CHAPTER 3 RESEARCH DESIGN AND METHOD
This chapter describes the research design and method that were used to conduct the study. A quantitative research approach is essential to develop the body of knowledge that helps to improve health care and public health. Concepts that are relevant to quantitative research are basic research, applied research, rigour and control. Basic research is a scientific investigation that involves the pursuit of knowledge for the sake of knowledge. The purpose of basic research is to generate and refine theory and build constructs. Hence, the findings might not be directly useful in practice. Basic research provides a basis for applied research.
Applied research is conducted directly to improve public health and clinical practice. The purpose is then to solve problems, make decisions, or to predict outcomes. Applied research is also used to test theory. Thus this study used applied research methodology to examine PMTCT services and its outcomes which are relevant to improving mothers’ and infants’ health and to reduce the incidence of MTCT.
Rigour implies striving for excellence in research and involves discipline, adherence to detail and accuracy. Rigour ensures precise measurement methods, representative samples and controlled design. The researcher in this study applied rigour through using appropriate data collection techniques, proper sample size determination and applying a cohort design. Control shows the application of rules by the researcher to avoid the possibility of errors and making sure that the findings are accurate reflections of reality. Controlling confounding variables, sampling criteria, selection of the research setting, data collection technique and analysis are some of the areas to which control applied during this study (Burns & Grove 2005:32-34).
In an observational study the aim of the researcher is to examine the relationship between characteristics and events without manipulating the variables. The characteristic can be a treatment, behaviours, environmental factors, laboratory measurements, or genetic markers. An event can be an outcome or the occurrence of a disease. In this study the outcomes are related to maternal health, child HIV status and patient satisfaction. Observational studies can be either prospective or retrospective. In a prospective study subjects are selected based on the presence or absence of specific characteristics. In a retrospective study subjects are selected based on the presence or absence of events or occurrences in their lives. Subjects affected by the event are cases and those not affected are controls (Bacchieri & Cioppa 2007: 21). In this study, HIV-infected women were selected provided they enrolled in the PMTCT programme.
This chapter describes the study population, sampling, research instruments, data collection procedures and analysis. The procedures followed for ensuring validity and reliability of the research findings are outlined. The ethical considerations, followed during the conduct of this study, are also described.
The selected research design directs the selection of the population, sampling procedure, methods of measurement, and a plan for data collection and analysis. It depends on the researcher’s expertise, the problem and purpose of the study. “A research design is a blue print for the conduct of a study that maximises control over factors that could interfere with the study’s desired outcome” (Burns & Grove 2005:40). The two broad categories of study design are observational and experimental. Observational studies can be analysed using descriptive and/or analytical/correlational techniques. Descriptive studies quantify the extent of a disease or factor in a population while analytical/correlational studies help to identify a cause or risk factor of a disease or an outcome. Both descriptive and analytical/correlational studies use cohort and cross-sectional designs. A cohort is a group of people who share certain characteristic and are followed over a period of time. A cohort study is also sometimes called a follow-up study. There are two types of cohort studies: prospective and retrospective. A retrospective study shortens the period of study as it makes use of previous data (Morroni & Myer 2007:77-81). This study used a retrospective cohort design. Both descriptive and correlational analyses were conducted during this study as elaborated in the data analysis section 3.3.3.
Cohort studies can be conducted using historic data to identify outcomes. A retrospective cohort design is helpful in studying multiple risk factors or multiple outcomes. Cohort designs are considered to be relatively powerful to assess the association between risk factors and outcomes. When designing such kinds of studies, the researcher should have developed clearly stated dependent and independent variables prior to commencing the study (Murphy 2009:20). In cohort studies the researcher selects a group of study subjects, based on their exposure status, and then examines the rate of disease or outcome occurrences. Cohort studies are often conducted retrospectively and they fulfill at least one major criterium for causality: the time sequence of cause and effect (Stommel & Wills 2004:144). At the baseline people are classified according to their exposure status. Overtime follow-up investigations are conducted to identify the development of the disease (Olsen et al 2010:8).
In retrospective cohort designs, researchers assemble a cohort by reviewing their exposure status in the past. Then they also measure the occurrence of outcomes to some defined point in time. Tracing most of the cohort members must be possible in order to measure the outcome of interest. Retrospective cohort designs have comparative advantages over prospective designs that help to accomplish the research within shorter periods of time, rendering the study less expensive (Burns & Grove 2005:308).
In this study exposure status was determined based on PMTCT services. HCT, ART and ARV prophylaxis, infant feeding counseling and health care practices were taken into consideration. A cohort was assembled among women and their children who used PMTCT services at health facilities in Addis Ababa within two years prior to the study. Along with this additional variables were measured such as factors affecting HIV status of babies and barriers and enablers of PMTCT service utilisation. Outcome variables that were studied included the health status of the mother, HIV status of the baby and patient satisfaction with PMTCT service. These characteristics were studied retrospectively by reviewing patients’ records and conducting interviews with mothers who had utilised the PMTCT services.
