QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Bhargava, A. Articles by Amialchuk, A. Search for Related Content PubMed PubMed Citation Articles by Bhargava, A. Articles by Amialchuk, A.

---

Community and International Nutrition

Added Sugars Displaced the Use of Vital Nutrients in the National Food Stamp Program Survey

Department of Economics, University of Houston, Houston, TX 77204-5019

^* To whom correspondence should be addressed. E-mail: bhargava{at}uh.edu.

	ABSTRACT

TOP ABSTRACT Introduction Materials and Methods Results Discussion LITERATURE CITED

 
Recent increases in consumption of added sugars in the US can reduce intake of vital nutrients and increase the overall energy intakes. A comprehensive framework is necessary for addressing the controversy surrounding the effects of added sugars on nutrient intakes while controlling for dietary knowledge, behavioral factors, and total energy intakes, especially among low-income households. This paper analyzed the proximate determinants of 1-wk availability (or use) of added sugars by 913 households participating in the National Food Stamp Program Survey conducted in 1996–97. In addition, households' use of protein; fiber; calcium; iron; ß-carotene; vitamins A, C, E, B-6, and B-12; folic acid; and potassium were explained by socioeconomic and behavioral factors and by added sugars use. Linear programming analyses were conducted for assessing the effects of food prices on added sugars use. Participation in Supplemental Food Program for Women, Infants and Children program, food stamp benefits, and consumption of low-fat diets were associated (P < 0.05) with lower added sugars use. Furthermore, greater use of added sugars significantly lowered households' uses of protein; iron; vitamins A, C, B-6, and B-12; and potassium. Behavioral and socioeconomic variables such as adopting a low-fat diet, consuming fruits and vegetables, saving money at grocery stores, frequency of shopping trips, skipping meals, and food insecurity were often important predictors of nutrient use. Overall, the results indicated that added sugars should be discouraged in dietary guidelines, because their adverse effects on diet quality were evident in this low-income population.

	Introduction

TOP ABSTRACT Introduction Materials and Methods Results Discussion LITERATURE CITED

The Food Stamp Program in the US provides many low-income households with benefits that can be used for purchasing foods from authorized retailers (9). Unlike the Supplemental Food Program for Women, Infants and Children (WIC)¹ that provides nutritious foods such as milk, cheese, and fruit juice to low-income women with children <5 y, food stamp benefits can be used to purchase most foods; many households receiving food stamp benefits also participate in the WIC program. A previous study using data from the National Food Stamp Program Survey (NFSPS) (5) found that moving away from date of receipt of food stamp benefits reduced households' energy use (or availability). Because food stamp benefits can be used to purchase foods high in added sugars, it is important to investigate their effects on diet quality using the NFSPS data. Moreover, NFSPS data were compiled over 7 consecutive days and hence were better indicators of "habitual" food use than data compiled for a single day. Also, these data contain extensive information on households' dietary knowledge and behavioral aspects and record prices paid for >2000 food items, so that one can analyze the effects of added sugars on nutrient use in a broad analytical framework.

Further, methodological issues are important for assessing the hypothesis of displacement of vital nutrients by energy from added sugars (EAS) (3). First, in some instances, expressing households' intakes of EAS and nutrients as ratios to total energy intakes could create spurious dependences (10). However, one can devise rigorous statistical tests for restrictions on model parameters to decide whether the variables should be transformed as ratios to energy intakes (11,7). Second, it is important for statistical models to reflect the special features of nutrient displacement hypothesis. For example, although foods high in added sugars generally contain lower quantities of vital nutrients, because of nutrient composition of foods, intakes of most nutrients increase with energy intakes. Moreover, food intakes are affected by households' dietary knowledge and behavioral factors. Thus, it is important to control for a diverse set of factors in analyzing the effects of added sugars on nutrient availability or use; robustness of inferences drawn from empirical analyses is likely to depend on adequacy of models postulated for the relations. From this viewpoint, Forshee and Storey (3) presented a somewhat narrow analysis of the effects of added sugars primarily on calcium intakes in a U.S. population; their conclusion that added sugars did not adversely affect diet quality needs to be reappraised in a broader methodological framework.

