Ownership, Coverage, Utilisation and Maintenance of Long-lasting insecticidal nets (LLINs) in Bamenda, Santa and Tiko Health Districts in Cameroon

Introduction The Bamenda, Santa and Tiko Health Districts are in the highest malaria transmission strata of Cameroon. The purpose of this study was to explore the indicators of ownership and utilisation as well as maintenance of long-lasting insecticidal nets (LLINs) in three health districts in Cameroon. Methods A cross-sectional household survey involving 1,251 households was conducted in the Tiko Health District (THD) in July and June 2017 and in Bamenda and Santa Health Districts in March to May 2018. A structured questionnaire was used to collect data on LLIN ownership, utilisation and maintenance as well as demographic characteristics. Results The average number of LLINs per household was higher in the Bamenda Health District (BHD) compared to the Tiko Health District (THD) (2.5±1.2 vs. 2.4±1.6) as well as the household ownership at least one LLIN (93.30% vs. 89.00%). The proportion of the de-facto population with universal utilisation was higher in BHD compared to THD (13.1% vs 0.2%). In multinomial regression analysis, households in the SHD (p = 0.007, OR; 2.8, 95% C.I; 1.3 – 5.8), were more likely to own at least one LLIN compared to those in THD. Conclusion Ownership of LLINs was low in SHD and THD in comparison to the goal of one for every two household members. Overall LLINs coverage and accessibility was still low after the free MDCs, as only 14.6% of children 0 – 5 years and 16.1% of the entire population used LLIN the night before the survey.


INTRODUCTION
Malaria is a preventable and curable disease transmitted by the bites of female Anopheles mosquitoes [1,2] and it is a serious global public health problem with an estimated 216 million cases in 91 countries in 2016 [2,3].
Africa is the most affected region, with 90% of all estimated malaria cases and 91% of deaths in 2016 and 15 African countries alone contributing 80% of all cases, Nigeria and the Democratic Republic of the Congo (DRC), being the top two contributors [1,3] provision of free LLINs to pregnant women at antenatal care (ANC) clinics since 2008 [4,5]. In 2011, the Cameroon MOH undertook a nationwide free LLINs distribution campaign from health facilities to all households, with the objective to provide a LLIN with a lifespan of five years, to all household beds or a LLIN for every two individuals per household, to a maximum of three LLINs per household [6,7]. Malaria continues to be endemic in Cameroon, with an estimated mortality rate of 11.6%, surpassing that of the African region of 10.4% as well as neighbouring countries [4]. It is the first major cause of morbidity and mortality among the most vulnerable groups [5,8,9]: children under five years and pregnant women, accounting respectively for 18% and 5% of the total population estimated at 19 million [10].
The main contemporary malaria control interventions are insecticide-treated bed nets (ITNs) and indoor residual spraying (IRS) [1,5,11,12]. Alliance for Malaria Prevention has been instrumental in keeping long-lasting insecticide net (LLIN) campaigns on track: between 2014 and 2016 about 582 million LLINs were delivered globally and in 2017 where there was the successful delivery of over 68 million nets to targeted recipients in Sub-Saharan Africa (SSA) and beyond [3]. Over 80% of all households have at least one mosquito net, up from 57% in 2011, still, only about 60% of these households have enough nets to cover everyone at night [13].
The proportion of people in SSA sleeping under a LLIN rose from less than 2% to over 50% between 2000-2015, preventing an estimated 450 million malaria cases [14].
Most studies in Cameroon have focused on various aspects of net ownership and utilisation. Effective LLIN use in the prevention of malaria in parts of Mezam Division, North West Region [15], Plasmodium falciparum infection in Rural and Semi-Urban Communities in the South West Region [6], predictive factors of ownership and utilisation in the Bamenda Health District (BHD) [16] and socio-demographic factors influencing the ownership and utilization among malaria vulnerable groups in the Buea Health District [17]. However, there is paucity information on the indicators of LLIN ownership and utilisation. This study examines the indicators of net ownership and utilisation as well as maintenance, through analysis of household survey data collected in three health districts in Cameroon and discuss their implications for programmatic interventions designed to increase LLINs ownership and use.

