Relationship Between Past Food Deprivation and Current Dietary Practices and Weight Status Among Cambodian Refugee Women in Lowell, MA

Abstract

Objectives. We investigated Cambodian refugee women's past food experiences and the relationship between those experiences and current food beliefs, dietary practices, and weight status.

Methods. Focus group participants (n = 11) described past food experiences and current health-related food beliefs and behaviors. We randomly selected survey participants (n = 133) from a comprehensive list of Cambodian households in Lowell, Massachusetts. We collected height, weight, 24-hour dietary recall, food beliefs, past food experience, and demographic information. We constructed a measure of past food deprivation from focus group and survey responses. We analyzed data with multivariate logistic and linear regression models.

Results. Participants experienced severe past food deprivation and insecurity. Those with higher past food-deprivation scores were more likely to currently report eating meat with fat (odds ratio [OR] = 1.14 for every point increase on the 9-to-27–point food-deprivation measure), and to be overweight or obese by Centers for Disease Control and Prevention (OR = 1.28) and World Health Organization (OR = 1.18) standards.

Conclusions. Refugees who experienced extensive food deprivation or insecurity may be more likely to engage in unhealthful eating practices and to be overweight or obese than are those who experienced less-extreme food deprivation or insecurity.

Since 2000, almost 500 000 refugees have resettled in the United States, with tens of thousands arriving annually.¹ In addition to their high rates of mental health disease resulting from the turmoil they are fleeing,^2–4 refugees have higher rates of heart disease, hypertension, and diabetes than do other immigrant groups and native-born Americans.^2,3,5,6 The high rates of chronic disease are likely related to multiple factors. Refugees may have suffered physiological damage during stress and war,⁷ and traumatic stress may have increased their risk of cardiovascular disease and stroke.⁸

The increased rates of chronic disease may also be related to changes in food consumption. In a postconflict environment with plentiful food, people may adopt harmful eating behaviors that affect health both directly and through increased weight.^9–14 World War II prisoners of war who experienced highest trauma and food deprivation also reported the highest rate of binge-eating behaviors 50 years after the war.¹⁵ Holocaust survivors reported lifelong binge eating and preoccupation with food, including worrying about food availability and hoarding.¹⁶

Uneven access to food is associated with higher rates of overweight and obesity and weight gain in the United States,^9–13 possibly because it may lead to excessive consumption of food in times of plenty.^9,11,13,14 Refugees who experienced food deprivation or insecurity and who currently have abundant access to food may approach food in ways that increase risk for overweight and obesity. African refugees reported eating high-status foods, such as meat and steak, more often in the United States than in their native countries.¹⁷ Hmong refugees indicated that they purchased and ate food they knew to be unhealthful because it was very affordable in the United States.¹⁸ Studies of Vietnamese, Hmong, and Cambodian refugees reported high preference for steak.^19–21 Although food security has been well-defined,²² to our knowledge, there is no existing quantitative measure of variation in the past food deprivation experiences of refugees.

Cambodian refugees stand out as a potential refugee model for examining how past experiences of food deprivation or food insecurity affect current food beliefs, dietary practices, and weight. Cambodian refugees survived high levels of trauma and food deprivation in their home countries,³ and both trauma²³ and food deprivation or insecurity are experienced by most refugees.²⁴ Cambodian refugees also have disproportionately high rates of chronic disease,⁵ as do other refugee groups.^3,6,25

Our research sheds some light on the food experiences of Cambodian refugees from 1975 through arrival in the United States (1980s through mid-1990s), develops and validates a measure of past food deprivation to allow measurement of potential effects on current dietary practices, and tests for relationships between severity of past food deprivation and current food beliefs. We also discuss implications for refugee communities.

METHODS

From 1999 to 2007, Lowell, Massachusetts, was the site of Cambodian Community Health 2010 (CCH 2010), a Centers for Disease Control and Prevention (CDC) program²² led by the Lowell Community Health Center. In the final year of CCH 2010, the Cambodian Mutual Assistance Association of Greater Lowell Inc, a CCH 2010 partner, conducted a study of Cambodian refugee women. All procedures were approved by the institutional review board of the University of Massachusetts, Lowell, to which the Tufts University institutional review board deferred. Informed consent was obtained from all participants.