We analyzed the effects of added sugars on households' uses of protein; calcium; iron; vitamins A, B-6, B-12, C, and E; ß-carotene; fiber; folic acid; and potassium using NFSPS data. Models were specified for nutrient uses in logarithms and where nutrients and EAS uses were expressed as ratios to total energy use. A statistical test was applied to each set of models and results were compared for the 2 formulations. The models accounted for behavioral and socioeconomic variables such as: whether respondents in the households reported looking for grocery specials, consumption of fruits and vegetables, frequency of shopping trips, skipped meals, and food insecurity. In addition, household incomes and food stamp benefits were introduced as separate explanatory variables to assess possible differential effects of food stamp benefits on nutrient use. A similar model was estimated for use of EAS to identify behavioral variables that can be influenced via nutrition education (12,13) and for assessing effects of poverty on diets (14). The comprehensive analysis of NFSPS data shed useful light on effects of added sugars on uses of vital nutrients among these low-income households.

	Materials and Methods

TOP ABSTRACT Introduction Materials and Methods Results Discussion LITERATURE CITED

 
    Subjects. The NFSPS was sponsored by the USDA and conducted in 1996–7 (9). From a national sample of 2142 food stamp participants, a random sample of 937 households was selected to complete 1 wk food use records. Food items purchased were converted into nutrient and energy use after taking into account factors such as weight of bones in the meat quantities used; adjusted food quantities were subsequently converted into energy and nutrients using Pyramid Serving Database for National Nutrition and Health Examination Survey III (15). Added sugars were calculated from food items assuming that a teaspoon of added sugars such as in sweeteners contained 16 kcal of dietary energy (1 kcal = 4.18 kJ) (6). Moreover, 2237 food items recorded in NFSPS were a subset of 4542 foods listed in the Pyramid Serving Database for NHANES III (15). Added sugars used by NFSPS households were calculated using a database (15) that breaks down food items into their component groups such as meat, fruits, vegetables, added sugars, etc. Prices paid for food items were recorded in the NFSPS making these data quite unique for linear programming analyses investigating the effects of food prices on food use. Complete data on households' uses of dietary energy, added sugars, and selected nutrients were analyzed.

One week prior to completion of food use records, primary respondents in the households were interviewed through a computer-assisted personal interview; a second interview took place 1 wk after completion of the records. Background, socioeconomic, behavioral, and dietary knowledge variables of primary respondents ("subjects") were measured. Ethnicity, age, education levels, occupation of head of household, and household size in terms of Adult Male Equivalents (AME) were recorded. Income from various sources was recorded. Although Hispanic ethnicity covered Mexican Americans, Chicano, Mexican, Puerto Rican, Cuban, Central or South American Hispanic, and other Hispanic households, these categories were not separately recorded in the data. Participation of household in the WIC program and presence of guests at meals, number of meals consumed outside the home, and meals skipped were recorded. Although reasons for skipping meals such as dieting, illness, or lack of food were not recorded in NFSPS, households' food security status in the previous 12 and 1 mo were investigated. Because food use data covered the previous 1 wk, the 5 main items investigating food security in the previous 1 mo were used; an indicator (0–1) variable was created that was 1 if households recorded an affirmative response to any of these items. More detailed classifications such as "food insecure, no hunger," "food insecure, moderate hunger," and "food insecure, severe hunger" that are often constructed using items for the past 12 mo were not feasible using items covering the previous 1 mo. Demographic features of metropolitan area were also included in the data set.

    Constructions of behavioral and dietary indices. In the diet and behavior module of the questionnaire, respondents were asked several questions relating to their dietary knowledge, eating preferences, and shopping practices. Because questions covered different dimensions of household characteristics and behavior, 3 indices were created to quantify the effects of behavioral and economic factors on food use (5).

First, because diets low in fat are less energy dense and their consumption is indicative of dietary knowledge (4), a low-fat diet index was created; this index assigned score 1 to affirmative answers to 6 questions regarding choices of low-fat foods. The questions were: if respondent's diet was low in fat and cholesterol, if eating habits had changed to reduce fat, if they were currently limiting the amount of fat, if they had limited fat in the diet in the past, if they were currently eating a low-fat diet, and if in the past month they had thought about changes that could decrease fat in the diet. The answers to these 6 questions were summed to create a low-fat diet index that ranged from 0 to 6 (5).

