PART 3

The Dietary Intervention

CHAPTER 10 PLANNING DIET STUDIES

CHAPTER 11 DESIGNING RESEARCH DIETS

CHAPTER 12 PRODUCING RESEARCH DIETS

CHAPTER 13 DELIVERING RESEARCH DIETS

CHAPTER 14 PLANNING AND PRODUCING FORMULA DIETS

CHAPTER 15 MEETING REQUIREMENTS FOR FLUIDS

CHAPTER 16 COMPARTMENTAL MODELING, STABLE ISOTOPES,

AND BALANCE STUDIES

CHAPTER 17 ENERGY NEEDS AND WEIGHT MAINTENANCE IN

CONTROLLED FEEDING STUDIES

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
CHAPTER 10

PLANNING DIET STUDIES

BEVERLY A. CLEVIDENCE, PHD; ALICE K. H. FONG, EDD, RD; KAREN TODD, MS, RD;
LINDA J. BRINKLEY, RD; CARLA R. HEISER, MS, RD; JANIS F. SWAIN, MS, RD;
HELEN RASMUSSEN, MS, RD, FADA; RITA TSAY, MS, RD; MARY JOAN OEXMANN, MS,
RD; ARLINE D. SALBE, PHD, RD; CYNTHIA SEIDMAN, MS, RD;
SUSAN LEARNER BARR, MS, RD; AND ABBY G. ERSHOW, SCD, RD

Types of Research Diets Multiple Dietary Treatments, Multiple Calorie Levels, and
General Classification of Research Diets Variety of Food Items
Formula Diets Concurrent Studies
Conventional Food Diets Length of Menu Cycles and Feeding Periods
Diets Used in Association with Clinical or Diagnostic Scheduling Diet Studies
Tests Tailoring the Study to the Physical Facility
The Diet Classification System of St Jeor and Bryan Pilot Studies
Estimated Diets Estimating the Cost of Diet Studies
Weighed Diets Labor
Controlled-Nutrient Diets Purchasing Food and Supplies
Constant Diets or Metabolic Balance Diets Food
Phase Components of Diet Studies Supplies and Materials
Planning Considerations Budget Planning for Research Diet Studies
Inpatient and Outpatient Settings Conclusion

Many types of diets and food preparation techniques are repeated daily without variation for the duration of the study,
used in clinical and metabolic investigations. Some diets offers the most rigid control and can be designed to accom-
must be strictly controlled in order to detect small changes modate the study of various nutrients (1). Liquid formula diets
in outcome variables. Other diets may provide more flexible have been particularly useful for metabolic balance studies in
eating patterns and acceptable food choices for participants. clinical research. Chapter 14, ‘‘Planning and Producing For-
Each has its place in nutrition research, but each has advan- mula Diets,’’ contains a comprehensive description of formula
tages and disadvantages that must be recognized and eval- diets and their application in human feeding studies.
uated. This chapter describes the various types of research
diets and their associated terminology, as well as the many Conventional Food Diets
factors that must be considered when study designers plan In contrast to liquid formula diets, mixed diets composed of
controlled diet studies. conventional food products permit a normal fashion of
eating and thus are generally more acceptable to study par-
TYPES OF RESEARCH DIETS ticipants. Conventional diets can be served according to sev-
eral meal plans, as described here.
Several classification systems can help to distinguish the A 24-hour menu consists of 3 to 6 meals repeated daily
types of research diets and dietary techniques that are used without variation for the duration of the study. Preparing and
for human feeding studies. serving the same diet day after day minimizes fluctuation in
nutrient composition, reduces the chance of error, and results
General Classification in less confusion on the part of laboratory and kitchen staff.
of Research Diets Metabolic balance studies usually require this type of diet
plan to minimize variability (2). Also, frozen metabolic or
Formula Diets constant diets that are provided to outpatients usually use
Formula diets are composed of basic foods or food com- this diet plan. However, the 24-hour menu plan may not be
ponents from a limited number of sources. A formula diet, suitable for long-term studies because of its repetitive nature.

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
 Planning Diet Studies 143

A rotation or cycle menu provides a specified number acid composition (eg, an enhanced proportion of behenic
of daily menus that are repeated in an established sequence acid). This load test meal is given at a designated time and
for the duration of the study. All of the menus have a similar blood is collected prior to and following the test meal to
nutrient content. Different menus have different foods, but determine clearance of plasma chylomicrons. Other exam-
each menu (and each food item within the menu) remains ples are the Oxalate Load Test Meal to test oxalate bioavail-
the same throughout the study. Menus typically are num- ability in foods (5) and the Calcium Load Test Meal to test
bered to indicate the order of rotation through the cycle. To calcium absorption in hypercalciuria (6).
avoid confusion and minimize the chance of error, all par-
ticipants are fed the same menu (eg, Menu Four) on the same
day, regardless of their dietary treatment or when they enter The Diet Classification System of
the study. Problems would arise, for example, if there were St Jeor and Bryan
‘‘rolling entry’’ into a study (ie, subjects starting the protocol
on different dates), and each participant began the study with The degree of control required for a diet study depends on
Menu One. study design, the type of study and its length, the emotional
Ad libitum diets, also called self-selected, free, or ha- and physiological tolerance of the participants, laboratory
bitual diets, can be used for certain research purposes. The and kitchen facilities, and the philosophy of the investigator.
food items and amounts eaten are documented in a food As St Jeor and Bryan (7) note, ‘‘The investigator and re-
record or food diary. This diet offers subjects the most va- search dietitian should be in agreement about the most ef-
riety, but it requires more computation time than other meal ficient and successful way to achieve desired study results.’’
plans in order to analyze the participants’ food records. In All too frequently there is confusion regarding the terms
some cases participants are asked to weigh each food item used to describe research diets. One of the most problematic,
to be consumed and to record the data, or the staff can weigh for example, is the tendency to refer to all diets as ‘‘meta-
the foods in the kitchen and enter the data immediately into bolic diets.’’ In an attempt to standardize the use of terms
a computer. Duplicate portions of each food item may be and define appropriate controls, St Jeor and Bryan (7) listed
collected and a composite of these foods analyzed for nutri- five basic types of diets—estimated, weighed, controlled-
ents of interest. nutrient, constant, and metabolic balance—used in clinical
research studies. These diets were named, described, and
Diets Used in Association with Clinical or classified based on various dietary techniques used for cal-
Diagnostic Tests culation, measurement of diet, food source, water source,
Test diets are meals of exact composition that are used to preparation procedures, food refusals, and need for labora-
prepare the participant for specific types of diagnostic or tory analysis.
research-oriented tests. These meals may be provided in an
inpatient or an outpatient setting. One to 3 days’ worth of Estimated Diets
meals may be given to a participant to eat at home prior to The estimated diet relies on a visual estimation of a subject’s
coming to the research center for testing. For example, meals food and fluid intake, which is then calculated for individual
of specific carbohydrate content may be served 1 day prior nutrients. This diet is the easiest to administer because it is
to an oral glucose tolerance test. Other examples of test diets closest to the self-selected or free diet. However, it is the
include vanillylmandelic acid excretion tests; 5-hydroxy- least reliable because it estimates the participant’s intake
indoleacetic acid (5-HIAA) and serotonin restricted diets, fat from equivalency lists or approximate nutrient values for
malabsorption test diets; and high-calcium diets to define foods measured in usual household (or hospital) portions.
calcium intake in screening for hypercalciuria (3). Another These diets may be planned by the research dietitian but are
test diet, one that has a specific aluminum content and is typically provided from a foodservice system other than the
devoid of citrus, is consumed for 24 hours before deferox- research center’s facility (eg, the hospital food service).
amine infusion, a procedure for indirect measurement of alu-
minum stores (4). For some tests, a single meal of specific
Weighed Diets
composition is required on the day of the test. For example,
one meal of known magnesium content may be necessary The weighed diet may also be served by the hospital food
for a 1-day study that requires intravenous administration of service; but food portions are weighed as served to the par-
magnesium followed by collection of blood and urine. ticipant, and refuse is reweighed on return to the research
Nutrient levels in test diets are not generally excessive kitchen. The difference in intake is then charted and nutrient
in any given meal. Conversely, load test meals contain a intake calculated. This type of diet may be useful for mac-
specific nutrient(s) that may be excessive in relation to the ronutrient studies testing effects of protein, fat, or carbohy-
amount typically consumed at any single meal. However, drate intake.
the load test meal does not have a nutrient amount that would
be considered excessive for an entire day’s intake. One ex- Controlled-Nutrient Diets
ample is the Fat Tolerance Test Meal, which contains a des- The controlled-nutrient diet requires study nutrients to be
ignated amount of fat (eg, 50 g) and typically a specific fatty maintained at a constant level throughout the study. Cooked

