Contents - Previous - Next

The domain of women's illnesses

Unit 4. Free Listing of Women's Health Problems in the Community

Purpose

1. To determine the items that make up the domain of "women's health problems." Biomedically defined illnesses are likely to comprise only part of this domain. Other items that may be included are: weakness, various aches and pains, and so forth. Identifying the full range of health problems is important, because it will later allow a picture of the relative importance women attribute to illnesses that can be addressed by each type of healer in the area.

2. To document locally used terminology about women's health problems, or how women talk about their health problems. As mentioned earlier, these terms will be the basis for further work, including more structured interviews using the "working vocabulary" of indigenous terms that are well understood by the local people. Also, the exploration of the systematic relationships among terms, the beliefs related to causes and solutions, and illness episode narratives provide the basis for construction of the local ethnomedical model(s) of women's health problems.

Data Collection

1. Prepare Task Materials.
Translate the following questions by using the appropriate local terms:

A. [Main Question] What are all the different illnesses that women experience in (name of community/area?)
_______________________________________
_______________________________________
_______________________________________
B. [Probing Question] Are there other kinds of (illness name)?
_______________________________________
_______________________________________
_______________________________________
C. [Secondary Question] What signs and symptoms go with (illness name)?
_______________________________________
_______________________________________
_______________________________________

Practice using these questions with two to three key informants and try to find out how people speak about illness. Is it best to ask about "health problems," "illness," or "sickness"? Or is it best to ask "...that women get?," "...that women suffer from?," "...that befall women?," or another local expression? Refine the questions as necessary.

2. Select Respondents: Generate a free list of illnesses (and symptoms) with 12 to 15 respondents. Select respondents who represent different sections of the population. These women should be drawn from the client population and should represent a range of ages, economic status, ethnic background, and/or other factors considered important. Respondents should be relatively knowledgeable and willing to participate.

3. Suggested Wording: There are several ways to begin a free listing of terms with informants. The interviewer may want to inquire first about the general term for "illness," following up with questions about specific types of women's illnesses. Alternatively, it may be useful to start with a general symptom term, such as "vaginal discharge," and ask about "kinds of vaginal discharge," after which the names of other signs and symptoms can be elicited. An interview that begins with the informant narrating an illness event (see below) will provide words for "illnesses," "signs," and "symptoms," as well as examples of each. Once these initial terms are elicited, a more systematic process of "free listing" can be carried our.

In most cultural groups the illnesses, signs, and symptoms do not constitute distinctive separate "lists." For example, in ordinary English language usage, "a headache" may be considered an illness under some circumstances. Yet a headache is also often defined as a symptom. The key informants and interviewees will undoubtedly offer a mixture of signs, symptoms, and illnesses.

Try to compile a moderately extensive list of illnesses/symptoms from each informant. After recording the list from the first interview(s), subsequent informants can be asked to add to the list. The interviewer might say, for example: "A few of the people in the village have told me the names of some women's illnesses. I have safed paani, khamjori, sir ghumyo, and pairjari. Could you add any names of illnesses to this list? Can you think of any other women's illnesses that I need to add to this list?"

Look for special words, or parts of words, that occur in several different illnesses. For example, in the previous paragraph there are the illnesses safed paani and khamjoori. Safed paani corresponds roughly to the English term "white water," and khamjoori to "weakness." The informant can be asked "Are there any other kinds of safed paani?" A similar question might be, "Can you tell me - list for me - all the different kinds of khamjoori (weakness) ?"

The next step is to ask for the kinds of symptoms that are associated with each of the illnesses. As noted above, some terms will be used as both signs/ symptoms and to designate an illness.

As mentioned previously, when doing key informant interviews, it is important to be flexible. For example, an informant may spontaneously go on to give both causes and treatment in connection with a single term, once she or he understands the type of information being sought. Remember this is a "free listing" process. Eventually it will be necessary to match up signs, symptoms, terms, causes, and treatment to create a systematic description, or "explanatory model," but that comes later in the process.

If the key informants are not accustomed to listing things, the interviewer may have to probe further. For example, "You said that X and Y are ways to treat safed paani. What are the other ways to treat this problem?" (As a general principle of good eliciting style, do not ask "Are there other ways to treat X?" but rather, "What are the other ways to treat X?")

