Malaria epidemiology and comparative reliability of diagnostic tools in Bannu; an endemic malaria focus in south of Khyber Pakhtunkhwa, Pakistan

ABSTRACT

The present study was aimed at elucidation of malaria epidemiology and comparing performance of several diagnostic procedures in Bannu, a highly endemic district of Khyber Pakhtunkhwa, Pakistan.

Dried blood spots were collected from patients suspected of malaria visiting a hospital and two private laboratories in district Bannu and processed for species-specific PCR (rRNA). Patients were also screened for malaria through microscopy and RDT. A well-structured questionnaire was used to collect patient information to assess risk factors for malaria.

Of 2033 individuals recruited, 21.1% (N = 429) were positive for malaria by at least one method. Overall, positivity detected by PCR was 30.5% (95/311) followed by 17.7% by microscopy (359/2033) and 16.4% by RDT (266/1618). Plasmodium vivax (16.9%, N = 343) was detected as the dominant species followed by Plasmodium falciparum (2.3%, N = 47) and mixed infections (1.2%, N = 39). Microscopy and RDT (Cohen’s kappa k = 0.968, p = <0.0001, McNemar test p = 0.069) displayed significant agreement with each other. Satisfactory health, sleeping indoors, presence of health-care facility in vicinity (at an accessible range from home), living in upper middle class and in concrete houses significantly reduced malaria risk; whereas, low literacy level, presence of domestic animals indoors and malaria diagnosis recommended by clinician increased the disease risk.

Overall, findings from the study provide reasonable basis for use of RDT as a cost-effective screening tool in field and for clinicians who can proceed with timely treatment of malaria patients. Appropriate management of identified risk factors could contribute to reduction of malaria prevalence in Bannu and its peripheries.

Introduction

Malaria in Pakistan remains the fourth largest cause of death among communicable diseases, and along with Afghanistan, Somalia, Sudan and Yemen, it accounts for more than 95% of the total regional malaria burden. The country has a reported National Annual Parasite Incidence (API) averaging at 1.66 per 1000 population [1–3]. Plasmodium vivax (accounting for >80% cases) and P. falciparum are the only known prevalent species for malaria in Pakistan, although figures from the past few decades indicate that there is a substantial rise in P. falciparum infections [4–6]. Federally Administered Tribal Areas (FATA), Baluchistan and Khyber Pakhtunkhwa (KP) provinces reportedly observe the highest APIs in Pakistan. According to the latest stratification, 66 districts and agencies have been categorized in the high endemicity stratum (API >5 per 1000) [1]. In KP, malaria incidence (particularly P. falciparum malaria) is unstable and fluctuates with climatic deviations, that adds to underlying risk for outbreaks as reported in the past [7,8]. Many factors have contributed toward increase of malaria cases over the years including warm autumns (that subsequently prolong the transmission period) [7], emergence of chloroquine resistance across the country [8] and a chronic drop in vector control activities.

Designing efficient control and prevention strategies requires taking into account demographic and climatic factors that probably influence malaria transmission in a region [9–12], e.g. rainfall, age and gender [13], human migration [14], type and location of housing, etc [15–17]. Geographical and chronological differences in malaria transmission require well-timed identification of crucial risk factors, so as to implement targeted control measures [18].

Falciparum malaria, having fatal implications, is increasingly reported from southern and northern districts of KP [19]. District Bannu (population: 1.1 million) has an API of approximately 1.6–3.5 per 1000 population, which is substantially above the national average (0.8 per 1,000 population). The district is of great economic significance for being the central hub in the South of KP, in addition to serving as a short route to markets in Central Asia [7,20].

Variations in both human and parasite population in Pakistan demand periodic surveillance of burden and distribution of malaria to ensure appropriate and timely case treatment, particularly in situations where diagnosis by microscopy or species-specific rapid diagnostic tests are not accessible [21]. Misdiagnosis and poor differential diagnosis is a common concern in most endemic countries that usually leads to treatment failure [22]. The present study aimed to investigate epidemiology of malaria in Bannu district, Khyber Pakhtunkhwa, Pakistan. Additionally, we set out to assess and compare performance of common diagnostic techniques in this region of Pakistan.