Focus Groups

Eleven women, recruited through word of mouth, participated in 2 focus groups. Discussion focused on their food experiences during the time of the Khmer Rouge, during the Vietnamese invasion and occupation of Cambodia, and during their travel to and time in Thai refugee camps, and included current perceptions about food and medical advice on food and health. A bilingual mental health liaison from the local Cambodian-focused Metta Health Center, a site of the Lowell Community Health Center, was present to provide referrals for participants who had difficulty with memories discussed in focus groups. Groups were conducted with an English-speaking moderator, a bilingual translator, and 2 bilingual observers providing clarification and correction. Discussions were audiotaped and transcribed verbatim into English and reviewed for accuracy.

Survey

We developed a preliminary survey based on community needs, a previous CCH 2010 survey, and published literature of similar issues in different populations. Modifications were based on focus groups and recommendations from the CCH 2010 team. A semifinal version was translated into Cambodian and backtranslated into English, with incongruities reconciled, then tested with survey administrators. The final version was tested on 10 community members, and no changes resulted. Survey sections relevant to this paper were measured height and weight, 24-hour dietary recall, demographic information, acculturation, and food experiences before coming to the United States.

The sampling frame consisted of Cambodian names identified from the 2007 Lowell City Census and the 2007 Verizon telephone book for Lowell; a list of registered Cambodian American voters maintained by the University of Massachusetts, Lowell; and lists of clients from community agencies. Households were randomly selected through Microsoft Excel's (Microsoft Office 2003, Microsoft, Seattle, WA) random number generator.

Selected households were contacted via letter in English and Cambodian, then in person. Surveys were collected from women aged 35 to 60 years to include a full generation who were likely to remember the Khmer Rouge and also likely starting to face food-related health issues. Eligibility was determined in person. Surveys were completed by a total of 160 women. The overall response rate was 81.6% of those contacted and eligible.

Survey data were double entered into Microsoft Access (Microsoft Office 2003, Microsoft, Seattle, WA). Dietary recalls were entered into FoodWorks (version 11.0, The Nutrition Company, Long Valley, NJ) for nutrient analysis. Data were analyzed with SPSS, version 16.0 (SPSS Inc, Chicago, IL, 2008).

Measure of Food Deprivation

We developed a measure of food deprivation, which we included in the survey to take into account 3 periods of social disruption: (1) time of the Khmer Rouge, (2) the Vietnamese invasion and occupation of Cambodia, and (3) time in Thai refugee camps. Only respondents who reported remembering all 3 periods (n = 133) were included in our analyses.

We asked a set of 3 identical questions for each of the periods about quantity and quality of food and disrupted eating patterns, which have previously been described as components of hunger²⁶ and food insecurity.^27,28 Question response choices were designed to apply to both episodic and chronic hunger.²⁶

Additional questions were asked on strategies to avoid starvation and signs of malnutrition as described in the focus groups, and acute food deprivation (Table 1). Strategies to avoid starvation have been shown to be related to hunger,²⁶ and signs of malnutrition are related to extreme, prolonged lack of food.²⁹

TABLE 1.

Food Deprivation Measure Components for Survey Based on Focus Group Descriptions of Experiences in Cambodia and Thailand, 1975 through the 1980s: Cambodian Refugee Women, Lowell, MA, 2007–2008

Component	Measurement Question	Score Assignment for Food Deprivation Measure
Quantity	Did you have enough food to eat often, sometimes, or never?	Often = 1
		Sometimes = 2
		Never = 3
Quality	Did you have the kinds of food you wanted to eat often, sometimes, or never?	Often = 1
		Sometimes = 2
		Never = 3
Disrupted eating patterns	How many meals did you usually eat each day?	≥3 = 1
		2.0–2.9 = 2
		0–1.9 = 3
Acute food deprivation	Did you ever go without food for at least a whole day because there was not enough food?	Not included in measure
Strategies to avoid starvation	Please tell me if you did any of the following things when there wasn't enough food to eat:	Not included in measure
	▪ Ate foods that you usually would not eat, like animal skins, or tadpoles, or water crabs
	▪ Ate things that are not food, like grass or tree roots
	▪ Secretly took food from the Khmer Rouge
	▪ Traded for food
Signs of malnutrition	I am going to read you a list. Please tell me if you experienced any of these things before coming to the United States:	Not included in measure
	▪ Weight loss
	▪ Hair loss
	▪ Swollen belly
	▪ Puffy arms or legs

