'{'J \FOOD PATTERNS OF SOUTHERN WEST VIRGINIA COAL MINERS/ Sarah Ellen,~owling 11. Thesis submitted to the Faculty of the Virginia

Transcription

1 1 '{'J \FOOD PATTERNS OF SOUTHERN WEST VIRGINIA COAL MINERS/ by Sarah Ellen,~owling 11 Thesis submitted to the Faulty of the Virginia Polytehni Institute and State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in HUMAN NUTRITION AND FOODS APPROVED: Mary K?"K'orslund, Chairma~s r I ~ Julein M. Axelson Deember, 1985 Blaksburg, Virginia

2 FOOD PATTERNS OF SOUTHERN WEST VIRGINIA COAL MINERS by Sarah E. Bowling (ABSTRACT) Fifty-one male oal miners between 18 and 65 years old (78% underground and 22% surfae miners) ompleted a self-administered questionnaire. Population and food pattern harateristis were desribed. Food frequenies were used to obtain a food sore whih was a weighted average of the frequeny of use eah month for eah food item as determined by the harateristis. Cluster analysis on the food sores were used to identify foods as ore, seondary or peripheral foods. Core foods were foods onsumed at least ten times in a 28 day period. 'nle food frequenies were us~d to alulate an estimated daily dietary sore for nine nutrients. These sores were ompared with the U.S. Dietary Guidelines. Thirty-two perent of the miners had worked in the mines longer than 15 years, 80% had a family history of oal mining, 50% were between the ages of 18 and 35, 52% worked the eight a.m. to four p.m. shift, and 63% had an grade eduation. Eating three t.imes a day was the most frequent eating pattern. When food hoies during non-work times were ompared to hoies when working, the kind and amount of food differed. Time of meals also differed. Analyses of variane indiated that the food sores were signifiantly different at the.05 level for the type of mining and for inome levels; foods sores for eduation levels,

3 work hours and age harateristis were signifiantly different at the 0.1 level. As a group, the miners had a high fat, sodium and protein intake, a medium total arbohydrate intake with a low fiber intake. In order for the miners to meet the U.S. Dietary Q.iidelines, nutrition eduators may need to fous edutional programs on the guidelines emphasizing a dietary inrease in starh and fiber, a derease in refined arbohydrate, a derease in total fat, holesterol, saturated fat and sodium.

4 ACKNOWLEDGMENTS I would like to thank all of the people who have assisted and supported me throughout this program. To the oal miners in the study, for taking the time to omplete and return the questionnaire. To my family, for the total support and enouragement to omplete this projet. To my parents, Kermit and Irene Bowling, for their love, enouragement and training whih made this possible. To Dr. Mary K. Korslund, for her patiene, understanding and advie. To Dr. Dan Kider, for his willingness to teah, share and brainstorm on the methodology and statistis. To Dr. Julein Axelson and Dr. Peggy Shifflett, for their willingness to partiipate on my ommittee, offering suggestions and understanding. To Saga Corporation, Larry Krouk, John Stephenson and Alan MaQueen, for their support and enouragement for this projet. To Wadine Williams, for her patiene in typing numerous hanges and for her understanding of various postponements. To my extended "family", for helping to deide the topi and enouraging me to ontinue. To God, for his promises. "Commit thy way unto the Lord; trust also in him; and he shall bring it to pass." Ps. 37:5. iv

5 TABLE OF CONTENTS Page ABSTRACT ACKNOWLEDQ1ENTS LIST OF TABLES LIST OF FIGURES.. i iv viii ix INTRODUCTION REVIEW OF LITERATURE COAL MINING AND COAL MINERS IN SOUTHERN WEST VIRGINIA FOOD AND/OR NUTRIENT INTAKE STUDIES OF MINERS FOOD PATTERNS Definition.. Components of Food Patterns... SELECTED APPROACHES TO FOOD PATTERN RESEARCH Diet Surveys Twenty-four Hour Realls.. Habit Changes Food Preferenes Cognitive Studies Buying and Food Preparation Praties Frequeny of Consumption METHODOLOGY Pilot Study Distribution Questionnaire Food Frequenies Food Sores Core, Seondary and Peripheral Foods. Estimated Dietary Sores. Diet Evaluation Statistial Analysis RESULTS POPULATION CHARACTERISTICS v

6 TABLE OF CONTENTS (ontinued) FOOD PATTERN CHARACTERISTICS FOOD SCORES, Underground and Surfae Miners (Figure l and 2) Inome Levels (Figures 3, 4, and S) Eduation Levels (Figures 6, 7, 8, and 9) Hours Worked (Figures 10, 11, 12, and 13) Age (Figures 14, 15, and 16) ESTIMATED DIETARY SCORES THE MINERS' DIET COMPARED TO THE U.S. DIETARY GUIDELINES (USDA, AND DDHS, 1980) DISCUSSION POPULATION CHARACTERISTICS.. Eduation Levels.... Hours Worked FOOD PATTERN CHARACTERISTICS FOOD SCORES CORE, SECONDARY AND PERIPHERAL FOODS... Total Group... Underground and Surf ae Miners Inome Leve ls age ESTIMATED DIETARY SCORES THE MINERS' DIET COMPARED TO THE U.S. DIETARY GUIDELINES (USDA, AND DDHS, 1980) Comments on the Study SUMMARY LITERATURE CITED APPENDICES A Food Patterns Questionnaire B Cluster Analysis of Food Sores Page vi