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
144 The Dietary Intervention

food items are weighed prior to consumption. If foods are can be tested only when the day-to-day coefficient of vari-
refused, they are weighed when returned, nutrient compo- ation in nutrient content is very small; other hypotheses can
sition of returned food is calculated, and nutrients are re- tolerate more oscillation. Therefore, when the study is being
placed. planned, the levels of control provided by the different types
Controlled-nutrient diets are more demanding of the of research diets should be evaluated and matched to the
dietitian’s time because nutrients must be maintained at the requirements of the protocol. Investigators may find that the
same level throughout the study. Accurate calculations of most rigorous levels of control are needed for only short
daily menus are required, and food preparation procedures periods of time, and that a constant diet can be first preceded
are controlled (but not as strictly as for the constant and and then followed by a less burdensome, less expensive es-
metabolic diets). This type of diet may be used for long-term timated or weighed diet.
studies requiring a specific diet pattern but not rigidly con- The degree of dietary control depends on specific re-
trolled nutrient intake. quirements of the study. Investigators adapt these classifi-
cations to accommodate the setting and the nutrient control
necessary for the study, frequently combining or overlapping
Constant Diets and Metabolic Balance Diets the methods to control intake.
Constant diets and metabolic balance diets are the most rig-
orous forms of the controlled-nutrient diet. Procedures for
diet calculation, food and beverage procurement, food prep- Phase Components of Diet Studies
aration, patient instruction, and verification of nutrient con-
tent can be essentially identical for these two types of diets; Results from an experimental diet are often compared to
they are equally demanding of the research kitchen. The dis- those measured during a baseline period. This is typically a
tinct terminology indicates the type of research design for period just prior to the dietary treatment during which par-
which the diets are used. Metabolic balance diets are used ticipants consume their usual (ie, self-selected) diets. Diet
specifically for one type of research: nutrient balance studies records are often taken during the baseline period so that
(described later). In contrast, a wide variety of research de- outcome variables measured in association with the baseline
signs, as well as diagnostic tests, may require a constant diet, period can be related to the habitual diet of the study popu-
which achieves constant (ie, minimally variable) nutrient in- lation. This use of the term baseline period should not be
take through particularly close attention to food sources and confused with the baseline period required by some drug
preparation techniques. trials, in which the diet is kept constant throughout each of
The metabolic balance diet usually requires that nutrient three phases: the ‘‘baseline period’’ (no drug), the ‘‘trial pe-
data correspond with specific time periods (or phases) des- riod’’ (drug treatment), and the ‘‘off period’’ (no drug).
ignated for collecting biological materials. The metabolic A run-in diet or stabilization diet is an initial diet phase
balance defines net gain or deficit of a particular nutrient or that provides a standardized intake for all participants. With
nutrients. The balance is said to be positive when dietary this design, all participants begin intervention from the same
content exceeds loss in urine and feces and negative if losses diet; and the effects of any one eating style are minimized
in urine and feces exceed dietary intake. For a more detailed prior to feeding the experimental diets. The run-in diet is
discussion, see Chapter 16, ‘‘Compartmental Modeling, often patterned after current dietary patterns (eg, a typical
Stable Isotopes, and Balance Studies.’’ American diet) or a reference diet (eg, the Step 1 diet of the
The fundamental principle involved in the metabolic National Cholesterol Education Program). This period adds
balance study is that of the ‘‘single variable.’’ By controlling another diet dimension for statistical comparison with sub-
all dietary intake and keeping the diet as constant as possible sequent feeding periods. (See Chapter 2, ‘‘Statistical Aspects
from day to day, any changes in balance may be attributed of Controlled Diet Studies,’’ and see Chapter 6, ‘‘Recruit-
to the disease, procedure, medication, or nutrient under ment and Screening of Study Participants,’’ for a discussion
study. The nutrient content of a metabolic balance diet is of use of the run-in diet during screening.)
assessed from a food composite made by preparing a dupli- When a response that is measured during one diet period
cate portion of the subject’s food and beverages for a single is influenced by the previous diet, there has been a carryover
day. The reliability of metabolic balance data depends on ca- effect or diet order effect. Preventing carryover effects is
loric constancy and weight maintenance, a diet that achieves commonly accomplished by incorporating washout periods
the desired nutrient composition, a fixed volume of food and between experimental diet periods. Washout periods are
fluid intake, stabilization on the diet, accurate urine and fecal nontreatment diet periods used to eliminate or minimize the
collections, and control of physical activity and other stresses. nutrient carryover effect. Washout diets may be controlled
Because of its rigorous demands, metabolic balance studies diets that are patterned after the typical American diet, or
are best done in the setting of a research center equipped they may be the participants’ self-selected diets. Although
with a research kitchen and an analytical laboratory (8). The washout periods between study diets can be used to reduce
study design determines the acceptable level of variability carryover effects, they may not be required when diets are
in the research diet and, accordingly, the necessary degree randomly assigned and dietary treatments are of sufficient
of control and ‘‘constancy.’’ For example, some hypotheses duration. In other words, the diet period must be long enough