We created a food-deprivation measure by adding the values for reported experiences by period (Table 1). Period 1 was the time of the Khmer Rouge, and for the typical respondents, period 2 was the Vietnamese occupation and period 3 was the time spent in Thai refugee camps. The periods of the Vietnamese occupation and the time spent in Thai refugee camps overlapped and were not experienced uniformly by all respondents but occurred during roughly the same time (1979 to mid- to late- 1980s). Some respondents left Cambodia immediately after the Vietnamese invasion and never experienced occupation; their experiences reported during their stays in Thai refugee camps were attributed to both periods 2 and 3. Other respondents never went to refugee camps and remained in Cambodia throughout the Vietnamese occupation; their experiences reported during the occupation were attributed to both periods 2 and 3. Each of the 3 periods had a total possible score range of 3 to 9; the final food deprivation measure had a range of 9 to 27. We used correlations between the food-deprivation measure and strategies to avoid starvation and signs of malnutrition to validate the food-deprivation measure.

Variables

Outcome variables measuring food beliefs were based on whether the respondent correctly identified eating a lot of meat with fat as bad for one's health, eating fast food as bad for one's health, and eating a lot of whole grains as good for one's health.

Outcome variables measuring dietary behaviors from the 24-hour dietary recall were based on whether the respondent had eaten meat classified as high fat, the number of times meat was eaten, whether whole grains were eaten, and whether intake met the dietary guidelines for percentage of kilocalories (kcal) from fat (20%–35% of kcal) and saturated fat (< 10% of kcal).³⁰ Because of small sample sizes, the statistical calculations were not sufficiently powered to detect even major effects for percentage of kilocalories from fat and saturated fat and are not included in tables.

Body mass index (BMI; weight in kilograms divided by height in meters squared) was calculated from measured weight (obtained with a Model 68978 Thinner digital scale, Conair, Stanford, CT) and height (obtained with a Seca 214 portable stadiometer; Seca, Hanover, MD). Weight categories were determined from BMI by using World Health Organization (WHO) and CDC standards. The CDC standards are based on increased US population risk for diseases including coronary heart disease, type 2 diabetes, some cancers, hypertension, and stroke at BMIs in the overweight and obese categories.³¹ Because Asians have increased risk of these diseases at lower BMIs, WHO recommends lower cutoffs for overweight and obesity for Asians³²; however, the CDC does not. The weight categories were: (1) overweight or obese (WHO; BMI ≥ 23 kg/m²), (2) overweight or obese (CDC; BMI ≥ 25 kg/m²); obese (WHO; BMI ≥ 25 kg/m²); and (3) obese (CDC; BMI ≥ 30 kg/m²).

Statistical Models

We used bivariate and multivariate binary logistic and linear regression models to test the hypotheses that those who experienced more severe food deprivation would be (1) less likely to correctly identify food–health relationships, (2) more likely to report less healthful eating practices, and (3) more likely to be overweight or obese by CDC and WHO standards. Multivariate models were controlled for age, education (≤ 1 year education, some grade school or high school, high-school graduate), and acculturation (1-to-5–point Psychological Acculturation Scale developed and validated by Tropp et al.³³ and previously validated in this population³⁴).

