Active community sweeps
2014 in Kedougou: During the 20 weeks of the intervention, 132 CHWs saw 11,844 patients during sweeps (mean 5 patients per sweep), of which 84% had fever, 9% had diarrhea, and 5% had cough. RDT positivity rate was 80% (8071/10,141) (Table 1). Treatment with ACT was given in 98.6% of RDT positive cases (29% of whom were < 5 years); 1.0% had signs of severity and 0.4% of the cases were referred because of drug stock-outs. Almost all diarrhea cases (1,031/1,035) were either appropriately treated by the CHWs with ORS and zinc or referred (3% referred for ORS or zinc stock out); as was the case for pneumonia (158 cases), though 40% of pneumonia cases had to be referred for stock out of antibiotics. In total, 2,317 patients were referred for care at health posts: 90% for a negative RDT, 4% for medication stock-out, 3% for severe malaria, 2% for pregnancy and 1% for age less than two months.
2015 in Kedougou and Kolda: For 21 weeks in Kedougou and 19 weeks in Kolda, 246 CHWs saw 27,621 patients during active sweeps (mean 8 patients per sweep), of which 83% had fever (96% of whom received an RDT), 5% had diarrhea, and 12% had cough. RDT positivity rate was 68%. Of those diagnosed with malaria (27% of whom were < 5 years), 96.1% received an ACT, while 2.2% were referred due to ACT stock-out and 1.3% were referred due to signs of severity. Among children diagnosed with pneumonia (n=1,768), 41% were referred for stock-out of amoxicillin, and among children diagnosed with diarrhea (n =1,379), 57% were referred due to stock out of ORS or zinc.
2016 in Kedougou, Kolda, Tambacounda, and Sedhiou: For 25 weeks in Kedougou, 22 weeks in Kolda, 19 weeks in Tambacounda, and16 weeks in Sedhiou, 708 CHWs saw 64,168 patients during active sweeps (mean 6 patients per sweep), of which 77% had fever (99% of whom received an RDT), 9% had diarrhea, and 13% had cough. RDT positivity rate was 48%. Of those diagnosed with malaria (21% of whom were < 5 years), 99% received an ACT. The remaining 1% were referred for signs of severity. The majority of children diagnosed with diarrhea or pneumonia had to be referred due to lack of availability of medication for diarrhea and pneumonia.
Table 1 summarizes the results of the weekly sweeps by region and year for 2014-2016.
Table 1: Weekly sweeps results for malaria case management by year and region during scale up (2014-2016)
Region
|
Year
|
Number of villages (CHWs)*
|
Number of weekly sweeps reported completed (% of expected)
|
Number of fever cases
|
Number of positive RDTs / RDTs performed
(test positivity rate)
|
Mean number of positive RDTs per sweep
|
Kedougou
|
2014
|
132
|
2,375 (90%)
|
9,970
|
8,071/10,141 (80%)
|
3.4
|
2015
|
144
|
1,898 (94%)
|
10,321
|
6,095/9,997 (61%)
|
3.2
|
2016
|
165
|
3,897 (99%)
|
15,382
|
9,378/15,260 (61%)
|
2.4
|
Kolda
|
2015
|
102
|
1,462 (87%)
|
12,517
|
8,973/11,918 (75%)
|
6.1
|
2016
|
105
|
2,372 (99%)
|
17,176
|
10,384/16,823 (62%)
|
4.6
|
Sedhiou
|
2016
|
150
|
2,400 (100%)
|
9,177
|
1,115/9,100 (12%)
|
0.5
|
Tambacounda
|
2016
|
288
|
5,009 (95%)
|
22,433
|
9,677/22,420 (43%)
|
1.9
|
* One CHW per village
PECADOM passively detected malaria cases
During the scale up of PECADOM Plus, cases detected and treated during passive work by CHWs (the standard PECADOM model with care being sought by community members outside of weekly sweeps) also increased among CHWs who did proactive sweeps. From the year prior to introduction to the first year of implementation in each region (2013 to 2014 in Kedougou, 2014 to 2015 in Kolda, and 2015 to 2016 in Sedhiou and Tambacounda), the number of RDTs performed by CHWs during passive work increased by 56% in Kedougou (from 4,473 to 6,959), 112% in Kolda (from 5,298 to 11,254), 110% in Tambacounda (from 13,283 to 27,840), and 139% in Sedhiou (from 3,490 to 8,346). During the same period, the number of cases diagnosed by CHWs during passive work increased by 61% in Kedougou (from 3,328 to 5,363), 149% in Kolda (from 3,126 to 7,787), 54% in Tambacounda (from 7,247 to 11,462), and 87% in Sedhiou (from 453 to 849) (Table 2). (The number of CHWs increased by 2% in Kedougou, decreased by 1% in Kolda, and increased by 12% and 10% in Tambacounda and Sedhiou, respectively.) While the largest increases were seen during the first year of implementation, the numbers of people receiving diagnostic tests and diagnoses of malaria in between sweeps continued to increase during years two and three in Kedougou and year two in Kolda.
