Survey on The Health Care Facilities Utilization for Influenza Like Illness in Two Sub-District Communities, Bandung District, West Java, Indonesia (3)

Statistical Analysis

We will develop a model for inclination to utilize the puskesmas and a separate model for actual utilization of the puskesmas. In the first model, we will develop an inclination score based on the questionnaire. Univariable analysis will be performed to identify a significant association of age, gender, and socioeconomic status, knowledge and disease severity perception, and geographical accessibility with this score using a t-test or Anova. Multivariable analysis for this model will be conducted by multiple linear regression. For the second model, since we will have actual utilization (yes or no), we will build a multiple logistic regression model of probability for seeking care to the puskesmas for ILI with knowledge and disease severity perception as explanatory variable while controlling for age and sex. We will also test the modifying effect of socioeconomic status and geographical accessibility in both models. We will use 95% confidence level for all statistical tests.  We will use computer software R version 2.12.2 for basic & regression analysis, AccessMod 3.0 for geographical accessibility calculations, and Arc GIS version 10 for mapping. 

Ethical Issues

Interviewees will sign informed consent forms approved by the ethical review committees in Indonesia and US. All data will be stored in secure database. Only de-identified data will be used for all analysis. Ethical clearance for this study will be obtained from Faculty of Medicine, Universitas Padjadjaran, Bandung, Indonesia and COMIRB, Aurora, Colorado, US.

References

1.     WHO. Cumulative number of confirmed human cases for avian influenza A(H5N1) reported to WHO, 2003-2012.  2012  [cited 2012 03/29]; Available from: http://www.who.int/influenza/human_animal_interface/avian_influenza/EN_GIP_20120326CumulativeNumberH5N1cases.pdf

2.     Kartasasmita CB, Agustian D, Murad C, Mutyara K, Tessa P, Uyeki TM, Simões EAF,. Community-based Influenza Surveillance in Two Rural Communities in West Java, Indonesia.  Options of The Control of Influenza VII. HongKong; 2010.

3.     Najnin N, Bennett CM, Luby SP. Inequalities in care-seeking for febrile illness of under-five children in urban Dhaka, Bangladesh. J Health Popul Nutr. 2011 Oct;29(5):523-31.

4.     Krishnaswamy S, Subramaniam K, Low WY, Aziz JA, Indran T, Ramachandran P, et al. Factors contributing to utilization of health care services in Malaysia: a population-based study. Asia Pac J Public Health. 2009 Oct;21(4):442-50.

5.     Taffa N, Chepngeno G. Determinants of health care seeking for childhood illnesses in Nairobi slums. Trop Med Int Health. 2005 Mar;10(3):240-5.

6.     Tinuade O, Iyabo RA, Durotoye O. Health-care-seeking behaviour for childhood illnesses in a resource-poor setting. J Paediatr Child Health. 2010 May;46(5):238-42.

7.     Chernichovsky D, Meesook OA. Utilization of health services in Indonesia. Soc Sci Med. 1986;23(6):611-20.

8.     Kahabuka C, Kvale G, Moland KM, Hinderaker SG. Why caretakers bypass Primary Health Care facilities for child care - a case from rural Tanzania. BMC Health Serv Res. 2011;11:315.

9.     McLafferty SL. GIS and health care. Annu Rev Public Health. 2003;24:25-42.

 

Survey on The Health Care Facilities Utilization for Influenza Like Illness in Two Sub-District Communities, Bandung District, West Java, Indonesia (2)

Methods

We will conduct a cross-sectional survey by door to door interview. This home visit interview will be performed by trained enumerator who will collect data on the health care facilities utilization pattern and perceived severity of ILI in the community using a pre-tested questionnaire. Socio-economic status such as education and occupation of household head, household size and structure, house condition, assets ownership, and health insurance status will be collected as well. The association of these variables and health care seeking behavior will be analyzed. Existing household databases from CIRAI and the high resolution map of the study area will be used to locate the randomly selected study households. The existing high resolution map with road network will be used to measure geographical accessibility of puskesmas from the house of respondent.  All surveyors will be trained on how to conduct interview and communicate effectively with the potential respondent regarding the purpose of the study and in obtaining informed consent. Only those subjects who agree to participate and sign the informed consent will be interviewed. 

The sample will be selected by stratified random sampling with probability proportionate to size. All households will be stratified into 8 categories based on the neighborhood health care utilization rate, neighborhood geographical network distance to puskesmas, and household socioeconomic status.  A neighborhood will be defined as a block of buildings that visually appear to be one community cluster without any physical or landscape barrier on the high resolution map using GIS. All neighborhoods will then be stratified into 4 classes based on the proportion of population who utilize puskesmas for ILI and the geographical distance between the centroid of the neighborhood boundary and puskesmas (see figure 2.). All household within each neighborhood class then will be stratified based on the household socioeconomic status into two categories. 

