Poor housing construction is associated with contracting malaria in a rural area south of Zimbabwe: A Case-control Study


 Background Ninety percent of the global annual malaria mortality cases emanate from the African region. About 80-90% of malaria transmissions in sub-Saharan Africa occur indoors during the night. In Zimbabwe, 79% of the population are at risk of contracting the disease. Although the country has made significant progress towards malaria elimination, isolated seasonal outbreaks persistently resurface. In 2017, Beitbridge District was experiencing a second malaria outbreak within twelve months prompting the need for investigating the outbreak. Methods An unmatched 1:1 case-control study was conducted to establish the risk factors associated with contracting malaria in Ward 6 of Beitbridge District from week 36 to week 44 of 2017. The sample size constituted of 75 randomly selected cases and 75 purposively selected controls. Data were collected using an interviewer-administered questionnaire and Epi Info version 7 was used to conduct descriptive, bivariate and multivariate analyses of the factors associated with contracting malaria. Results Fifty-two percent of the cases were females and the mean age of cases was 29±13 years. Cases were diagnosed using rapid diagnosed tests. Sleeping in a house with open eaves (OR: 2.97; 95%CI: 1.44-6.16; p<0.01), spending the evenings outdoors (OR: 2.24; 95%CI: 1.04-4.85; p=0.037) and sleeping in a poorly constructed house (OR: 4.33; 95%CI: 1.97-9.51; p<0.01) were significantly associated with contracting malaria while closing eaves was protective (OR:0.45; 95%CI: 0.20-1.02; p=0.055). After using backward stepwise logistic regression, sleeping in a poorly constructed house was still associated with getting sick from malaria (AOR: 5.88; 95%CI: 1.11-31.30; p=0.038). Those who had mosquito nets did not use them consistently. The district health team and the rural health center were well prepared for an outbreak response despite having limited human resources. Conclusion Health promotion messages should emphasize the importance of closing the entry points of the malaria vector, and the construction of better houses in the future. Residents had to be educated in the importance of consistent use of mosquito nets. The district had to improve malaria preventive measures like distribution of mosquito nets and lobby for more human resources to assist with malaria surveillance thus, curbing the recurrence of malaria outbreaks.


Background
Nearly half the global population is exposed to malaria [1] and in 2018, an estimated 228 million malaria cases occurred globally (95% Confidence interval (CI): 206-258 million) [2] up from 212 million in 2015 [3]. 405 000 malaria deaths were recorded in 2018 with the sub-Saharan region and India carrying 85% of the global malaria burden [2]. Malaria disease burden in Africa is very high and 90% of the global annual malaria mortality cases emanate from the African region [3]. In sub-Saharan Africa, 80-90% of malaria transmissions primarily occur indoors during the night [4].
In Zimbabwe, malaria is a major public health problem affecting all age-groups and according to the World Health Organization (WHO) malaria report of 2016, 79% of the population are at risk of contracting the disease. Malaria incidence in the country was 139 per 1 000 population at risk in 2013 [5]. Transmission of the disease happens in seasonal epidemics largely occurring during the summer months of November to April [6].
The Zimbabwe National Malaria Control Program and Roll Back Malaria Program have realized major successes in combating the diseases over the last 13 years and this led to the reorientation of focus from malaria control to elimination [7]. The progress in malaria elimination was tenaciously challenged by vector and parasite resistance to insecticides and anti-malarial medicines, outdoor malaria transmission, dynamic vector behavior and invasion of new areas by vectors. Climate change, erratic funding, local economic and political disturbances, as well as increased cross border population movements, have also contributed to the regression of malaria control programs [7].
These militating factors have caused sporadic malaria outbreaks throughout Zimbabwe.
In 2017, the Beitbridge District was experiencing a second malaria outbreak in a space of fewer than twelve months. The first malaria outbreak in the district started in late 2016 and was declared over in week 22 of 2017. This earlier malaria outbreak recorded over 800 cases resulted in five fatalities, three children and two adults [8]. The recent malaria outbreak in Beitbridge District had resulted in one death in Beitbridge town during week 42 of 2017 [9].
Situated in Ward 6 of Beitbridge District, Mtetengwe Clinic started to record a rapid increase in malaria cases during week 36, 2017 and subsequent weeks. Malaria threshold and alert levels for the clinic could not be calculated because the health facility was established in 2016 however, malaria threshold levels for Makakabule Clinic which is in Ward 7 were used. Makakabule Clinic used to serve most of the population in Ward 6 before Mtetengwe Clinic was constructed. This study aimed to determine the factors associated with contracting malaria in Ward 6 of Beitbridge and to investigate the malaria epidemic preparedness and response in the district. Understanding the risk factors associated with the occurrence of a malaria outbreak enables early and effective initiation of prevention interventions and control measures towards the elimination of the disease in the affected area. in houses made of mud and wood or bricks roofed using thatch, asbestos or corrugated sheets.