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
 Planning Diet Studies 145

for changes that had been induced by the previous diet to Inpatient and Outpatient Settings
cease (Table 10-1) and for changes associated with the new
diet to take effect. (See Chapter 2, ‘‘Statistical Aspects of Research that has a feeding component can be undertaken
Controlled Diet Studies.’’) in different settings. Residential feeding studies where con-
Depletion diets and repletion diets may be used to de- tinuous supervision is possible are most often conducted in
termine nutrient requirements or to determine the optimal hospitals, but boarding schools, convents, and monasteries
effective dosage of specific medications. For example, a diet have also been used. Prisons, although theoretically well-
low in magnesium, sodium, and potassium may be required situated for feeding studies, are usually not suitable; com-
to deplete blood levels of these nutrients in preparation pliance is likely to be poor, and the element of coercion
for comparing the effectiveness of three preparations of cannot be ruled out. (In some states, there is legislation that
potassium-magnesium citrate in correcting hydrochlorothia- specifically prevents the enrollment of prisoners in research
zide induced hypokalemia and hypomagnesemia. protocols.)
The advantages and disadvantages of all types of study
settings must be carefully considered during protocol devel-
PLANNING CONSIDERATIONS opment. Balance studies or studies requiring frequent mea-
Whether evaluating the feasibility of an existing protocol or surements are best conducted on an inpatient basis because
planning a diet study from conception, the research dietitian it is easy to control exercise, supervise food and fluid intake,
and the principal investigator should give careful consider- collect urine and stool specimens, and carry out timed blood
ation to physical, financial, and human resources as they collections. During inpatient studies, subjects have fewer
relate to the number, length, and order of dietary treatments; temptations to eat unauthorized foods, often share a spirit of
the range of calorie levels; and the recruitment of partici- camaraderie with other subjects, and find it advantageous to
pants as inpatients or outpatients. Among the research team have everything prepared for them. Investigators can easily
members, the dietitian is the best judge of how intricate the verify compliance.
feeding phase can be without creating error-prone situations Outpatient studies, on the other hand, are much less ex-
in production and delivery of meals. pensive to conduct and much less disruptive to the subjects.

TABLE 10-1
Length of Feeding Periods or Washout Periods for Selected Nutrients
Nutrient Amount per Day Outcome Variable Study Population Study Length Reference
Calcium 400 mg Steady state or hypercalciuria Normal adults 4d 9
Calcium 400 mg Intestinal adaptation Healthy adults 4 wk 9
Calcium 400 mg Steady state Adults with osteomalacia, 3d 10
osteoporosis, or
hypercalciuria
Calcium 300 mg Calcium retention and Adult women 2 wk 11
stabilization of PTH and
1,25-(OH)2-D3
Chromium ⬍20 lg Impaired glucose tolerance Adult men and women 4 wk 12
Chromium 200 lg Glucose and insulin Adult men and women 4-8 wk 12, 13
Chromium 250 lg Blood lipids Adult men and women 7 mo 14
Fats ⬍30 en% Lipoprotein levels Adults over 40 yr 4 wk1 15
Fats 40 en% Lipoprotein levels Adult men and women 6 wk2,3 16
b-carotene 30 mg Carotenodermia Healthy men 25-42 d 17
Fish oils 10 g Stabilization of platelet and Health adults 3 mo 18
MaxEPA erythrocyte phospholipid
fatty acids
Fish oils 5 g fish oil Ⳮ Persistence of x-3 fatty acids Healthy men ⬎18 wks4 19
1 g 20:5 n-3 Ⳮ in erythrocyte membranes
1 g 22:6 n-3 after discontinuation of x-3
fatty acids

Rationale or implication:
1
Plasma lipids, apolipoproteins, and fatty acid patterns stabilized by week 4 and remained constant thereafter through week 24.
2
Lipoprotein levels continued to change between week 3 and week 6 of feeding.
3
Double the turnover time of the slowest LDL component, which is about 3 weeks (20, 21).
4
Longitudinal rather than crossover studies are advised because of the protracted washout periods (⬎18 wk).

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
146 The Dietary Intervention

Often the participant pool is larger for outpatient studies be- • Select foods that will be readily available throughout the
cause people can go about their lives fairly normally. The study. Some foods are seasonal and are available only at
nature of the study as such may dictate that outpatient studies a high cost during off-season. For example, although fresh
are more suitable because people are allowed a ‘‘more strawberries may be available in the market in November,
normal lifestyle.’’ For this reason, the results may also be they will cost too much to comply with many purchasing
more generalizable to the population at large. contracts.
Present-day research goals do not always require a re- • Use foods that are easy to reheat in the containers in which
strictive inpatient setting. Confinement to a metabolic ward they are packaged; for example, mixed dishes reheat better
is almost always a major interference with the lifestyle of than do whole pieces of meat. It may be necessary to pur-
the research participants; and, for some studies, it is not in chase microwave ovens for participants who do not have
the best interest of the research to disrupt the participants’ them.
daily activities and social habits. Also, inpatient studies are • Use foods that do not require multiple transfers from con-
expensive. The high cost of human feeding studies is in- tainer to pan to dish. This may result in losses from spills
creasingly important to funding agencies, which are encour- or from food adhering to multiple surfaces.
aging scientists to develop more efficient and cost-effective • Make adjustments to accommodate routines of the various
research approaches. One of these approaches is to conduct volunteers. Although dinner trays for residents may be de-
studies using free-living participants. livered at a specified time, it may be necessary for a facility
Outpatient nutrition studies should be considered when to stay open for a window of time, perhaps 4:30 PM to
there is no other reason to keep subjects confined to the 6:30 PM, to accommodate work schedules of free-living
research unit except to provide a ‘‘constant’’ diet, when the adults.
costs of hospitalization or institutional overhead are a con- • Pack take-out meals in insulated coolers with sealed, re-
sideration, or when the participant is unable to reside on the freezable ice packs to keep foods cold during transport.
unit. A combination study is often the ideal; meals are pro- On the other hand, a bag lunch for a salesperson or student
vided on an outpatient basis for most of the study and then may require a thermos to keep stew hot.
an abbreviated inpatient phase is employed for critical data • Be sure the participants have adequate refrigerator and/or
collection or testing procedures. freezer space to store food and beverages for take-out
Outpatient studies are more difficult than inpatient meals. As a rule, no more than 2 to 3 days’ worth of fresh
studies to standardize and follow because subjects are living food should be provided at once. Meals, snacks, and bev-
out of the ‘‘controlled environment.’’ To ensure compliance, erages for a single day often fill a grocery bag; a 2-day
some units require that all subjects eat 2 out of 3 meals per supply fills a large cooler. The ability to understand food
day in the research dining room, 5 to 7 days a week. This storage safety should be one of the selection criteria for
gives the participants frequent contact with the research die- participants, and adequate cold storage space is funda-
titian, dietary staff, nursing staff, investigators, and other mental to this issue.
subjects and keeps motivation high. (Also see Chapter 7,
‘‘Managing Participants and Maximizing Compliance.’’) Estimating energy requirements correctly and main-
Outpatient studies are labor intensive for the research taining body weights during an outpatient study are crucial
kitchen staff because meals must be safely packed for trans- to the success of most projects. (See Chapter 17, ‘‘Energy
port, storage, and convenient reheating. Specific considera- Needs and Weight Maintenance in Controlled Diet
tions for planning outpatient studies are listed here: Studies.’’) If weight maintenance is important, frequent in-
teraction with the dietitian may be critical. This includes
• If possible, plan a cycle menu of at least 2 to 3 days to frequent weighing as well as monitoring of exercise patterns,
give variety. A 1-day menu is common for short studies; sometimes daily but usually once or twice per week.
however, some variety is best for outpatient studies. Outpatient studies require more detailed participant in-
Longer menu cycles are generally used with longer out- struction than when participants are provided all the study
patient feeding studies (see Length of Menu Cycles and food on site. ‘‘Free foods’’ and any limits on these items
Feeding Periods). need to be specified. Clear, written instructions about safe
• Select foods and recipes that are easy for the research food handling, storage, refrigeration, and reheating of foods
kitchen to prepare in bulk and store. Foods that are easy to should be given along with directions for handling emer-
prepare, freeze, store, and reheat and that are acceptable to gency situations.
most people include pasta with sauce (prepare and weigh
pasta, sauces, and meat separately), meatloaf, Salisbury
steak, hamburger patties with gravy, cookies/cakes/muffins Multiple Dietary Treatments, Multiple
and brownies (frozen brand names may be appropriate), Calorie Levels, and Variety of Food
rice, mashed potatoes, baked potatoes, and noodles. Items
• Use foods that ‘‘travel’’ well; eg, canned and frozen prod-
ucts. Limit the use of fresh produce because of the risk of A study providing a common background diet to all partic-
spoilage. ipants is easier to conduct than a study with multiple dietary