RESULTS

Ten of the 11 focus group participants reported experiencing severe food restrictions during the time of the Khmer Rouge. Eating conditions were described as thin rice soup (mostly water with just a handful of rice) with some Asian watercress, allowed only 2 times per day, and in a group setting with very limited portions. Participants described being forbidden to grow their own food while being forced to work in rice fields and to harvest rice that was taken by the Khmer Rouge. Many family members and friends died from starvation and illness. Participants described finding covert ways to survive: eating nonpreferred foods (tadpoles, small fish, crickets, and frogs), eating nonfoods (banana tree roots, grass, and tubers), and taking food secretly from the Khmer Rouge, which was extremely risky because those who were caught were killed. Conditions worsened over time.

Participants noted a marked improvement in food availability when the Vietnamese army invaded and occupied Cambodia in stages through 1989. The Vietnamese army let villagers grow and eat their own food, but the army sometimes stole the food. The food situation remained volatile because retreating Khmer Rouge soldiers burned rice and used civilians as moving shields against attack by the invading Vietnamese army.

Travel to the refugee camps in Thailand was marked by almost total lack of food or water. Participants described traveling for 2 to 3 days almost nonstop, foraging for food and water, encountering contaminated water, and avoiding land mines, wild animals, and hostile Vietnamese and Khmer Rouge soldiers. Participants described acute lack of provisions, constant movement, and hiding throughout the dangerous journey. Many travelers died during the flight.

The refugee camp experience, although described as much better than the experiences during the Khmer Rouge period and the Vietnamese occupation, varied by individual and camp experience. Pregnant women or young or adolescent orphans had good conditions with special accommodations. Other participants reported that food and water were always limited and inadequate but that no one in the camps was starving.

Overall, the food situation was described as very poor and deteriorating with little variation during the time of the Khmer Rouge, and as improved but variable during both the Vietnamese occupation and the time spent in Thai refugee camps. For all 3 periods, eating patterns were at least partially disrupted, with the omission of a full meal daily during the time of the Khmer Rouge and uneven access to 3 eating episodes per day for the Vietnamese occupation and the times spent in Thai refugee camps.

Physical effects of malnutrition were described as worst during the time of the Khmer Rouge, and included edema, swollen legs, weight loss or skinniness, diarrhea, hair loss, and big or swollen heads.

Current Food Beliefs

Participants said they were not worried about current access to food in the United States because they were confident of its consistent availability. They noted that people here did not appear to have food preoccupations.

Participants said they would make dietary changes for health reasons if recommended by a health care provider, but felt that many community members would not voluntarily limit their steak, grilled chicken with skin, and other cultural dishes made with high-fat meats even if they received advice to moderate intake of those foods. Participants said it was surprising, after nearly starving, to learn that some foods now have the potential to make them sick. Eating the fat from meat products had literally been a lifesaving measure, and to learn that it could now threaten health seemed unnatural to them. Participants also noted that family and friends often would not listen to advice on changing or moderating food intake. One participant expressed group sentiments: “I see a lot of people overweight, or people who die of heart attack… . I learn … [how] to take better care of myself. But I wish I could make my husband eat healthier.”

When asked how health care providers could sensitively encourage dietary change for health reasons, participants suggested focusing on healthy choices. One participant recommended: “Say: ‘There's plenty of food that you can eat. If you cannot eat apple, you can eat orange … and don't worry about starving.’” Additionally, participants suggested that health care providers should show greater understanding of the patient's years of near-starvation when discussing voluntarily limiting preferred foods such as steak.

Survey Results

Most survey respondents had less than a high-school education (82%), and 27% had 1 year of school or less. Average respondent age was 48 years, and average time in the United States was 19 years. Average acculturation score was 2 on a 1-to-5 scale, indicating high identification with Cambodian culture and low identification with American culture. For those refugees who spent time in Thai refugee camps, average length of stay was 4 years (Table 2).

TABLE 2.