Study area
The study area consisted of the BHD with an estimated 350,000 residents and the Santa Health District (SHD), 35 Km from the BHD with 73,406 residents in the North West Region and the Tiko Health District (THD), 351 Km from the BHD with an estimated 134,649 residents in the South West Region of Cameroon [18]. Generally, malaria in Cameroon is caused mainly by Plasmodium falciparum, with Anopheles gambiae as the major vector [5,10].
The BHD (a semi-urban community) and the SHD (a rural community) are in the high western plateau altitude malaria geographical strata of Cameroon, where malaria transmission is permanent, occurring all year long, sometimes lessened by altitude though never totally absent [10,19]. It is one of the most densely populated regions of Cameroon [5,10].
The THD (urban and rural communities) is in the coastal strata, zone of dense hygrophile forest and mangrove swamp with the highest transmission of malaria in the country [5,10]. Like the Buea health district, the THD has a constant variation in the trends of malaria prevalence allround the year [20,21].

Sampling design
This cross-sectional household survey conducted in the THD from June to July 2017 and BHD and SHD from March to May 2018 utilised a stratified multi-stage cluster sampling design.
A study sampling frame included all health areas (HAs) in the study area, except those that were inaccessible for security reasons. Within each HA: urban, rural or semi-urban localities were subdivided into quarters (the primary sampling units or clusters in our study). On average, each HA had about five quarters. Sampled HAs in the THD had about 2,089 (35.58%) of the sampled population.
First stage: we randomly selected four HAs in the THD and by probability proportionate to size (PPS) and conveniently sampled one each from the BHD and SHD.
Second stage: within each selected HA we randomly selected at least three quarters and at most eight quarters also by PPS, thus totalling 31 quarters (Figure 1) in the sample.
Third stage: within each selected quarter, the survey team mapped and enumerated all households and selected households in each cluster by systematic random sampling (that is, a random start and interval to cover the entire quarter). The estimated number of households in each quarter was obtained from the quarter leader to determine the sampling interval to select the households. Where: n = minimum sample size, Z = 1.96, critical Z value at 95% confidence interval (95% C.I.), p = 50% estimated population of households owning mosquito bed nets = 0.5, q = 1 -p = 0.5, (p)(q) = (0.5) 2 = 0.25, d = acceptable margin of error for proportion being estimated = 0.05.

Recruitment procedures and measures
Interviewers explained the purpose of the study and obtained verbal informed consent from the head of the household or spouse. In cases where neither household head was available, any elderly person who has lived in the house for at least the last 12 months replaced him/ her.

Outcome variables
The main LLIN outcome variables were;

LLIN utilisation indicators:
Household universal utilisation: proportion of population that slept under a LLIN the previous night [8,23]. By the vulnerable population in the household: proportion of children under five (or pregnant women) that slept under a LLIN the previous night [8]. Regularly sleeping under bed nets: household heads who reported habitually using nets on a daily basis [24]. Household head slept under a LLIN last night: proportion of households in which the household head slept under a LLIN last night, where the numerator comprises the number of households surveyed wherein the household head slept under a LLIN last night and the denominator, the total number of households surveyed.

Independent variables (IV)
considered for association with LLIN ownership, use and maintenance were age, gender, marital status, education, occupation, health district, house type and household composition.

Statistical analysis
We entered data into, and analysed with IBM-SPSS Statistics 21.0 for windows (IBM-SPSS Corp., Chicago USA). The Chi square (χ 2 ) test was used to compare socio-demographic characteristics with the health districts and multivariate logistic regression to identify significant correlates of the main outcomes. p values less than 0.05 were considered significant.

Ethics statement
The study, obtained approval from the Institutional Review Board of the Faculty of Health Sciences, University of Buea (N o : 624-05). Administrative authorisation was obtained from the South West Regional Delegation of Public Health. Written informed consent was obtained from all participants and confidentiality was maintained at all steps of data collection.

