Therefore, specific arrangements are required in order to select a number of pastoralist clusters from Somali region according to the population size. It was estimated that, out of the 8 pastoralist clusters, 4 clusters from the Somali region are needed. 5. Screening strategy Individual interviews and chest x-ray screening will be used to screen study participants. Participants presenting with no TB symptoms and a normal chest x-ray will not have to submit sputum samples, and will be defined as NOT having pulmonary TB. Individuals with TB symptoms, specifically cough for at least 14 days, and/or abnormal chest x-ray will be considered as eligible for sputum examinations and will be asked to provide two sputum specimens for smear microscopy and culture. TB symptom screening: Eligible individuals attending the survey site will be informed about the study procedures and will be asked to sign the informed consent form. Those who consent will be registered as study participants, and they will be requested to answer study questions from a pre-tested questionnaire translated into and explained to them in their local language. The inteview will be conducted by a qualified health worker. The questionnaire includes: Personal data, including study ID, name, age, sex. (Note: Name will be removed for data entry.) Questions concerning the presence of TB symptoms at the time of the survey. The 14

symptoms considered suggestive for TB according to the National TB Manual, are the following: o Cough and duration. For persons reporting cough for 14 days or more, the participant will be asked to answer additional questions at the end of the survey regarding health seeking behavior. o Presence of the following symptom(s): Expectoration, blood contained sputum, fever, chest pain, shortness of breath, night sweats, loss of appetite, body weight loss. o Enlargement of neck lymph nodes. Each study participant will also be examined by a health care worker to assess the absence or presence of swollen cervical lymph nodes. Past or present anti-tb treatment. In the case of history for TB or ongoing anti-tb treatment, the participant will be asked to answer a second set of questions at the end of the survey to explore in detail past TB history, previous treatment, outcomes, health care facilities, etc. Persons reporting cough for 14 days or more will be eligible for sputum exams. However, any of the TB symptoms may be used for eligibility for sputum exams for persons who do not participate in a chest x-ray. It is not anticipated that there will be many of these cases. Persons reporting cough for 14 days or more and/or neck lymph node swelling detected by an interviewer will be assessed and re-interviewed by a physician during the second interview. Chest x-ray screening All participants who completed the screening interview will proceed to the chest x-ray unit where an upright posteroanterior chest x-ray (full size film of 40x40cm) will be taken on the spot using a portable x-ray machine with an automatic processor. Instruction, including a review of the risks and benefits of the chest x-ray, will be given to the participant by a healthworker. Consent for the chest x-ray examination is included in the survey consent form. The film will be immediately developed and then read in the field by one experienced or trained physician An average of 150 persons per day are expected to receive chest x-ray examinations. Two x-ray technicians (radiograpers) and an assistant will be assigned to carry out chest x-ray examinations in the field. 15

The chest x-ray results will be classified as one of the following: Normal Any abnormality in lung field or mediastinum, including cavities, infiltrates, pleural effusion, hilar or mediastinal lymphadenopathy, pulmonary nodules, interstitial abnormalities and healed TB. Other abnormalities (cardio vascular diseases, goiter, injury, etc) If there are any abnormalities found in the lung field or mediastinum, participants will be requested to submit two sputum samples, regardless of the presence or absence of TB symptoms (cough for 14 days or more). In cases where the chest x-ray results are either normal or other abnormalities are found, and the participant didn t report cough for at least 14 days, he/she will not be requested to submit sputum samples. If abnormalities that require medical attention are detected, such as cardiovascular diseases, pneumonia, pneumothorax, etc, an efficient referral system to routine health care will be put in place. After the field reading, chest x-ray films will be sent to a radiologist in Addis Ababa for a central reading (St Paul Hospital, Radiology Department) for internal quality control and further classification. Bacteriological screening Two sputum specimens, one immediately (spot) and a second one early the next morning, will be collected from everyone with either TB symptoms, specifically cough for at least 14 days, and/or an abnormal chest x-ray: Persons who did not participate in the chest x-ray exam and who have any TB symptom are also eligible for sputum examination. At the survey site, participants will be provided with a labeled sputum container and will be requested to provide sputum on the spot in a pre-specified sputum collection corner. Participants will be clearly instructed to ensure that at least 3 ml of good quality sputum (rather than saliva) is produced. Supervision by field workers will be provided. Poor spot 16