Sample Characteristics, Strategies to Avoid Starvation, Signs of Malnutrition, and Acute or Severe Food Deprivation: Cambodian Refugee Women, Lowell, MA, 2007–2008

	% (No.) or Mean ±SD
Sample characteristics
Education (n = 133)
≤ 1 y	27.1 (36)
Some primary school or high school	54.9 (73)
High-school graduate	18.0 (24)
Age, y, (n = 133)	48.2 ±7.2
Time, y, spent in United States (n = 133)	18.8 ±7.5
Acculturation scorea (n = 127)	2.0 ±0.6
Time, y, spent in refugee campb (n = 96)	4.2 ±2.1
Strategies to avoid starvation (n = 133)
Traded for food	35.7 (46)
Ate inferior foods	77.5 (100)
Ate nonfoods	44.2 (57)
Secretly took food from Khmer Rouge	30.2 (39)
Employed any coping strategy	87.6 (113)
No. of coping strategies	1.9 ±1.1
Signs of malnutrition (n = 133)
Weight loss	21.6 (27)
Swollen belly	9.6 (12)
Swollen arms or legs	10.4 (13)
Reported any sign of malnutrition	29.6 (37)
No. of signs of malnutrition	0.5 ±1.0
Acute or severe food deprivation (n = 133)
Whole day without food	14.6 (19)
No. of times, whole day without food	2 ±1

^aMeasured as score on 1-to-5–point Psychological Acculturation Scale.³²

^bTime spent in refugee camp is reported only for those who spent any time in a refugee camp.

Survey responses about the 3 historical periods showed that a large majority (87.6%) of respondents reported engaging in at least 1 food coping strategy before coming to the United States, with 77% having eaten inferior foods and nearly half (44%) eating nonfoods to survive. Thirty percent reported at least 1 sign of malnutrition and 14% reported going without food for at least 1 full day (Table 2).

Food Deprivation Measure

Responses to the 3 sets of 3 survey questions in the food deprivation measure showed strong consistency with focus group experiences. The time of the Khmer Rouge was most extreme and least variable (mean score = 7.9 ±0.9; range = 3–9). The experience was milder but still severe and more variable during the Vietnamese occupation (5.5 ±1.6; range = 3–9), and mildest but still high and somewhat variable in the Thai refugee camps (4.9 ±1.3; range = 3–9; Table 3).

TABLE 3.

Food Deprivation Experience Descriptive Information for Experiences in Cambodia and Thailand, 1975 through 1980s: Cambodian Refugee Women, Lowell, MA, 2007–2008

Period	Quantity and Quality of Food			Meals per Day			Period Total,a Mean ±SD
	Often, % (No.)	Sometimes, % (No.)	Never, % (No.)	≥3, % (No.)	2.0–2.9, % (No.)	0–1.9, % (No.)
Time of the Khmer Rouge							7.9 ±0.9
Enough food to eat	2.3 (3)	16.5 (22)	81.2 (108)
Kinds of food wanted to eat	0.8 (1)	10.5 (14)	88.7 (118)
Meals per day				1.5 (2)	76.7 (102)	21.8 (29)
The Vietnamese invasion and occupation							5.5 ±1.6
Enough food to eat	24.8 (33)	60.9 (81)	13.5 (18)
Kinds of food wanted to eat	23.3 (31)	51.9 (69)	24.1 (32)
Meals per day				45.1 (60)	48.9 (65)	5.3 (7)
Thai refugee camps							4.9 ±1.3
Enough food to eat	38.3 (51)	59.4 (79)	2.3 (3)
Kinds of food wanted to eat	31.6 (42)	57.9 (77)	10.5 (14)
Meals per day				55.6 (74)	42.1 (56)	2.3 (3)

Note. The sample size was n = 133. For those who never left Cambodia, data for the Vietnamese invasion and occupation are attributed to the Thai refugee camps period (n = 36); for those who left Cambodia immediately after the Vietnamese invasion, data for the Thai refugee camps are attributed to the Vietnamese invasion and occupation (n = 25).

^aPossible range for each time period is 3–9, with 3 representing least deprivation and 9 greatest deprivation.

The food deprivation measure had a high correlation with number of strategies employed to avoid starvation (r = 0.24; P < .01), including the riskiest strategy of all: secretly taking food from the Khmer Rouge (r = 0.18; P < .05). The food deprivation measure did not have a statistically significant relationship with whether the participant reported any sign of malnutrition (r = 0.13) or the number of signs of malnutrition reported (r = 0.09). This may be because participants were not able to accurately gauge whether they had lost weight or had a changed appearance; scales and mirrors were not available, and participants may not have perceived personal changes in comparison with others in a similar situation.

