Author(s): Doris Piccinin, MS, RD and Katie Lai, PharmD, CDE

Reviewer(s): Ellen Howard, MSL; Christine Wilson Owens

Date Authored: June 1, 2002

Ethiopian food
Ethiopian food. Photo by Avry (cc license).


Here at Harborview Medical Center we are faced with educating diabetics in a variety of different cultures. The majority of the patients we serve in our clinic are SE Asian or East African. How do we educate our patients effectively and encourage self-management and embrace preventive care? We are delighted to share with you the multicultural diabetes classes that we offered to the Vietnamese and Ethiopian patients.

At Harborview part of our mission includes serving the non-English speaking population. One of our strengths is that the International Medicine Clinic serves approximately 20 different ethnic communities, having the availability of an interdisciplinary team within the clinic: nutritionist, pharmacist, social worker, psychiatrist, physiologist, acupuncturist, nurses and certified cultural interpreters.

Educating about self-management includes encouraging compliance with medication use and behavior modification. Behavior modification is especially important since it is the backbone of preventive care. The caseworker cultural mediator (CCM) at our institution has several vital roles including:

  1. providing continuity of interpretation for high-risk patients and family;
  2. educating providers about cultural health beliefs and understanding;
  3. extending community outreach and resource availability to their cultural groups; and,
  4. advocating for institutional change to support and facilitate communication between health care workers and the communities they represent.

Target Populations

Two-thirds of the patients we see from the International Medicine Clinic are from Southeast Asia. We are blessed with having an excellent Interpreter Services that is able to provide interpreters for approximately 72 different languages. Common languages in Southeast Asia that we encounter are: Mien, Hmong, Vietnamese, Cambodian, Laotian, Cantonese, Cha Fo, Mandarin, Champ, Thai, Filipino, Tagolag, Korean and Hindi. Because Vietnamese is the most popular of them all, we chose to target our Vietnamese population first.

Approximately one-third of our patients are from East Africa with few patients from West Africa. Languages spoken among this population include: Somali, Sudan, Fulani, Amharic, Tigrinya, Oromiffa, Kikuju and Congolese.


CCMs were able to attend our bi-monthly planning meetings to help develop and translate culturally appropriate educational materials. The ratio of one Vietnamese or Amharic speaker to three patients helped us to effectively incorporate lecture, discussion, demonstration, games, and printed materials as learning strategies.

We used different teaching styles based on the cultural acceptance of diabetes in the different groups that we worked with. For example, since social implications of poor diabetes control carries more weight than personal empowerment in the Vietnamese group, we used the compliance-based approach for education with that group. The choice of checking feet daily versus having an amputation was an example used in this group that made an impact on avoiding negative societal viewpoints. It was important to discuss what the Vietnamese community would think of someone with a foot amputation. Avoiding amputation would then encourage patients to embrace preventive care to avoid the negative viewpoint. In Southeast Asian culture, emotions are often kept hidden or are not shared with others. Thus, to encourage sharing about obstacles to caring for diabetes, games were developed. Games were also used to assess understanding of key concepts and helped promote learner participation.

Methods in the Amharic class included utilizing an empowerment-based approach due to the group’s deep spiritual beliefs. If you tell the Ethiopian patient about checking their feet to avoid amputation, they’ll say, “If God is going to take it, He’s going to take it,” almost to the point that they have no control over the long term consequences of the disease. Luckily for us, one of the educators was our own physician assistant who is Ethiopian and speaks Amharic. She was able to advocate the message that maybe God sent them to the class to learn about checking their feet, and therefore empower them to prevent amputation. This style of delivery was well accepted and was able to prompt more discussion which is the format that was preferred by this group.

The format of the classes varied slightly in length due to our availability of bilingual health care providers. Since both our nurse and pharmacist were Vietnamese speaking, the Vietnamese class series was a total of eight hours of class (two sessions, four hours each). Two separate classes were conducted with a total of 18 patients. Since we had a limited number of Amharic-speaking providers, the Amharic diabetes class was a total of 12 hours in length (three sessions, four hours each) in which six Harborview patients and one non-Harborview patient attended.


Translating and Culturally Sensitizing Education Materials and Teaching Approach

The next phase of our project was diabetes translation. From pre-existing English language materials about hypo and hyperglycemia, two charts were translated into Vietnamese and then back into English to ensure adequacy. These materials have been used numerous times with our patients, but have not yet been formally field tested.