Second, a fruits and vegetables index was based on responses to 5 questions: if fruits and vegetables were a regular part of the diet, if respondents had ever changed their eating habits to increase consumption of fruits and vegetables, if they were eating more fruits and vegetables than previously, if they had been eating more fruits and vegetables in the last year, and if they chose 5 or more servings of fruits and vegetables. Affirmative answers were assigned score 1 and summed for the 5 questions; the fruits and vegetables index ranged from 0 to 5.

Third, a save money index was based on responses to 6 questions regarding shopping practices: how often respondents looked in newspapers for grocery specials, used store discount coupons, stocked up on items when there were bargains, compared prices at different supermarkets, visited other food stores for advertised specials, and used a shopping list. Affirmative answers were scored as 1 and the save money index ranged from 0 to 6. Overall, because of missing observations on 24 participating households, complete data on 913 households were used in the analyses.

    Models for households' nutrient use taking into account EAS and the total energy use. The model for households' nutrient use (Model 1) was given by equation (1) (i = 1,...,n):

(Eq. 1)

Model 1 embodied salient features of nutrient use by NFSPS households. Indicator variables for ethnicity, college degree, guests present during meals, presence of elderly and children, and participation in the WIC program reflected background information. Socioeconomic variables were household income, food stamp benefits, and shopping practices; the indicator variable for households reporting being food insecure in the past 1 mo and percentage of skipped meals also reflect poverty. Behavioral variables covered the reported consumption of a low-fat diet and fruits and vegetables, saving money at grocery stores, and shopping trips to grocery stores. Because number of monthly shopping trips ranged from 1 to 4, this variable was not transformed into logarithms. Model 1 also included AME, AME-squared, and percentages of meals consumed outside the home to control for household size and eating patterns. The u_i were random errors that were assumed to be normally distributed with 0 mean and constant variance.

An alternative model (Model 2), expressing uses of nutrients and EAS as ratios to total energy use, was derived from Model 1 provided that restrictions on the parameters:

(Eq. 2)

were accepted by the data (7). A statistical test described below was used to test restrictions in Equation 2 against the alternative hypothesis that:

(Eq. 3)

If this null hypothesis were accepted by the data, it would be appropriate to express nutrient and EAS as ratios to total energy use, i.e. as densities. Furthermore, one can replace total energy use in Equation 1 by energy from all sources other than added sugars (3). However, the 2 formulations were mathematically equivalent if variables were specified in levels. Moreover, even in natural logarithms, the 2 formulations led to very similar results, although maximized values of log-likelihood functions indicated a preference for models that included total energy use. We report some of the results for Models 1 and 2, although these were quite similar because of the nutrient composition of foods.

    Statistical methods. The Cronbach (16) were computed for low-fat diet, fruits and vegetables, and save money indices. These statistics measured internal consistency of variables combined to form the indices. Bivariate (Pearson) correlation coefficients were computed between ratios of EAS to total energy and nutrient densities uses. Models 1 and 2 were estimated using PROC MIXED of statistical package SAS (17). Maximized values of log-likelihood functions were compared for models containing the variables total energy use and energy from all sources other than added sugars. Moreover, restrictions in Equation 2 were tested against the alternative hypothesis in Equation 3 using an F-test. The 0 values of indices were increased to 0.00001 prior to logarithmic transformations. Coefficients of the variables in logarithms were "elasticities" (percentage change in the dependent variable resulting from a 1% change in the explanatory variable). Zero values of proportion of meals consumed outside the home were set to 0.001 prior to the logarithmic transformation; sensitivity analyses were performed and the results were quite robust. Lastly, for linear programming analysis, PROC LP available in SAS (17) was used.

	Results

TOP ABSTRACT Introduction Materials and Methods Results Discussion LITERATURE CITED

 
    Descriptive statistics and bivariate correlations. The sample means of variables used in the analysis are in Table 1. Mean household size was 2.98 and it was 2.16 in terms of AME. Approximately 40% of the households were black, 44% were white, and 12% were Hispanic. Only 7% of the heads of households had a college degree; 33% of the households were participating in the WIC program. Mean monthly income was $666 and mean monthly food stamp benefit was $165.