On the basis of strong agreement with reported experiences and strong correlation with strategies to avoid starvation, the sum variable was considered appropriate to use as a measure of extent of food deprivation over the 3 periods.

Food Deprivation, Food Beliefs, and Dietary Behavior Relationships

There were no statistical associations between any outcomes measuring food beliefs and the food deprivation measure in either bivariate or multivariate models (Table 4).

TABLE 4.

Relationships Between Extent of Food Deprivation and Food Beliefs, Dietary Behaviors, and Weight Status: Cambodian Refugee Women, Lowell, MA, 2007–2008

Dependent Variables	% (No.) or Mean ±SD	Bivariate Model, OR (95% CI) or b (95% CI)	Multivariate Model, OR (95% CI) or b (95% CI)
Food beliefsa
A lot of meat with fat
Did not identify as bad for health (Ref)	17.3 (22)	1.00	1.00
Identified as bad for health	82.7 (110)	1.05 (0.90, 1.23)	1.05 (0.88, 1.26)
A lot of fast food
Did not identify as bad for health (Ref)	32.3 (42)	1.00	1.00
Identified as bad for health	67.7 (90)	1.13 (0.99, 1.29)	1.13 (0.98, 1.32)
A lot of whole grains
Did not identify as good for health (Ref)	34.6 (45)	1.00	1.00
Identified as good for health	65.4 (87)	1.01 (0.89, 1.14)	1.01 (0.87, 1.18)
Behaviors in reference 24-hour periodb
Meat with fat
Did not consume (Ref)	56.8 (74)	1.00	1.00
Consumed	43.2 (57)	1.17* (1.03, 1.33)	1.14 (0.99, 1.30)
Whole grains
Did not consume (Ref)	93.6 (123)	1.00	1.00
Consumed	6.1 (8)	1.00 (0.78, 1.28)	1.09 (0.82, 1.46)
Times ate meatc	1.1 ±0.7	0.00 (–0.05, 0.05)	–0.01 (–0.05, 0.04)
Weight statusd
BMIc(kg/m2)	25.9 ±4.0	0.15 (–0.10, 0.39)	0.13 (–0.13, 0.39)
Overweight or obese (WHO standards)
BMI < 23 kg/m2 (Ref)	27.9 (36)	1.00	1.00
BMI ≥ 23 kg/m2	72.1 (93)	1.23** (1.06, 1.43)	1.28** (1.08, 1.52)
Overweight or obese (CDC standards); obese (WHO standards)
BMI < 25 kg/m2 (Ref)	42.1 (56)	1.00	1.00
BMI ≥ 25 kg/m2	56.9 (74)	1.18* (1.03, 1.35)	1.18* (1.02, 1.37)
Obese (CDC standards)
BMI < 30 kg/m2 (Ref)	83.5 (111)	1.00	1.00
BMI ≥ 30 kg/m2	14.0 (18)	0.89 (0.74, 1.08)	0.88 (0.73, 1.08)

Notes. BMI = body mass index; CDC = Centers for Disease Control and Prevention; CI = confidence interval; OR = odds ratio; WHO = World Health Organization. The multivariate model controlled for acculturation, education, and age. ORs are for every 1-point increase on the 9–27-point food-deprivation scale for each binary outcome variable. Unstandardized parameter estimates are for every 1-point increase on the 9–27-point food deprivation scale for each continuous outcome variable.

^aFor the bivariate model, n = 132; for the multivariate model, n = 126.

^bFor the bivariate model, n = 131; for the multivariate model, n = 125.

^cContinuous variable.

^dFor the bivariate model, n = 128; for the multivariate model, n = 122.

*P < .05; **P < .01.

Eating meat with fat was significantly related to the food deprivation measure in the bivariate model and nearly significant in the multivariate model (P = .06). For each 1-point increase on the 9-to-27–point food deprivation measure, the odds were 1.14 times higher for reporting having eaten meat classified as high in fat. No other eating practice varied by food deprivation measure (Table 4).