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
 Planning Diet Studies 147

treatments. Extra control measures must be implemented to calorie levels, typically of 200-kcal to 400-kcal increments;
avoid potential mix-ups when multiple dietary treatments are each participant is assigned to the calorie level that most
delivered concurrently. In practice, this may mean that less closely matches his or her estimated requirements. Partici-
experienced foodservice workers could be assigned to pre- pants’ calorie level assignments are then changed, if neces-
pare foods that will be served to all participants regardless sary, to maintain their initial body weights. This method is
of treatment, whereas a reliable and experienced employee easily used in conjunction with a modified cafeteria style of
might be assigned to prepare or portion foods that are ve- foodservice. Participants proceed down the food line se-
hicles for those nutrients critical to the study. There are many lecting designated food items according to their dietary treat-
ways to accurately prepare and deliver multiple treatments, ment and calorie level.
but they all require a high degree of organization and su- Menus are typically calculated at an ‘‘average’’ calorie
pervision. For example, in some facilities one employee may level, then scaled so that each food item is fed in proportion
be assigned the task of portioning and labeling muffins of to calorie requirements. This means that food items from the
varying composition (but identical appearance) for four dif- average calorie level are of normal serving sizes. Those at
ferent diets. Another facility may prefer to have different the extreme ends may look unusual; for example, a person
foodservice workers portion and label each type of muffin. eating foods of a low calorie level may receive many food
In both situations, records should be maintained to identify items but less of each item than he or she is accustomed to
all individuals involved in the preparation and portioning of eating. Serving unusual portions of food is a particular
each food item so that any procedural error can be traced problem when large men and small women are included in
and corrected. the same study. For example, if menus are written so that
As the number of individual food items and the com- small women can meet their calcium requirement, then large
plexity of the items increases, more time is necessary to pre- men may receive quantities of dairy products that they find
pare and portion these items. Many facilities find it more objectionable.
efficient and less costly to purchase commercially prepared Some studies have a fixed enrollment date for all vol-
food items than to produce equivalent items. This option unteers. Others may have a ‘‘rolling’’ or sequential admis-
should be investigated within the context of the facility as sion, in which the subjects may be engaged in different
well as that of the study protocol. When foods must be pre- phases of the study at any given time. The latter strategy
pared at the facility, it is advisable to use relatively simple facilitates recruiting participants and has the advantage of
recipes to ensure accuracy in production. Accuracy of de- keeping the number of participants manageable. It poses
livery can be improved by limiting the number of items to problems associated with consistency of the dietary treat-
be distributed for a given meal. Errors of omission or du- ments across time, however, has the potential to confound
plication are easier to catch during the tray-check process the study.
when trays are not cluttered.
Although the amount of work involved in food prepa- Concurrent Studies
ration and distribution would seem to be directly related to
the number of study participants being fed at a given time, Extra precautions are required when multiple studies are
this relationship is not one of direct proportion. Much of the conducted simultaneously in a facility. By coordinating ac-
effort of producing and portioning is in setup and cleanup tivities and planning for space and equipment needs, con-
rather than in cooking and weighing. There is consequently flicts about scarce resources can be avoided. To ensure op-
an economy of scale associated with increasing numbers of timal use of the facility, this extra coordination should take
study participants—to a point. The limit of this economy of place among the principal investigators of the studies as well
scale is reached when the burden of additional participants as among members of the dietary staff and should occur
begins to confuse operations. early in the planning process. In some facilities each study
The number of participants in a study has an effect on is assigned separate counter space, cooking equipment, re-
the number of calorie levels that must be used. When each frigerators, freezers, serving areas, and dining areas, and
serving of food is weighed individually in proportion to en- little is shared except equipment. In other facilities, all
ergy requirements, portioning is made more complicated by studies must share all resources. This means that facility-
greater numbers of treatments, participants, and calorie wide quality control measures must be instituted to ensure
levels. In studies having few participants, the use of individ- proper delivery of test foods to the intended subjects.
ually tailored calorie levels is common; the number of cal- Tracking expenditures for each study is more complex
orie levels to be prepared may equal the number of partici- when more than one study is in process. The dietary staff
pants in the study. The energy requirement of each will most likely purchase food and supplies in common for
participant is assessed as accurately as possible, and daily all ongoing studies and subsequently allocate costs to each
menus are prepared to meet these calorie requirements. Typ- study. Although this process makes bookkeeping more dif-
ically, the kitchen staff assembles an individualized tray for ficult, food costs may be lower if requirements for quantity
each participant. discounts are met by combining orders.
With larger numbers of participants, it is more common Finally, subject morale can be affected when partici-
to use calorie increments. Menus are prepared at designated pants with high and low levels of study ‘‘burden’’ share a

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
148 The Dietary Intervention

