Contents - Previous - Next
Methodology and design
Appendix I: ARI interview guide
By Ruth P. Wilson, 'Mamochaki M. Shale, and Kathleen A. Parker
Ruth P. Wilson is a medical anthropologist at the Centers for Disease Control, International Health Program Office, Technical Support Division, Atlanta, GA. 'Mamochaki M. Shale is Head of the Department of Sociology and Anthropology, National University of Lesotho, Roma, Lesotho. Kathleen A. Parker is a Health Education Specialist at the Centers for Disease Control, International Health Program Office, Technical Support Division, Atlanta, GA.
In this paper, Rapid Anthropological Procedures were used to determine local terms, categories, and treatment practices associated with acute respiratory infections (ARI). It reports the findings from one phase of a study aimed at obtaining data on ARI among children under the age of five in Lesotho. Data were collected using participant observation, observations, and informal and formal interviews with individuals and groups living in a rural area. The authors conclude home treatments, traditional and religious healers play a part along with over-the-counter medicines, health providers, and modern medicines in caring for children with ARI. This paper represents a highly professional use of RAP and may serve as input for developing similar studies and for training. - Eds.
ACUTE RESPIRATORY INFECTIONS (ART) are a major child hood health problem in developing countries. Although coughs, simple colds, ear infections, bronchiolitis, and croup are part of the ARI complex, most deaths from ARI in young children result from pneumonia, or acute lower respiratory tract infection (ALRI).1 In 1987, ALRI accounted for 24% of the deaths in hospitalized children (unpublished data, Statistical and Planning Unit, Lesotho Ministry of Health). In 1988, the Lesotho Ministry of Health decided to address the problem nationally and requested assistance from the Centers for Disease Control (CDC). The International Health Program Office (IHPO) at CDC responded by assisting with baseline studies for policy development.2 The anthropological data reported here were presented, discussed, and used to outline plans for the health education and training component of the ARI Working Session held in Lesotho, December 1989. These data, combined with clinical and epidemiologic information, were used by the Lesotho Ministry of Health to develop a national ARI control policy and to initiate the first phase of the programme in three health service areas during 1990.
Lesotho, located in southern Africa, is the home of the Basotho people. They were united into a nation by the legendary figure Moshoeshoe I against Zulu war bands during the reign of Chaka. Today, the Basotho are a homogeneous ethnic group, speaking the same language, having similar, but not identical, cultural traditions.3 They are a patriarchal, patrilineal, and virilocal people who live predominantly in the rural villages throughout the mountains, valleys, and flatlands of Lesotho.4 In such societies, males have the authority to make final decisions in the public, and often domestic, spheres of life; ancestry is traced only through paternal relatives; and upon marriage, a woman resides in her husband's village or homestead with his kinsmen.
Urban residents may use public transportation to reach health services. If they must walk to reach a health facility, the walk is short, unlike the long walk taken by villagers living in the mountainous areas.
In each of
the two pert-urban sites selected for this study, villagers had
access to a church-sponsored clinic participating in the
government's village health worker programme. The clinics were
centrally located to a cluster of villages composing the health
catchment area. and were equipped with vehicles to transport
critically-ill patients to a nearby hospital. This factor, access
to modern health services, can be decisive in saving the lives of
infants with pneumonia, one of the most severe diseases
associated with ARI.
This paper reports the findings from Phase One of a two-phased anthropological study designed to obtain community baseline data on acute respiratory infections (ARI) in Lesotho. The primary purpose of Phase One was to determine if there was a specific comprehension of ARI among village health workers and among caretakers5 of children under five years of age in Lesotho. The assessment of rapid anthropological methods in the collection of social and cultural data on ARI was the secondary purpose of Phase One. The effectiveness of rapid anthropological methods6 in reducing the amount of time required to collect sociocultural data in primary health care and nutrition (Scrimshaw and Hurtado 1987) and diarrhoeal diseases (Bentley, Pelto, Straus, et al. 1988; Yoder 1988) has been noted in the literature. Although the use of rapid anthropological methods in eliciting ARI data at the time of this study was innovative, it seemed appropriate to the needs of Lesotho's Ministry of Health. Finally, data from Phase One were to be used to design a culturally appropriate survey instrument for a second study - a quantitative, national survey that would validate and qualify the Phase One data. This first study was conducted February 10 - March 5. 1989.
Methods have recently been developed for the rapid collection of cultural data for use in policy and programme development in developing countries (Scrimshaw and Hurtado 1987; Bentley, et al. 1988; Yoder 1988). When using rapid anthropological methods, a number of techniques (individual interviews, group interviews, focus group interviews, observation, and participant observation) are used. The methods developed by Scrimshaw and Hurtado (1987) and Herman and Bentley (personal communication, 1988) demand that data be collected and analyzed concurrently. This allows for the development of important questions, a culturally appropriate survey instrument or health education approach, and the inclusion of key local terms in a translated instrument. In this study, we modified the research techniques used by Herman and Scrimshaw and Hurtado and designed a study that was successfully implemented in Lesotho.
Site selection and sample size
The Ministry selected two study sites, Bethany and Matukeng. Criteria used to select the sites were accessibility and the presence of an active village health worker programme.
Originally, we envisaged interviewing a total of 20-24 caretakers of small children (about 10-12 per health catchment area) and two groups of village health workers (5-6 per village).
