RESEARCH
Research and Professional Briefs
Using a Visual Plate Waste Study to Monitor
Menu Performance
PRISCILLA L. CONNORS, PhD, RD; SARAH B. ROZELL, RD
ABSTRACT
Two visual plate waste studies were conducted in 1-week
phases over a 1-year period in an acute care hospital. A total
of 383 trays were evaluated in the first phase and 467 in the
second. Food items were ranked for consumption from a low
(1) to high (6) score, with a score of 4.0 set as the benchmark
denoting a minimum level of acceptable consumption. In
the first phase two entrees, four starches, all of the vegeta-
bles, sliced white bread, and skim milk scored below the
benchmark. As a result six menu items were replaced and
one was modified. In the second phase all entrees scored at
or above 4.0, as did seven vegetables, and a dinner roll that
replaced sliced white bread. Skim milk continued to score
below the benchmark. A visual plate waste study assists in
benchmarking performance, planning menu changes, and
assessing effectiveness.
J Am Diet Assoc. 2004;104:94-96.
A
visual plate waste study provides valuable informa-
tion that can be used as a benchmark of menu
performance and to guide activities designed to im-
prove menu response and efficiency. By minimizing waste
an effective menu optimizes patient nutrition, sponsors
positive perceptions, and establishes a foundation for
meeting financial objectives (1). As delineated in the
Standards of Professional Practice established by the
American Dietetic Association for dietetics professionals
in management and foodservice, a routine function of the
profession is evaluating services and analyzing practices
(2). A systematic review of the menu to determine its
effectiveness falls within the scope of practice for dietetics
professionals who have access to professional resources,
such as journals, for guidance (3). A structured approach
for menu review avoids overreliance on anecdotal evi-
dence and supplies objective information for decision
making. One method for evaluating menu performance is
a plate waste study. Previously reported studies have
P. L. Connors is an assistant professor, Hospitality
Management, University of North Texas, and S. B. Roz-
ell is chief clinical dietitian, Nutritional Services, Den-
ton Community Hospital, Denton, TX.
Address correspondence to: Priscilla L. Connors, PhD,
RD, Hospitality Management, University of North
Texas, PO Box 311100, Denton, TX 76203-1100.
E-mail: pconnors@unt.edu
Copyright © 2004 by the American Dietetic
Association.
0002-8223/04/10401-0005$30.00/0
doi: 10.1016/j.jada.2003.10.012
94 Journal of THE AMERICAN DIETETIC ASSOCIATION
indicated that patients waste an average of 16% of calo-
ries served and 21% of foods served (4,5).
Plate waste refers to the volume or percentage of food
discarded and can be used to measure menu effectiveness
(6). Three approaches have been reported: weighed or
physical measurement, 24-hour recall, and visual. Al-
though the weighed method optimizes accuracy, it is la-
bor intensive, requires space for holding soiled trays and
measuring waste, and becomes increasingly unmanage-
able as sample size grows (6,7). In addition when the
objective is simply to compare meal consumption with a
predetermined standard, the high level of accuracy pro-
vided by a physical measurement may not be necessary.
The recall method asks patients to rate the amount of
food they discarded over the past 24 hours (8,9). This
method is inexpensive and avoids tray handling, how-
ever, it depends on retrospective self-reported data and
may not provide enough information for specific foods.
The third method is a visual plate waste study. For the
purpose of taking a benchmark measurement of food
items wasted on a cycle menu, a visual plate waste study
provides adequate detail yet does not significantly alter
foodservice activities.
Although Kirks and colleagues caution that for aggre-
gated food groups a visual plate waste study may provide
misleading results, if the purpose is to highlight those
items generating the greatest amount of waste, this type
of study provides useful information (10). In addition, it
does not require additional space for holding soiled trays,
adds little to labor costs, collects data on actual rather
than intended behavior, does not require interaction with
patients, and, in keeping with the precepts of universal
precautions, minimizes handling of soiled trays. The pur-
pose of this article is to describe how a visual plate waste
study can be used to take benchmark measurements of
food items and guide menu modifications.