single dining facility. It is only natural that participants in Seasonal effects are another factor to consider in sched-
differing studies will compare their protocols for degree of uling diet studies. Nutrients in specific foods can change
effort in relation to benefits. with the season (an example is the carotenoid and vitamin
C content of summer vs winter tomatoes). Outcome vari-
ables may be affected by colds and flu, which are common
Length of Menu Cycles and in winter, or by allergies, which are commonly induced in
Feeding Periods spring and fall by pollen. The summer months are problem-
atic for conducting feeding studies because of changes in
The decision to use a long or short menu cycle depends on physical activity and conflicts with vacations. Unwelcome
a number of factors specific to the facility and the study snow storms may interrupt diet studies at critical points of
design. However, the length of the study is a major factor in dietary intake and sample collection. Elderly subjects may
determining the length of the menu cycle. Short-term met- be less interested in studies conducted in winter months be-
abolic balance studies typically use menu cycles of 3 days cause of anticipation of poor driving conditions and the
or less. Long feeding studies generally use longer menu cy- safety issues that accompany shortened daylight hours.
cles of 1 or 2 weeks to add variety and make the diets more
acceptable. The disadvantages of long cycles include the
added time in menu preparation, the requirement for more
labor in food handling, and a greater variability in nutrient Tailoring the Study to the
content among menus.
Menu cycles frequently are based on 7 days, as in a Physical Facility
7-day or 14-day menu cycle. An advantage of this system is Architectural plans for a metabolic research kitchen must
that certain menus can be tailored to the special requirements address a variety of special needs. Ideally, these include
of carry-out meals (eg, weekend meals). These menus are cooking facilities that allow for multiple modes of prepara-
often written with convenience and food safety issues in tion; workstations that are efficient and well equipped for
mind. Some dietitians and investigators, however, prefer weighing, packaging, and preparing food; storage space for
menu cycles that are based on a number other than seven. refrigerated, frozen, and dry goods; and dishwashing facili-
This approach has the advantage of avoiding the predicta- ties. Office space, including computer workstations for die-
bility and boredom that are inherent when menus are stan- titians and technicians, as well as a dining room for outpa-
dardized to the day of the week (eg, ‘‘It must be Tuesday; tients, may also be included. A setting that provides for all
ham sandwiches for lunch again!’’) of these needs can successfully accommodate multiple
The length of the feeding periods is selected according studies and large numbers of subjects consuming complex
to estimates of how much time is needed for the outcome diets for long periods of time either at the facility or as out-
variable of interest to stabilize in response to a change in the patients.
nutrient variable of interest. This time period may depend Rarely, however, is this ideal setting available. Usually
on the level of the nutrient in the diet; the degree of change the study must be tailored to an existing physical facility.
from the previous diet; the level of other dietary compo- (Also see Chapter 19, ‘‘Facilities and Equipment for the Re-
nents; and characteristics of the study population, including search Kitchen.’’)
nutritional status, age, gender, and body composition. Table Because space restrictions lead to equipment limita-
10-1 provides general estimates of the typical length of tions, frequently only one piece of each type of equipment
feeding periods for various nutrients and associated outcome is available. Consequently, contingency plans must be avail-
measurements. (Note: There is seldom universal agreement able at all times to provide for emergency situations such as
on the appropriate length of feeding any particular nutrient breakdowns, repairs, and replacement of equipment. Most
for any outcome variable.) research kitchens that do on-site cooking should have both
a conventional range/oven and a microwave oven. A double-
Scheduling Diet Studies oven household range can provide assurance that if one oven
breaks down, the other will still be in working order. If a
In scheduling diet studies, it is prudent to avoid holidays that microwave oven breaks down, the range and oven can pro-
might interfere with the normal feeding of participants. In vide a method of reheating. A portable hotplate can also
the United States, this generally rules out the time period serve this purpose. If a dishwasher is out of order, paper
between Thanksgiving and New Year’s Day. For Lent and supplies should be readily available for meal service. Most
Passover, it is sometimes possible to calculate appropriate institutions are required by their state public health depart-
alternate meals that will be acceptable to participants ob- ment to meet sanitation codes governing dish washing and
serving these holidays. Local events should be taken into do not allow washing dishes by hand without a disinfectant.
account (eg, Mardi Gras in New Orleans). Dates that are Alternative refrigeration and freezer space should be iden-
likely to be important to a particular study population should tified for emergency food storage when such equipment fails.
be considered (eg, college breaks in a population of young Ideally, important refrigerators and freezers will be con-
women). nected to an emergency generator.

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
 Planning Diet Studies 149

To conduct successful feeding studies under circum- becomes available for the delivery and storage of food for
stances of limited storage space, it may be necessary to study the afternoon or evening pickups. To use a work schedule
fewer participants at a time, to extend the study over a longer of this kind, perishable foods and some precooked foods
time period, and to offer fewer food choices. As long as may need to be procured from another foodservice area and
sufficient quantities of key food items can be stored for the brought to the unit on a meal-by-meal basis. These arrange-
participant at any given time, intrasubject variability in nu- ments can frequently be made through the institution’s food
trient intake over long periods of time can be minimized. In services.
addition, chemical analysis of diet composites can provide Ultimately, the limitations of space, equipment, and per-
a measure of variability in nutrient intake across time. sonnel within a research kitchen must be addressed individ-
Meals may also be simplified so that only those foods ually with emphasis on the ability to maintain accuracy of
containing key nutrients are kept in storage. Other food items food preparation and service and continuity of function by
may then be procured as needed. For example, in a study a trained staff. Careful consideration of the necessary dietary
investigating the effects of vitamin C on the incidence of controls will address most scientific and logistical issues in
colds, the diet can be designed so that frozen orange juice is advance. Often, food preparation procedures can be simpli-
the only significant source of vitamin C in the diet. Frozen fied, resulting in improved efficiency and accuracy. With
orange juice would then be purchased in a single lot shipment thoughtful creativity and advance planning, well-controlled
and would be the only food needing storage for the length of feeding studies can be conducted in almost any facility.
the study. The remainder of the diet, composed of foods low
in vitamin C, could be replenished on a regular basis.
Food procurement, preparation, storage, and delivery Pilot Studies
from another foodservice department within the institution
can decrease storage needs and reduce on-site food produc- A pilot study is a small-scale study that is conducted to fa-
tion. This arrangement can be particularly helpful if cooking cilitate the successful completion of a full-scale study. The
facilities and personnel are limited. Under these circum- many reasons for conducting pilot studies include: gathering
stances diets might be designed around portion-controlled, preliminary information prior to committing scarce re-
precooked items that may even be available as frozen entrees sources; assessing the feasibility of a feeding regimen; and
and preportioned foods such as juice, cereals, and snacks. It honing techniques for the analysis, storage, and distribution
is necessary, however, to evaluate in advance whether the of foods and samples. Pilot studies also are used to refine
composition and weight of such food items are sufficiently protocols as well as to make comparisons and choose among
accurate and constant for the purposes of the study. alternative approaches.
This approach has been effectively used by McCullough It is important that treatments, subject populations, and
et al (22) in studying electrolyte balance. Inpatients con- study conditions used in a pilot study closely parallel those
suming rigorously controlled constant diets were compared planned for the subsequent, full-scale study. Investigators
to outpatients consuming less exactingly controlled diets must resist the tendency to use inadequate dietary control
consisting of manufacturers’ preportioned foods and other groups or treatments; controls are as essential for inter-
foods prepared and weighed in the hospital’s central food preting the results of a pilot study as they are for interpreting
production area. Analyzed diet composition for electrolytes a full-scale study. In addition, the subjects participating in a
and for 24-hour urine collections showed no statistical dif- pilot study should resemble those who will participate in the
ference between the two methods of dietary control. subsequent, full-scale study; otherwise, the data gathered
The number of employees that can effectively and ef- may be misleading. For example, procedures and menus
ficiently work in a given work area is likewise governed planned for a study in elderly individuals should be pilot-
by the amount of space and equipment available. Simpli- tested on persons from that age group, not on young students
fying the meals, food selections, and meal rotations can de- or middle-aged adults.
crease the number of food items that need weighing and For the dietary staff, recipe development is an important
preparation. This can decrease personnel requirements. The aspect of pilot studies. Some recipes that initially appear to
maximum capacity for the types and numbers of research be successful may not hold up under the close scrutiny of a
diets produced should be established in advance. Creative pilot study. During recipe development, the food is taste-
staffing patterns, sometimes extending into the late evening, tested for quality and acceptability with several people
can address the problems of space limitations and increased eating a whole portion as part of a meal. Tasting a spoonful
production needs. The use of temporary personnel can also is often not enough; eating the same amount that the partic-
effectively meet the study’s needs for short periods of time. ipants are expected to consume can lead to a different eval-
The efficiency of a research kitchen is maximized in uation of the food item. Portions of the recipe are stored as
studies that supply food for home consumption because the the food item will be for the study (refrigerated/frozen) and
meals do not have to be prepared and served three times a reheated at a later date to evaluate how these treatments af-
day. Food pickups by the participants can be arranged for fect appearance and quality of the product. Any changes that
both morning and afternoon to free up refrigeration space. are made to the recipe are incorporated, and the evaluation
Once the morning food is picked up, refrigeration space then process is repeated.

