ࡱ > Z \ Y R
0 bjbj 4J ff' + t t t t t 8 L 0 ( : b x x x T Q0 S0 S0 S0 S0 S0 S0 $ 2 5 D w0 t , , , w0 t t x x 4 0 T" T" T" , F t x t x {( T" , Q0 T" T" $ $ x ̊]1 r R $ g( 0 0 0 $ , 5 5 $ 6 % J 5 t S% , , T" , , , , , w0 w0 L , , , 0 , , , , 5 , , , , , , , , , b : : Supplementary materials (SM)
Effectiveness of the innovative 1,7-malaria Reactive Community-based Testing and Response (1, 7-mRCTR) Approach on malaria burden reduction in Southeastern Tanzania
Selection of study areas
The design of this project started with workshops and kick-off meeting held by technical personnel from potential partners followed by field visits in Tanzania. The field visits involved consultations with central and local government authorities for the identification of project sites. Based on the availability of prior data, epidemiological parameters, and logistical convenience, Rufiji district was selected. The selected study area covered four wards. Of four wards, two were assigned to the intervention arm and the other two were controls. The identification of the project sites was followed by a baseline survey to establish parameters upon which the impact of implementing the project would be evaluated. After the baseline survey, the implementation of the project started. Intervention package of the project involved the application of a modified Chinese 1-3-7 ADDIN EN.CITE Zhou2015168[1]16816817Zhou, Shui-SenZhang, Shao-SenZhang, LiRietveld, Aafje ECRamsay, Andrew RZachariah, RonyBissell, KarenVan den Bergh, RafaelXia, Zhi-GuiZhou, Xiao-NongChinas 1-3-7 surveillance and response strategy for malaria elimination: Is case reporting, investigation and foci response happening according to plan?Infect Dis PovertyInfect Dis Poverty554120152049-9957[1] model for malaria surveillance and response in combination with the WHO-T3 Initiative and the local resources. The local resources here included a platform of health system infrastructures, manpower, funds, and supplies that existed to provide the base of the intervention.
Stakeholder engagement
At the national level, the study team was composed of members from the Ifakara Health Institute (IHI-Tanzania) and National Institute of Parasitic Disease Chines Centre for Disease Control and Prevention (NIPD) together with the National Malaria Control Programme (NMCP) and the National Institute of Medical Research (NIMR). The approved protocol was shared with NMCP and NIMR for discussion and agreement on the proposed intervention. Subsequently, at the district level, the local government officers (District Medical Officer, District Malaria focal person, Vector control officer, and Council Health Management Team) were consulted and the study design was discussed in detail.
At the local community level, engagement activities were conducted before and during the screening campaigns to broaden and strengthen community awareness, to raise the general knowledge of malaria and to promote the intervention. Meetings involving; community leaders, school teachers, and children and key informants at the district, wards, villages, and sub-village levels were held. Over 200 government stakeholders in 36 administrative villages were involved.
As part of working collaboration between Chinese and Tanzania, during field implementation, at least 36 Chinese teams of malaria expertise (epidemiologist, medical entomologist, laboratory scientist, and anthropologist) were deployed at the local community to provide malaria technical and scientific assistance on study design, feasibility, and practical workflow coordination. Both malaria experts from China and Tanzania worked cooperatively on this community-based pilot project to provide the platform for expanding Chinese malaria experiences gained and transferred into innovative 1,7-RTCR approach development and integration in strengthening community engagement and mobilization.
For implementing the 1,7-mRCTR, the project recruited and trained 35 CHCWs on malaria surveillance and treatment activities. Detailed training included the fundamental skills for malaria case diagnosis, and treatment, vector control, and health education. The CHCWs were divided into four teams, which were referred to as surveillance response teams. Each surveillance response team comprising of at least a two-laboratory technician, a clinician, a nurse, two field interviewers, and a field supervisor. The teams were deployed to the two intervention wards to conduct 1,7-mRCTR. Furthermore, a community sensitization team was formed from each respective village to raise the communitys awareness and compliance with a 1,7-RCTR. The size of the sensitization team depended on the number of sub-villages units, including the hamlet leader and village community health volunteers (VCHV).
Statistical Analysis
To examine whether the 1,7-mRCTR affected the number of health facility cases in the treated wards, we considered the total number of cases diagnosed at health facilities during each week for each ward and computed case ratios (the number of cases at the HF per population base). We divided the year into the high season (weeks 18-31 of the year, roughly May to July) and low season (all other weeks) based on case count. The weekly case ratios per population base were considered as replicates of the situations within the wards. GEE models of the annual effects (2017 vs 2016 - the only full years in the project) controlled for the season were fitted. To make sure that our modeling methods did not unduly influence the results, we modelled the logarithmic case ratios (with identity link) using both independent and exchangeable correlation structures. We also modelled the case ratios using identity link and independent and exchangeable correlation structures. Thus, we had four ways of examining the outcomes. Moreover, we considered the possible interaction of year and season, i.e. whether or not the approach tested was more or less effective depending on the season.
To examine the length of time, the case ratio per population was suppressed after a village was treated, we modelled the logarithm of the case ratio using mixed models with the identity link, the exchangeable working covariance structure, and the empirical variance. We considered each village-week to be a replicate within the village (random effect). The exposure of interest was the (integer) number of weeks post-treatment. If the time since the last treatment exceeded 13 weeks, the time was reset to pre-treatment. If a village was treated within the 13 weeks, the clock restarted. In addition to ward and season (which was modeled harmonically), a linear term for time since the project started was included, as were interaction terms between the ward and this term as well as the season variables. Since the treatment was not allocated randomly but based on being a hotspot, we needed to weight the observations to reflect the probability that an individual village would be treated in a particular week. We followed the method of Hernan ADDIN EN.CITE Hernan200049[2]494917Hernan, M. A.Brumback, B.Robins, J. M.Department of Epidemiology, Harvard School of Public Health, Boston, MA 02115, USA.Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive menEpidemiologyEpidemiology561-701152000/08/24Anti-HIV Agents/*therapeutic useCD4 Lymphocyte CountCausalityConfounding Factors (Epidemiology)Epidemiologic MethodsHIV Infections/*drug therapy/*mortalityHumansMale*Models, StatisticalProportional Hazards ModelsSurvival AnalysisTime FactorsTreatment OutcomeUnited States/epidemiologyZidovudine/*therapeutic use2000Sep1044-3983 (Print)
1044-3983 (Linking)10955409https://www.ncbi.nlm.nih.gov/pubmed/10955409[2] to produce stabilized weights using season, ward, and linear time for the numerator model and adding the number of times a village had been previously treated, and the previous weeks incidence ratio for the denominator model. Since the stabilized weights had a large range, we assigned .005 to observations with computed weight below 0.01 and 101 to observations with computed weight over 100.
References
ADDIN EN.REFLIST 1. Zhou S-S, Zhang S-S, Zhang L, Rietveld AE, Ramsay AR, Zachariah R, et al: Chinas 1-3-7 surveillance and response strategy for malaria elimination: Is case reporting, investigation and foci response happening according to plan? Infect Dis Poverty 2015, 4:55.
2. Hernan MA, Brumback B, Robins JM: Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology 2000, 11:561-570.
PAGE \* MERGEFORMAT 1
h x ~ E N ɻxxiZiZZZKZi h, h@O OJ QJ mH sH h, ht OJ QJ mH sH h, h2s OJ QJ mH sH h, ht 5OJ QJ mH sH h, h2s 5OJ QJ h, h7 5OJ PJ QJ h&