7 TABLE OF CONTENTS (ontinued) APPENDICES (ontinued) Standard Serving Sizes Used to Calulate Estimated Nutrient Sores 92 D Food Sore Identifiers for Family History of Coal Mining. 94 E Food Sore Identifiers for Year Worked in Coal Mine. 95 F Food Sore Identifiers for Substane Use 96 G Food Sore Identifiers for History of Medial Problems.. 97 H Estimated Dietary Sores for Individual Miners VITA Page vii

8 LIST OF TABLES TABLE Page 1 Population Charateristis of Surveyed Coal Miners 34 2 Food Pattern Charateristis of Surveyed Coal Miners Total Group Sores and Identifiers Core Foods For Surveyed Coal Miners ANOVA for Food Sores for Eah Charateristis of Subjets and Signifiane Level Food Category Response to Consumption and Sample Foods Identified Compared to Food Sore Identifier Derived from a Food Frequeny Cheklist 61 7 Inidene of Estimated Dietary Sore Identifiers for Eah Nutrient 63 viii

9 LIST OF FIGURES FIGURE Page 1 Food Sore ldentif iers for Underground Coal Miners Food Sore!dent if iers for Surf ae Coal Miners Food Sore Identifiers for Coal Miners with an Inome Between $20,935-$22, Food Sore Identifiers for Coal Miners with an Inome Between $22,250-$22, Food Sore Identifiers for Coal Miners with an Inome Between $22,864-$24, Food Sore!dent if iers for Coal Miners with 6-8 Grade Eduation Food Sore Identifiers for Coal Miners with 9-10 Grade Eduation Food Sore Identifiers for Coal Miners with Grade Eduation Food Sore Identifiers for Coal Miners with a College Eduation Food Sore Identifiers for Coal Miners Working Between 8 a.m. and 4. p.m Food Sore Identifiers for Coal Miners Working Between 4 p.m. and 12 a.m Food Sore Identifiers for Coal Miners Working Between 12 a.m. and 8 a.m Food Sore Identifiers for Coal Miners Working a Swing Shift Food Sore Identifiers for Coal Miners Between the Age of Years Food Sore Identifiers for Coal Miners Between the Age of Years Food Sore!dent if iers for Coal Miners Between the Age of Years ix

10 INTRODUCTION Food patterns are one of the onsiderations for evaluating nutrition eduation needs of a group. Food patterns of various ultures and groups of people have been studied (Sanjur, 1982). Sine atual dietary praties or food patterns are influened by environmental, ultural, eonomi, psyhologial, and emotional fators (Rozin, 1980; MKenzie, 1980), knowledge of nutrition alone is insuffiient to indue hanges in dietary praties. The key influenes of nutrient intake are these food patterns. Thus, in order to hange them and hene nutrient intake, a thorough understanding of present food patterns beomes essential. Titis understanding should inlude onditions and irumstanes that surround them in order to provide insight into appropriate strategies for teahing nutrition. Titese food patterns should be investigated prior to the design of any program attempting hange in nutrient intake patterns. Sine knowledge of food patterns is needed to assist with the design of nutrition eduation programs, any health programs with a nutrition omponent would probably benefit from this information. Health linis in southern West Virginia have Blak Lung Clinis assoiated with them. MDowell County offered a health ativation program promoting health eduation. Both programs have a nutrition omponent. Underground oal mining is the hief oupation of workers in southern West Virginia. However, the food intake studies of miners have onentrated on total alories, vitamin A or vitamin C intake. Vitamin A intake has been onsidered to have a possible link to a oal 1

11 2 miner's suseptibility to pulmonary problems assoiated with hroni exposure to oal dust (Swartz, et al, 1978). Studies of atual intake or food patterns of long-term ative oal miners are not readily available in the literature. Most ommunities in southern West Virginia are dependent upon the oal industry for their growth. People in the Appalahian oalfields have been desribed as having an unique ulture differing from others groups (Weller, 1965; Erikson, 1976). Coal miners in Appalahian work in an unique environment and have an unique ulture. Aording to Rozin (1980) and MKenzie (1980), three major fators appear to influene the variane in food patterns: ulture, individual experiene and environment. Therefore, oal miners may have unique food patterns as influened by their ulture and environment. The purpose of this study was to desribe the praties of the oal miners regarding food and to determine major fators whih relate to their food hoies. A gross evaluation of these praties was onduted to assist the eduator in identifying needs for nutrition eduation programs. The information obtained from this study may provide insight into their food patterns and thus may assist the nutrition eduator in identifying fous areas and potential problem areas related to the hange proess.

12 REVIEW OF LITERATURE COAL MINING AND COAL MINERS IN SOUTHERN WEST VIRGINIA The Poahontas Coal Field has been alled the "greatest oalfield the world has known, or of ourse, will ever know" (Mauk, 1959). The largest part of this field is in southern West Virginia and extends to southwest Virginia. Commerial mining began in this area in Southern West Virginia and portions of southwest Virginia tended to be poor farming ountry with very little tillable soil. The Poahontas oal fields have been named "The Billion Dollar Coal Field", "The Coal Bin of the World", "The Nation's Coal Bin", and the "Blak Diamond Empire". 'Welh, the ounty seat of MDowell County, in Southern West Virginia has been ommonly desribed as "in the heart of the nation's oal bin" (Mauk, 1959). The growth of this area an be traed to the growth of the oal industry. The oal industry and the resultant oal ommunities are known for experiening a series of "boom and bust" yles. Coal towns or ommunities developed and existed as a result of loal mines. the mines shut down, the ommunity tended to beome a ghost town. If Coal ommunities depend on the oal industry for employment opportunities and onsequently suffer from flutuations in the single industry. Sine oal is the prinipal means of support the rugged Appalahian land offers, opportunities for other employment are extremely limited. Muh has been written on the influene of the oal ompany on these ommunities (Campbell, 1981; Lee, 1969; Densmore, 1977). It was 3