Methods

Study area and sampling

Bannu (32°43′–33°06′ N; 70°22′–57′ E) is one of the 26 districts in the south of Khyber Pakhtunkhwa province of Pakistan with an area of about 1227 sq. km and a population of 1,167,892 (Pakistan Bureau of Statistics, 2017), majority of which is rural. Bannu is divided into two tehsils Domel/Bannu-1 and Bannu-II, comprising 40 union councils [23]. About 80 public health centers including Medical Teaching Institutes (MTI), District Headquarter Hospitals (DHQ), Regional Health Centers (RHC), Basic Health Units (BHU), Civil Dispensaries (CDs) and Mother & Child Health Centers (MCH) along with 15 private centers operate in Bannu District. In these health centers, first-line treatment for unconfirmed malaria is Chloroquine, for P. vivax is Chloroquine-Primaquine, for uncomplicated P. falciparum malaria is Artesunate/Sulfadoxine-Pyrimethamine (AS+SP), whereas Artesunate, Artemether or quinine is recommended for treating severe P. falciparum malaria or cases with treatment failures [24]. Control interventions being conducted here are run and sponsored by ‘Integrated vector control/malaria control program Khyber Pakhtunkhwa’ (IVC/MCP-KP) and ‘Frontier Primary Health Care’ (FPHC). These include regular trainings for focal persons in basic microscopy, RDTs use, malaria case management, outbreaks etc. Other essential prevention strategies include indoor residual spraying, mass distribution of bed nets to every household, antenatal care (ANC) bed net distribution (only pregnant women receive a bed net during her visit to hospital for antenatal care) and community education and mobilization campaigns (Personal communication with IVC/MCP-KP).

Blood samples were collected from patients suspected of malaria (patients manifesting classical nonspecific clinical malaria symptoms in an endemic area including fever, malaise, headache, myalgias, jaundice and sometimes nausea, vomiting and diarrhoea) [25] visiting Malaria Model Laboratory, Malaria Lab (Women & Children Hospital Bannu) and Siddique Laboratory Bannu from March to October 2013.

All participants gave written informed consent to participate in the study. Ethics Committee at University of Peshawar (ref. no.06/EC-16/Pharm, March 2016) approved the study documentation. To acquire demographic and other relevant information for assessing malaria risk factors, a questionnaire was administered to these suspected patients.

Sample processing and nested-PCR diagnosis

Finger-prick blood spots on filter paper from patients were processed with RDT and microscopy on-site (thick and thin smears) [26]. Briefly thin and thick slides were made on a clean, grease-free microscope slide and allowed to air dry. The films were stained with 10% Giemsa solution, allowed to air dry and then examined by microscopists at the labs/clinics with light microscopy using an oil immersion objective lens. A slide was declared negative only after observing 100 microscopic fields without finding parasites. Thick films were examined first for the detection of malaria parasites. Thick films in which malaria parasites were identified, were subsequently examined for species identification on their thin films as per standards described by WHO [27].

Rapid tests were conducted by using malaria rapid diagnosis (First response® Malaria Ag, pLDH/HRP2 Combo Card test kit, Cat. No 116FRC30). About 5 μl fresh blood was taken using a disposable pipette, dispensed into the RDT sample well and processed following manufacturer’s instructions.

Coarse porosity filter paper discs (Fisher Scientific, Loughborough, UK) were used to obtain blood spots from each patient. Filter papers were wrapped individually in properly labeled airtight sealed bags with silica gel and stored at 4°C until further processing [28]. A standardized disc section of the dried blood spot was punched using a Harris Uni-Core hole punch (2.5 mm diameter) and reconstituted in sodium azide (1 g per 1 L PBS/Tween) as described in [28]. DNA was later extracted using a resin-based Chelex® method [29] and then processed by a two-step Nested-PCR targeting conserved rRNA genes for discriminating Plasmodium species (using genus and species-specific primers) as described earlier (Table 1) [30]. In nest2, P. falciparum produces a 205 bp PCR product while P. vivax a 120 bp product.