METHODS
Five nutrition professionals reviewed procedures, includ-
ing a tick sheet with a six-point scale that was based on
one designed by Comstock and colleagues (7). The tick
sheet was preprinted with food items listed in a left-hand
column adjacent to horizontal rows of circles designed to
reflect an ascending scale of none consumed (1) to all
consumed (6) (Figure). Plate waste data were recorded
directly on the tick sheet, with a separate form used for
each tray. The same investigator was responsible for data
collection in both phases of the study. Before collecting
data the investigator tested plate waste estimation accu-
racy by scoring waste on 20 soiled lunch trays and com-
paring results to those obtained by using weighed plate
waste. Scoring strategy was adjusted until an accuracy
© 2004 by the American Dietetic Association
Figure. Tick sheet for plate waste study.
level of 80% was achieved. For each meal in both phases
of the study the investigator monitored tray assembly for
15 minutes to observe a minimum of 10 trays and estab-
lish an accurate image of a standard portion for eventual
comparison with waste on soiled patient trays. Although
portion sizes were standardized as 3 oz meat and one-half
cup starch, fruit, or vegetable, the shape and texture of
each food item influenced portion appearance. Familiar-
ity with the plated portion enhanced accuracy of subse-
quent waste estimation.
At the conclusion of each meal, soiled trays were indi-
vidually collected at bedside by nursing staff and placed
on patient carts. Carts were transported for visual inspec-
tion in the dish washing room. The investigator used
menu slips on returned trays as a reference to original
contents and inspected each tray before removing it from
the cart and placing it on a tray rack. Trays without menu
slips were included only if the remaining foods clearly
suggested the original contents. The extra step of trans-
ferring trays to a rack before washing did not change the
hour and a half normally required to clean meal trays,
and data collection was completed in this time. For each
phase of the study, plate waste data were collected for all
lunches and dinners over 7 consecutive days. Breakfast
was excluded because of its uncomplicated menu and pop-
ularity with patients. Trays for patients receiving liquid or
pureed diets as well as trays that were delivered but not
served were excluded from the study. For each meal a mean
score was tabulated for the entree, starch, vegetable, fruit,
bread or roll, and milk that was served to the majority of
patients. It was determined that any item with a mean
score below a benchmark of 4.0 was producing excessive
waste and would be scrutinized for possible replacement.
RESULTS AND DISCUSSION
A North Texas acute care hospital was the site for the
study. A daily average of 66 inpatients were served meals
during the first phase (August 2001), and 68 were served
during the second phase (August 2002). Average length of
stay was 4.5 and 4.0 days, respectively. The population
was evenly divided between male and female patients.
Approximately 82% of patients were white, 10% were
Hispanic, and 8% were Asian or African American. No
single diagnostic related group or age range predomi-
nated. The nutritional services department operated a
conventional food production system with a 7-day cycle
menu. Although patients had the option to make menu
choices, most received nonselect trays. A hazard analysis
critical control point system was in place, and all foods
met minimum temperature standards as outlined in the
hospital policies and procedures manual.
Fourteen entrees, seven starches, seven vegetables,
and six fruits were observed during the first phase, con-
ducted for 1 week in August 2001. A total of 383 trays
were observed (190 lunch and 193 dinner), with an aver-
age of 27 trays per meal. All fruits served met or exceeded
the benchmark value. Twelve entrees exceeded the
benchmark value, including beef and macaroni (5.0),
southwest chicken (4.7), and turkey with gravy (4.7). The
only entrees to score below the benchmark value were pot
roast (3.6) and Salisbury steak (2.8). Nursing staff re-
ported that patients who had difficulty cutting disliked
the large chunks in the pot roast, and patients with poor
dentition found the Salisbury steak difficult to masticate.
Three starches, mashed potatoes (4.5), baked potato
chips (4.3), and dressing (4.3), scored above the bench-
mark value, although baked potato (3.8), noodles (3.7),
rice (3.5), and au gratin potatoes (3.2) scored below. All
vegetables, including mixed vegetables (3.6), peas (3.5),
broccoli (3.4), carrots (3.4), green beans (3.4), corn (2.8),
and spinach (2.8), scored below the benchmark, as did
sliced whole wheat bread (3.6) and skim milk (3.6). The
following changes were made based on the evidence of
unpopularity that was shown by the first phase of the
Journal of THE AMERICAN DIETETIC ASSOCIATION 95
 Table. Effect of changes on plate waste by menu category
Phase 1 Phase 2
Category Observations Mean SDa Observations Mean SD t Pb

Entree 328 4.3 1.8 373 4.9 1.5 ⫺04.60 .000c

Starch 229 3.9 1.9 374 4.6 1.6 ⫺04.51 .000c
Vegetable 357 3.5 2.2 371 4.0 2.0 ⫺02.91 .004d
Fruit 210 4.5 2.0 304 4.6 2.1 ⫺0.702 .483
Bread/roll 262 3.4 2.3 309 4.6 2.1 ⫺06.50 .000c
Skim milk 295 3.6 2.3 304 3.8 2.3 ⫺0.788 .431
a
SD⫽standard deviation.
b
Two-tailed significance.
c
Significant at .001.
d
Significant at .005.