The sample mean of households' energy use for the week was 187.31 MJ and 96.84 MJ in terms of AME. Mean of EAS was 17.96 MJ and the ratio of EAS to total energy (expressed as a percentage) was 9.6% that was comparable to estimates from other data sets for the US (3). Bivariate correlations between the ratio of EAS to total energy use and ratios of nutrients to total energy use were negative for all nutrients listed in Table 1. Correlation coefficients between the ratio of EAS to total energy and densities of protein; iron; ß-carotene; vitamins A, C, B-6, and B-12; fiber; and potassium were –0.38, –0.13, –0.07, –0.09, –0.07, –0.13, –0.16, –0.08, and –0.20, respectively; these correlations were significant at the 5% level, indicating that higher EAS was likely to have displaced uses of these nutrients. However, results from comprehensive Models 1 and 2 for nutrient use controlling for socioeconomic and behavioral factors are essential for concluding whether added sugars in fact displaced the uses of vital nutrients.

    Results for households' use of added sugars. The results from Models 1 and 2 for households' use of EAS are presented in Table 2; results in the last column combined household income and food stamp benefits into a single variable. The F-test rejected restrictions in Equation 2 (P < 0.0001), thereby indicating a preference for Model 1 that included total energy use as an explanatory variable. This was not surprising, because the estimated coefficient of total energy use was significantly >1.

The percentage of skipped meals was positively and significantly associated with EAS use. Although the coefficient of indicator variable for food insecurity in the past 1 mo was positive, it was not significant at the 5% level in Model 1 but approached importance in Model 2 (P < 0.06). Number of shopping trips and the index of low-fat diet were significantly negatively associated with EAS use. Lastly, the R² adjusted for number of parameters for Model 1 was 0.54, indicating that explanatory variables accounted for 54% of variation in the dependent variable.

    Results for the households' nutrient uses. The results for households' uses of protein, calcium, and iron from Models 1 and 2 are in Table 3. F-test statistics were significant for the models for protein and calcium uses, so that Model 1 was the preferred specification for these nutrients. In contrast, restrictions in Equation 2 were accepted for iron use and, hence, the coefficients in Model 2 were more precisely estimated. Overall, differences between the estimated coefficients from Models 1 and 2 were small and this was presumably due to the fact that estimated coefficients of total energy use were close to 1 for all 3 nutrients in Table 3.

The results for calcium use showed that EAS was not a significant predictor, which may be due to many factors. Black households had significantly lower calcium use. Households participating in the WIC program had significantly higher calcium use, which was not surprising because milk and cheese was provided through this program. Also, households with children <18 y had higher calcium use. In households where members reported skipping meals, calcium use was significantly lower. However, calcium use was not significantly lower among food insecure households. Furthermore, households scoring high on the fruits and vegetables index and reporting greater number of shopping trips had significantly higher calcium use. The R² was slightly lower for calcium use than for protein use.

The results for iron use showed negative and significant effects of EAS. Moreover, black and Hispanic households had significantly lower iron use. Iron use was significantly higher among households with children <18 y and also for households participating in the WIC program. Coefficients of household size in terms of AME showed nonlinear effects; both AME and its square were statistically significant. Coefficient of household income was positive and significant although that of food stamp benefits was not significant. Households reporting skipping meals and those that were food insecure had significantly lower iron use. Fruits and vegetable index and number of shopping trips were positive and significant predictors of iron use. Because the F-test accepted restrictions in Equation 2 for iron use, it was not surprising that the estimated coefficients for Models 1 and 2 were extremely close.

Table 4 presents the results for uses of ß-carotene, vitamin A, and vitamin C for the model (Model 1 or 2) accepted by the data. Because F-test accepted restrictions for all 3 nutrients, estimated parameters from Model 2 are presented in Table 4. The EAS was estimated with negative signs and these coefficients were significant in the models for vitamin A and C uses. In comparison with white households, black households had significantly lower uses of ß-carotene and vitamin A, and Hispanic households had higher vitamin C use. Coefficients of indicator variables for members >60 y and for children <18 y were estimated with positive signs and were significant for all 3 nutrients. Also, AME and its square were estimated with significant coefficients in the models for ß-carotene and vitamin C uses.