The chosen research paradigm for this study is quantitative. According to Burns and Grove (2005:23) ”…quantitative research is a formal, objective, systematic process in which numerical data are used to obtain information about the world”. A quantitative approach deals with variables and might explore the association or cause and effect relationships between them. Stommel and Wills (2004:178) mentioned the features of quantitative research as being orientated toward theory testing, focusing on predetermined variables, preference for statistical and numerical summaries or descriptions, being analytical and providing explanations in terms of causal theories.
This study identified relevant theories (the HBM and Donabedian’s Model) that had been used for other studies. Variables were selected prior to data collection and analysis. The selected variables were helpful for numerical and statistical measurements and analyses. Both descriptive and analytical analysis techniques were applied.
Population and Sampling
In probability (random) sampling techniques every element of the population has a greater than zero chance of being included in the sample. Probability samples are more likely to be representative of the population as compared to the non-probability samples. Most commonly probability sampling techniques are used in quantitative research. There are four kinds of probability sampling techniques: simple random sampling, stratified sampling, cluster sampling and systematic sampling. Stratified sampling is used in situations where the researcher knows some of the variables that are critical to achieve representativeness. Variables which are commonly used for stratification are age, gender, ethnicity, socioeconomic status, diagnosis, geographical region, type of institution, type of care and site of care. Subjects are then randomly selected from the chosen strata. Stratified sampling can be disproportionate sampling or proportionate sampling. In disproportionate sampling each stratum should have an equivalent number of subjects in the sample. During proportionate sampling each stratum should have a number of subjects in proportion to their occurrence in the population (Burns & Grove 2005:346-348).
Probability sampling is considered in selecting a representative sample. In stratified random sampling the population is divided into sub-populations that are individually more homogeneous than the total population. These different sub-populations are called strata. Hence, strata are formed based on common characteristics. They are formed based on the past experience and personal judgment of the researcher. Usually the size of a sample from each stratum is proportional to the size of the relevant stratum. If P represents the proportion of population in the stratum i and n represents the sample size, then the number of elements selected from stratum i is n*P. To select elements from each stratum usually simple random sampling is applied (Kothari 2004:62-64).
A population constitutes the universe of interest and hence it is a universe of all of the units or elements to which the researcher wants to generalise the results obtained during a specific study. A researcher further defines these elements or units. This process of developing units or elements is referred to as the definition of the target population. Usually target populations are infinite or open ended. Practically researchers draw their samples from accessible populations that are often finite. Accessible populations may be defined in terms of their geographic locations or institutional affiliations or personal characteristics. Accessible populations are defined by observation periods chosen by the researcher. A sampling frame is a symbolic representation of study population. It is a list of all members of the target population or it is a set of procedures that allow the researcher to access members of the target population (Stommel & Wills 2004:297-306). Population, sometimes referred to as target population, comprises the entire set of individuals or elements who meet the sampling criteria.
An accessible population is a portion of a target population to which the researcher has reasonable access. Samples are obtained from accessible populations and hence, the generalisations made are applied to the accessible population and then to the target population. In order for each person in the target or accessible population to have a chance of being included in the sample, a sampling frame must be identified and random sampling must be implemented. A list of every member of the population must be acquired for this sampling frame. Hence, the sampling frame is this list of the population. The researcher then can select members from this list using the sample plan (Burns & Grove 2005:342-346). A sampling frame is a source list from which a sample is drawn. It should be comprehensive, correct, reliable and appropriate. It should represent the target population (Kothari 2004:56).
For this study the population comprised 8 365 HIV-positive women and their babies utilising PMTCT services in Ethiopia. The target population comprised 1 689 HIV-positive women and their babies who used PMTCT service in Addis Ababa, Ethiopia. The sampling frame comprised 796 mother-infant pairs who used PMTCT services at the selected 12 hospitals and health centers in Addis Ababa. The accessible population comprised those mother-infant pairs who received PMTCT services at selected public hospitals, private hospitals and health centers in Addis Ababa. Hence, inclusion criteria included mother-infant pairs who used PMTCT services, infants’ HIV known test results and the mothers who agreed to sign consent for themselves and on behalf of their babies. Exclusion criteria were mother-infant pairs who discontinued PMTCT services, pregnant women at the time of data collection, babies whose HIV test results were unknown and those who used PMTCT services at facilities outside Addis Ababa or at facilities that did not participate in this study.