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
150 The Dietary Intervention

To test for potential problems associated with scaling for • Data from the pilot study can be used in a power calcu-
quantity preparation, the recipe is made in amounts that will lation to estimate the number of subjects required in the
ultimately be used for the study. All procedures and details of full-scale study.
preparing the recipe are documented. The insights of kitchen • Data from pilot studies will often convince a funding
staff can be tapped to assess the ease of preparation, avail- agency to support a major study. Data from pilot studies
ability of foodstuffs, and other pertinent considerations. are often publishable.
In preparation for the pilot study, the kitchen staff,
nursing staff, laboratory staff, and study coordinator develop
study-specific flow sheets, ordering forms, and quality con- Estimating the Cost of Diet Studies
trol guidelines. (See Chapter 3, ‘‘Computer Applications in
Controlled Diet Studies,’’ and Chapter 18, ‘‘Documentation, Producing research diets is an expensive endeavor—so
Record Keeping, and Recipes,’’ for examples of forms.) As much so that their cost is a major consideration in the design
with a full-scale study, a statistician should be included on and implementation of human feeding studies. Injudicious
the research team. cost-cutting maneuvers, however, can jeopardize the integ-
rity of a study. The research dietitian thus walks a fine line
With these steps completed, the pilot study can begin.
in an effort to contain expenses while committing the re-
At this time, many aspects of the study can be fine-tuned,
sources required to achieve the primary goal of producing
comments collected, and data gathered. Besides cooking and
high-quality diets of specified composition.
tasting recipes, serving the food, and checking the compo-
Various feeding patterns have been successfully used in
sition of menus, many other activities and procedures can
research studies, depending on the level of control required,
be pilot-tested. These include all laboratory analytical tech-
the facility, and the research question to be addressed. As
niques, sample processing and storage procedures, record-
shown in Table 10-2, studies that require a high degree of
keeping procedures, forms, interviewing techniques, recruit-
dietary control usually are more expensive. Such studies typ-
ment strategies, nursing techniques such as blood pressure
ically require longer hours of facility operation, more su-
measurements, and data processing. Pilot studies can be ex-
pervision of research participants, and more personnel, all
tremely useful, not only because many preliminary discov-
adding to the cost of conducting a study. The goal is to pro-
eries are made at this step, but also because the design, diet
vide the degree of dietary control required for a particular
plan, or procedures can be ameliorated before sponsors in-
research question without spending additional resources
vest effort and money on a larger scale study.
when control is not required.
A final evaluation of the pilot study is done at this stage
Liquid formula diets probably are the most cost-
and plans are revised. If enough participants were studied
effective of all the research diets. However, ingredients, even
and the outcome variables were measured after a time course
when purchased in large quantities, are still expensive. The
similar to the one planned for the major study, data collected
food costs for some liquid formula diets may match the food
during the pilot study may be used by the study’s statistician
cost of a natural food diet, but the labor cost likely will be
in a power analysis. This calculation can provide valuable
much lower. For example, in a typical facility an employee
data to project the number of participants that will be re- can prepare 80 liters of a single formula each day that sup-
quired for the full-scale study in order to determine statis- plies 64,000 kcal. This is enough to feed a participant for a
tically significant differences among treatments. month, and there is little waste.
There are many advantages to conducting a pilot study: The cost of producing research diets varies among
• Data collected is likely to predict the success of a full- studies and among facilities but is based, in large part, on
scale study. the complexity of the study, including level of labor; number
• Rough spots and failures in the recruitment process, in the of meals served and hours of facility operation; the type of
kitchen, and in the research laboratory can be corrected diet, including the expense of dietary components; the nature
prior to outlay of major resources. of the measurements made on participants; and payments
• New staff members can be trained under less pressure than and other charges associated with enrolling study partici-
during the full-scale study. pants.
• Recruitment strategies for subjects can be improved.
• Recipes and menus can be improved based on input from Labor
study participants and kitchen staff. Labor is the major operating expense for most feeding fa-
• Sample handling and techniques for laboratory analysis cilities. A major responsibility of the research dietitian is to
can be evaluated. assess, within the framework of each new study, the skills
• Foods can be composited and chemically analyzed to pre- and the time requirements of each position and to make as-
dict accuracy of nutrient content for the full-scale study. signments accordingly. Additional employees frequently are
• The number of persons, the time commitment, and the cost needed for days or hours of peak workload; it can be less
required to carry out the full-scale study are more easily expensive to make appropriate use of part-time and overtime
predictable. labor than to hire additional full-time employees. Some fa-

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
 Planning Diet Studies 151

TABLE 10-2
Direct Relationship between Degree of Dietary Control and Expense of Conducting a Study
Level of
Dietary Control Cost Features
Highest Highest • Metabolic ward
• Formula and conventional food diets
• All meals supervised
• No alcohol
• All foods analyzed for selected nutrients (23)
High High • Free-living participants
• Conventional foods
• Two supervised meals Monday–Friday
• Lunch and weekend meals packed for take-out
• No alcohol
• Diet composites analyzed for selected nutrients (24)
Medium Medium • Free-living participants
• Conventional foods
• Two supervised meals Monday–Friday
• Two self-selected weekend meals (within guidelines)
• Alcohol consumption allowed within limits
• Diet composites analyzed for selected nutrients (25)
Medium Low • Free-living participants
• Self-selected diets
• Food records
• Fiber-rich cereals prepackaged for home consumption (26)
Low Lowest • Free-living participants
• Self-selected diets based on flexible, individualized diet plan
• Food records
• Food frequency questionnaire
• Diet counseling (27, 28)

cilities successfully employ students, especially college- In some facilities, free-living participants perform lim-
level dietetic or foodservice management majors, or sec- ited but time-saving functions as part of the study protocol.
ondary school students who are enrolled in work training For example, labor hours are saved by having participants
programs. The personnel department can provide advice rather than staff members assemble the items required for a
about institutional regulations and liabilities regarding the take-out lunch: participants can collect their lunch items
use of students and volunteers in the research kitchen. In along with breakfast items as they progress along a cafeteria
some settings the kitchen staff may be hired only in accord line, then pack lunch items for take-out. The time required
with strict union contract rules. for a staff member to check the accuracy of participants’
It often is advisable to plan the staffing budget on a year- selections can be considerably less than the time required
round basis even if the study participants are actively en- for a staff member to assemble the trays. This must be done
gaged in the feeding protocol for only part of the year. This without burdening the participants and without breaking the
recommendation acknowledges the complete time sequence blinding of the study. (Also see Chapter 20, ‘‘Staffing Needs
of the research protocol. For example, before the study be- for Research Diet Studies.’’)
gins the nutrition staff must be trained and must try out the
experimental menus to ensure they meet the study design Purchasing Food and Supplies
goals. There is a great deal of preparation work before and
between feeding periods. Also, if the investigation comprises Food
a series of studies, it is inefficient to lay off the trained staff The research dietitian must view food selection and pur-
and hire new recruits a few months later. Studies are not part chasing within the context of three major goals: (1) to control
of the usual hospital foodservice activity, and highly skilled, nutrient content of the diets, (2) to aim for participant com-
well-motivated workers are needed to ensure that the pro- pliance, and (3) to follow principles of cost containment
tocols are being followed correctly. within the parameters of the research protocol. This means