Table 1.

Primers and reaction conditions used in diagnostic PCR.

PCR reaction	Forward primers(F)/reverse primer(R) (5ʹ-3ʹ)		PCR mixture	Cycling conditions
Nest 1 (genus specific primers)	F: TTAAAATTGTTGCAGTTAAAACG R: CCTGTTGTTGCCTTAAACTTC		20 µl reaction; 5 µl DNA from filter paper, 10.5 µl dH2O, 1X NH4SO4, 2 mM MgCl2, 250 µM dNTPs, 250 nM primer, 1 unit Taq polymerase	95°C, 5 minutes 25 cycles 58°C, 2 minutes 72°C, 2 minutes 94°C, 1 minute 1 cycle 58°C, 1 minute 72°C, 5 minutes 25°C, 10 minutes
Nest 2 (species specific primers)	rV1V2 rV1V2	F: ACTTCCAAGCCGAAGCAAAGAAAGTCCTTA R: CGCTTCTAGCTTAATCCACATAACTGATAC	20 µl reaction; 1 µl nest 1 product, 14.5 µl dH2O, 1X NH4SO4, 2 mM MgCl2, 250 µM dNTPs, 250 nm primer, 1 unit Taq polymerase	95°C, 5 minutes 30 cycles 58°C, 1 minutes 72°C, 1 minutes 94°C, 1 minute 1 cycle 58°C, 1 minute 72°C, 5 minutes 25°C, 10 minutes
	Rfal2 Rfal1	F: ACACAATGAACTCAATCATGACTACCCGTC R: TTAAACTGGTTTGGGAAAACCAAATATATT

Data analysis

Due to financial constraints, PCR was performed only on a subset of samples screened by microscopy but not by RDT. For the reason mentioned above, no sample was screened for all the three methods simultaneously, i.e. some samples were screened for Microscopy & RDT (N = 1618) and another set for Microscopy & PCR (N = 311). Therefore, the performance of RDT was compared against microscopy as the ‘gold standard’, whereas for microscopy, PCR was taken as ‘gold standard’ since PCR demonstrates high analytical sensitivity and is especially recommended for detecting low-grade infections (chronic, afebrile or mildly symptomatic, <100 parasites/μL) [31,32] in epidemiological research and surveys particularly in low-transmission areas [33].

Cross-tabulation and tests for comparison of diagnostic techniques were carried out in IBM SPSS Statistics 20. Kappa statistic (frequently used to test interrater reliability) [34–36] sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and McNemar’s marginal homogeneity were also calculated for each diagnostic method against a selected true positive standard [37,38].

All malaria positives were matched by gender and age (<5, 5–20, >20 years) to obtain matched case–control pairs in Stata v13 (Statacorp, 2013). Matching leads to a balanced number of cases and controls across levels of the selected matching variables leading to reduced variance in the parameters of interest, which ultimately improves statistical efficiency [39]. Using this paired data, univariable and step-wise multivariable conditional logistic regression was used to investigate socioeconomic and demographic risk factors for malaria in the study population.

Results

Samples were collected from 2033 individuals suspected of malaria visiting 3 major laboratories of Bannu. Of the total, 429 (21.1%) were positive for malaria by at least one diagnostic technique (Table 2). Prevalence among males (21.2%) and females (20.8%) were comparable. Similarly, prevalence among all age groups was almost similar, i.e. ≤5 (20.7%, N = 143), 5–25 yrs (22.5%, N = 170) and >25 (19.8%, N = 116).

Table 2.

Region-wise distribution of Plasmodium species.