13 4 not unommon for the ompany to build row houses, operate ompany stores, provide ompany dotors and ontrol shools and hurhes. Most of the land is owned by people and/or land ompanies loated outside the area. Wardell, et al (1979) has attributed the nature of the boom and bust yles and the ditatorial ontrol of the oal operators as leading to the development of the United Mine Workers of Ameria union. Southern West Virginia tends to be rural, nonf arm ommunities dependent on the single industry, oal, for its eonomy. 'nlus, oal and oal ompanies have had a great influene on the people of this area. Southern West Virginia is part of Appalahia and exemplifies many harateristis desribed by Weller (1965). Rupert Vane suggested that the mountain isolation beame mental and ultural isolation and resulted in ertain attitudes whih now operate in these areas (Weller, 1965). Aordingly, the influene of television has been noted as a major hange agent in the Appalahian ulture. 'nle ulture of Appalahia has been desribed by many people (Weller, 1965; Caudill, 1962; Maurer, 1977). 'nle mountaineer has been desribed as individualisti, traditionalisti, fatalisti, ation-seeking, somewhat fear motivated, and oriented to people rather than objets. 'nle ativity and potential danger of underground oal mining appeals to the mountaineer. Medial are is sought mainly in response to risis with little attention given to preventive mediine. Dentists frequently omment about the patients' desire to have teeth pulled rather than

14 s have the are neessary to preserve them. Shifflett and Noel (1979) ommented on the use of the onsultation of the referene group and folk mediine in addition to the physiian onsultation. The soial life of the mountaineer is entered in referene groups omposed of members of the same sex. The family is adult-entered but losely knitted. Weller (1965) desribed the resistene to hange so ommonly enountered in mountain ommunities. "Outsiders" are viewed with suspiion. When problems are identified and possible solutions evolve, "outsiders" ommonly enountered this resistane to hange. Of interest to the nutrition eduator is the observation by Weller of the tendeny of the mountaineer to prefer individual instrution to group lasses. With individual instrution, the burden of responsibility for the program is plaed on the offerer rather than the partiipant. As a group, mountaineers are not interested in planning, saving or budgeting. They are oriented to existene rather than progress. One miner summarized the Appalahian oal miner's life as follows: (Brooks and Shepherb, 1980) The oal miner's life is a funny thing. I guess they are the best-hearted people in the world. The biggest part of them are the easiest going people in the world. They never aumulate a great deal of money. They make a lot, but they give a lot also. We don't want any of our hildren to ever go in the mines or around the mines if we ould help it. We've seen too muh of it. It seems like a miner's money doesn't do him any good. I don't know why. It's probably management. Ninety perent of the oal miners just live from day to day. If he don't work that day, he don't have it. That's just the way it is. They don't aumulate muh at all.

15 6 FOOD AND/OR NUTRIENT INTAKE STUDIES OF MINERS Studies relating to food or nutrient intake of oal miners are limited. Some related studies are available on gold miners. Most relate to speifi nutrients: alories, vitamin A or vitamin. Satyanarayana, et al (1972) studied the effet of the addition of 500 alories to the diet of Indian gold mine workers and the effet on nutritional status and effiieny as measured by produtivity and attendane. The authors justified their study due to the general poor nutritional status of the Indian workers but were unable to show a signifiant differene between the supplemented group and the ontrol group. Several authors have indiated interest in studying various nutrient levels of miners. Previous studies by Van Graan had shown the average level of illumination of underground gold mining operations was muh lower than that of artifially illuminated surfae operations. A safety hazard may develop if there is any impairment in the ability to adapt and maintain optimal visual auity underground. Van Graan, et al (1975) followed up by studying the vitamin A status and the dark adaptation of thirty-five blak miners upon arrival to the mines and after four to six months of ontinuous work underground in the gold mine. Results indiated that a highly signifiant derease ourred in the miners' serum vitamin A level between the time of arrival at the mine and the seond examination four to six months later. This derease was aompanied by an inrease in mean dark adaptation time. The authors onluded that the findir.gs were indiative of an inadequate

16 7 dietary intake of vitamin A during the period when the miners were on the mine diet. The study did not give details of the "mine diet". In a letter to the editor, Furman (1973) ritiized the use of serum vitamin A levels to predit dietary vitamin A intake. He reommended the use of dark adaptation, as used by Van Graan, et al (1975). A desription of the vitamin A ontent of the onsumed diet would add redibility to the findings. Visagie, et al (1974) investigated hanges in vitamin A and C levels in blak mine-workers. Serum vitamin A and C and arotene levels were obtained at time points: new arrivals, 1 month, 3 months, 6 months, 12 months, and 24 months. The dietary intake was estimated by determining the total food purhases ompared to the number of people being fed. A survey determined the food items whih were onsistently being disarded by the workers at the self-servied kithens. The results indiated vitamin A levels were not signifiantly dereased. However, serum vitamin C levels were lower at later points in time. This was found in spite of an "apparently adequate vitamin C intake". Sine the dietary intake estimate was based on food purhases, a more thorough estimate of vitamin C intake of the individual study partiipants seems to be indiated. Forty-four retired oal miners and two ative oal miners were the subjets of a study by Swartz, et al (1978). The miners had worked in the mines an average of thirty-two years with a range of l to 53 years. All had some degree of respiratory insuffiieny. Mean daily intakes of 11 nutrients and alories were alulated from three 24-hour reall reords maintained by the subjet. Bakground information was obtained