The research instrument
The preliminary interview guide, developed through discussions with staff from the Ministry of Health and the Centers for Disease Control, used cultural and biomedical knowledge to structure questions that would elicit a clear understanding of ARI in Lesotho. After two days of pretesting this interview guide in the field setting, the research team revised and developed a final instrument (Appendix 1) which elicited five categories of information from respondents:
1. illness terms related to general childhood illnesses (Question 1);
2. categories of severity for elicited illness terms (Question 2);
3. illness terms associated with breathing difficulty (Question 3);
4. narratives of ARI-related illnesses that explain how Basotho define, diagnose, and respond to ARI episodes in young children (Questions 49, 12); and
5. suggestions for appropriate health education interventions (Questions 10-11).
All of the questions in the interview guide were designed to elicit qualitative data. The emphasis was placed on the quality of the data, and a good interview provided understanding and depth in at least one dimension of ARI beliefs and behaviours. As such, it was more important for researchers to obtain good narratives from respondents about an ARI-related event in young children than to receive a brief answer to every question in the interview guide. The ability to obtain a good narrative during an interview was the core criterion for selecting the interviewers.
A research group was assembled consisting of the authors and two interviewers. The interviewers were selected from a list of applicants who had conducted survey research, and who could read and write English and Sesotho. The final selection criteria were the capability of developing rapport, conducting an interview with rural and urban Basotho women knowledgeable of health practices, and being able to take good notes in Sesotho. We assessed the applicants' skills during a series of mock interviews in English and Sesotho.
The research group worked in two teams - interviewing respondents, making observations, tape recording the interviews, translating and coding the data, and analyzing and discussing the findings. When possible, interviewers were accompanied by a village health worker assigned to the selected village who assisted by helping us gain entrance to the communities (through meeting with the village chiefs) and obtain permission to conduct the interviews.
From households identified as including caretakers of children under the age of five, a small purposive sample was selected to represent the differing socio-economic groups in the community. As each household with potential respondents was identified (in most cases by the village chief or health worker) observations of the household surroundings (material conditions) were the basis for deciding socio-economic status. The criteria included the type of construction of the house (e.g., mud-smeared vs. cinder-block construction, tin vs. thatched roof) and the material objects near the house (automobiles, oxcars, bicycles, etc.) were used as rough estimators of socioeconomic conditions. Demographic variables such as sex and generation were also considered. The selection of small purposive samples like this is common in ethnographic research and requires that researchers who are knowledgeable of the culture make quick decisions after asking a few direct questions7. Respondents were interviewed individually or in a group.8
Because distance from a health centre is often a determinant of health behaviour, the team interviewed respondents in villages nearby and far from the health centre. In both health catchment areas, the farthest village where respondents were interviewed was a 15-minute jeep ride, or a three- to four-hour walk from the health centre.
By including people from different social strata, we were able to elicit qualitative, community-based ARI data using a small sample, and construct a culturally valid instrument for Phase Two, the quantitative study. As a precaution, we planned that the second phase would use probability sampling techniques to correct bias introduced during Phase One. In this way, our key findings from the Phase One study could be tested quantitatively (Wilson and Kimane 1990).
Analyses of the interviews were conducted in several stages: during the course of data collection in the field, immediately after the completion of all of the data collection, and after the first draft of the report had been written. This process is similar to that suggested by both Herman (E. Herman and M. Bentley, personal communication, 1988) and Scrimshaw and Hurtado (1987).
During the first stage of the analysis, the tape and text from each interview were reviewed and discussed by the team members at the end of each work day. These discussions helped the co-investigators build and revise a hypothetical explanatory model for ARI from the perspective of the respondents; in anthropological terms, this is called an emic, or respondent's, perspective. The data from each interview tested the hypothetical model, and served to refine and classify the new illness term added to the model as interviews were completed each day. The daily discussions were also instrumental in bringing a consensus to the process of making approximate English translations of illness terms from the tape-recorded and written Sesotho text.
After these discussions, interviewers expanded their field notes, resulting in a more detailed version of the interview in Sesotho (or English). The discussion sessions also assisted interviewers to learn key terms and phrases that triggered respondents' narratives of ARI episodes. Thus, each day, interviewers became more confident in obtaining meaningful data regardless of the time constraints in field conditions. By listening to the tape-recorded interviews, the team learned which Sesotho phrases or terms were likely to successfully trigger a respondent to provide good, qualitative data on ARI. Providing a detailed response to a narrative would allow the respondents to answer most of the questions on the interview guide without further prompts.
The second stage of analysis was summative. Immediately after the data collection was completed, the team met to review the interviews as a collective unit. The social scientists constructed a code book for some of the variables in the instrument, and the team reviewed each interview, and coded the interview data. Then the team hand-tallied frequencies for specific, easily coded variables, (e.g., illness terms associated with ARI; whether or not traditional healers cured children with ARI in the village). These frequency data provided an overview of the characteristics of the respondents and some general trends in the data.
third stage of the analysis, the English translated texts were
typed into WordPerfect files, coded, and analyzed. First, the
"search/find" and "move" commands in the
WordPerfect software were used to identify, and group, responses
to similar questions together in separate files. Second, the
responses to each of the segments of the interview guide were
analyzed separately; similarities and dissimilarities in
responses were noted as were the conditions under which
dissimilar responses were given. This process helped refine and
reconstruct the explanatory model of ARI provided by the
Contents - Previous - Next