In 2011 there were 1 445 hospitals and health centers providing PMTCT services in Ethiopia (FMOH/HAPCO 2012:30). The accessible population for this study encompassed PMTCT service users at health facilities in Addis Ababa during 2011-2013. This area was chosen because Addis Ababa has the highest HIV prevalence in the country and more PMTCT attendees were found in this area than anywhere else in Ethiopia. In 2012 PMTCT services were provided at 61 hospitals and health centers in Addis Ababa, comprising the site target population for this study. Thus the accessible sites comprised 10 public hospitals, 15 private hospitals and 36 public health centers providing PMTCT services In Addis Ababa. All these health facilities are located within Addis Ababa. These sites provide PMTCT services and they were accessible for data collection purposes.
Proportionate stratified random sampling was used to select health care facilities. The types of health facility served as strata. Public hospitals, private hospitals and health centers were selected to represent each category. These health facilities were categorised, given consecutive numbers within each category and from each group, sites were selected randomly. From each selected site, the respondents were selected using simple random sampling to ensure representativeness. Patients’ ART/pre-ART/medical record numbers used. Then a table of random number was used to select women attending PMTCT services. Those selected women were contacted during their follow-up visits to the PMTCT service sites. The number of mother-infant pairs selected was proportional to the number of patients treated at each health facility.
Twelve (19.7%) health facilities, out of a total of 61 health care institutions, were selected for this study. To ensure representativeness from each category of health facilities (private hospitals, public hospitals and health centers), approximately 20% of the facilities in each stratum were selected. The names of all facilities in each of the three strata were written on individual slips of paper, folded and placed into three different containers. Someone, not associated with this study, blindly selected two slips of paper indicating which public hospitals would participate in the study. The same procedure was followed to select three private hospitals and seven health centers. This amounted to stratified random selection without replacement. Subgroups of the health facilities, hospitals and health centers, were mutually exclusive. The selected public hospitals are Zewditu Hospital (Kirkos subcity) and Yekatit Hospital (Arada subcity). The selected private hospitals are Bethezata Hospital (Kirkose subcity), Myungsung Christian Medical Center (MCM) Hospital (Bole subcity) and Kadisco Hospital (Bole subcity). The public health centers selected are Addis Ketema (Addis ketema subcity), Bole17 (Bole subcity), Kolfe (kolfekeraneo subcity), Yeka (Yeka subcity), Meshualekia (Kirkose subcity), Kotebe (Yeka subcity) and Gulele (Arada subcity) health centers.
CHAPTER 1 ORIENTATION TO THE STUDY
1.2 BACKGROUND INFORMATION ABOUT THE RESEARCH PROBLEM
1.3 PMTCT AND VIRUAL ELIMINATION HIV
1.4 PMTCT IN ETHIOPIA
1.5 STATEMENT OF THE RESEARCH PROBLEM
1.6 AIM OF THE STUDY
1.8 DEFINITIONS OF KEY CONCEPTS
1.9 OPERATIONAL DEFINITIONS
1.10 THEORETICAL FOUNDATIONS OF THE STUDY
1.11 RESEARCH DESIGN AND METHOD
1.13 STRUCTURE OF THE THESIS
CHAPTER 2 LITERATURE REVIEW
2.2 DESCRIPTION OF THE THEORETICAL FRAMEWORKS
2.3 FACTORS AFFECTING THE UTILISATION OF PREVENTION OF MOTHER TO CHILD TRANSMISSION OF HIV SERVICES
2.4 PREVENTION OF MOTHER TO CHILD TRANSMISSION OF HIV
2.5 FACTORS AFFECTING MOTHER TO CHILD TRANSMISSION OF HIV
2.6 OUTCOMES OF PREVENTION OF MOTHER TO CHILD TRANSMISSION INTERVENTIONS
CHAPTER 3 RESEARCH DESIGN AND METHOD
3.2 RESEARCH DESIGN
3.3 RESEARCH METHOD
3.4 RELIABILITY OF THE STUDY
3.5 VALIDITY OF THE STUDY
3.6 ETHICAL CONSIDERATIONS
CHAPTER 4 ANALYSIS, RESULT AND DISCUSSION OF RESEARCH FINDINGS
4.2 RESPONDENTS’ SOCIO DEMOGRAPHIC CHARACTERISTICS AND IMMUNOLOGICAL CONDITIONS
4.3 FACTORS AFFECTING THE UTILISATION OF PMTCT SERVICES IN ADDIS ABABA
4.4 PROCESS OF PMTCT SERVICES IN ADDIS ABABA
4.5 OUTCOMES OF PMTCT SERVICES
CHAPTER 5 CONCLUSIONS, LIMITATIONS AND RECOMMENDATIONS
5.2 SUMMARY OF RESEARCH FINDINGS
5.3.2 Process of PMTCT services in Addis Ababa
5.6 LIMITATIONS OF THE STUDY
5.7 CONTRIBUTIONS OF THE STUDY
5.8 CONCLUDING REMARKS
List of references
GET THE COMPLETE PROJECT
REVIEW OF PREVENTION OF MOTHER TO CHILD TRANSMISSION OF HIV IN ADDIS ABABA, ETHIOPIA