17 8 by an interviewer. A omprehensive medial examination, various pulmonary funtion tests and seleted hematologial tests were performed on eah subjet. Thirty-eight of the subjets were sixty years or older. The authors onluded the results should be examined more as an elderly problem than a "miner" problem. intakes were low as ompared to the RDA. Vitamin A and C and alium Sine the majority of the subjets were retired, these results annot be onsidered refletive of younger, ative oal miners. There appears to be a lak of desription of food patterns of this group of workers. A study of ative oal miners food patterns seems to be needed. FOOD PATTERNS Definition: Any disussion of food patterns must begin with the onept learly defined. Kolasa (1981) has identified a need for preise terminology. The omplex and diverse nature of food patterns is illustrated by the number of definitions and varied terminology assoiated with them. Food patterns have also been termed food habits (Committee on Food Habits, 1945), eating patterns (Shwerin et al, 1981), nutritional patterns (De Garine, 1972), dietary praties (Jalso et al, 1974), and food onsumption praties (Rizek and Jakson 1980). Food preferenes, food use patterns, food ways (Kandel et al, 1980), food seletion (Rozin, 1980) and food meanings (Reaburn et al, 1979) have been assoiated with the onept of food patterns. The Committee on Food Habits (1945) has defined food habits as "the way in whih indi~iduals in response to soial and ultural pressures selet, onsume and utilize portions of

18 9 the available food supply." Lowenberg (1974) prefers the use of the term "food patterns" to food habits in order to indiate that food praties are patterned rather than fixed. Food patterns is the pref erred term in this disussion to indiate this gradual hange proess and the potential for intervention to our. Components of Food Patterns: People selet food rather than nutrients for their food patterns. Nutrient needs are only one aspet of food hoies. Other aspets have been desribed in various ways. Rozin (1980) suggests we divide the determinants into biologial, ultural, and individual fators. Biologial fators inlude geneti and metaboli fators; ultural fators are the onstraints imposed by ulture on food availability, seletion, and preparation; and individual fators are the results of the individual's interations with his environment, within the onstraints of biology and ulture. He stressed the interation of these fators. Jerome, et al (1980) proposed an eologially based model for nutritional anthropology in whih individual requirements for nutrients is entral. Five peripheral omponents of the model inlude: (1) Physial environment - limate, water resoures, soil harateristi and related features for prodution; (2) soial environment; (3) soial organization - eonomi and politial strutures; (4) tehnology; and (S) ultural and ideologial systems whih inlude ideas about the role of food in health, religious beliefs involving food, food preferenes and restritions, and the use of food in soial interations. De Garine

19 10 (1972) divided soioulture influenes or fators into material ulture and intangible influenes. Material ulture inludes food prodution, tehnology, food proessing, food onsumption, and transportation. Intangible influenes inlude the general attitude towards food, body image, ooking and gastronomy, nutrition, prohibitions and taboos, use of time, "hot" and "old" foods, the supernatural, magi, religion, soial organization, eonomis and soial lass. Sanjur (1982) reviewed a lassi study by Passim and Bennett in 1943 whih resulted in a proposed sheme of food ategories. Foods were ategorized by frequeny and importane. Core foods were the routinely onsumed foods, the staple, universal, important foods. Seondary foods were "foods in widespread use but not universally onsumed." Peripheral foods were not ommonly onsumed and were harateristi of an individual and not groups. Shifflett and Noel (1979) desribed non-nutritional uses of food in Appalahian ulture and the influenes of folklore. This study suggested the importane of referene group influene on food use patterns. Khan and Hakler (1981) reviewed fators involved with food habits and final food onsumption and relationship with food preferenes. Eide (1982) suggested a framework whih separates influenes on food preferenes and available produts whih interat for food patterns. Thus, basi available produts interplay with preferred produts to produe food patterns. The preferred produts are derived from the

20 11 pereived value of food. The resultant food patterns produe nutritional onditions. Oikens (1965) disussed the relationship of ultural, soial, personal and situational fators to food preferenes. She defined ulture as the total man-made part of man's environment and inludes the ideas and beliefs of man, the material items with set limits on kinds and types of food seleted, and food preparation methods. Tile soial motivation for food preferenes ditate that few deisions onerning food are made without regard for others. Personal fators involve personal harateristis inluding age, years of shool ompleted, and seleted psyhologial harateristis, suh as mental flexibility. Situational fators of food preferenes are those desribing the situation in whih the family finds itself at the time of the buying deision. Lowenberg (1974) observed the interation of influenes on food patterns whih annot be distint entities at all times. In other words, the past food patterns of family members interat with the environment. The two are so losely interwoven they annot be disussed separately. Other influenes affet food patterns simultaneously. Sanjur (1982) stressed the synergisti ation of strutural and funtional fators whih appear to be more signifiant than any single independent fator. Food patterns are onstantly evolving sine influenes are multiple and individuals are onstantly hanging. Aordingly, MKenzie (1980) defined food habits as the sum total of food hoie over a period of time.