Methods
The Ward 6 of Beitbridge District is primarily rural and it is located in natural region V which is semiarid with erratic rains and very hot climate. Due to poor annual rainfalls, the population depends on the buying and selling of commodities purchased from South Africa. Furthermore, the district is also known for its thriving goat, cattle, and game ranching activities. Beitbridge District offices are in Beitbridge town 22 km east of Ward 6. According to the Matabeleland South Generic Report of June 2017, the district had an estimated population of 128 454.

Study Population
The study population consisted of all adults and children residing in Ward 6 of Beitbridge District between week 36 and week 44 of 2017. A case was a patient who resided within Ward 6, of Beitbridge District who presented with signs and symptoms of malaria and tested positive on the random diagnostic test (RDT) for malaria from week 36 to week 44 of 2017. Symptoms of malaria referred to one or a combination of fever, vomiting, headache, general body malaise and rigors/chills. Controls who saved as the comparison group were individuals who resided with or near a case and did not contract malaria during the period under study.

Sampling
Using Stat-Cal and assuming that engaging in outdoor activities during the evening was a significant risk factor for contracting malaria with an odds ratio (OR) of 8.5 and 81.6% of controls having been exposed. A power of 80% and a 95% confidence interval (CI) gave the minimum required sample size of 63 cases and 63 controls [11]. With an expected 20% attrition rate, the sample size was approximately equal to 79 thus, the final sample size was 79 cases and 79 controls.
The Mtetengwe Clinic line list which was used as the sampling frame for this study had 109 malaria cases and Ward 6 was purposively selected as the study site in this investigation. A simple random sampling of cases was done by allocating numbers to all the 109 cases on the Mtetengwe Clinic line list and putting the numbered cards in a hat. After mixing, 79 card numbers were blindly picked by a nurse from the clinic without replacement until the calculated sample size was reached. Controls were purposively selected from individuals residing with or within the neighbourhood of cases. A case who was not found at home or a case who refused to participate was replaced by the next case as selected from the line list.

Data Collection And Analysis
Data collection was conducted over a period of two weeks using structured and pretested intervieweradministered questionnaires. These were translated from English to the local language Shona and back-translated to English to ensure comprehension of the questions. Treatment records, clinic registers as well as healthcare workers were part of the study population. The participant questionnaire aimed to determine the demographic characteristics, the knowledge levels on the transmission of malaria and practices used for protection against the disease. Checklists were also used to assess environmental health risk factors and the availability of resources essential to mount an outbreak response. Healthcare workers were also interviewed to assess the district's preparedness to respond to a malaria outbreak. Data collection tools were pretested using ten participants from a village in Gwanda District.
The researcher and his assistant camped at Mtetengwe Clinic from the beginning of week 44 to the end of week 46 of 2017. Local VHWs were used to assist in locating selected cases using the clinic line list during community visits which were conducted during weekdays from 8 am to 4 pm. The researchers were fluent in both Shona and English visited interviewees at their home and used the structured questionnaire to collect data on demographic characteristics, risk factors associated with contracting malaria, knowledge, and practices of malaria prevention. Appointments for interviewing key informants were made in advance to minimize inconveniences. Interviewed key informants included the two nurses working at the clinic, the Medical Officer, Environmental Health Officer, District Laboratory Scientist, District Pharmacist, and local VHWs.
Data were entered into Microsoft Excel before it was exported into Epi Info version 7 (CDC, USA) for data cleaning, coding, and analysis. The statistical software was used to calculate frequencies, means, proportions, odds ratios, and p-values at 95% CI. Backward stepwise multivariate logistic regression analysis was done to control for any confounding variables.
Permission to conduct the study was sought from Provincial Medical Director of the Matabeleland South Province, and District Medical Officer Beitbridge District. Ethical review and approval were granted by the Africa University Research Ethics Committee and the approval number was 266/17.
Written informed consent was obtained from the study participants before data was collected.
Participation in this study was voluntary.

