Research Article, Vector Biol J Vol: 2 Issue: 1

Predicting Malaria Risk and Utilization of Health Care Services among the Population in Uganda

Muwanika FR^*, Atuhaire LK and Ocaya B

School of Statistics and Planning College Of Business and Management Science, Makerere University, Uganda

*Corresponding Author : Muwanika Fred Roland
School of Statistics and Planning College Of Business and Management Science, Makerere University, Uganda
Tel: +256779604453
E-mail: frmuwanika@yahoo.co.uk

Received: September 07, 2017 Accepted: September 14, 2017 Published: September 21, 2017

Citation: Muwanika FR, Atuhaire LK, Ocaya B (2017) Predicting Malaria Risk and Utilization of Health Care Services among the Population in Uganda. Vector Biol J 2:1. doi: 10.4172/2473-4810.1000119

Abstract

Keywords: Population attributable risk; Population attributable risk fraction; Relative risk

Introduction

Malaria remains a leading cause of morbidity, mortality and non-fatal disability in Sub-Saharan Africa and in Uganda, especially among children, pregnant women and the poor. Ugandan population is estimated to be approximately 30.7 million people with over 80% of the population still living in rural areas [1]. With the growing population and varying needs, the government has developed several policies to improve the health status and lives of the people. According to UBOS [1] the prevalence of malaria was reported to be 52% in the general population. This suggests that malaria is still prevalent and is associated with high mortality and morbidity in Uganda, which could be associated with the methods of prevention and management of the disease. The country within its National Development Plan (NDP) of 2010 [2] aims at reducing the prevalence and incidence of communicable diseases by at least 50 percent, as part of its efforts towards achieving MDG 6. This achievement can only be observed through well designed health policies that are pivoted around integrating both health cost as well as the effectiveness of the services received by the beneficiaries. Modelling the dynamics in the intensity rate does not only provide information on the severity of the disease but also on its dynamics with varying populations. The estimates of the malaria incidences are based on household surveys which are conducted after every five years. It is also extremely hard for these surveys to justify if the reported fever was actually malaria but it gives an indication of the burden in absence of more accurate methods. Therefore this study modelled the intensity of malaria burden over time based on routine data obtained in normal health seeking environment to establish the risk and utilization of malaria care services and use it to explain the dynamics of the disease in the population.

One of the main concerns with choices for scaling up prevention and treatment of malaria in Uganda and other African countries has been the cost [3]. National and international policy-makers require information on which strategies are best for prevention, improving treatment and determining the effectiveness of investments in malaria control compared with other health-care interventions.

Most studies conducted are based on data collected during trials, or use models to synthesize data from many different sources. Evaluations of routine service delivery are very rare. Country coverage is haphazard and heavily related to the location of the main research institutions. These few studies provide some guidance to decision-makers, and indicate that many of our current control measures can be highly cost-effective relative to many other healthcare interventions [4]. However, the potential of these studies to inform policy debates is limited by the lack of evidence on the costs and effects of packages of measures. In addition there are problems in generalizing or comparing studies that relate to specific settings, and use different methodologies and measures of outcome. While more studies particularly those focusing on new interventions and delivery modes in operational settings are needed this takes time and it will never be possible to perform studies in an ideal situation. The development of modeling approaches, which can provide a systematic analysis of variation in cost-effectiveness across different epidemiological and economic zones, may be a much cheaper and feasible option which this study attempts to do [5]. The suggested approach estimates the risk of malaria in the general population and utilization of malaria care services at the health facilities. This helps not only in establishing the intensity of the disease in the population but also the extent to which the various interventions can eliminate malaria from the country.

Methods

Data source

The materials used in this study were obtained from two secondary data sources. Malaria data were obtained from Ministry of Health Uganda (MOH) while the population projection data were from Uganda Bureau of statistics (UBOS). The Ministry of health Uganda through its health facilities collects routine data using a standard Health Management Information System (HMIS) form. The HMIS 105 form is provided at all the health facilities across the country. The HMIS form 105 is structured in seven section collecting data on different services offered at the facilities. This study utilized data collected in section one of the form. Section one of the form collects data on outpatients (OPD) attendance, outward referrals to higher levels of service delivery, and outpatient diagnoses. Under the Outpatient Diagnoses Section monthly observed cases of infectious/communicable diseases like malaria, noncommunicable diseases as well as other neglected diseases are recorded. The population at risk of malaria was estimated from the population projections that were estimated by Uganda Bureau of statistics for the period 2002-2017.