District total no. of patients screened	No. of positive cases N (%)	P. vivax N (%)	P. falciparum N (%)	Mixed infections N (%)
Bannu N = 1566	349 (22.3%)	288 (18.4%)	34 (2.17%)	27 (1.7%)
FATA N = 320	52 (16.3%)	40 (12.5%)	10 (3.1%)	2 (0.6%)
Karak N = 19	4 (21.1%)	2 (1.5%)	0	2 (10.5%)
Lakki Marwat N = 128	24 (18.7%)	13 (10.2%)	3 (2.3%)	8 (6.3%)
Total N = 2033	429 (21.1%)	343 (16.9%)	47 (2.3%)	39 (1.9%)

In Bannu, Plasmodium vivax (14.2%) was the dominant species followed by P. falciparum (1.7%) and mixed infections (1.3%). Even though the largest number of cases were reported in Bannu, the positivity rate did not seem to differ between areas from where the patients originated (Table 2). Patients that belonged to neighboring districts possibly had to travel to Bannu due to non-availability of local diagnostic services or health facilities. Here, health centers in Bannu received maximum patients originating from agricultural belts of Bannu, Sarai Naurang, Domel and Lakki Marwat (Figure 1). Within Bannu, patient visits were more frequent from Bannu II which is predominantly urban (Table 3).

Figure 1.

Origin of malaria patients in District Bannu and its vicinities a) Abundance of malaria patients b) Abundance on land cover map.

Table 3.

Tehsil-wise distribution of Plasmodium species in district Bannu.

Tehsil total no. of patients screened	No. positive cases N (%)	P. vivax N (%)	P. falciparum N (%)	Mixed infections N (%)
Bannu 1 N = 468	91 (19.4)	74 (15.8)	9 (1.9)	8 (1.7)
Bannu 2 N = 1098	258 (23.5)	214 (19.5)	25 (2.3)	19 (1.7)
Total N = 1566	349 (22.3)	288 (18.4)	34 (2.2)	27 (1.7)

Apparently, the prevalence of P. vivax cases was lowest in March followed by a steady increase until it peaked in August. On the other hand, P. falciparum infections were consistently low throughout the year with an abrupt rise in October where these infections peaked. No cases were reported in August and September by us. Similarly, mixed infections displayed constant low rates that peaked in October (Figure 2).

Figure 2.

Seasonal abundance of Plasmodium infection.

PCR outperformed the other two diagnostic methods by displaying highest positivity (30.5%) followed by microscopy (17.7%) and RDT (16.4%) (Table 4). With PCR as a true positive test, microscopy presented a poor sensitivity value of 41% and a reasonable specificity of 73%. Sensitivity and specificity of RDT was 79% and 99%, respectively, using microscopy as a reference test. Significantly strong agreement between microscopy and RDT was observed (Cohen’s kappa k = 0.968, p = <0.001, McNemar test p = 0.595). On the other hand, there was a moderate agreement between PCR and microscopic procedures (Cohen’s kappa k = 0.14, p = 0.026. McNemar test p = 0.069) (Table 5).

Table 4.

Plasmodium species detected using three diagnostic techniques.

	No. of malarial infections N (% positivity)
Diagnostic Method no. of patients screened for each method	P. vivax	P. falciparum	Mixed	No. of total positive N (% positivity)
Microscopy N = 2033	290 (14.3)	30 (1.5)	39 (1.9)	359 (17.7)
RDT N = 1618	197 (12.2)	30 (1.9)	39 (2.4)	266 (16.4)
PCR N = 311	79 (25.4)	16 (5.1)	0 (0)a	95 (30.5)

^aSubset of samples screened for PCR had no positive mixed infection by microscopy.

Table 5.

Comparative performance of diagnostic methods.

Diagnostic Method No. of patients screened for each method	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	McNemar marginal homogeneity test (p-value)	Cohen’s Kappa value (p-value)
Microscopy a N = 311	41	73	41	73	0.069	0.14 (0.026)
RDT b N = 1618	79	99	98	99	0.593	0.968 (<0.001)

PCR: polymerase chain reaction; RDT: rapid diagnostic test; PPV: positive predictive value; NPV: negative predictive value.