Operational Definitions
Well-constructed house: a modern house with closed eaves, brick walls, asbestos or corrugated roof, conventional windows, and doors that can be closed without leaving gaps for mosquito entry [12].
Poorly constructed house: a house with no conventional windows or unscreened windows, multiple openings, open eaves or doors that leave gaps when closed.
Eaves: gaps between the top of the wall and the overhanging roof [12].

Description of the malaria outbreak by person
A total of 109 cases of malaria were reported from week 36 to week 44 of 2017 and no deaths occurred in the study area during this period. The response rate for interviews was 95% because four of the selected cases had traveled outside Ward 6 therefore 75 cases and 75 controls were interviewed. 52% (n = 39) of the cases were females and the mean age of the malaria cases was 29 ± 13 years while most (35%) cases were in the age-group 20-29 years.
Participants who had long-lasting insecticide-treated nets (LLTNs) (n = 85) reported that the LLTNs were donated in 2015 and researchers noted that some of them had since developed holes which are potential entry points for the malaria vector. Of those who had LLTNs, 22% (n = 19) did not use them because they were suffocating (n = 11), itchy (n = 4), or individuals were not interested in using the nets (n = 7). The key informants reported that women and children in Ward 6 usually spend evenings outside cooking.
Although the community knowledge of malaria transmission was not significantly associated with contracting malaria, the majority (97%) managed to recall the use of LLTNs while 35% mentioned IRS as a way of protecting self from malaria. Eighty-three percent of participants agreed that one can protect themselves from getting malaria and all the participants reported that malaria disease is curable and they would visit the clinic upon suspecting any signs and symptoms of malaria. 93% of participants correctly identified that malaria is more common during the hot and wet season. 12% of the cases had traveled outside Ward 6 approximately four weeks before they were diagnosed with malaria while none of the controls had traveled outside Ward 6. The mean number of days taken from the onset of malaria symptoms to visiting the clinic by the malaria cases was 3.0.    individuals who slept in a poorly constructed houses were five times more likely to contract malaria than individuals sleeping in a well-constructed house.

Malaria Case Management
Due to the unavailability of microscopy tests and microscopists in the rural areas, health workers rely on malaria rapid diagnostic tests (RDT) to confirm diagnosis and this test can be conducted by nurses and village health workers (VHWs). The VHWs were equipped with RDT kits for rapid testing as well as artemisinin-based combination therapy (ACT) treatment for uncomplicated cases. Uncomplicated cases who visited the clinic were also given ACT with the first dose being directly observed by a healthcare worker before they were discharged home. All cases reported having completed the full course of antimalarial medication. The complicated cases were referred to Beitbridge District Hospital for further management. The two cases who were referred during this study period, recovered well. At Mtetengwe Clinic, malaria treatment and diagnostic testing are free for all age groups.