Model development

This study demonstrates how the proposed model can be used to study the malaria incidence rate or intensity in the general population. Incidence rate is defined as the number of new events in a population at risk per unit person per unit time. From the calculus point of view the incidence rate is the limit of the number of events observed in the time interval ( t, t +t divided by the population at risk.

(1)

Mathematically the incidence rate is expressed as; Therefore differentiating the proposed model Muwanika FR, Atuhaire and Ocaya [6] with respect to time gives the expression for the incidence rate of malaria in the population given by 2 equation below.

(2)

The model 7.5 predicts the monthly malaria risk in the next month caused by infectious individuals from the previous month. Therefore the expected number of new cases of malaria in the next month is given by

(3)

This model is in agreement with defined for predicting the number of new cases of an infectious disease in the susceptible- Infectious-susceptible (SIS) modeling framework. This model can be used to predict the expected number of new cases of malaria in the next month. One of the advantages of this model is that the intensity of malaria varies with time which is more realistic as opposed to constant incidence rate which is assumed when researchers apply the Poisson model in the analysis of incidence rate.

We demonstrate how the proposed model for predicting monthly incidence rate could be used to assess access and utilization of malaria health services in the general population. Let λ_T (t) denote observed number of new cases of malaria that sought treatment for malaria at the government health facilities. The expected malaria number of new cases of malaria in the population denoted by λ_P (t) was obtained from our proposed model 3.

The difference between the expected and observed malaria incidence gives the estimate of the malaria incidence attributable to infected individuals who never sought treatment for malaria from the government health facilities. In epidemiology the difference in monthly malaria incidence among the population and the population seeking medical care at the health facilities is commonly referred to as population attributable risk.

Population attributable risk (PAR) is defined as the portion of the incidence of a disease in the population that is due to exposure. In this study the monthly malaria incidence estimated in the population is denoted by λ_T (t) . The study therefore assesses the effect of seeking treatment (exposure) on reducing monthly malaria incidence among the population. Therefore the population infected with malaria exposed to treatment is denoted by λ_T (t) The monthly malaria incidence attributable to infected population not seeking medical treatment from the health facilities is given by equation 4.

(4)

From equation 4 the population attributable risk percentage is computed. Population attributable risk percent (PARF) is the percentage of the incidence of a disease in the population that is due to exposure. The population attributable risk percentage (PARF) is given by equation 5.

(5)

Lastly we compute the relative risk of monthly malaria incidence among the population that sought treatment and that, which never sought malaria treatment. The relative risk (RR) is a measure of association between a disease or condition and a factor under study in our case the factor is seeking medical care for malaria. The relative risk was computed using equation 6.

(6)

Utilization of malaria health services in Uganda

The average estimates of the population attributable risk, population attributable risk percentage and relative risk among the individuals seek medical treatment for malaria and those who never sought treatment are summarized in Table 1. On average 340 new cases of malaria per 10000 people were observed at the various government health facilities across the country. However the expected number of new malaria cases was estimated to be 876 cases per 10000 people. This estimate was far higher than what was observed because not all the malaria cases in the country are regularly diagnosed and treated at the government health facilities. The excess of 538 new cases of malaria per 10000 people either sought treatment at private facilities or never sought treatment from any health facility. The excess cases accounts for 61% of the monthly malaria incidence in the population Table 1. Failure to monitor the 60% of the new cases to establish whether they actually received treatment for the disease might be responsible for the observed stagnation in malaria incidence in the country. Our findings are comparable to the findings of the Uganda malaria indicator survey that established that about 52 percent of the children had malaria parasites in their blood although they had not yet shown clinical symptoms [6]. This implies that if adults were also to be targeted for the malaria test the proportion of asymptomatic cases in the population could even be higher. Our estimate of 39% of the new malaria cases observed at the government facilities are higher than the UBOS estimate of 24.7% simply because the UBOS estimate is based on prevalence while our estimate is based on incidence. It should be noted that estimation of disease risk based on prevalence always underestimates the risk. Furthermore it was established that 42.7% and 32.7% sought treatment from private health facilities and drug shops respectively. This implies that 26.6% (233 new cases of malaria per 10000 people) would be eliminated monthly if all infected individuals seek treatment either from government or non-government health facilities. These finding shed more light on potential use of routine data collected at the health facility level. This data can be used to design better interventions to prevent, manage and control malaria incidence in the population. The analyses show that cases observed at the government facilities can be used not only for predictions within the facilities, but also measure the risk in the population that is seeking treatment from non-government facilities. Therefore improved collection of the routine data can provide timely statistics for better monitoring and design of interventions that will reduce malaria incidence in Uganda. Utilization of malaria health care services also varies across the regions in the country where some regions have fewer health facilities as compared to others. Therefore we extended the analysis to assess the difference in utilization of malaria health services in the four regions of the country. The estimates of the malaria incidence in the four regions are summarized in Table 2.