Table 5 presents the results for the uses of fiber, vitamin E, and potassium; F-test rejected the restrictions implied by Model 2 for vitamin E and potassium. EAS was negatively and significantly associated with potassium use. Black households had significantly lower uses of fiber and potassium, whereas Hispanic households had significantly higher fiber and vitamin E uses. Households with members >60 y had significantly higher fiber and potassium uses, whereas households with children <18 y had higher vitamin E use. Household incomes, food stamp benefits, and WIC participation were not significant predictors of fiber, vitamin E, and potassium uses. Percentage of skipped meals was negatively and significantly associated with fiber use; fruits and vegetables index was positively and significantly associated with fiber and potassium uses.

	Discussion

TOP ABSTRACT Introduction Materials and Methods Results Discussion LITERATURE CITED

Further, linear programming analyses have shown that higher prices of nutritious foods can lower the quality of diets (18), although food prices in some of these analyses were averages for regions from different time points. Using actual data on food prices paid by NFSPS households, use of added sugars also increased in our linear programming analysis when more stringent cost constraints were imposed (results not shown). However, linear programming analyses are mainly suggestive of households' food purchasing behavior. By contrast, models for households' uses of energy and nutrients estimated using NFSPS data reflect actual decisions, controlling for socioeconomic and behavioral factors. For example, the empirical results showed small but significant positive effects of households' incomes on uses of iron, folic acid, and vitamins B-6 and B-12. Interpretation of these coefficients as "income elasticities" of nutrients (19) may be limited by the study design, because all households had incomes below certain thresholds to qualify for food stamp benefits.

The variable skipped meals was positively and significantly associated with added sugars use and the indicator variable reflecting food insecurity was also positively associated although it did not attain significance. Furthermore, percentages of skipped meals were negatively and significantly associated with uses of calcium, iron, vitamin C, fiber, folic acid, and vitamin B-6 that are critical nutrients; adequate calcium intake is essential for maintaining bone density and iron is especially important for pregnant and lactating women and for young children. Moreover, food insecurity was negative and significantly associated with protein and iron uses. Although NFSPS did not explicitly investigate the reasons for skipping meals, these results taken together support the view that poverty in the US is likely to lower diet quality (14). This phenomenon is ubiquitous in developing countries such as India (20), Philippines (11), and Bangladesh (19).

Finally, analysis of food use data from NFSPS showed the importance of behavioral factors such as respondents reporting eating low-fat diets and fruits and vegetables and number of shopping trips to grocery stores for improving diet quality. Moreover, WIC participation increased uses of calcium and iron and decreased use of EAS. Thus, for households with poor dietary knowledge and unhealthy behaviors, there might be a case from a nutritional standpoint for restricting purchase of unhealthy foods such as those high in added sugars using food stamp benefits. However, food stamp benefits were significantly negatively associated with EAS use in NFSPS data (Table 2), thereby not supporting placing restrictions on purchases of such foods. Even so, these data were collected in 1996–97; with the burgeoning obesity epidemic and its effects on prevalence of noninsulin dependent diabetes mellitus, such issues merit closer examination in experimental and other studies (21–23). For example, it may be useful to devise intervention studies that subsidize grocery stores for discounting prices of fruits and vegetables for food stamp recipients to encourage higher consumption. Economic incentives are likely to play an important role in improving diet quality especially among low-income households that often exhibit poor dietary behaviors.

	ACKNOWLEDGMENTS

 
While retaining the responsibility for views in this article, the authors would like to thank Margaret Andrews, Peter Basiotis, Elise Golan, Patricia McKinney, and Jim Ohls for their help and advice.

	FOOTNOTES

 
¹ Abbreviations used: AME, adult male equivalent; EAS, energy from added sugars; NFSPS, National Food Stamp Program Survey; WIC, Supplemental Food Program for Women, Infants and Children.

Manuscript received 25 July 2006. Initial review completed 4 October 2006. Revision accepted 15 November 2006.