21 12 Due to interonnetion of fators, identifiation of struture or models varies in terminology but in general the major omponents of food patterns appear to be ulture, environment, and personal onsiderations. Although authors vary on the influenes emphasized, the interation of various fators appears to play an important role in the atual food pattern of an individual. However, awareness of ways in whih different people respond to environment and ultural influenes on food patterns is ritial to planned hange. SELECTED APPROACHES TO FOOD PATTERN RESEARCH Food pattern studies may be lassified in numerous ways. Hertzler, et al (1982) reviewed ultural food patterns by ontent and ontext or diversity lassifiation. Another approah has been used by Kandel, et al (1980). They identified four "lines" of food researh: diet survey, food habits, ognitive aspets, and eologial studies. Both approahes suggest nutritionists and soiologists are beginning to oordinate, efforts in order to study food use with the multiple interative omponents. The present interest is in preditive models whih an eliit information on aeptable dietary ombinations. Diet Surveys: One of the most involved approahes to food pattern researh has been the diet survey studies. Three major surveys have been onduted. The Department of Health Eduation and Welfare olleted information on onsu.'llptior. praties in two studies: the Ter.-State Nutritional Survey (June 1968 to July 1970) ai:d the Health and Nutrition Examir~ation Survey

22 13 (HANES - April 1971 to June 1974). The most reent survey was onduted by the U.S. Department of Agriulture. The Nationwide Food Consumption Survey (NFCS) olleted information from April 1977 to Marh 1978 (Croetti and Guthrie, 1981). All three surveys involved over 60,000 respondents and utilized 24-hour realls for diet information. utilized two day food diaries in addition to reall information. The NFCS HANES and Ten-State Nutritional Survey inluded assessment of nutritional status by biohemial and anthropometri measurements. Youland and Engle (1976) desribed some of the praties and problems with HANES. They noted the frequeny with whih ertain foods were eaten was useful in establishing general food patterns by age, sex, inome level and analyzing nutrient inadequaies. Shwerin, et al (1981) reexamined the data from the Ten-State Survey and HANES. Fator analysis on the data foused attention on eating patterns as a means of relating food intake to health. Seven statistially different eating patterns resulted and were haraterized by varying levels of onsumption of the different food groups. Signifiant differenes were found between the seven eating patterns and the absene of linial symptoms and biohemial defiienies. Oae (1981) ritiized this appliation of fator analysis beause the eating patterns do not of fer a lear piture of the total diet and the study failed to standardize the data for age. Oae felt there was a need to do more than define nutritional health as the absene of a subset of symptom.s. She noted the eating patterns were not really patterns but groupings of persons who have unusually high or low onsumption of two

23 14 or three types of food and ignores the onsumption of the other twelve or thirteen food groupings. Twenty-four Hour Realls: Two reent studies by Pao, et al (l985) and Guthrie and Croetti (1985) examined data from the Nationwide Food Consumption Survey (NFCS) to determine the usefulness of single twenty-four hour realls for desribing nutrient intakes of population groups. Pao, et al (1985) onluded little differene was observed between three day mean intakes and day one mean intakes for most nutrients studied. The authors onluded the statistially signifiant differenes that were observed for energy, fat and arbohydrate ould have been the result of hane. They used regression analysis whih indiated the first day's intake reord was "strongest as a preditor of intake on subsequent days for energy, arbohydrate, alium, magnesium and phosphorus; less strong for protein, fat, iron, thiamin, riboflavin, niain, vitamin B-6, and vitamin C; and poorest for vitamin A." Guthrie and Croetti (1985) found intakes varied by more than 25 perent from the mean for vitamins A and C over a three day period. They onluded one day food reords alone are of limited value in assessing nutrient adequay of an individual and inrease the magnitude of the variane over a three day period. In a ommentary, Woteki (1985) interpreted the differenes in these studies to onlude that single 24-hour realls may be adequate for nutrition monitoring sine the realls reflet mean nutrient intakes for population groups. However, 24-hour realls should not be

24 15 used for epidemiologial studies in whih the usual nutrient intake of an individual is important. This interpretation is in agreement with the findings of other researhers. Sempos, et al (1985) obtained 5115 daily diet reords from 151 women over a two year period. The subjets reorded their food intake on two randomly seleted days per sampling month. Calulations were made of yearly estimates of the ratios of intraindividual variane omponents on fifteen nutrients and dietary intake plus vitamin-mineral supplements. The authors inluded two days per sample and 31 reords. Measuring total intake may be needed for eah individual to insure the estimate of the population orrelation between an individual's total intake of a dietary risk fator and an individual's mean or usual level of a physiologi risk fator. In addition, Axelson (1984) studied groups of teenagers divided aording to body weight. The groups ompleted two 24 hour realls. The author alulated nutrient intakes from the reall to demonstrate that variane in measurement may lead to inorret onlusions in program evaluations of nutrition eduation. St. Jeor, et al (1983) analyzed food logs maintained by 16 paid volunteers for a 12 week study period. The study showed no signifiant differenes in the intake of energy or seven seleted nutrients from week to week ompared to a four-week period. Individual differenes varied widely. Twenty-four hour realls have been used in various ways. Koh (1984) used the 24 hour reall to estimate energy intake, alohol and affeine onsumption as well as soures of energy for 384 people in