specimens may be replaced with an immediate spot collection. A second labeled sputum container will be given to all persons eligible for sputum examinations and they will be requested to collect their morning sputum and to bring the container to the survey site the next morning. Alternatively, the samples may be collected from their homes by survey team members and placed in an ice box as soon as possible. If the morning sputum is of poor quality or quantity, an additional early morning sputum sample may be collected. The sputum samples will be placed immediately upon receipt into cold boxes at 4 degrees and then transported to the reference laboratory, preferably within 3 days (at most 5 days) and culture will be done no more than 7 days after sputum collection. Dispatch logs will be maintained for documentation of sputum transport details. The bacteriological examination will be performed in the National Reference Laboratory (EHNRI) located in Addis Ababa. The laboratory at Saint Peter s Hospital will be used as a back up when there is lack of space in EHNRI. EHNRI is experienced in performing TB diagnosis with microscopy (fluorescence and light), and culture (solid and liquid media). The institution is currently the country s research center for nutritional and public health research, training center for laboratory professionals and mandated to control the overall laboratory activities performed in the country. The institution is responsible for conducting quality assurance of regional laboratories. The first national multidrug resistant (MDR) TB survey was conducted by this institution. Laboratory procedures Sputum smear microscopy: Sputum samples received at the laboratory will be prepared on two frost-end labeled glass slides from each spot and each morning sputum. One slide from each sputum sample will be air dried, fixed and stained with auramine for fluorescence microscopy. The result of fluorescence microscopy will be recorded as the study result. The second slide from each sputum sample will be air dried, fixed and stored for staining with Ziehl Neelsen (ZN) only if the twin slide is found positive by fluorescence microscopy. The ZN staining will be examined by light microscopy and findings reported with grading of the concentration of acid fast bacilli in the sputum for purpose of comparing the results used in routine service (using the IUATLD 17

scale). Positive results will be confirmed by a supervisor. Culture: All morning sputum samples will be processed for culture. Samples will be processed according to standard procedures using modified Petroff s method for decontamination. Each processed sputum sample will be inoculated on to a slant of Lowenstein-Jensen (LJ) with glycerol and a second slant of LJ medium supplemented with pyruvate and incubated at 37 degrees with weekly examination for growth. Specimens that do not have colonies at 8 weeks will be defined as negative. Specimens with any growth will be examined for AFB using ZN staining and positive isolates will be characterized by deletion typing for speciation of mycobacteria (with polymerase chain reaction on heat killed cells using appropriate deletion primers according to standard procedures). Labratory results, smear, culture and identification will be recorded per specimen. Storage: Stained and unstained slides with sputum smears will be stored and made available for subsequent examination. All morning samples will be discarded after processing for culture. All spot sputum samples will be discarded without being processed for culture after the sputum smears have been read for AFB and are confirmed to be negative. Appropriate safety precautions will be taken during specimen processing and waste disposal. All sputum samples will be autoclaved before discarding or will be incinerated in closed cups in tightly sealed autoclave bags. 6. Case definitions Prevalent TB cases will be diagnosed according to the survey case definitions recommended by WHO task Force. Bacteriologically definite TB case: Study participant with one culture positive specimen with mycobacterium tuberculosis AND at least one of following conditions: 18

AFB-smear positive Culture positive in another specimen Chest x-ray positive from audited central reading TB symptom(s) positive The chance of cross contamination should be carefully assessed when a culture is positive with only one test tube with less than five colonies. If this is the case and when the possibility of cross contamination cannot be negated, this should not be categorized as a culture positive TB case unless there are also TB consistent chest x-ray findings (e.g. normal, smear negative, and culture positive with 3 colonies: not categorized as culture positive TB) Sputum smear-positive case: AFB-smear positive confirmed case, can be definite or probable study TB cases. Study participant with one AFB-S positive specimen AND at least one of following conditions: Culture positive TB ---definite case AFB-smear positive in another specimen AND no culture (TB) positive AND no isolation of MOTT --probable case Chest x-ray (audited reading) TB consistent AND no CTB positive AND no isolation of MOTT - probable case When culture/identification of mycobacterium tuberculosis is available, the definition of smear positive case may be different from the clinical definition by the routine programme. Isolation of mycobacterium other than tuberculosis is not rare in survey specimens. Two positive smear slides with MOTT isolates may not be considered as TB case in the study analysis. Bacteriologically confirmed TB case Study participants can be either a definite or probable case. The table below shows summary of the definition of prevalent TB cases in the survey. 19