Table 1: Utilization of malaria health services.

Table 2: Utilization of malaria health services.

Utilization of malaria health care services in central region

In the central region 339 new cases of malaria per 10000 people were observed across the government health facilities. The predicted number of new cases of malaria was estimated to 808 cases per 10000 people. This implies that on average about 447 new cases of malaria per 10000 people either sought treatment from non-government health facilities or never sought treatment for malaria at all. This difference translates into a 57.8% of the monthly malaria intensity in the central region likely either to seek treatment from non-government health facilities or never sought treatment at all. Further analysis shows that 32.7% (253 new cases of malaria per 10000 people) were attributable to non-government health facilities. A monthly reduction of 25.1% (194 new cases of malaria per 10000 people) would be eliminated if the infected individuals sought treatment either from government or non-government health facilities. Therefore self-medication seems to be playing a big role in explaining the persistently high incidences of malaria recorded in the region.

Utilization of malaria health care services in eastern region

In the eastern region 391 new cases of malaria per 10000 people were observed at the various government health facilities. Our estimate of the malaria intensity in the population was 903 new cases of malaria per 10000 people. The average monthly incidence of 515 new cases of malaria per 10000 people can be attributed to non-government health facilities or self medication. According to UBOS [1] 42.6% of the people who were infected with malaria/fever sought treatment from non-government facilities while 32.7% sought from drug shops and pharmacies an indication of self medication. This implies that 31.7% (291 new cases of malaria per 10000 people) sought treatment from the non-government facilities. Therefore in the eastern region 24.3% (224 new cases of malaria per 10000 people) of the monthly malaria incidence would have been eliminated if all infected individuals had sought proper malaria medication either from government or non-government health facilities.

Utilization of malaria health care services in northern region

The northern region had the lowest number of observed monthly malaria incidence. On average the monthly malaria incidence observed at the health facilities was 315 new cases per 10000 people. The predicted monthly malaria incidence was estimated to be 1155 new cases of malaria per 10000 people. Therefore the difference of about 845 new cases of malaria per 10000 people is attributable to individuals that either sought treatment from the private health facilities or self-medication which accounted for about 72.7% of the monthly incidence. But 42.6% and 32.7% sought treatment from private health facilities and self-medication respectively [7]. Therefore 31.6 %( 367 new cases of malaria per 10000 people) of the monthly malaria incidence would have been eliminated if all infected individuals had sought treatment either from government or private health facilities.

Utilization of malaria health care services in western region

In this region about 336 new cases of malaria were observed at the various government health facilities. On average the estimated monthly malaria incidence was about 940 new cases per 10000 people in the region. The difference of 63.5 %( 601 new cases of malaria per 10000 people) were attributed to self-medication or infected individuals who either sought treatment from non-government health facilities. Furthermore 27.6% (261 new cases of malaria per 10000 people) of the monthly malaria incidence would have been eliminated if all the infected individuals sought treatment from government or private health facilities rather than self-medication. Lastly we assess the difference in utilization of malaria health care services across the demographic characteristics of the population. The demographic characteristics considered in this study are age and sex. The estimates of the attributable incidence for each of the demographic groups are summarized in Table 3.

Table 3: Utilization of malaria health services by demographic factors.

Utilization of malaria health care services and age

We assessed our prediction of malaria incidence across the different age groups since age is a risk factor with children below five years being at a higher risk. Among the children below five years the observed monthly malaria incidence was estimated as 643 new cases of malaria per 10000 people. The model predicted average monthly malaria incidence was estimated to be 1289 new cases of malaria per 10000 people. The average incidence that was attributed to either self-medication or seeking treatment from non-government health facilities was estimated to be 646 new cases of malaria per 10000 people. According to UBOS and ICF Marco (2010), 56% of the children who had suffered from malaria were taken to non-government health facilities for treatment. This implies that 21.9% (284 new cases of malaria per 10000 people) of the monthly malaria incidence among children below five years would have been eliminated if the children were presented to government or private health facilities rather than self-medication by their parents.