25 16 Mississippi. The results indiated a signifiant raial differene in energy ontent and soures in the diet. Dennis, et al (1985) studied mean energy intake and its omponents for 4,568 white adults between years of age in a Lipid Researh Center. Twenty-four hour realls were obtained and omparisons were made with the data from the USDA Nationwide Food Consumption Survey and HANES I and with the Reommended Dietary Allowanes and the Dietary Goals. Sex and age-related differenes varied for eah maronutrient. The Lipid Researh Center group mean intakes exeeded the Dietary Goals for protein and fat and fell below the reommendation for total arbohydrate. Randall, et al (1985) studied the relationship between diet diversity (the number of foods eaten), alori intake and nutrient density values. Nutrient density value is the total nutrient intake divided by alori intake multiplied by They used the 24-hour reall data from NHANES II for 1,747 white men and 1,898 white women between years of age. Nutrient intake was diretly related to diet diversity, total alori intake and nutrient density. The authors onlude diet diversity ould be useful in estimating intakes of some nutrients (fat, saturated fat, potassium) while nutrient-speifi density indexes may be needed for holesterol, alium, and vitamin A. Sodium intake appears to be influened by both alori intake and nutrient density. Evans and Gines (1985) used three methods to obtain 24-hour dietary realls. They found no signifiant differene between the

26 17 results obtained by an unaided oral reall, a printed questionnaire obtained orally and an interview with atual-size olor pitures of foods. Age, sex, eduational level, obesity, plae of residene and food preparer did not have an influene in the study. Habit Changes: Food patterns have been investigated in relation to disease, suh as heart disease (i.e., the Framingham study) and aner (Hunter, et al, 1980). Titey have been desribed in terms of hanges from one ulture to another. Yang and Fox (1979) studied food habit hanges of Chinese persons living in Linoln, Nebraska. Quantitative food intake was assessed in terms of inreased or dereased intake of 47 food items while qualitative food intake was assessed by 24-hour reall and one day reall of a typial menu onsumed in the native ountry. Change in intake was determined by a omposite adaptation sore (CAS). Tite greatest hange was observed in the lunh meal. Food Preferenes: Food patterns have also been desribed in terms of food preferenes. Piggott (1979) obtained a ompleted food preferene questionnaire from 1304 people. Considerable differenes in food preferenes were related to age, sex, and area of residene. Five different lusters of foods showing similar preferene patterns were identified. Another type of exploratory study involves the identifiation of favorite and disliked foods. Worsley and Leith (1981) studied students' pereptions of favorite and disliked foods by a rating

27 18 system of 35 pairs of antonyms arranged in seven point rating sales. Students wrote down a favorite food and a disliked food rated aordingly. these thirty-five desriptors were redued by prinipal fator analysis to a few lusters of related desriptors or fators aording to their underlying meaning. the results indiated little differene between students' oneptions of their favorite and most disliked food. Again lues to the reasons for food hoie were identified. Another similar type of system has been used to lassify food use. Shutz, et al (1979) used appropriateness of use as the basis for rating the food-use ombinations. Two hundred women, 50 from eah of four regions of the ountry, were asked to identify appropriateness by a rating sale of one as never to seven as always appropriate. Fortyeight food uses were listed and partiipants were asked to rate appropriateness of eah use for 56 foods. Regional deviations were found. Five food fators were identified: high alorie treat, speialty meal item, ommon meal item foods, refreshing healthy food, inexpensive filling foods. the authors identified four fators for use: asual, utilitarian, satiating and soial. this study differs from preferene studies beause appropriateness of food for various uses was identified ompared to overall preferene. Einstein and Hornstein (1970) studied food preferenes of approximately 50,000 ollege students. Foods were ranked as liked, disliked, or do not know. Nutritional impliations of food preferenes were also examined. the results indiated the best soures of vitamin A were among the most disliked in the survey.

28 19 Cognitive Studies: Study of ognitive aspets of food patterns is another approah. Numerous studies have been onduted on attitudes and food hoie. Food patterns were identified as the food that a student athlete was likely to eat, avoid, or feel indifferent toward in various situations as defined by Werblow, et al (1978) in their study of athletes. "Food patterns" rather than nutrient intake was studied in order to ompare similarity of food onsumption in various eating situations. Food pattern similarity sores were developed in whih foods were ompared for similarity of onsumption in various eating situations. A high sore ould be due to eating more of the same foods, avoiding the same foods and/or feeling indifferent about the same foods. Correlation oeffiients were omputed between the knowledge, attitude, and food pattern similarity sores. Nutritional knowledge and attitudes were not orrelated with frequeny of food onsumption. Perron and Endres (1985) studied the relationship between the nutrition knowledge and attitudes and atual dietary praties of adolesent female athletes. They used a 24-hour reall, a 48-hour food reord and a self-administered knowledge and attitude questionnaire of Werblow, et al (1978). positively orrelated. Nutrition knowledge and attitudes were No signifiant orrelation was found between nutrition knowledge or attitude and dietary intake. Shwartz (1975) also studied nutritional knowledge, attitudes and praties of high shool graduates. Nutritional praties were defined as frequeny of intake during a three-day period of foods in 17 food groups as ompared

29 20 with the basi four pattern. Frequeny data were oded for eah food group with numerial sores assigned aording to adherene to the basi four. Desriptive information indiative of nutritional praties suh as use of nutrient supplements, types of fats and salt used, and adherene to modified diets was sought as part of the questionnaire. There was not a signifiant orrelation between nutritional knowledge and praties. Other areas of ognitive aspets of food patterns involves nonnutritional use of food suh as magi, folklore, and home remedies. Of partiular interest is the study by Shifflett and Noel (1979) of nonnutritional uses of food in traditional Appalahian ulture. ntey stressed the importane of referene groups in Appalahin ulture as desribed by Weller (1965). nte influene of referene groups should be onsidered as part of the examination of food patterns. Buying and Food Preparation Praties: Buying and food preparation patterns are another approah to the desription of food patterns. Wheeler and Haider (1979) studied ghetto blaks and Hispanis in Brooklyn. Food patterns were desribed in terms of planning, budgeting and shopping for food; type of stores from whih the food was purhased and the frequeny of trips; and ooking methods. Frequeny of Consumption: Frequeny of onsumption is a ommon tehnique utilized to identify food patterns. Koh and Caple (1979) studied frequeny of