TABLE. Overview of classifications according to culture and AFB-smear positive results Study TB case Definite Probable Classification of definite (CTB positive) TB cases according to AFB-S AFB-S negative, CTB positive AFB-S unknown, CTB positive AFB-S positive, CTB positive Classification of AFB-S positives Laboratory results CTB positive AND AFB-S negative CTB positive AND AFB-S unknown CTB: Mycobacterium TB Complex culture positive *Culture results missing, unknown or contaminated AFB-S positive TB confirmed case CTB positive AND AFB-S positive CTB indecisive* AND AFB-S positive Bacteriologically negative, but active TB suggested by chest x-ray: No evidence of bacteriologically positive TB, but chest x-ray strongly suggests TB disease, and the final decision is made by a central medical panel, consisting of at least three chest physicians and/or radiologist Possible TB cases will not be counted as prevalent TB cases in this study. However, they will be followed up according to advice by the central medical panel. The panel will review collected data and examination results of participants with any positive results (chest x-ray, smear, culture) to classify final study results and to provide advice for treatment or further medical interventions. 7. Organization Steering committee The steering committee is composed by FMOH directors, NTP managers, and representatives of national institutes, international organizations and funding agencies. The 20

steering committee has the primary responsibility for selecting the survey implementing organization and the principal investigator, designing the study, eliciting funding, ensuring the quality of survey implementation and disseminating the survey results. One senior staff of the Ethiopian Health and Nutrition Research Institute has been nominated as the Principal Investigator (PI), and he is the chair of the Steering Committee (SC). Specific terms of reference for the PI have been developed and approved by the SC. The SC meets on a monthly basis as well as when it is necessary to monitor progress and to provide support to survey teams. The Steering Committee is advised by a Technical Advisory Group (TAG), and delegates tasks and responsibilities regarding the day-to-day execution and management of the survey to the Survey Coordinating Team. Members of the SC may participate in monitoring and supervisory activities directly both in the field and at the central level. Survey Coordinating Team / National Survey Coordinator The Survey Coordinating Team (SCT) is composed of chiefs for the Lab, Radiography, Statistics and Logistic teams as well as the leaders of field teams (figure below). The SCT is responsible for carrying out the survey and to report to the SC. The Survey Coordinator is the chair of the SCT and is assisted by a Deputy Survey Coordinator. The Survey Coordinator is an NTP staff member engaged full time for the whole duration of the survey. He has responsibility for day-to-day management and is Executive Secretary. His terms of reference include: Day-to-day management of survey preparation and implementation Organizing preparation, training, pilots, field work implementation, data management and monitoring of progress and data quality Chair of the Survey Coordinating Team Report to the Steering Committee on progress and general monitoring issues 21

The Technical Advisory Group, or TAG, consists of national TB experts as well as international experts in survey, lab, radiology and epidemiology to provide timely technical advice to the survey coordinator and central units. The TAG also assists the SC and PI in training and quality assurance activities. The director of Medical/Diagnostic panel is in charge of medical decisions. It can be called upon by field teams, and can overrule the team s decisions on case management. The panel will be responsible for reviewing documents, x-rays and lab results for all suspected cases. Field teams Five survey teams will work in rotation; three teams will be on duty in the field to implement the survey and collect data, while two teams will be back at the center to summarize and submit field data, get maintenance, if needed, prepare logistics and rest. Each team consists of a fixed and a flexible part. The fixed part consists of staff from the 22

central level and its team members remain the same for the different clusters. The flexible part consists of staff from the Regional, Zonal and Woreda levels and differs for each cluster. Local health workers and community volunteers should be recruited to assist the field operations. For each field activity (team leadership, census, interviews, X-ray, lab) standard operating procedures (SOPs) will be laid out in the field manual. These will describe, for each field day, the tasks and responsibilities of the field team members. The fixed part of the field team consists of the following members: One team leader (physician or senior health professional) Three census takers/interviewers Two radio-technicians One radiologist or physician as a chest x-ray screening reader One lab assistant Two clerks Three drivers The flexible part of the team includes local staff from the selected Kebele, and includes at least 4 members. Training workshops will be held for regional and Woreda staff. Health centre staff and health workers will receive a brief training during the preparation visit by the central team approximately one month before the field data collection: Regional and/or Zonal TB coordinator 1-2 staff Woreda Health office 1 staff Hospital/Health center 1 staff Health extension worker: 1 staff Community volunteers will be recruited to assist field operations. Instruction will be given during the preparation visit and on the arrival of the survey team: Community Chief 1 staff Assistants/Volunteers 3 staff Interpreter 1 staff (depending on local conditions) 23