Discussion
This study sought to determine the factors associated with contracting malaria and to investigate the malaria epidemic preparedness and response in Ward 6 of Beitbridge District. The majority of the malaria cases were between the ages of 20-30 years which is contrary to Drakeley, et al., (2017)'s study findings in which the disease was more prevalent in the age group 5-19 years [13]. This may be attributed to the nature of the population in the study area. Many people come to Ward 6 for economic opportunities particularly seeking employment in the farming estates and selling goods from South Africa. Most of the farmworkers and traders are between the ages of 20 to 30 years.
Our study revealed that sleeping in a house with open eaves and sleeping in a poorly constructed house were significantly associated with contracting malaria while closing eaves was protective. A similar study in Gambia closing eaves halved the prevalence of anemia caused by malaria in children [14]. A systematic study of five case-control studies and two cohort studies showed that both modern walls made of bricks closed eaves reduced the odds of contracting malaria infection by a quarter [15].
Well-constructed houses protect inhabitants from malaria by preventing entry of the malaria vectors.
Modern walls and roofs also limit resting places for mosquitoes and reduce the attractiveness of the internal environment to the vectors when compared to the traditional mud/pole and thatch houses [15]. Although the use of LLTNs and IRS are equally important, housing improvements have significantly contributed to a decline in malaria cases and the elimination of the disease in many countries [16].
Sleeping in a house with open eaves was also found to significantly increase malaria transmission.
This echoed with Wanzirah (2015) were they cited that improving house facilities had the potential to reduce malaria transmission in sub-Saharan Africa since 80% of malaria transmission occurred indoors [17]. Two studies on malaria outbreak in Zimbabwe also cited that the presence of open eaves increased the likelihood of contracting malaria [18; 19]. The majority of the remote Ward 6 residents lived in poorly constructed houses and similar settings of poverty and remoteness were found to be risky with regards to contracting malaria in Amazon Villages [20].
Spending time outdoors after sunset was significantly associated with contracting malaria and this finding was similar to another study in which overnight hunting and spending night time outdoors were significant malaria transmission risk factors in French Guiana [21]. Zimbabwean studies in Chipinge and Mberengwa also cited that routine outdoor activities such as bathing and night fishing were significantly associated with contracting malaria [18; 22]. Mothers and children in Ward 6 were known to spend evening time outdoors sitting around a fire, cooking and exposing themselves to mosquito bites thus increasing their chances of contracting malaria.
The community had fairly good knowledge of malaria disease since the majority of participants managed to identify infected mosquitoes as the vector for the disease although a few still believed that eating infected fruits cause malaria. Most of the cases and controls could identify the signs and symptoms of malaria and all cases reported to have completed the treatment course. A similar study also recorded high knowledge of the signs and symptoms as well as the prevention of the disease [18]. This finding was also consistent with the malaria survey which also cited reasonably good knowledge about malaria causes, signs and symptoms and prevention among the rural community of Participants reported that malaria was curable and they would visit a healthcare center upon suspecting that they had the disease. Health seeking behaviour in Ward 6 was good, a finding that was contrary to what Onwjekwe and colleagues, found out in Nigeria were populations consulted herbalists, used spiritual/ritual water or just pray ignoring the malaria symptoms [24]. This difference in results can be attributed to increased sensitization on the disease in Beitbridge District since this was a second malaria outbreak within the same year.
Having an LLTN does not directly translate to its consistent use in malaria prevention; this was noted in 42% of LLTN owners being malaria cases. A study by Msellemu, et al., (2017) in Tanzania also made the same conclusion after finding out that LLTNs were often reserved for children and visitors exposing other family members to the vector [25]. The study participants who had LLTNs in Ward 6 reported that the LLTNs were donated in 2015 and on observation some nets had developed holes.
Such damages to the preventive barrier can still expose individuals to mosquito even though they used the nets. Some LLTN owners mentioned reasons like presumed suffocation and itchiness as barriers to LLTN use causing inconsistencies in their use as malaria preventive measures.
The use of the EPR team to mount a response against the outbreak was an effective way of combating the malaria outbreak. Unlike in 2015 when the Beitbridge Emergency Response team did not consistently conduct EPR meetings [19], the team was sitting every Wednesday for the past six months and there were meeting minutes to support this. All the nurses at Mtetengwe Clinic were trained in integrated disease surveillance and response, a finding that was contrary to a similar study done in Shamva, Zimbabwe [26]. Repeated malaria outbreak attacks had prepared the Beitbridge District team to be on high alert for a possible increase in cases of the disease. This encouraged prompt outbreak notification and response to prevent further proliferation of the disease in Ward 6.
Overall, both the healthcare center and district were well prepared to respond to the outbreak because most essential stocks were available and there were no stock-outs during the course of the outbreak.
Zimbabwe faces a critical shortage of human resources for health and the shortage of nurses, EHTs and microscopists can has a negative impact on the control of malaria in Zimbabwe [26].

Study Limitations
All cases used in this study were diagnosed using RDT and this may have resulted in the recruitment of participants infected outside the duration of the outbreak particularly during the early stages of the outbreak. We obtained self-reported data which could be affected by recall bias on exposure as well as social desirability when responding to interview questions. The sample size for the study was fairly small which could affect the power to identify significant changes in some exposure variables. This study did not include those patients who had a positive malaria test, received ACT treatment from local VHW and recovered before visiting the clinic.

Conclusion
Health promotion messages should emphasize the importance of closing eaves and other entry points of the malaria vector, and constructing better houses in the future. The residents should also be educated on the importance of consistent use of LLTNs. The district had to improve on preventive measures like distribution of LLTNs and lobby for more human resources to assist with malaria surveillance thus, curbing the recurrence of malaria outbreaks in the future. To counter the possibility of another malaria outbreak, these measures had to be instituted promptly.

Ethics approval and consent to participate
The permission to carry out this study was sought from Matabeleland South Medical Director and ethical clearance was obtained from Africa University Research Ethics Committee. Approval number 266/17. All participants voluntarily signed written consent before taking part in the study.

Consent to publish
Not applicable associated with contracting malaria in Ward 29 of Shamva District, Zimbabwe, 2014.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.