Similarly, the hypo and hyperglycemia charts were translated into Amharic. When developing non-English educational classes, it is important to not only focus on straight material translations but to look beyond that and remember the culture and individuality of each person.

Equally as important as materials translation, the diversity and complexity of the food our patients consume is of critical importance in the success of the multicultural diabetes classes. There were several challenges faced in identifying appropriate nutrition educational materials for the Vietnamese and Amharic communities:

  • Lack of available translated materials was one of the first challenges we faced. To our knowledge, there were no educational materials specific to diet and diabetes in the Amharic language.
  • Obtaining or producing translated materials at the appropriate literacy level was another challenge we faced. Since many of our patients are illiterate in their native languages, translated materials with pictures to assist in learning key concepts are important in the educational process. The hospital’s goal for appropriate reading level of patient educational materials is the 4th-6th grade reading level and many translated materials are at much higher reading levels.
  • The third challenge we faced was nutrition educational materials with culturally appropriate foods. How can patients follow the diabetic diet without reference to specific foods that they are consuming at home?
  • In addition to the inclusion of culturally appropriate foods, very few nutrition education materials take into account portions of foods in terms that patients can understand or that are relevant to their culture. The article by E. Harris-Davis and B. Haughton, “Model for multicultural nutrition counseling competencies,” J Am Diet Assoc 2000; 1178-1185, stresses the importance of knowledge regarding cultural eating patterns and family traditions such as core foods and traditional celebrations as important contributors for successful nutrition counseling. It also emphasizes the identification of ethnic food stores or other ethnic community resources that may influence eating patterns.
  • Rarely do nutrition education materials take into account eating practices or social factors influencing food selection as discussed in the article by Harris-Davis. This was of particular importance with the Amharic groups as they eat from the same plate and identifying educational materials that would take this into account had to be developed.

Since the English speaking patients seen at Harborview Medical Center are predominately indigent or of low literacy levels, the English diabetes classes developed by the Hospital incorporated the Idaho Plate Method for the nutrition component of the class. This method has been demonstrated to be one of the most effective tools for instructing patients on their diabetic diet. (Camdon, K. J Am Diet Assoc 1998) It is accepted by both the American Diabetes and American Dietetic Associations.

The plate method helps patients understand portions of food groups to help control their blood sugars. A 9 inch plate is divided into quarters, where one quarter consists of meats and starch and two quarters or a half consists of vegetables. Adjacent to the plate is a fruit serving and a dairy serving in a one cup portion.

Along with the plate method, the National Dairy Council’s “Seven Ways to Size Up Your Serving” is an easy tool to help visualize appropriate portion sizes. The most challenging aspect of working with nutrition management of the diabetic patient is helping patients understand portion control. This tool is a very simple way to educate patients about serving size and helping them conceptualize portions using common body parts or tools to teach portion control. For example, a fist may be used to describe a cup of cooked rice or pasta, a deck of cards related to a 3 ounce portion of meat and 4 die equal an ounce of cheese. This handout was used for both the Vietnamese and Amharic diabetes class.

The Dairy Council’s “Seven Ways to Size Up Your Servings” was not evaluated with focus groups nor was it field tested prior to our classes. In the future we will field test these materials to see if they are suitable in these cultures.

In our Vietnamese diabetes class, every time we tried to use a fist to help describe a portion of one cup of cooked rice, the patients would convert this to a “cupped hand of rice”. Perhaps when we field test the Dairy Council handout, it may be more appropriate to describe a serving size as the “cupped hand” method or the “ying and yang of portion control.” In the Asian culture, balance in life is important and since we are trying to teach balance in diet for diabetes control, the ying and yang concept may be appropriate in meeting this goal. It will be interesting to field test the material to view the validity of this statement.

It is important that the cupped hand be no wider than 1 1/2 ” or a thumb size, otherwise it does not correlate to one cup of cooked rice, an appropriate carbohydrate choice. In addition to the “cupped hand method”, to assist in understanding portion control for our Vietnamese patients, Asian food models and rice and pho soup noodle bowls were purchased. Patients were asked about quantities of commonly consumed foods. Asian bowls and food models proved to be a useful strategy to educate patients on appropriate portions for blood sugar control.

The International Diabetes Center has developed a nutrition education pamphlet both in English and Spanish. The Spanish version has been modified to incorporate foods commonly consumed by people in the Spanish culture. This is exactly the type of nutrition education material we were looking for in the Vietnamese and Amharic languages but it did not exist. The Spanish materials are low literacy with photos of culturally appropriate foods for patients that do not read their native language. Portions of food are described in ways that are related to the way they are handled, prepared, served and consumed by people in that culture.