Similarly among the adults above 5 years of age the observed cases were estimated to be 254 new cases of malaria per 10000 people. The model predicted monthly average incidence among adult population was 977 new cases of malaria per 10000 people. The average monthly incidence attributed to self-medication and seeking treatment from non-government facilities was 726 new cases of malaria per 10000 people representing about 73.9% of the monthly incidences. Therefore 32.1% (315 new cases of malaria per 10000 people) of the monthly malaria incidence among adults would have been eliminated if all the infected individuals had sought treatment from government or non-government facilities rather than self-medication.

Utilization of malaria health care services and sex

Among the male population the observed monthly malaria incidence was about 281 new cases of malaria per 10000 people. The model predicted incidence among this population was 877 new cases of malaria per 10000 people. The average monthly incidence attributed to either self-medication or seeking medication from private facilities was about 598 new cases of malaria per 10000 people representing about 67.8% of the total malaria incidence among the male population. Therefore 29.4% (260 new cases of malaria per 10000 people) of the monthly malaria incidence among the male population would have been eliminated if self-medication was avoided.

Similarly among the female population the observed monthly malaria incidence was about 366 new cases of malaria per 10000 people. The model predicted incidence among this population was 854 new cases of malaria per 10000 people. The average monthly incidence attributed to either self-medication or seeking medication from non-government facilities was about 490 new cases of malaria per 10000 people representing about 57% of the total malaria incidence among the female population. Therefore 24.7% (213 new cases of malaria per 10000 people) of the monthly malaria incidence among the male population would have been eliminated if selfmedication was avoided.

Conclusions

In this study we assessed how the proposed model could be used to reduce malaria incidence among the Ugandan population. In particular this paper explains how the predicted monthly incidence could be used to identify malaria incidence attributable to selfmedication and to individuals who might have sought treatment from non-government health facilities. The observed monthly incidence at the government health facilities were less than what the model predicted because not all new cases of malaria were commonly and regularly diagnosed at the government health facilities.

About 60% of new cases of malaria are more likely to either seek treatment from non-government health facilities or from drug shops. Seeking treatment from the drug shops, pharmacies without a medical prescription is a sign of self-medication. The study established that about 27% of the new cases of malaria incidence would have been eliminated if all infected individuals sought proper treatment from either government or private health facilities. However the high proportion of patients who sought treatment from non-government health facilities was attributed to absence of government owned health facilities in the communities.

It is advisable that the practice of seeking treatment from drug shops and pharmacies should be done after the infected individuals have been properly diagnosed with malaria to reduce the risk taking anti-malarials medication without presence of the disease in the host. Self-medication does not only increase the risk of malaria in the population but also the likelihood of future treatment failure. This study shows how the routine data collected at the health facilities could be used to model risk and have timely statistics on how to reduce disease incidence in the population. The findings from this study can be used to strengthen the Health Management Information and how the data therein can be utilized beyond routine reporting. However further research is required to modify the model to include other factors that may affect disease incidence

References

1. UBOS (2010) Socio-economic report based on the Uganda National Household Survey of 2009/2010.
2. National Planning Authority (2010) National Development Plan (2010/11-2014/15)
3. Hanson K, Goodman C, Lines J, Meek S, Bradley D, et al. (2000) The Economics of Malaria Control Interventions. Global Forum for Health Research. Geneva, Switzerland.
4. Goodman C, Kara Hanson K, Anne Mills A, Virginia Wiseman V, Eve Worrall E (2003) The Economics of Malaria And Its Control. WHO/TDR Scientific Working Group on Malaria, Geneva, Switzerland.
5. Goodman CA, Mills AJ (1999) The evidence base on the cost-effectiveness of malaria control measures in Africa. Health Policy Plan 14:301-312
6. Muwanika FR, Atuhaire LK, Ocaya B (2017) Prediction of Monthly Malaria Incidence in Uganda and its Implications for Preventive Interventions. J Med Diagn Meth 6: 248
7. Uganda Bureau of Statistics (UBOS) and ICF Macro (2010) Uganda indicator survey 2009. Calverton, Maryland, USA.