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
152 The Dietary Intervention

that food costs will vary considerably among studies, de- and during the conduct of the protocol (monitoring phase).
pending on the design. Ingredients for a simple formula diet The cost of chemically analyzing a commercially prepared
may cost as little as $5 per day, whereas the components for mixed dish such as baked lasagna may be less expensive
constructing whole-food diets with specially manufactured than the labor involved in preparing the dish. Many com-
ingredients may cost as much as $20 per day. The decision mercially prepared mixed dishes are highly standardized and
to purchase a brand-name cookie at a higher price than a companies typically provide nutrient analysis of the product.
nonbrand-name counterpart may be justified if it promotes (Some food companies are willing to share available nutrient
compliance or if nutrients are more consistent among lots. data that is not included on their nutrient data sheets.) How-
Similarly, a particular brand of reduced-sodium and reduced- ever, it is important to determine the variation among por-
fat salad dressing may best suit the requirements of the re- tions of commercially prepared dishes; within-batch or
searcher and the palates of the participants. among-batch variation may be unacceptably high.
To minimize waste and save money, the dietitian must
order each food in the quantities needed. This is no simple
Supplies and Materials
task, because each food item must be assessed separately
considering its rate of use, shelf-life, and stability once Just a generation ago, the term paper supplies referred
opened. Additionally, greater waste is associated with largely to paper napkins and cups, waxed paper, and alu-
weighed food portions. For example, when a hard roll or minum foil. Today supplies include a substantial proportion
bagel is portioned and precisely weighed, a large part may of Styrofoam, plastic wrap, and other disposable containers
be lost to maintain the aesthetic appeal of the served portion. and utensils. Although the cost of paper supplies varies
Buying in large quantities of single batch lots not only among facilities, as rule-of-thumb estimate, the expense of
ensures consistency in nutrient composition for the entirety paper supplies is approximately 10% of food costs.
of the study but also has the benefit of discounted prices if Disposable plastic and paper containers have become a
the supplier’s required minimum order is met. Although pur- mainstay in the preparation and service of research diets.
chasing in bulk can be cost-effective, it is not appropriate in Frequently, foods are weighed in advance of the study, por-
all situations. For example, when a recipe calls for 10 g more tioned into high-quality disposable containers, frozen until
than is supplied in an institutional pack of tuna, it is cost- ready for use, and then assembled on permanent ware for
effective to stock small cans as well. The amount of storage service in the hospital setting or kept in the original container
space available also influences whether items are purchased when a packed meal is ordered. Food needing to be reheated
in bulk. If storage space is limited, additional emphasis will prior to service can be microwaved in these same containers
be placed on selecting a food item that is readily available if plastic or paper is used.
and provides consistency of nutrients across lots. Although disposable containers are convenient, a high
Buying in bulk is an appropriate method of purchasing price is paid for their purchase and disposal. Disposing of
ingredients for liquid formula diets. However, special con- solid waste will be an increasing concern for foodservice
sideration must be given to the shelf stability of the ingre- managers as disposal costs accelerate. Studies are needed to
dients, including the fat and protein sources, which may de- document the amount and type of solid waste generated in
teriorate over time. This is particularly important for dried feeding facilities.
milk powders and other ingredients that undergo Maillard
(browning) reactions and other changes affecting taste. Budget Planning for Research Diet Studies
Storage represents another food-associated cost in diet As earlier discussions make clear, feeding studies create
studies. Bulk dry storage at room temperature requires large some unique budgetary pressures that other types of research
amounts of floor space, usually in locked rooms; the institu- are spared. Although categories vary according to institu-
tion where the study is conducted may charge for this space. tional policies, the general categories that must be consid-
Walk-in refrigerators and freezers may be needed. It also may ered when researchers develop the study budget are:
be necessary to rent space for bulk frozen or refrigerated items
in specialty warehouses. These facilities usually are located Personnel (includes salaries, wages, overtime, and fringe
at a distance from the research center. Beyond the rental fees, benefits)
additional travel and time costs are incurred when staff go to Principal investigator
refrigeration sites to retrieve the food. Coprincipal investigator
In deciding whether to make a food product ‘‘from Research dietitian
scratch’’ or to purchase a commercially available product, Medical officer
considerations for controlling nutrient content of diets over- Study coordinator
ride principles of cost containment. However, in some cases Secretary/administrative support
commercially available foods can be used appropriately and Cooks
offer a cost savings compared to food and labor expenses in- Foodservice workers
curred with the in-house production of a comparable product. Nursing support
It is wise to allot funds for food composition analysis Laboratory technicians
both before the study (during validation and pilot phases) Phlebotomist