After extensive interviewing with case worker cultural mediators, interpreters and informal interviews with patients, the “Vietnamese Healthy Bowl” was developed. Similar to the plate method, the bowl method describes portions of starch, vegetables, meats and fruits. The bowl builds on a base of starch. Rice is the most commonly eaten starch or core food in the Vietnamese diet and rather than discouraging rice consumption, we focused on portion control. Instead of instructing patients to switch to brown rice, we focused on the total amount of rice eaten with each meal. We used Asian cups and bowls or samples of Vietnamese foods (instant noodles) to help patients understand appropriate portions in measurements with which they were familiar.

Since small amounts of meat or proteins are eaten by the Vietnamese and very little oil is used as compared to Chinese cooking, we encouraged a larger meat serving to balance the excess rice consumption. We changed the name of the Dairy Group to “Calcium food sources” emphasizing the inclusion of these foods for “strong bones”. Culturally appropriate foods in this group included tofu, soy milk, dried fish, bok choy and broccoli as food sources to ensure adequate calcium intake.

Both photos for illiterate patients and Vietnamese words with the English translation were included for the clinician and family members that may understand English.

Photos taken from local Asian markets were used in addition to food models. Commonly consumed vegetables including bitter melon, bok choy, and Chinese broccoli were emphasized. As vegetables are a common part of the Vietnamese diet, we wanted to encourage vegetable consumption. With acculturation of the Vietnamese population, there is less emphasis on vegetables and we wanted to make sure patients understood the importance of maintaining their culture.

Fruits including lychee, longan and durian were discussed along with adequate portions. A photo of durian fruit was used as a quiz question to help patients understand appropriate portions of fruits (1 cup of durian per meal). Quizzes and games incorporating food from their culture were used to assess patient understanding.

Packaged noodles were used to demonstrate portion and sodium control, especially for those patients with high blood pressure.

We were able to find some supplemental educational materials in both Vietnamese and Amharic language. Although these materials were not specific to diabetes and diet, they had been field tested and did incorporate culturally appropriate foods. They were included to supplement the nutrition component of the classes.

The Vietnamese Food guide pyramid accents foods that are commonly consumed by Vietnamese patients, such as rice and noodles in the starch group. The fruit group included mango and papaya. Carrots and bok choy were added to the vegetable group. Similarly, the dairy group includes tofu or soy milk in addition to dried fish, Chinese broccoli or bok choy to meet patients calcium needs.

Sweetened condensed milk often used for desserts and coffee in the Vietnamese culture was discussed to be sure patients limit the amounts consumed in the diet to maintain blood sugar control.

Since we knew so little about the content of foods consumed by the Amharic speaking Ethiopians, our efforts were concentrated on researching the ingredients and analyzing the recipes of commonly consumed foods in this culture. The Ethiopian traditional recipes were used as a source for identifying common foods. Because there was a lack of information about many of the ingredients, patients, interpreters and cultural mediators were extensively interviewed about the foods prior to the development of the class.

The most commonly consumed starch or core food in the Amharic diet in injera. Injera is to Ethiopians what rice is to the Asian culture. It is a flat, sour tasting, bread, eaten at almost every meal. Injera is usually made from teff, the smallest grain in the world which is grown primarily in Ethiopia. Since there is a considerable difference in the quality of the teff grown in the U.S. as compared to Ethiopia, a variety of different grains need to be incorporated in the injera recipe in order to achieve the consistency and taste of the traditional bread. Patients described two different recipes used in the U.S. One included white flour, self-rising yeast and teff. The other incorporated additional grain including barley, millet, oatmeal, corn, rice flour, whole wheat flour, self-rising yeast and teff. These two recipes were analyzed to determine starch content.

If the recipe for injera used more grains, a larger portion (3/4 piece) could be consumed with meals. Typically a piece of injera looks like a giant pita bread, about 12″ – 16″ in diameter. Patients were instructed that ideally, injera made primarily from teff would be the best choice for helping to control blood sugar. The role of fiber and its effect on blood sugar were discussed. Patients were instructed to use more grains to match more closely the fiber content of the traditional injera made from Ethiopian teff.

Patients who were purchasing their injera from local ethnic food stores rather than making their own, were instructed to have a smaller portion of injera (1/2 piece per meal). Variations depending upon gender, age, height and weight, were discussed, with an emphasis on the importance of developing individual meal plans with the assistance of a registered dietitian.