Characteristics of the study participants
A total of 1,251 household heads was sampled with 5,870 de-facto residents across six health areas in three health districts. Of the total household residents counted, 1,267 (21.6%) were children 0 -5 years old and 93 (1.6%) were expectant mothers. There were generally more female (68.0%) household heads than males, with mean (±SD) age of participants of 36.1±10.8.    (Table 3), where households in the THD significantly (p = 0.007) owned few nets, while those in the BHD significantly (p < 0.001) had more coverage than other district. However, after adjusting with educational and marital status, association to heath districts was still significant. Coverage was also associated with gender of the household head and household size (Table 3), where households headed by females (p = 0.005) and those of household size of 1 -4 members (p = 0.002) significantly influenced coverage than the others. Secondary educational and unskilled occupational status significantly influenced household ownership of nets (p < 0.05).
Household accessibility (Table 3) to bednets was associated to gender of the household head and health districts, where household residents in housed headed by females and those in the BHD significantly (p < 0.001) had more access to LLINs than the other groups. After adjusting with educational and marital status, the significance between accessibility and gender and health district was maintained. 4,058 (69.1%) of the de-facto population, from 865 (69.1%) of the 1,251 households sampled, had access to LLINs in the household.

Use of LLINs
Of the 1,251 households sampled, 520 (41.6%) and 256 (20.5%) were those in which all children 0 -5 years and those in which all who slept home last night used bednets, respectively representing 859 (14.6%) and 942 (16.0%) of the 5,870 de facto population that slept home last night (Table 2). Bednet utilisation in households by all children 0 -5 years and the entire family (Table 4), was associated to age and health district where more households with household heads in the 21 -30 years age group (p = 0.021) and in the BHD (p < 0.001) significantly used nets than the other groups. The household utilisation of bednets by all children 0 -5 years old (Table   4), was also associated to the gender and educational status of the household head where more households with female heads and those with primary and secondary educational status significantly (p < 0.05) used the nets last night than the other groups. Bednet utilisation by the entire family (Table 4) was associated to the composition of the household, where more households with no children < 5 and with fewer members (1 -4) significantly (p < 0.05) used nets than the other groups. week, while 350 (28.0%) had their household heads using bednets last night ( Table 2). The use of bednets on all nights of the week and consequently last night (Table 5) by the household head was associated to the age of the household head as well as to the health districts, where more household heads in the 21 -30 and 31 -40 age groups, and 20 and 41 -50 age groups significantly (p < 0.05) used bednets regularly and last night than the other age groups. Also, more household heads in the SHD and BHD significantly (p < 0.05) used bednets on all nights of the week and last night respectively (Table 5). After adjustments of all utilisation indicators with educational and marital status, significance was maintained. The other uses, "out of norms", of LLINs are summarised in These uses ranged from being used as goal post nets by children; 2.8% (95% C.I; 2.0 -3.9), to yard fences; 22.7% (95% C.I; 20.5 -25.1). With the exception of harvesting and drying of melon seeds (egussi), all the other "out of norm" uses of LLINs were significantly (p < 0.05) associated to the health districts. The question of washing bednets or not (Table 7), was associated to the gender of the household head, where households with females heads significantly (p = 0.027) washed them compared to those headed by males. The WHO recommended washing frequency was associated to age of the household head and health district, where households with heads in the 31 -40 years age group (p = 0.018) and those in the BHD (p < 0.001) abided more significantly to the recommended washing frequency than those in the other age groups and health districts respectively. On the recommended LLINs washing frequency, heads in the BHD (p < 0.001, OR; 2.4, 95% C.I; 1.7 -3.5) were significantly more likely, while those in the SHD (p = 0.321, OR; 0.8, 95% C.I; 0.6 -1.2) were insignificantly less likely to respect the recommended LLINs washing frequency compared to those in the THD (Table 6).

DISCUSSION
This study examined the indicators of LLIN ownership, utilization and maintenance in the Bamenda, Santa and Tiko Health Districts. Overall, 92.5% and 20.5% of households interviewed owned at least one LLIN per household and utilisation by entire household last night respectively.