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
 Planning Diet Studies 153

Statistician Although it may seem that costs in all categories would

Consultants be proportional to the number of participants enrolled, this
Administrative Expenses is not the case. Any study, regardless of size, requires a cer-
Travel (out-of-town meetings, local mileage) tain minimal number of staff; increments about this, how-
Telephone/other communication ever, may be irregular. The necessary level of labor for di-
Publications etetic and kitchen staff is particularly high; it tends to range
Shipping from 0.2 to 0.5 full-time equivalents (FTE) per participant,
Postage depending on the complexity of the protocol. In addition to
Office supplies the nutrition personnel, scientific, medical, statistical, man-
Facilities charges agerial, and laboratory effort is needed; when summed, this
Nutrient database software effort can vary from 2 FTE to 6 FTE.
Printing and photocopying Study subjects often receive a stipend or financial com-
Foodservice Expenses pensation for participating in the study. This is a common,
Food but not universal, practice. Some investigators believe that
Paper goods and disposables a financial incentive is a disservice to participants because
Kitchen supplies, equipment, and maintenance it can pressure them to continue with a study when they
Coolers/cold packs prefer to withdraw. This consideration is compounded when
Off-site freezer/refrigerator storage payment is partially or fully linked with completing the
Laundry, linens, and uniforms study. Others believe that the study will be taken more se-
Laboratory Expenses riously when there is payment involved and that this incen-
Equipment tive promotes compliance and willingness to complete long
Clinical laboratory tests studies. Offering a financial incentive also can broaden the
Laboratory supplies recruitment pool, which is often an important consideration
Diet composite analysis when participants with specific characteristics are being
Service contracts on laboratory equipment sought. (See Chapter 5, ‘‘Ethical Considerations in Dietary
Blood collection and processing supplies Studies,’’ for further discussion of stipends.)
Sample storage
Subject Costs
Subject payment
Recruitment costs/advertising CONCLUSION
Parking fees/public transportation/local mileage Controlled human feeding studies are labor intensive, ex-
pensive, and require a great deal of attention to numerous
The exact assignment of these costs to specific catego- details, but they can yield information that cannot be gained
ries (for example, is food a ‘‘supply’’ cost or does it fall by other types of research. Advance planning for all aspects
under ‘‘other’’ costs?) will vary among research centers, de- of human feeding studies can lead to cost savings, greater
pending on factors such as the policies of the funding agency accuracy in data collection, and a more harmonious setting
and the budget format preferred by the institution. for study participants as well as dietary and research staff.
The financial structure of studies differs among research
centers. For example, the nature of institutional support
varies considerably: some investigators may need to request
funds for nursing services that elsewhere may be provided REFERENCES
without charge. Another highly variable factor is the indirect 1. Aherns EH Jr. The use of liquid formula diets in meta-
cost or overhead rate. In some situations this is minimal (ie, bolic studies: 15 years experience. Adv Metabol Disord.
10%), whereas in others it can be high enough (ie, 80%) to 1970;4:297–332.
effectively double the cost of the project. 2. Sampson AG, Sprague RG, Wollaeger EE. Dietary tech-
Costs may be calculated on a per-study, per-subject, per- niques for metabolic balance studies. J Am Diet Assoc.
meal, or per-annum basis. This basis will vary with the cate- 1952;28:912–916.
gory: 3. American Dietetic Association. Handbook of Clinical
Dietetics. (Part H) Miscellaneous, Test Diets. New
• Small equipment (such as kitchen scales) is usually pur- Haven, Conn: Yale University Press; 1981:H15-H18.
chased once at the outset of the study. 4. Malluche HH, Smith AJ, Abreo K, et al. The use of
• Financial compensation or stipends would be paid in fixed deferoxamine in the management of aluminum accu-
amount to each participant. mulation in bone in patients with renal failure. N Engl
• Parking fees may be charged each time the participant J Med. 1984;311:140–144.
drives to the center for a meal. 5. Brinkley LJ, Gregory J, Pak CYC. A further study of ox-
• Salary support for a principal investigator usually repre- alate bioavailability in foods. J Urol. 1990;144:94–96.
sents a percentage of full-time effort per year. 6. Pak CYC, Kaplan RA, Bone H, et al. A simple test for

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.
154 The Dietary Intervention

the diagnosis of absorptive, resorptive, and renal hyper- 18. Sanders TAB. Influence of moderate intakes of fish oil
calciurias. N Engl J Med. 1975;292:497–500. on blood lipids. In: Lands WEM, ed. Proceedings of the
7. St Jeor ST, Bryan GT. Clinical research diets: definition AOCS Short Course on Polyunsaturated Fatty Acids and
of terms. J Am Diet Assoc. 1973;62:47–51. Eicosanoids. Champaign, Ill: American Oil Chemists’
8. Brinkley L, Pak, CYC. Metabolic balance regimen and Society; 1987.
nutritional aspects of clinical research. In: Pak CYC, 19. Brown AJ, Pang E, Roberts D. Persistent changes in the
Adams PM, eds. Techniques of Patient-Oriented Re- fatty acid composition of erythrocyte membranes after
search. Dallas, Tex: University of Texas Southwestern moderate intake of n-3 polyunsaturated fatty acids:
Medical Center at Dallas; 1992:108–120. study design implications. Am J Clin Nutr. 1991;54:
9. Leichsenring JM, Norris LM, Lameson SA, et al. The 668–673.
effect of level of intake on calcium and phosphorus me- 20. Egusa G, Beltz WF, Grundy SM, et al. Influence of obe-
tabolism in college women. J Nutr. 1951;45:407–419. sity on the metabolism of apolipoprotein B in humans.
10. Pak CYC, Stewart A, Raskin P, et al. A simple and re- J Clin Invest. 1985;76:596–603.
liable method for calcium balance using combined pe- 21. Beltz WF, Kesaniemi A, Howard BV, et al. Develop-
riod and continuous fecal markers. Metabolism. 1980; ment of an integrated model for analysis of the kinetics
29:793–796. of apolipoprotein B in plasma very low-density lipopro-
11. Dawson-Hughes B, Harris S, Kramich C, et al. Calcium teins, intermediate-density lipoproteins, and low-density
retention and hormone levels on high and after adapta- lipoproteins. J Clin Invest. 1985;76:575–585.
tion to low calcium diets in black and white women. J 22. McCullough ML, Swain JF, Malarick C, et al. Feasi-
Bone Miner Res. 1993;8:779–87. bility of outpatient electrolyte balance studies. J Am Coll
12. Anderson RA, Polansky MM, Bryden NA, et al. Nutr. 1991;10:140–148.
Supplemental-chromium effects on glucose, insulin, 23. Ribaya-Mercado JD, Russell RM, Sahyoun N, et al. Vi-
glucagon, and urinary chromium losses in subjects con- tamin B-6 requirements of elderly men and women. J
suming controlled low-chromium diets. Am J Clin Nutr. Nutr. 1991;121:1062–1074.
1991;54:909–916. 24. Clevidence BA, Judd JT, Schatzkin A, et al. Plasma lipid
13. Anderson RA, Polansky MM, Bryden NA, et al. Chro- and lipoprotein concentrations of men consuming a low-
mium supplementation of human subjects: effects on fat, high-fiber diet. Am J Clin Nutr. 1992;55:689–694.
glucose, insulin, and lipid variables. Metabolism. 1983; 25. Barr SL, Ramakrishnam R, Johnson C, et al. Reducing
32:894–899. total dietary fat without reducing saturated fatty acids
14. Abraham AS, Brooks BA, Eylath U. The effects of chro- does not significantly lower total plasma cholesterol
mium supplementation on serum glucose and lipids in concentrations in normal males. Am J Clin Nutr.
patients with and without non-insulin-dependent dia- 1992;55:675–681.
betes. Metabolism. 1992;41;768–771. 26. Whyte J, McCarthur R, Topping D, et al. Oat bran
15. Lichtenstein AH, Millar J, McNamara JR, et al. Long lowers plasma cholesterol levels in mildly hypercholes-
term lipoprotein response to the NCEP Step 2 diet en- terolemic men. J Am Diet Assoc. 1992;92:446–449.
riched in n-3 fatty acids. Circulation. 1990;82:III-475. 27. Henderson MM, Kushi LH, Thompson DJ, et al. Fea-
Abstract. sibility of a randomized trial of a low-fat diet for the
16. Judd JT, Clevidence BA, Muesing RA, et al. Dietary prevention of breast cancer: dietary compliance in the
trans fatty acids and plasma lipids and lipoproteins of Women’s Health Trial Vanguard Study. Prev Med.
healthy adult men and women fed a controlled diet. Am 1990;19:115–133.
J Clin Nutr. 1994;59:861–868. 28. Kristal AR, White E, Shattuck AL, et al. Long-term
17. Micozzi MS, Brown ED, Taylor PR, et al. Caroteno- maintenance of a low-fat diet: durability of fat-related
dermia in men with elevated carotenoid intake from dietary habits in the Women’s Health Trial. J Am Diet
foods and b-carotene supplements. Am J Clin Nutr. Assoc. 1992;92:553–559.
1988;48:1061–1064.

Source: "Well-Controlled Diet Studies in Humans, A Practical Guide

to Design and Management", American Dietetic Association, © 1999.