The important piece in educating the Amharic diabetic patient was helping establish portion control when patients ate from a shared plate. In Ethiopian culture, people eating together share food from a plate. Injera, rather than a fork, is used as a utensil to scoop out the food.

In addition to injera as a common starch source, lentils and split peas are commonly consumed during fasting periods. Many Ethiopians are Orthodox Christians. Fasting prior to Christmas and Easter involves abstention from all meat products, poultry, fish and dairy products. Our Amharic physician assistant was instrumental in educating our staff and patients about the importance of diet control for diabetes management during fasting. It was emphasized that during fasting more attention needs to be paid to appropriate portion sizes of injera, lentils and split peas.

Using measuring utensils when cooking is a foreign concept in the Amharic culture as it is in many cultures. As stated earlier, the Dairy Council’s “Seven Ways to Size Up Your Servings” was used to help educate patients on portion control. Since we were unable to run focus groups or field test our diabetes education materials, we do not know if the Dairy Council pamphlet was an effective tool.

In future focus group efforts it will be interesting to see if adapting cups and ladles that are common measuring utensils in Ethiopia would be more effective than the Dairy Council illustrations in demonstrating portion control.

Unlike the Vietnamese culture, the Amharic culture cooks with a lot of oil. Oil and ghee (clarified butter) are used in making meals and desserts. Large amounts of oil are used during celebrations or when honoring special guests. During observation of cooking a lentil dish, 24 ounces of oil were used. T. Kassm-Khamis, et al (J Hum Nutr Diet 3/2000) observed persons from South Asian and East African cultures pouring oil into soups and stews while they were cooking. This was also noted when observing the Amharic cooking practices. The importance of portion control of oils to prevent cardiovascular disease, a common complication of diabetes, was discussed in the class. Two tablespoons of oil were poured into a cup to demonstrate total fat requirements for the day.

In addition to the use of photos taken at local Ethiopian restaurants, photos of cuts of meat were used to help educate patients about fat and cholesterol control. The Amharic speaking people primarily consume beef, lamb and goat. Generally chicken is reserved for holiday meals. As patients often butcher their own animals, they are familiar with cuts of meat and their anatomical origins. Generally meat from the lower back and leg is the most exercised muscles and contains the least fat. Patients were instructed to consume these cuts more often in order to meet their dietary fat goals for cardiovascular health.

The Amharic diabetes plate was modified to reflect foods from the culture including lentils, beans, potatoes, pasta and injera for the starch serving. Pasta, the second most commonly consumed starch in the Amharic diet is a consequence of the influence of the Italian colonization of Ethiopia. Tropical fruits such as bananas, mangoes and oranges were included in the fruit group. Commonly consumed vegetables included tomatoes, carrots, lettuce, cabbage and collard greens. Goat, beef, lamb, poultry and fish were listed in the meat group as discussed earlier. Since milk and cheese are commonly consumed in this culture, no modifications for the dairy group were included except for the emphasis on low fat and non-fat dairy alternatives.

No photos are available from clip art files for Amharic foods. With limited funding we were unable to provide drawings or photos of food for this pamphlet. The Amharic food guide pyramid, although not specific to diet and diabetes, was used to help patients understand which food were included in the different food groups.

Some of the important areas that we were unable to focus on in these classes included food selection patterns, processing of foods, and how food selection and processing would affect our patients’ blood sugar control.

Measuring Outcomes

For the Vietnamese diabetes classes, the average age of the patients was 65 years old and 87% felt that the diabetic educators listened to their concerns. Ninety-three percent of the patients would recommend the class to their friends which demonstrates excellent satisfaction with the classes. Eighty-three percent of patients were checking feet daily, which shows that the class was able to promote behavior change. For the education assessment 80-90% of patients were able to appropriately answer questions on hypo/hyperglycemia management and goals for blood sugar testing, but only 73% were able to understand sick day guidelines for diabetes care, an area that could use improvement in teaching.

The hemoglobin A1c (HbA1C) is a measure of diabetes control that is used as a standard by the American Diabetes Association. It is a blood test that measures average blood sugars over a period of 4-6 weeks. The goal for tight diabetes control is less than 7.0%.

The pre and post HbA1C before and after the class as shown in this slide, shows that there were no changes for patients with HbA1C <8.5%, but there were improvements for poorly controlled diabetics.