It is important to carefully record the answers to the free listing results. If possible, these interviews should also be tape recorded, so that the written notes can be checked for correct pronunciation and transcriptions of terms. An example of a data collection form follows this explanation, and blank forms can be found in Appendix D for replication and modification.

In the initial free listing exercise, record as extensive a list as possible in order to cover the full domain of terms and concepts that pertain to women's illnesses. Within this broader domain, it is critical to identify the list of illnesses suffered by most women. Do not prematurely "censor" or restrict the list until a round of interviews with key informants is completed. It may be that the category "women's illnesses" is not sharply defined in the culture. Perhaps the types of illnesses are divided in some other manner. In addition, a number of terms will be mentioned by all the informants, while others will be volunteered less frequently.

4. Data Recording: Use Form 4.1 to record free lists of women's health problems.

Data Analysis

Microcomputer Application (optional) These data can be analyzed using the Free List feature in ANTHROPAC

Forms 4.2 and 4.3 provide examples of data analysis sheets that can be constructed to assist in Data Analysis. Blank forms can be found in Appendix D.

1. Tabulate women's health problems for the entire sample using Form 4.2. Usually this will be a mixture of illness terms, signs, and symptoms.

2. For each item on the list, count the number of informants who mentioned the term.

3. Arrange the items in order, based on the frequency of response from most frequent to least frequently mentioned.

4. Prepare a Signs and Symptoms Tabulation Form (Form 4.3) for each of the most commonly mentioned health problems, those mentioned by 50% or more of informants. A manifestation is what women respondents say happens when they suffer a particular type of illness.

DATA COLLECTION FORM 4.1
Free Listing of Women's Health Problems (sample)

A. What are the illnesses that women suffer from in (name of community/area)?

Interpretation Suggestions

1. Examine the distribution of responses. Items that are mentioned by all of the informants are probably very common, and/or centrally important to that domain. Items that are mentioned by only one or two people are likely to be less widely known and used in the community. Use the 10 to 15 most common words in the pretest of the pile sorting task (Unit 5). This list may need to be reduced further for the informant interviews.

2. Illness/symptoms mentioned earlier are likely to be more "culturally cognizant." Does this observation make sense to those who know the community well?

3. Examine Form 4.2, which has the full list of terms and the number of respondents who volunteered each term. Before setting these results aside, reexamine the list of terms that were eliminated, and check to see which informants offered them. Although the numbers for some terms might be very small, they may offer clues for later follow up. For example, it may be the case that some terms are known only by older women or some only by the more well-educated. Even if these terms are not used in the pile sorting exercises, it would be valuable to examine them further during the course of the study. It may be important for the community health practitioners to know more about them for communicating with families of patients.

EXAMPLE
Sample list of 14 women's health problems from data collected in India (Baroda Citizen's Council, 1990)

DATA ANALYSIS FORM 4.2
Free Listing Tabulation of Women's Illnesses

DATA ANALYSIS FORM 4.3
Free Listing Tabulation of Signs and Symptoms of a Women's Illness (sample)

Signs and Symptoms Manifestation

General Class: Khamjoori

Unit 5. Free Pile Sort: Grouping Women's Illnesses

Purpose

This unit collects systematic data on the perceived inter-relationships among different illness terms in relation to perceived similarities. It also permits assessment of the degree to which individual signs and symptoms of women's health problems are perceived to cluster together. Another outcome of the procedure is an assessment of the extent to which women in the community share a common classification system for women's illnesses.

Data Collection

1. Select a Sub-sample of Key Illnesses/Symptom Terms. Select the 10 to 15 most commonly recognized women's health problems based on the free listing exercise in Unit 4. These problems may include symptoms as well as illnesses (combination of tabulations in Forms 4.2 and 4.3). The terms selected should include the most important or representative signs and symptoms elicited from key informants, such as "cough," "weakness," and "fever." In general, select the most frequently mentioned items. A good rule of thumb is to select items mentioned by 40 percent or more of respondents. On the other hand, be sure to include terms that reflect reproductive problems and symptoms even if community people do not perceive them to be related to women and even if they are not so commonly mentioned.