30 21 food onsumption of low-inome families in southwestern Mississippi. The authors used frequeny of use to evaluate ontribution of the different food groups in order to have a rapid assessment of the food patterns of a larger number of households. Other items onsidered in this study inluded food resoures, shopping praties and preparation skills. Inome had a greater effet on low nutrient intake than eduation. The authors onluded that meal patterns of the households were not able to provide nutrient levels high enough for individual family members' needs. Margetts, et al (1981) onduted interviews onerning the frequeny of onsumption of 76 food items. Three different ethni groups were interviewed. Italians, Australian- Italians, and Australians. Multiple disriminant analysis was used to identify food items whih best separated the three groups. Initial analysis of all 76 food items showed that many of these items were eaten in approximately equal frequeny by all groups. ontrasts were in types of vegetables being onsumed. The major The authors suggested the onept of "marker foods" to distinguish different groups. Bailey et al (1984) obtained food frequeny data for 372 adolesents from urban and rural low-inome households. The food frequeny questionnaire inluded 24 food groups ategorized aording to nutrient ontribution and was ompleted for eah subjet by a trained interviewer. The onsumption was reorded by (a) one or more times a day, (b) four to six times a week, () one to three times a week and (d) seldom or never. The researhers used the frequeny data to identify

31 22 frequeny of onsumption of folain-dense foods. They found urban blaks onsumed a higher frequeny of these foods than urban Hispanis or rural blaks. Rural whites onsumed a higher frequeny of these foods than rural blaks. These findings orrespond to inidene of poor folain status in urban Hispanis ompared to urban blaks and in rural blak ompared to urban blaks. Axelson and Csernus (1983) tested the reliability and validity of a food frequeny heklist. Reliability of a food frequeny heklist was tested by omparing present and past food frequeny intake as a measure of hange in food onsumption. Validity of the heklist was established by omparing mean frequeny of intake of speifi food groups at two periods in their lives with that of "an age ohort" in omparable age groupings of USDA Nationwide Food Consumption Surveys (NFCS). All frequenies were onverted to frequeny per week. High orrelations were found in the two administrations of the heklist. The authors suggested that the omparison of food intake between two different periods prompted memory of relative frequeny. The food frequeny heklist was onsidered reliable and the validity of the heklist was tested by omparing it with another. The authors onluded the food frequeny heklist would be useful in the evaluation of the influene of dietary intake. Frequeny of onsumption of 35 foods and food groups and patterns of food exlusion were used by Williams and Penfield (1985) to de'lelop an instrument for haraterizing food-related behavior. The instrument was validated for ontent and preditive ability.

32 23 Roberge, et al (1984) identified food frequeny as a tehnique appropriate in epidemiologial approahes and evaluation of eduational programs. Aordingly, Dennis and Shifflett (1985) used food frequenies as one omponent of a oneptual and methodogial model for studying dietary habits in the ommunity. In this model, dietary habit was defined to represent a harateristi of an individual revealed through reurring behavior in different situations. To desribe dietary habit, the authors suggested use of one day's food intake, frequeny with whih foods are onsumed and long-term food preferenes. The review of literature indiates various approahes have been taken to identify and desribe food patterns, studies of diets of ative oal miners are limited, and oal miners seem to have an unique ulture. All three inferenes are linked by the need to identify and desribe the food patterns of ative oal miners along with ultural harateristis. Food frequenies and questions relating to population and food pattern harateristis seem to be the appropriate tool to aomplish this goal.

33 METHODOLOGY This study was designed to explore the food patterns of oal miners by desribing population and food pattern harateristis, identifying ore, seondary and peripheral foods onsumed by this population group, and translating this information into estimated dietary sores to evaluate the overall diet. The purpose of the study was to identify the key harateristis influening food sores. A self-administered questionnaire was ompleted by 51 male oal miners between 18 and 65 years old. Two different groups of miners were studied: underground miners who worked in underground mines and surf ae miners involved in strip or auger ntining or preparation (oal leaning) plant operation. This instrument was used to identify population and food pattern harateristis. Food frequenies were used to obtain food sores and identifiers in addition to estimated dietary sores. Sample foods listed were ompared to the food frequenies. The approah utilized for this study was a questionnaire onerning food onsumption and desriptive information involving food praties. Warwik and Lininger (1975) disussed the advantages of the self-administered questionnaire ompared to the interview. Sine the influene of peer or referene groups appears to be very strong in the Appalahian ulture, the researhers felt the subjets would be more willing to answer questions if approahed by a peer ompared to an "outsider". Thus, a self-administered questionnaire to be distributed 24