There were minimal changes in overall HbA1C before and after the class, which may be related to a more elderly patient population. With an elderly patient population, physicians tend to have their diabetics maintained at an overall higher HbA1C to prevent hypoglycemic or low blood sugar reactions that are more common with the elderly. Outliers may also explain the data. A relatively small number of patients may have skewed the results. In addition, at least two patients were either hospitalized or went to a nursing home after the class which may explain poor HbA1C outcomes. Another reason for poor results with the Vietnamese class may be due to poor return to classes, with only 56% of the Vietnamese patients returning to both sessions.

Post class results of HbA1C for the Amahric group were much more favorable. This may be due to better attendance of the classes with over 70% of patients attending all three sessions. With the Amharic group, we were unable to see any changes in blood sugar control for patients with HbA1c <7.5%.

The change in HbA1c for the Amharic group before and after the diabetes classes showed excellent results with an overall 2% drop. However, these results only included 6 patients and this is not statistically significant.

The average age of the Amharic diabetic patients was much younger than the Vietnamese patients; the average age was 47 years. We were very pleased to see the results of the patient satisfaction surveys for the Amharic class with 80-100% of the patients satisfied with the class. Eighty percent of patients were checking their feet daily while 80-100% of patients were able to appropriately verbalize hypo/hyperglycemia management, blood sugar goals and sick day guidelines.

There were several factors that affected our outcomes. First of all, the number of years patients lived in the U.S. may affect their health and attitude towards preventive health. We unfortunately did not have access to this information prior to the classes. Secondly, as stated earlier, our results for the Amharic group may be better than the Vietnamese group due to a younger patient population as opposed to the Vietnamese group. The younger patients may be more motivated to change behavior as compared to an elderly patient population. Also, more participation in the classes demonstrated by greater attendance and the format of the class for the Amharic group may have produced more favorable results. The Amharic group was much more involved in active discussion and managing their disease by the use of the empowerment approach to diabetes self-management. All these factors may have contributed to an overall greater success in the Amharic versus the Vietnamese diabetes class results.


Upon reflection of the class, there were several things we would do differently if we were starting the process again. A pre-assessment prior to the patients attending the class would be important to obtain information on patient’s prior understanding of diabetes would be an important piece in measuring outcomes. The use of focus groups would be crucial to assist in planning the class and gearing information towards the needs of the patients. Support groups as a follow-up to the diabetes class could assist patients in forming peer group support and maintaining long-term diabetes management. Unfortunately, we did not have funding to perform focus groups or for support groups.

The things that we thought went well with our classes were good patient satisfaction surveys and excellent patient attendance, especially with the Amharic group which shows the importance patients place on a voluntary group education class to manage their disease. And finally, the behavior outcomes were good in both groups demonstrating motivation of patients to self-control their diabetes which is ultimately the key to diabetes care.

We faced several challenges along the way. We discussed the lack of available translated materials and certainly the cultural appropriateness of the handouts that were available. Without focus groups or field testing, we do not know how applicable our translated language materials were. Also, there is the challenge of the quality of the message we were attempting to deliver with the use of certified interpreters in which the additive complexity is unknown. No studies indicated the optimal number of interpreters to patient ratio to get the message across. There was limited published data to draw upon regarding the educational approach to use for the different cultures in diabetes care.

In the future, we propose to develop Somali, Cambodian, Tigrinya and Spanish diabetes classes. During the development of these classes we will include a pre-assessment and will add exclusion criteria to obtain improved outcomes. For example, we will not offer the class to unstable patients or try to include patients with broader ranges of HbA1C control. Putting our information on EthnoMed will help to communicate with health care providers about cultural influences affecting diabetes and other health conditions, and to share culturally appropriate translated materials.

In the future, we will attempt to include information on number of years patients have resided in the U.S., age, gender, family and/or caretaker support and the use of herbals in treating diabetes and how these factors affect outcomes. We are presently looking into funding sources for conducting focus groups for validating our materials and to further develop classes.

Finally, in closing and summary we would like to say, that as we work through the process of developing patient education classes and materials, we need to consider the diversity of the culture and their differing cultural health beliefs. We also need to keep in mind the foods from the culture and the individuality of each person rather than just focusing on translating materials.


We would like to acknowledge a number of people for making these classes successful: Our CCMs from the Vietnamese and Amharic culture whose input in planning these classes was invaluable; our certified interpreters who assisted with translation and throughout the class; our medical director, Dr. Carey Jackson, who provided ongoing support and advice in ensuring that we were promoting cultural acceptance of the classes; our clinic support staff who assisted with scheduling and advertising the classes to our patients; and, the EthnoMed website which hosts our translated materials.