	LITERATURE CITED

TOP ABSTRACT Introduction Materials and Methods Results Discussion LITERATURE CITED

 

1. Institute of Medicine, Food and Nutrition Board. Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein and amino acids. Washington: National Academy Press; 2005.

2. Murphy S, Johnson R. The scientific basis of recent U.S. guidance on sugars intake. Am J Clin Nutr. 2003;78 Suppl: S827–33.[Abstract/Free Full Text]

3. Forshee R, Storey M. Controversies and statistical issues in the use of nutrient densities in assessing diet quality. J Nutr. 2004;134:2733–7.[Abstract/Free Full Text]

4. Bhargava A, Hays J. Behavioral variables and education are predictors of dietary change in the Women's Health Trial: feasibility study in minority populations. Prev Med. 2004;38:442–51.[Medline]

5. Bhargava A. Socio-economic and behavioral factors are predictors of food use in the national Food Stamp Program Survey. Br J Nutr. 2004;92:497–506.[Medline]

6. Guthrie J, Morton J. Food sources of added sweeteners in the diets of Americans. J Am Diet Assoc. 2000;100:43–51.[Medline]

7. Bhargava A, Guthrie JF. Unhealthy eating habits, physical exercise and macronutrient intakes are predictors of anthropometric indicators in the women's health trial: feasibility study in minority population. Br J Nutr. 2002;88:719–28.[Medline]

8. Willett W. Nutritional epidemiology. 2nd ed. Oxford: Oxford University Press; 1998.

9. Cohen B, Ohls J, Andrews M, Ponza M, Moreno L, Zambrowski A, Cohen R. Food stamp participants' food security and nutrient availability. Technical report. Washington: Economic Research Service, USDA; 1999.

10. Kronmal R. Spurious correlation and the fallacy of the ratio standard revisited. J R Stat Soc A. 1993;156:379–92.

11. Bhargava A. Modelling the health of Filipino children. J R Stat Soc A. 1994;157:417–32.

12. USDA/DHSS. Nutrition and your health: dietary guidelines for Americans. 5th ed. Washington: Department of Health and Human Services; 2000.

13. National Cancer Institute. 5 a day for better health program. Publication No. 01–5019. Bethesda (MD): NIH; 2001.

14. Drewnowski A, Specter S. Poverty and obesity: the role of energy density and energy costs. Am J Clin Nutr. 2004;79:6–16.[Abstract/Free Full Text]

15. National Cancer Institute. Pyramid Servings Database (PSDB) for NHANES III: http://riskfactor.cancer.gov/pyramid/database/; accessed 07/13/2006.

16. Cronbach LJ. Essentials of psychological testing. 4th ed., New York: Harper & Row; 1984.

17. SAS. SAS version 9. Cary (NC): SAS Institute Inc; 2000.

18. Darmon N, Ferguson E, Briend A. A cost constraint alone has adverse effects on food selection and nutrient density: an analysis of human diets by linear programming. J Nutr. 2002;132:3764–71.[Abstract/Free Full Text]

19. Bhargava A, Bouis H, Scrimshaw N. Dietary intakes and socioeconomic factors are associated with the hemoglobin concentration of Bangladeshi women. J Nutr. 2001;131:758–64.[Abstract/Free Full Text]

20. Bhargava A. Estimating short and long run income elasticities of foods and nutrients for rural south India. J R Stat Soc A. 1991;154:417–32.

21. Johnson R, Frary C. Choose beverages and foods to moderate your intake of sugars: the 2000 dietary guidelines for Americans: what the fuss is all about. J Nutr. 2001;131:S2766–2771.[Abstract/Free Full Text]

22. Bray G, Nielsen S, Popkin B. Consumption of high-fructose corn syrup in beverages may play a role in the epidemic of obesity. Am J Clin Nutr. 2004;79:537–43.[Abstract/Free Full Text]

23. Basiotis P, Guenther P, Lino M, Britten P. Americans consume too many calories from solid fat, alcohol and added sugar. Nutrition insight 33. Alexandria (VA): Center for Nutrition and Public Policy, USDA; 2006.

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Bhargava, A. Articles by Amialchuk, A. Search for Related Content PubMed PubMed Citation Articles by Bhargava, A. Articles by Amialchuk, A.