Table 2: PECADOM passively detected malaria cases by region and year
Region
|
Phase
|
Year
|
Number of villages (CHWs)*
|
Fever cases
|
Positive RDTs/
RDTs performed
(test positivity rate)
|
% of all cases detected by DSDOM† that were detected during passive work
|
Kedougou
|
Pre (Trial)
|
2013
|
129
|
4,507
|
3,328/4,473 (74%)
|
all
|
Year 1
|
2014
|
132
|
7,206
|
5,363/6,959 (77%)
|
40%
|
Year 2
|
2015
|
144
|
8,078
|
5,074/7,633 (66%)
|
46%
|
Year 3
|
2016
|
165
|
13,435
|
8,981/13,974 (64%)
|
49%
|
Kolda
|
Pre
|
2014
|
104
|
5,935
|
3,126/5,298 (59%)
|
all
|
Year 1
|
2015
|
102
|
11,930
|
7,787/11,254 (69%)
|
47%
|
Year 2
|
2016
|
105
|
18,892
|
10,321/17,969 (57%)
|
50%
|
Tambacounda
|
Pre
|
2015
|
258
|
15,116
|
7,247/13,283 (55%)
|
all
|
Year 1
|
2016
|
288
|
27,750
|
11,162/27,840 (41%)
|
54%
|
Sedhiou
|
Pre
|
2015
|
136
|
3,523
|
453/3,490 (13%)
|
all
|
Year 1
|
2016
|
150
|
8,390
|
849/8,346 (10%)
|
43%
|
*One CHW per village
† actively and passively
Total malaria diagnoses by CHWs
The increase in the total number of malaria cases detected (actively and passively) by CHWs from the year prior to implementation to the first year of implementation was 304% in Kedougou (2013 to 2014) and 436% in Kolda (2014 to 2015). Increases of 399% in the Sedhiou region and 232% in the Tambacounda region were seen from 2015 to 2016 (Figure 3 and 4). The dramatic increase among cases diagnosed occurred among both children < 5 years and residents ≥ 5 years (Figure 4). Overall, from the year prior to implementation to the first year of implementation, while the number of CHWs increased by 7% in the four regions, the mean number of malaria cases diagnosed per CHW during the year increased from 23 to 79, an increase of 252%. The most pronounced increase was in the Kolda region, where the number of CHWs stayed constant, but the mean number of malaria cases diagnosed per CHW increased from 30 to 164. Increases in numbers of cases diagnosed by CHWs were also noted from 2014 to 2015 in Tambacounda (172%) and Sedhiou (15%), the two regions that did not introduce PECADOM Plus until 2016, as efforts to improve the availability of malaria commodities at the community level for the PECADOM Plus regions were applied to other regions as well. In 2016, in the regions in which PECADOM Plus was implemented, 40% of all malaria cases reported were detected at the community level (at health huts and by CHWs), while in regions in which PECADOM Plus was not implemented, only 8% of reported malaria cases were detected at the community level.
Learning lessons and addressing challenges
At the end of each season, the NMCP hosted an evaluation meeting in each region attended by regional and district health representatives and community stake holders, in which results from each district were presented, and feedback was solicited from stakeholders at all levels of the health system and community members. PECADOM Plus was positively received by communities, health post nurses, and DHMTs. Nurses and district medical officers stated that they saw fewer cases of severe malaria. The implementation of weekly sweeps greatly increased the visibility of CHWs in their communities and increased community engagement. Populations with high poverty and low access, with relatively little exposure to modern medicine, were able to observe the efficacy of malaria diagnosis and treatment in their communities. In addition to case management, CHWs used the visits as opportunities to talk about use of insecticide treated nets, intermittent treatment in pregnancy, and SMC. There was a dramatic increase in symptomatic, parasitologically confirmed cases diagnosed by CHWs when the proactive component was introduced. The relative contributions of improved care seeking due to increased community engagement, RDT and ACT availability due to enhanced supply chain support, and enhanced supervision with improved reporting to the observed increases in reported cases are not possible to ascertain, though all likely played a role. However, it is likely that a large number of malaria cases had previously gone undetected and untreated in the community.