34 25 by volunteer oal miners was seleted as the appropriate researh instrument. Pilot Study: The initial questionnaire onsisted of 26 questions whih inluded 22 losed response, two ompletion and a food use frequeny heklist. Ten questionnaires were distributed to identify potential problems with the design. Tile researher was available to note areas of diffiulty but did not assist with its ompletion. On the basis of the pilot study, some hanges were made. Tile wording was hanged on Question 25 from "Do you eat the same foods as your friends?" to "Do you and your friends eat similar foods?". Question 12 onerning hanges of food intake relating to shift hanges was deleted due to diffiulty in answering. The researhers felt this type of question might require personal interview with the subjet. Question 27 was added to eliit information onerning the person ompleting the questionnaire. The revised questionnaire used in the study is in Appendix A. Distribution: Questionnaires were distributed to six volunteer oal m1ners who in turn distributed them to friends and fellow workers. Subjets were asked to omplete the questionnaire immediately. Two weeks were allowed for their return. Tilree of the six volunteers returned questionnaires. Fifty-one male oal miners between the ages of 18 and 62 partiipated in the study. Tile majority of the miners were working

35 26 in MDowell County, West Virginia. Others worked in Merer, Wyoming and Raleigh Counties of Southern West Virginia. Questionnaire: Tile questionnaire was designed to ask both diret and indiret questions to obtain information relating to food patterns. Questions utilized to identify population harateristis and for omparison of signifiane of differene between food sores were oded in the following fashion: Years in the oal mines were divided into five ategories: 0-5, 6-10, 11-15, 16-20, and 20 or more. Job title was translated into inome information. Tile National Bituminous Coal Wage Agreement of 1981 was used to identify job grade and hene base inome. Base inome levels were divided into four ategories: (l) 20,935-21,082; (2) 21,162-22,079; (3) 22,250-22,407; and (4) 22,864-24,471. For the analysis, ategories one and two were ombined due to small sample size in these groups. Work hour information was used to obtain differenes in food patterns with different shifts. Some miners hange shifts every two weeks and thus were identified as working swing shift. Information on age was divided into three ategories: (1) 18-35; (2) 36-45; and (3) Family history of oal mining was asertained by identifying whether their fathers were oal miners. Substane use was identified by questions onerning smoking, hewing tobao or dipping snaff. Medial history indiators were obtained by questions onerning diagnosed diseases as well as urrent mediation. Peer influene on food patterns was somewhat indiated by

36 27 the question "Do you and your friends eat similar foods?". Spouse estimates of husbands' food intake was noted by the identifiation of the person ompleting the questionnaire. Studies by Kolonel and Lee (1981) and Marshall, et al (1980) indiate information obtained from a spouse has good orrespondene for a majority of food items exept for foods eaten away from the home. Thus, for the type of information obtained in this study, questionnaires ompleted by their wife were treated the same as questionnaires ompleted by the miner. Questions were asked to aid in the desription of food patterns. These inluded: the number of times eating eah day, eating during work, food whih annot be eaten when working, food preparation, ommon food intake, vitamin usage, and feelings toward the healthiness of their diet. Food Frequenies: Frequenies of hoies were used as indiators of major nutrient hoies. The food frequenies were oded for approximate monthly usage: 56 = use more than one a day; 28 use almost every day; 4 use almost every week; 1 - use at least one a month and 0 = do not eat. The numbers seleted were hosen to represent an approximate frequeny use per month. No response was oded separately to differentiate from the do not eat reponse. Question 17 was oded to indiate salting of food before tasting 56; after tasting ~ 28; and rarely or never ~ l. Question 18 indiated the number of foods to whih sugar was added. This question was used for desriptive purposes and as a hek for simple arbonhydrate.

37 28 Food Sores: Food sores were developed from the food frequenies as a measure of frequeny of use. Reaburn, et al (1979) used food sores derived from food frequeny data. Food use frequeny was measured using a 5-point fixed alternative sale. The food sore formula was: Food Sore m where Item n S Sale frequeny rating (l never to 5 daily) R = % Respondents seleting rating n = maximum sale rating The food sore as used by Reaburn, et al (1979) represents a rank sale. The above food sore was adapted to inorporate a ratio sale rather than a rank sale sine this study utilized more speifi frequeny information, i.e. 56 more than one a day to 0 Do not eat. The resulting food sore is a measure of the frequeny of use eah month. Food Sore Item S = Sale frequeny rating (56, 28, 4, 1, O) R = Proportion of respondents seleting rating n = maximum sale rating This modifiation is the weighted average of the frequeny onsumption for eah item as determined by the harateristi, i.e., total group or underground oal miners.

38 29 Core, Seondary and Peripheral Foods: The food sores derived from the food frequenies were used to identify individual foods as ore, seondary, and peripheral foods (Jerome, et al 1980). Core foods are defined as universal, staple foods with a very high frequeny of use. moderately high frequeny of use. Seondary foods are foods with a Peripheral foods have a moderately low frequeny of use. The food sore for the miners as a group was first determined. The initial parameters for the ore diet inluded food sores greater than or equal to 28 whih indiated a use greater than almost everyday. Seondary foods were food sores greater than or equal to four whih indiated a use almost every week. Peripheral foods were foods with a sore less than four indiating a use of one a month or less. This assumption led to the identifiation of three foods (whole milk, offee/ tea, white bread) as ore foods. This method was unsatisfatory sine a very limited number of foods were identified as ore foods. Therefore, a statistially based parameter was used to identify ore, seondary, and peripheral foods. Cluster analysis (Sneath and Sokal, 1973) using the single linkage lustering or nearest neighbor tehnique was performed on the food sores. between the sores alulated. Food sores were ranked and differenes Clustering was made with the nearest neighbor. This analysis indiated the division for the three groups should be greater than ten equal to ore foods. Six to ten equal seondary foods and less than 6 equal peripheral foods. The luster analysis is shown in Appendix B. Thus, ore foods were foods onsumed at least ten times during the month (28 days) or one every three days.