Higher food-deprivation scores were associated with an increased likelihood of being overweight or obese by both CDC and WHO standards and obese by WHO standards. In multivariate models, for every 1-point increase on the 9-to-27–point food-deprivation measure, the odds were 1.28 times higher of being overweight or obese by WHO recommendations and 1.18 times higher of being overweight or obese by CDC standards and obese by WHO standards. There was no statistical difference in the likelihood of being obese by CDC standards (Table 4).

DISCUSSION

The severity and length of food deprivation and insecurity had a profound impact on the ways that adult female Cambodian refugees approach food and health in the abundant US food environment. The food deprivation questions in the survey reflected the variation expressed in focus groups and allowed a nuanced quantitative description of food quantity, food quality, and disrupted eating patterns. The food experience questions may be useful in future research investigating effects of food deprivation and insecurity in other refugee groups.

Food beliefs and behaviors measured in the survey paralleled those expressed in the focus groups. Focus group participants noted that community members would eat meat with fat and chicken with skin, despite understanding potential harm to health, and variation in past food deprivation was related to likelihood of eating meat with fat but not related to the belief that a lot of meat with fat was bad for one's health. These findings are similar to findings from research hypothesizing that sporadic access to resources can lead to overconsumption of foods that are rich and palatable when resources are available.¹¹ The focus group results resonate with recent research documenting links between reported childhood food deprivation and adult food behaviors such as bingeing and preferences for calorie-dense, low-nutrient foods.³⁵ The focus group results are also similar to the previously discussed reports of Hmong refugees who ate steak despite understanding its potential harm to health¹⁸ and to studies citing preference for steak among Hmong, Vietnamese, and Cambodian refugees.^19–21

Preoccupation with food availability was not a concern for focus group members, who felt confident that the food supply in the United States was abundant and reliable. The consistent availability of inexpensive animal products may be increasing health risks if people who experienced food deprivation are overconsuming high-fat meats³⁶ in response to their past lack of meat and other food deprivation or insecurity.

The finding that those who experienced higher levels of food deprivation were more likely to be overweight or obese fits into the food insecurity literature documenting an association between higher rates of overweight and obesity and food insecurity in some US populations.^9–13 Past food deprivation may indirectly be contributing to increased health problems in the Cambodian refugee community through higher risk of overweight and obesity. The difference between WHO and CDC weight cutoffs for Asians deserves special attention because almost three quarters of respondents were overweight or obese according to WHO guidelines, but only 57% were overweight or obese according to CDC standards. Because of the increased health risks for Asians at this lower BMI,³² health care providers and program developers should consider using the lower WHO cutoff for overweight and obesity for Asians to accurately assess risk based on weight.

Limitations

Because the survey was cross-sectional, relationships cannot be assumed to show causality. All participants were women, and results cannot be assumed to apply to men. Dietary information was calculated by using single 24-hour recalls, which may distort amounts through measurement error and standardized recipes.³⁷ To minimize distortion, we focused on times the food was eaten per day rather than on serving amounts, and no standardized recipes were used. Our models used the single 24-hour recall to estimate group intakes, a method for which a single 24-hour recall may be used.³⁸

Implications

Many refugees come from refugee camps and tenuous situations where food security is limited.²⁴ The results of the Cambodian experience of food deprivation and food insecurity may thus be applicable to other refugees.

Health care providers and program developers working with refugees must recognize that patients may have suffered severe or prolonged food deprivation or insecurity, which likely affects the way they approach eating, and thus potentially affects their weight and health. Individual experiences should be assessed, and consideration given to a possible history of food deprivation. Providers should go beyond recommending dietary changes for health to refugees by showing sympathy for patients' past experiences and focusing on helping patients see how dietary changes will improve their current quality of life. One focus group member urged providers to be “gentle, caring, to help them, to make them understand.” Her advice was simple and direct:

Say that I know you've been through a lot, that you didn't have enough meat to eat, enough food to eat in general … however, now is a different time, and because of that, because there's plenty of food to eat, there can also be a health problem for you.