^aUsing PCR as a true positive test.

^bUsing microscopy as a true positive test.

Paired cases and controls were utilized to evaluate risk factors for malaria (Supplementary Table 1). Assessment of risk factors for malaria revealed satisfactory health (OR = 0.24, 95% CI 0.16–0.35), sleeping inside the room (OR = 0.62, 95% CI 0.46–0.84), presence of health-care facility in vicinity (OR = 0.65, 95% CI 0.47–0.88), an upper middle class (OR = 0.68 0.46–0.99) and living in concrete house type (OR = 0.54, 95% CI 0.29–1.01) as statistically significant factors for reducing the disease risk. Other exposure variables that were identified to increase risk of malaria exposure included primary literacy level (OR = 1.97, 95% CI 1.18–3.30), presence of domestic animals within houses (OR = 1.36, 95% CI 1.03–1.78) and malaria diagnosis recommended by clinician (OR = 1.84, 95% CI 1.39–2.43) (Supplementary Table 1). Step-wise multivariate analysis for risk factors, as assessed for all diagnostic methods and microscopy identified primary literacy level (OR = 1.98, 95% CI 0.16–0.36), a diagnosis recommended by clinician (OR = 1.47, 95% CI 1.09–2.00), weak and satisfactory health status (All positives OR(weak) = 0.56, 95% CI 0.37–0.84 and OR (satisfactory) = 0.18, 95% CI 0.12–0.26; Microscopy OR (satisfactory) = 0.24, 95% CI 0.16–0.36), sleeping indoors (All positives OR = 0.72, 95% CI 0.53–0.96; Microscopy OR = 0.69, 95%CI 0.50–0.95), body aches (OR = 0.66, 95% CI 0.47–0.93) and presence of health-care facility in vicinity (OR = 0.61, 95% CI 0.44–0.85) as factors crucial for malaria transmission and control (Table 6).

Table 6.

Risk factors for malaria based on multivariate conditional logistic regression of variables from case–control pairs.

Diagnostic method	MicroscopyN = 718	Positive by any methodN = 858
Variables	Odds ratio (95% CI)	Odds ratio (95% CI)
General health (compared with poor health)
Weak  Satisfactory	0.87 (0.57–1.32) 0.24* (0.16–0.36)	0.56* (0.37–0.84) 0.18* (0.12–0.26)
Literacy (compared to illiterate)
Primary  Secondary & above	–	1.98* (1.13–3.47) 1.25 (0.89–1.76)
Indoor sleeping habit (compared to outdoors)	0.69* (0.50–0.95)	0.72* (0.53–0.96)
Presence of health care facility in living area	0.61* (0.44–0.85)	–
Having body aches	–	0.66 (0.47–0.93)
Malaria diagnosis recommended by clinician (compared self-diagnosis without recommendation by clinician)	–	1.47* (1.09–2.00)

CI: confidence interval.

*Factors with p-value <0.05.

Discussion

The study was one of the first to provide a comprehensive understanding of malaria epidemiology in Bannu district and to assess the comparative efficiency of different diagnostic methods in this endemic focus.

Overall, 21.1% of the cases were positive by at least one method. Bannu was considered as one of the most malarious areas of the province [40]. Internal displacement of FATA population in 2012 (about 700,000 people from neighboring North Waziristan) was suspected to have added to the malaria burden in Bannu since FATA regions are known to exhibit highest rates of disease incidence in the country [21,41,42]. Corroborating previous studies, P. vivax infections were more prevalent compared to P. falciparum and mixed infections [43–45]. Dominance of one malaria species over the other is primarily determined by the parasite’s biology and by an area’s climatic and seasonal variations [46]. Larger influx of P. vivax is possible because true relapses do not occur in P. falciparum, whereas in P. vivax, relapses are common due to prolonged survival of Vivax hypnozoites in liver cells [47–49].