We plan to collect data from about 500 households in the study area, with the target of 90% response rate. This number is calculated based on the 95% confidence level, 80% power and 5% precision, with an estimate that the proportion of population using the puskesmas for ILI is 20%. 

Read more for statistical analysis...

Survey on The Health Care Facilities Utilization for Influenza Like Illness in Two Sub-District Communities, Bandung District, West Java, Indonesia (1)

Project Summary

Influenza A/H5N1 is posing a continuous threat of global pandemic due to its' high fatality rate. For early detection and treatment, health seeking behavior in the community is become very important to reduce the morbidity and mortality. From year 2008 – 2010 we conducted a surveillance study in community health care center in two sub-districts communities, Bandung, Indonesia. Using Geographical Information System, our preliminary analysis demonstrated that in some area people were less likely utilized health care for Influenza Like Illness (ILI). Therefore we propose a study to explore factors that associated with health care facilities utilization for ILI in the area using behavioral, epidemiologic, and geographic methodologies. The hypothesis is that the health care utilization for ILI is associated with knowledge, ILI severity perception, and perceive of puskesmas’s quality of care, and this association is modified by socio-economic status and geographical accessibility. For that aim, we will select subjects by stratified random sampling and conduct home visit interview. Model for health care utilization attitude and practice will be developed by linear and logistic regression. Geographical distribution of the patients will be visualized in maps.  

Project Description (Scientific view)

Introduction

Transmission of Highly Pathogenic Avian Influenza (HPAI) viruses from birds to human has resulted in over 598 infections with 352 fatalities globally¹, with Indonesia being at the epicenter of this phenomenon, a posing a serious threat of a global pandemic. In this context, care seeking behavior, in the area where humans interact with backyard poultry and birds such as in Indonesia, has become one of the crucial issues for reducing morbidity and mortality, by early detection and treatment. 
For this reason, and to determine the extent of avian to human transmission of HPAI, between 2008 to 2011, we conducted a passive surveillance study of Influenza Like Illness in community primary health care centers (Puskesmas),in two sub-districts community, in Bandung District, West Java, Indonesia(total population about 212,000)². Using Geographical Information Systems, we geocoded all houses of all residents in the catchment area, and of all patients who sought care at the Puskesmas in both areas over 3 years. Based on our preliminary analysis, it can be seen that in some areas (red circle), people were less likely to seek care and this was not explained merely by geographical proximity to the Puskesmas (Figure 1). 

Previous studies have shown that age, sex, socioeconomic status and perceived severity of illness influence health care seeking behavior^3-5. A study in Nigeria showed that for childhood illnesses, most caretakers sought care within the home⁶. A nationwide study in Indonesia illustrated that low household income was a barrier to health care utilization, even when health care was free⁷. Finally perceived poor quality of healthcare  service was also associated with lower utilization of primary health care services⁸. Thus health care utilization is a multidimensional concept that is related to the characteristics of the population, their behavior, and the environment where they live⁹.  While some of the determinants such as age, sex, and socioeconomic are not amenable to change, there are factors that amenable to interventions, such as knowledge, perceived severity of influenza like illness, and quality of service delivery.  A study that can inform public health authorities on the magnitude of these factors affecting care seeking behavior of the people, help in designing an effective program to improve health care delivery. Based on these observations, our preliminary study and the pre-existing geo-mapping of 2 entire sub-districts, I propose to study the  health care utilization behavior for ILI in this well defined community, incorporating social, behavioral, epidemiologic and geographic methodologies.  

The objective of this study is to explore factors that are associated with community health care facilities (puskesmas) utilization for ILI in the two sub-district communities in Bandung, West Java, Indonesia. The hypothesis is that puskesmas utilization for ILI is associated with perceived severity of illness and perceived quality of care at the puskesmas. The effects of socioeconomic status and geographical accessibility, on these modifiable behaviors will be studied. 

Read more for methods...

 

Reference for Rare Disease Mapping

1.

Gómez-Rubio V, López-Quílez A. Rare Diseases Epidemiology. Springer Netherlands; 2010. p. 151–71. Available from: http://dx.doi.org/10.1007/978-90-481-9485-8_10

2.

Wakefield J. Disease mapping and spatial regression with count data. Biostatistics. 2007 Apr 1;8(2):158 –183.

 

Project 3. Mapping for Rare Disease

The objective of this page is to describe rather a technical procedures in rare disease mapping rather than conceptual or theoretical. I will refer to some papers for those of you who might more interest in some theoretical background of it. However I also suggest that you could read these papers before getting into technical step, so you will be able to grasp the principle of rare disease mapping and apply appropriate procedures in relevant situation. 
Steps:
1. Create shape file with attributes of population at risk and number of cases per region (county for example)
2. Import the shape file into OpenGeoda.
3. Calculate rate
4. Create raw rate maps
5. Create smoothed rate maps