While much of this should have been seen in the passive model, however, without the added effort in supervision and supply chain management necessitated by the proactive community sweep component, as well as the higher visibility provided by the weekly visits, the potential of the passive model was not reached. As proactive detection of symptomatic cases led to detection and treatment of far more cases of malaria, it exposed weaknesses in the PECADOM program that required resolution, most critically supply chain management, supervision, compensation of CHWs, coordination, and reporting.
Supply chain challenges were evident from the beginning of the first pilot in 2012, with insufficient RDTs and ACTs to support the PECADOM Plus activities, even if stock outs at health facilities were uncommon. It rapidly became apparent that CHWs had been routinely experiencing frequent and prolonged shortages of commodities, and emergency deliveries were required on numerous occasions as the NMCP and DHMT worked to develop more sustainable solutions, such as planned deliveries of commodities at the beginning of the malaria transmission season sufficient to support the community level. Partners such as Peace Corps played instrumental roles in assisting the community to alert district health officials to shortages and helping redeploy commodities. The improvement of the supply chain may also have been at least partially responsible for the apparent increase in care seeking at the community level. While the NMCP was able to address malaria commodities shortages, this was not the case for diarrhea and pneumonia treatments during the time frame of this scale-up, as other partners were responsible for these. Except for diarrhea treatments in Kedougou in 2014, approximately half of children diagnosed with diarrhea or pneumonia had to be referred for treatment, which compromised the credibility of the integrated approach and discouraged the CHWs, some of whom reported that they stopped trying to identify cases of diarrhea and pneumonia when they had no treatment to offer. Fortunately, the attention drawn to the shortage of non-malaria commodities at the community level later enabled resolution of these shortages.
Health post nurses had the responsibility to supervise CHWs, but the distances over difficult paths and their additional workload resulted in spotty supervision. The NMCP developed a cadre of community supervisors trained to directly supervise CHWs on a weekly basis, and to report to the health post nurse. These community supervisors were also instrumental in collecting data for program monitoring and ensuring that the CHWs had sufficient commodities.
Previous community health projects did not financially compensate CHWs, based on the principle that the community should support them, however, the degree to which this happened was highly variable. Successful implementation of PECADOM Plus required compensation for CHWs. Artisanal gold mining is a common income-generating activity in this zone, and prior to the introduction of PECADOM Plus, some CHWs regularly left their villages to mine. Under PECADOM Plus, CHWs received a stipend (USD 5) for each day that they performed an active sweep of the village (approximately 20 days per year), equivalent to the amount they typically receive per day for work during public health campaigns (vaccination, deworming, etc.). While a few CHWs continued their mining activity away from the village on non-sweeps days, and occasionally missed a sweep, even the small stipend was sufficient to induce the majority of CHWs to stay in their communities and provide care. However, in cases in which the stipends were not paid in a timely manner, completion of sweeps sometimes suffered. To assure adequate supervision and participation in the monthly coordination meetings, resources were also needed for phone credit and transportation. Community supervisors also received a small stipend and funds for transportation and phone credit.
Effective coordination and monitoring required a great deal of time and support. Numerous partners assisted at every level of implementation. Regular coordination meetings at the health post and district levels were critical to rapid problem solving, and, at the central level, NMCP staff provided support with problem resolution. Within health post catchment zones, frequent communication between community supervisors and CHWs was required, necessitating cell phone credit.
Obtaining quality data from the community level is a persistent challenge, but one not new to the active model. Many CHWs, with very basic levels of literacy, found filling out registers and summarizing data extremely challenging. While data collection forms were designed in such a way that data entry and collation involved making and counting checkmarks, data were still fraught with inconsistencies, and required time and attention from community supervisors and other partners to address. Other ongoing challenges included timely integration of community level data and reporting of sweep data, and assuring that the population covered by each CHW was small enough in numbers and geographic spread to allow a complete sweep each week