In our study, P. vivax cases peaked in the summer month (August) while P. falciparum and mixed infections in winter (October). Similar distribution trends are known to exist in certain areas of neighboring Afghanistan that share similar climatic settings [20,50,51]. Generally, in Pakistan, P. falciparum transmission starts in the summer monsoon (July) when the temperature and humidity are optimum and it prevails until the end of the year when the temperature falls below the critical value (December). P. vivax usually observes two transmission peaks, one is its early transmission period during the wet months of spring (probably facilitated by true relapses) and the other is with the P. falciparum (monsoon) [7,20,40,52].

Unlike the findings of our study, age has earlier been associated with malaria acquisition [53,54] where children ≤5 years are shown to carry a comparatively higher risk [55–57]. The acquired immunity is both exposure- and age-dependent, and the older children are likely to have developed some degree of immunity because of repeated infections [58,59]. There also exists a possibility of extended exposure due to inattentiveness of parents/guardians toward protective/treatment measures [56]. Further, we documented no significant association with gender. Although, several earlier studies in Khyber Pakhtunkhwa and elsewhere prove males to be at a higher risk of malaria compared to females [43,45,60–68] suggesting increased exposure to mosquito bites since males are more likely to work outdoors and are not traditionally well covered as females (usually veiled) [52]. Cultural and social norms might have stemmed low influx of female patients to health-care facilities.

In the present study, microscopy displayed poor sensitivity when compared to rRNA PCR. Studies suggest that microscopy, when compared to molecular diagnostic methods like PCR, often exhibit considerable discrepancies (usually showing lower sensitivity to mixed infections), especially in submicroscopic infections [69–72]. Parasite densities measured by microscopy correlate with parasite gene copy number in quantitative PCRs and studies indicate that in submicroscopic infections lower copy numbers are expected in low transmission settings than those in the high transmission. These deviations eventually impact diagnostic outcomes [73,74]. Moreover, mixed infections pose challenges to microscopists since similarities exist between developmental stages of different Plasmodium species. Microscopy and RDT are prone to miss infections with density less than 100 parasites/μL, while the detection limit of PCR is generally <5 parasites/μL [75,76]. Other factors like lack of expertise and lack of standard good quality blood films also often contribute to such incongruities [71,77]. On the other hand, PCR will rarely miss microscopy positives due to poor DNA quality, poor quality standards in techniques or missed priming due to mutations [78]. Although studies demonstrate superior performance of PCR traditional diagnostics means in field [79–81], PCR cannot be used as a routine diagnostic tool due to its technologically advanced lab requirements and expenditures [82–84].

WHO recommends RDTs and microscopy as primary methods for diagnosis of clinically suspected malaria in all epidemiological settings, including low transmission areas like Khyber Pakhtunkhwa [85]. In our study, among 1618 samples screened through RDT, 266 (16.4) were detected positive for Plasmodium infection. RDT presented satisfactory sensitivity (79%) when compared to microscopy. Overall findings from the study provide reasonable basis for use of RDT as a screening tool in field and for clinicians to proceed with timely treatment of malaria patients. Studies also support the superior performance of RDT compared to microscopy in routine clinical settings and especially in remote locales where medical units face deficit in resources like contained labs, electricity, trained microscopists, etc. [71,86]. RDTs can serve as a cost-effective method for clinical diagnosis, mainly due to improved treatment and health outcomes for non-malaria febrile disease [87]. Incorrect diagnoses are a common observation in labs like the ones selected in our study, largely due to huge numbers of patients availing for free diagnosis. Furthermore, we observed that many of the patients visited the labs for diagnosis without referral by clinicians.