plate waste study. Pot roast and Salisbury steak were
replaced with steak fingers and chopped steak. Noodles,
rice, and au gratin potatoes were replaced with red pota-
toes and macaroni and cheese. A dinner roll that was
baked on site replaced sliced whole wheat bread.
In August 2002 the second phase of the study was
conducted. A total of 467 trays (231 lunch, 236 dinner)
were observed, with an average of 33 trays per meal.
Fourteen entrees, seven starches, eight vegetables, and
six fruits were observed. All entrees, including steak fin-
gers (5.4) and chopped steak (4.0), as well as all starches,
including baked potato (4.4), macaroni and cheese (4.6),
and red potatoes (4.1), cleared the benchmark value.
Mixed vegetables (4.4), peas (4.3), broccoli (4.2), carrots
(4.1), corn (4.1), and peas and carrots (4.0) met or ex-
ceeded the benchmark, although green beans (3.9) and
tossed salad (3.8) scored below it. Dinner rolls (4.6) out-
performed the previously offered sliced bread, but skim
milk (3.8) continued to score low. The findings of this
study are consistent with previous reports that sliced
bread and milk are among the least popular food items for
patient meals (4,5).
To determine the impact of menu changes on plate
waste, a series of t tests were conducted to compare phase
one and two by meal category (Table). Plate waste scores
for entree, starch, vegetable, and bread/roll were signifi-
cantly higher after the changes, although those for fruit
and skim milk were not. This outcome is not surprising
because menu changes (replacement of two entrees, three
starches, two vegetables, and sliced bread) did not
equally impact all categories. More importantly, results
suggest that menu changes based on plate waste evidence
are an effective way to decrease waste and improve pa-
tient consumption of meals.
To better understand those factors that influenced food
waste, a visual plate waste study was conducted to pro-
vide objective information for decision making. Action
taken based on initial results improved menu appeal and
made food production activities more effective. A fol-
low-up study confirmed that patients were eating more of
the foods provided.
APPLICATIONS
This article summarizes how the nutritional services in
one hospital successfully carried out a two-phase plate
waste study. Exploration of causative factors behind un-
desirable waste and poor patient consumption high-
lighted correctable problems and provided justification
for expenditures targeted at improving performance.
Measuring menu performance is an important task. A
visual plate waste study can provide valuable informa-
tion for determining the desirability of each food item on
the menu. A popular menu improves consumption and
enhances the nutritional status of patients.
References
1. Folio D, O’Sullivan-Maillet J, Touger-Decker R. The
spoken menu concept of patient foodservice delivery
systems increases overall patient satisfaction, thera-
peutic and tray accuracy, and is cost neutral for food
and labor. J Am Diet Assoc. 2001;102:546-548.
2. Griffin B, Dunn JM, Irvin J, Speranza IF. Standards
of professional practice for dietetics professionals in
management and foodservice settings. J Am Diet As-
soc. 2001;101:944-946.
3. Glore S. Show me the science. J Am Diet Assoc. 2001;
101:186.
4. Frankes RM, Arjmandi BH, Halling JF. Plate waste
in a hospital cook-freeze production system. J Am
Diet Assoc. 1986;86:941-942.
5. Hirsch KM, Hassanein RS, Nelson SJ. Factors influ-
encing plate waste by the hospitalized patient. J Am
Diet Assoc. 1979;75:270-273.
6. Buzby JC, Guthrie JF. US Department of Agricul-
ture, Economic Research Service. Plate waste in
school nutrition programs: Final report to congress
E-FAN-02.009. Washington, DC: US Government
Printing Office; 2002.
7. Comstock EM, Pierre RG, Mackiernan YD. Measur-
ing individual plate waste in school lunches. J Am
Diet Assoc. 1981;79:290-296.
8. Weisberg K. A remedy for patient meals: Memorial
Sloan-Kettering aims to increase meal intake. Food-
service Director. 2002;15:32.
9. O’Hara PA, Harper DW, Kangas M, Dubeau J, Bor-
sutzky C, Lemire N. Taste, temperature, and presen-
tation predict satisfaction with foodservices in a Ca-
nadian continuing-care hospital. J Am Diet Assoc.
1997;97:401-405.
10. Kirks B, Wolff H. A comparison of methods for plate
waste determination. J Am Diet Assoc. 1985;85:328-
331.

96 January 2004 Volume 104 Number 1