The analysis from the free listing may result in a fairly long list of terms collected from all informants. Many of these will represent minor variations in general classes of symptoms. In order to select 10 to 15 items from the larger list, it may be necessary to question selected key informants further. Ask them to identify items that are practically synonymous; compare their answers and make adjustments as necessary.

2. Construct the Pile Sort Cards.

Make a card for each illness, writing the name of the illness on one side and a letter or number code on the other. Do not put more than one term (i.e., illness, sign, or symptom name) on a card. The purpose of the code is to facilitate rapid recording of the responses during the interview. For sample card, see below.

Figure

Of course, illiterate informants will not be able to use such cards. With illiterate informants pictorial representations of the illnesses/symptom terms may be used, when possible. Such pictures may need to be drawn based on suggestions from key informants. The total number of items will usually need to be fewer when pictures are used, because pictures often prove harder for respondents to remember and keep track of. Pile sort cards should be carefully tested on other key informants for clarity.

3. Select Respondents for the Pile Sort. Use approximately 12 to 15 of the same or similar respondents as those used for Unit 4.

4. Suggested Wording and Recording Procedure: Pile sorting and card sorting tasks are often more enjoyable for respondents if it is suggested to them that the exercise is "like a game." Depending on the local culture and language, referring to this part of the interview as a "kind of game" can be explored.

Suggested wording to introduce the task:

a. "In order to help me understand more about women's illnesses in this community, I'd like to ask you to tell me which illnesses go with each other."

b. "I've made a list of some illnesses that can happen to women. Here they are." [Show the cards and say the names aloud. If using pictures, make sure the respondent understands what each card stands for.]

c. "Please sort these illnesses into piles, in whatever way you think is best, in as many piles you wish to see. There is no right way or wrong way to do this." Record these data in the Cards columns of Recording Form 5.1. The numbers on the back of each card should be written down, where each row on FORM 5.1 includes all the cards placed in a single pile.

d. Once the data have been recorded, pick up the cards in the first pile and ask, "Now please tell me why you placed (first card in first pile), (second card in first pile), and (third card in first pile, etc.) together in the same pile? Perhaps you could give this group a name?" Record these data in the Explanation column of Recording Form 5.1. Repeat this process with each pile.

e. Be sure to shuffle the illness cards between visits to respondents so that they do not all carry out the task in the same order.

Data Analysis

Microcomputer Application (optional)

Pile Sort data can be analyzed using the ANTHROPAC computer program to construct a Multi-Dimensional Scale (MDS) (See Appendix B-1). This procedure is highly recommended to help make sense of the data.

1. The results of the pile sort should be tabulated using a proximities matrix (Form 5.2). If two items are in the same pile, place a 'hatch' mark (|) in the appropriate intersection box in the proximities matrix. For example, from pile # 1 on the previous page, you would place marks at intersections of 1&2, 1&4, 1&6, 2&4, 2&6, and 4&6.

2. Continue this process for all respondents' data, tabulating on the same aggregate proximities matrix form.

3. The data may be further refined by dividing the number of hatch marks in each matrix square by the total number of informants and multiplying the result by 100. This gives the percent of informants that grouped the two illnesses together.

4. Based on the proximities matrix, prepare a tabulation of the "most close" (higher numbers) illnesses and "most distant" (lower numbers) illnesses (Form 5.3). This will be based on the percentage of informants who put each pair of illnesses together.

Interpretation Suggestions

1. Illnesses that are grouped together (have higher percentages) by more informants have some kind of "connection." Those that are not grouped together (have lower percentages) are conceptually different.

2. Examine qualitative responses to determine this connection. Examples of common explanations might be:

a. Illness X leads to illness Y (progression).

b. Illness X and illness Y occur at the same time (one is a sign/symptom of the other).

c. Illness X and illness Y have the same cause (causal linkages).

d. Illness X and illness Y should be treated the same way.

3. Are some illnesses/signs/symptoms close to each other that you would not expect from a biomedical model of illnesses? What local explanations are provided for these relationships?

Continue

Contents - Previous - Next