Guide 1 staff (depending on local conditions) Security staff 2 staff (depending on local conditions) Field team leaders & deputies Each field team will have a field team leader and a deputy field team leader. The field team leaders have direct responsibility for the implementation of the field work. Their terms of reference include: Lead the second pre-survey visits Responsible for logistics and organization of the field work Co-ordinate the day-to-day field work Communicate with local, Woreda and Zonal/provincial authorities on issues regarding the field work Responsible for completion of the field report The deputy field team leader will assist the field team leader and replace the field team leader when required. Terms of references and standard operating procedures for each unit and individuals in the organogram will be prepared in the survey operation manual. 8. Training In order to prepare the survey activities, the training of staff is necessary. The training will be organized in different steps and will include in-house training as well as field visit experience. Events will be organized at all levels to inform and orient the local authorities and personnel about the survey activities, including: Two week workshop to develop SOPs and to develop one field team to test survey instruments. Participants of the workshop will become trainers for other field team members. In-house 1.5-2 days workshop to give general guidance to all staff members. Specific capacity building training for lab technicians, x-ray techs, census interviewers, data management part by part training, also to include training on the forms, survey- 24

specific rules, census and interview procedures. Training will last for one-two weeks for each group. After the in-house staff training, an integrated field test will be arranged, in order to reproduce and gain experience in all field activities, including checking equipment (chest x-ray), collecting lab samples, sample storage and transportation, form filling, etc. The field test will take place near Addis Ababa and some modifications to the forms or to the data and participant s flow may be needed after the field test. After lessons learned from the field test are integrated into the new procedures, these will be pilot tested in the field. The timeline for these field trainings can happen during two months time. Pilot testing will be done by 2-3 teams, including all team leaders as well as some observers. The pilot survey will be performed in one cluster, which will not be included in the study sample. Instructional sessions will be organized for local health workers and volunteers during the preparation visit approximately one month before the field operations and on arrival day of the survey team. All SOPs should be developed before starting the training. 9. Survey operations The survey will be conducted using SOPs based on this protocol. The following SOPs will be developed: SOP Pre-survey visit including sampling step to designate survey census area SOP Guidance to local health workers and volunteers SOP Field data collection, including check list of necessary items for each step o SOP Field Management for team leader including reporting to local authority (Compile all field SOPs and define the role of team leader) o SOP Survey Census o SOP Enrolling participants and informed consent + poster and leaflet o SOP Group Instruction: standard instruction 25

o SOP Receptionist o SOP Screening interview o SOP Chest X ray at field level o SOP Data Checker o SOP Medical management and referral o SOP Re- interview o SOP Laboratory field procedures including collection and transportation instruction to produce quality sputum SOP Guidance to local o SOP Logistics support and administration including check list of necessary items o SOP Data management o SOP Chest X ray reading at central level o SOP Laboratory procedures for central level o SOP Monitoring and Supervision o SOP Central Panel to classify and manage detected cases o SOP Feedback the survey results (individual and cluster reports) Procedures before field survey Before the start of the field work, all Regional, Zonal, Woreda and Kebele authorities involved will be informed about the objectives and outline of the survey and clearance, approval for secondment of staff, and co-operation will be requested. Furthermore, a two day sensitization workshop for Regional Health Bureau staff and other partners working on TB will be held, and a two day orientation meeting will be organized addressing the Woreda authorities (District Health Office and TB focal persons). Pre-visit Once the Woredas have been selected for the survey sample, the central team will contact the local authorities requesting the list of Kebeles within that Woreda for the cluster sampling, using Kebele as the sample unit. After selection of the Kebele, the study team leader will conduct a pre-visit to the Kebele. The pre-visit takes place two months before the survey, and the aims of the pre-visit are: 26