Indicators of household LLINs ownership
Currently, targets in national strategic plans for all three LLINs coverage indicators are usually set for all people at risk of malaria [1,2], to ≥ 80%. Household ownership of at least one LLIN per household in this study is higher than rates reported elsewhere in Cameroon [6,7,9,13,[15][16][17] and out of Cameroon [24][25][26][27][28][29][30][31][32][33]. It was however, lower than proportions reported in Uganda and Myanmar [34,35] and in line with the 93.5% reported in Madagascar [27]. The high proportion of owning at least a LLIN per household in these health districts could be attributed to the free LLINs mass distribution campaigns (MDC) [6,10].
The universal household coverage of 66.8% (overall LLIN: Person ratio of 0.50) though within the WHO range of 39 -75 % [36], was lower than rates reported elsewhere in Cameroon and Myanmar [13,35]. It was however, higher compared to rates in Madagascar and Uganda [27,37] as well as a host of eight African countries [29].
Access to LLINs in the household of 69.1% in this study was lower compared to results reported elsewhere [27,34], higher than 21% reported in Batwa [38], within the 57.3 -78.8% in eight African countries [29] and 32.3 -81.3% reported in a multi-country study [39]. The low household universal coverage and (versus) accessibility in this study, could be attributed to the significant differences amongst the health districts: 86.4% vs 83.5% for the BHD, 55.6% vs 55.6% for the SHD and 56.2% vs 66.3% for the THD; χ 2 (2, N = 1,251) = 120.457, p < 0.001 vs χ 2 (2, N = 1,251) = 77.969, p < 0.001 and differences in family size vs gender of household head.

Household utilisation of LLINs
Household universal LLINs utilisation of 20.5% (16.0% of the de facto population) was very low compared to to previous studies elsewhere in Cameroon [16,17] and out of Cameroon [24,27,31,34,35,37,40]. This was however high compared to the 6.9 -15.3% reported in Myanmar [26]. The very low household LLINs utilisation could be attributed to the significant differences amongst the health districts: 43.1% in the BHD, 1.0% in the SHD and 14.1% in the THD; χ 2 (2, N = 1,251) = 240.400, p < 0.001 as well as household composition and the installation of LLINs on all beds in the household (p < 0.05). It could also be due to inadequate education on LLINs utilisation, the socio-political tensions and differences in the different study designs.
Bednet utilisation by all children 0 -5 years and expectant mothers in the household of 14.6% and 63.4% respectively, is low compared to 63% vs 60% reported in the BHD [16], 52% vs 58% in the national territory [7] and elsewhere in the world [25,27,28,31,35,37,40]. The low LLIN utilisation by all children 0 -5 years old could be attributed to significant differences in the health districts, age and gender of household heads, educational status of household head as well as the presence or absence of bushes or water pools around dwellings (p < 0.05).
Use of LLINs by household head last night of 28.0% was low compared to 58.3% reported in Rural and Semi-Urban communities in the South West Region of Cameroon [6] and 47.2% in China [24]. Meanwhile the regular use of LLINs of 38.7% was low compared to 48.0% reported in China [24]. The low use of LLINs last night by household head and regular use of LLINs, could significantly be attributed to differences in the health districts and ages of the household heads (p < 0.05).
LLIN misuse of 2.3 -22.7% was also similar to the 18.2% reported in Mezam Division [15] and 21% in Kenya [41]. The use of LLINs for other purposes, other than the prevention of mosquito bites could be attributed to: inadequate education on utilisation, lack of good playgrounds, as 2.8% (95% C.I, 2.0 -3.9) of the households admitted that children used as football goal post nets. The recommended LLIN wash frequency reported in this study was similar to the 52.0% reported in Kenya [41]. The optimal LLIN washing frequency could be attributed to the age of the household head as well as the health district.

RECOMMENDATIONS
The populations of the three health districts should be properly educated by community health workers and stakeholders on the regular utilisation of LLINs and by all household occupants.
The MOH should sustain another free MDC since those distributed in 2015 -2016 will be worn out and ineffective in preventing malaria by 2019.

Strengths
The data used in this study was collected by trained surveyors, who had mastery of all the HAs as they are responsible for the coding of houses during the Expanded Programme on Immunisation (EPI) and MDC campaigns. All the health district offices were consulted for the mapping of the HAs, quarters and census list of households used in the last MDC and EPI campaigns. In Cameroon, the MOH carries out seasonal EPI campaigns. The quality of data collected was assured through the multistage sampling strategy to minimize bias and pretesting of questionnaires.

Limitations
This was a cross sectional study, representing the snapshot of the population within the study period and does not show cause and effect since the predictor and outcome variable were

Ethics approval and consent to participate
Ethical clearance was obtained from the IRB-FHS of the University of Buea.
S2 file. Raw data and sampling files (Excel).

Funding
Financial assistance was received from the Proprietors of the Atlantic Medical Foundation Tiko road -Mutengene and Solidarity Hospital Buea.