Malaria risk was also investigated through multilevel analysis of individual and household level factors. Keeping livestock indoors was seen as a risk factor for acquiring malaria as described previously [65,88]. Cattle may attract mosquitoes, either by serving as bait or by creating mosquito breeding and resting sites near livestock pens [56,89–91]. Cattle provide an easy source of blood to mosquitoes, which results in increased vector populations, a significant proportion of which is attracted to feed on individuals sleeping outdoors near animals [92]. Risk of mosquito bites could be reduced due to zoo-prophylaxis [93] especially in situations where vector species predominantly display zoophagic foraging tendencies. However, the benefit of keeping livestock close to human dwellings has been refuted by many authors [94,95]. In Pakistan, Anopheles culicifacies and A. stephensi are confirmed as malaria vectors [96] and are chiefly Pakistan’s zoophilic [97]. A. culicifacies is the primary vector in rural areas while A. stephensi is considered to be an important vector for malaria transmission in urban areas [98,99]. Livestock are thought to be largely responsible for generating the high mosquito densities in the region.A strong positive correlation between cattle:man ratio and malaria incidence was reported in northern KP province of Pakistan [88].

Frequent outdoor activities also increased probability of infection as mosquitoes are generally active between dusk and dawn [100]. Risk of malaria transmission always exists where mosquito host-seeking behavior coincides with place and times of human presence [101]. In the study, herein, most of the individuals had evening activities while sleeping in open grounds was a common practice.

Sleeping habits especially clustering in sleeping rooms was established as a crucial risk factor. Clustering at household level appears to increase malaria risk probably due to increased release of human chemo-attractants [102,103]. The study area experiences long hours of power outage in summers that aggravates the biting rate.

Concrete houses were observed to reduce risk of malaria since they can decrease mosquito contact by limiting their entry and reducing their resting places [104]. Higher malaria incidence rate is commonly reported from poorly constructed houses (incomplete, mud, or palm walls and palm thatched roofs) compared to well-constructed houses (bricks/stones) since poor housing likely leads to an increase in human–vector contact [15,17,95,104–106]. On the contrary, several others suggest no such association between malaria incidence and housing quality [16,56,107–110].

Living in an upper middle class was protective against malaria as these households have the amenities to adopt better preventive measures compared to the poor and underprivileged [105,111–116]. Socioeconomic factors may directly or indirectly affect malaria transmission [57,117,118] and failure to consider effects of socioeconomic elements might jeopardize the success of control programs [119].

In Bannu, acquiring only primary education by patients or guardian of patients seemed to increase the risk of malaria. There is a dire need to disseminate information for preventive and control strategies in the study area as majority of the patients denied using any precautionary measures. Commonly, individuals availing government health services belong to lower socioeconomic backgrounds generally with low education levels. These have been associated with poor knowledge regarding utilization of preventative methods and malaria treatment among them [91]. Achieving higher education is known to be protective against malaria [120].

Apparent health status of the individual was indicative of malaria possibly because many patients were referred by clinicians for malaria tests due to their anemic status expected in malaria infections. If left untreated (or inadequately treated), malaria may result in several weeks or months of poor health following repeated attacks of fever and anemia [107,121–123].

Overtreatment of malaria is not only an economic concern, but it has been proposed that restricting antimalarial use to laboratory-confirmed cases will also delay the emergence and spread of resistance to artemisinin derivatives and their partner drugs [124]. The labeling of all febrile patients as having malaria can have severe consequences as the underlying disease would not be properly identified and treated [125]. In our study, we observed that although many of the patients were recommended by clinicians for diagnosis, there were several patients that visited the facility on their own behalf. Many had a history of malaria with self-medication while others were prescribed antimalarials without diagnosis. Many of the private practitioners and labs also prescribed antimalarial drugs without confirmatory malaria tests.

Conclusion

Level of malaria endemicity, the urgency of diagnosis, experience of clinician, the effectiveness of health-care workers and budget resources are all some factors affecting the choice of malaria-diagnostic method. In this study, RDT was identified as a suitable diagnostic test for its decent sensitivity and specificity; as well as its inter-procedure agreement for detecting parasite compared to microscopy. Population of Bannu, especially the rural lot lacks suitable health infrastructure, which is principal for disease management. Ultimately, targeting effective malaria control programs require essential understanding of epidemiology of malaria in a region.

Acknowledgments

We are sincerely thankful to the laboratory and hospital staff for their extended support during the survey.

Disclosure statement

No potential conflict of interest was reported by the authors.