Checking accessibility of the Kebele (security, military area, roads, distance from main road, ) Drafting a map of the area Assessing infrastructure (health structures nearby, school, water and power supply, hotels, etc) Assessing the most convenient time when to conduct the survey, taking into account seasonality, mobile pastoralist populations, etc. Identifying a reference lab where to send the specimen Designating survey areas (survey camp, chest x-ray site, etc.) Identifying diplomatic compounds/hospitals/boarding schools/jails/refugee camps/etc. in the Kebele to be excluded from the survey Meeting the local authorities and local health workers Informing the local authorities about the survey contents and procedures Asking local authorities for collaboration for community involvement If the Kebele is accessible, all the above mentioned activities are carried out in the pre-visit. If the Kebele is not accessible and the team decides not to conduct the survey in that selected Kebele, the exact reasons should be recorded and another Kebele in the same Woreda will be selected to replace the first one. Inaccessibility by car cannot be a reason to exclude the selected Kebele. Special arrangements will be put in place to reach Kebeles not accessible by road. A report of the pre-visit will be written by the team, including a map of the area. Second visit Approximately one month before the survey, the team leader performs the second visit to the selected Kebele. The aims of the second visit are: Meeting the local authorities, local leaders and health workers and informing them about the survey procedures Designating household blocks to be included in the survey census Sensitizing the community about the survey Informing the community about the exact time of the survey 27

Creating linkages with health facilities nearby for clinical management of identified patients/sick people Asking the local authorities/health workers to prepare the population list and identify local health worker/volunteers who will assist during survey Training local health workers and providing instruction to volunteers Deciding on the survey base (school, community centre, religious facility, house, open space, etc.) and accommodations of the survey team. Developing/revising of map of designated survey areas The community will be informed and sensitized during the second visit of the team. With the support of local authorities and local health workers, the survey procedures will be explained. Field survey procedures Day 1: Census taking The first day of the survey will be dedicated to the census; the census takers (3) of the team and local volunteers will visit the households where the survey will be undertaken under the guidance of the team leader. The SOPs for the household census will be developed to include: Providing a household number to every household in the designated census area Checking the pre- list which was to have been already completed by local health staff and completing the census form to add household members and visitors not already listed in the pre-list Confirming study eligibility of household members and visitors and issuing survey invitation cards with Individual Survey Number (ISN ) to those eligible Explaining the survey procedures to households members The team will mark the identified households with a household number. The census form (also known as the household registry) (see Annex) will capture all individuals living in the selected households, including children < 15 years old, based on the population list provided by the local authorities. The purpose of registering even children < 15 years is to identify the proportion of children in each cluster. Through household visits and rapid interviews with the household head/members on the census day, the pre-list of registered persons produced by local authorities will be updated. Each registered inhabitant will receive an Individual Survey 28

Number (a 7 digit ISN). In the Census form, an Individual Survey Number should be allocated to all family members and also to absentees (eligible but not available on the census day and for whole survey period) and also to those who refuse to participate to the survey. For each registered person, the eligibility criteria will be checked; eligible persons will be included in the study. People who have died or who left the household more than 2 weeks before (according to exclusion criteria) will be marked as not eligible on the census form, while additional eligible individuals who meet the inclusion criteria will be added (e.g. temporary inhabitants who have been in the household for at least 2 weeks). All the eligible persons will receive an invitation card, with name, ISN and appointment at the survey base (date/time). The finalized list of eligible persons will then be compiled. A composite Individual Survey Number (ISN) which serves as the survey individual number, will be used throughout the survey. It will consist of: the cluster number (2 digits) the number of the household as allocated during the census (3 digits). an individual serial number allocated during the census (2 digits) The household number will be given to every household by the census team prior to the household visit. The individual serial number will be assigned by the census team during the household visit. This allows the use of individual numbering for both adults and children living in the survey area. The ISN will be used on all questionnaires, results and request forms, in the survey laboratory register and on all slides, sputum samples and x-ray films. The census forms will be used as the primary source of ISNs and thus entered in the computer first. Other data collected include, name, age, sex, residential status, eligibility and presence at the time of the survey (to obtain the denominator for the prevalence estimate). Each registered and present person receives an A5 format double-side printed form on which will be filled in: name, ISN, and date, time and place for the interview and X-ray examination (see Annex). The person will keep this and bring it when he/she reports to the survey base for examination. Although names will not be entered into the survey database, they will be required during data collection in the field in order to avoid identification error when survey results for detected TB cases need to be given 29