R

UMP

ATpos(dtbboorbbaJ

ApapmSAip(epsfdma

PMet

RT

A

9

RESEARCH

esearch and Professional Briefs

sing a Visual Plate Waste Study to Monitorenu Performance

RISCILLA L. CONNORS, PhD, RD; SARAH B. ROZELL, RD

ir

d(ptbmaopvTfmemTpipf

gmiodatpppsf

MFioscrcdecdrp

BSTRACTwo visual plate waste studies were conducted in 1-weekhases over a 1-year period in an acute care hospital. A totalf 383 trays were evaluated in the first phase and 467 in theecond. Food items were ranked for consumption from a low1) to high (6) score, with a score of 4.0 set as the benchmarkenoting a minimum level of acceptable consumption. Inhe first phase two entrees, four starches, all of the vegeta-les, sliced white bread, and skim milk scored below theenchmark. As a result six menu items were replaced andne was modified. In the second phase all entrees scored atr above 4.0, as did seven vegetables, and a dinner roll thateplaced sliced white bread. Skim milk continued to scoreelow the benchmark. A visual plate waste study assists inenchmarking performance, planning menu changes, andssessing effectiveness.Am Diet Assoc. 2004;104:94-96.

visual plate waste study provides valuable informa-tion that can be used as a benchmark of menuperformance and to guide activities designed to im-

rove menu response and efficiency. By minimizing wasten effective menu optimizes patient nutrition, sponsorsositive perceptions, and establishes a foundation foreeting financial objectives (1). As delineated in thetandards of Professional Practice established by themerican Dietetic Association for dietetics professionals

n management and foodservice, a routine function of therofession is evaluating services and analyzing practices2). A systematic review of the menu to determine itsffectiveness falls within the scope of practice for dieteticsrofessionals who have access to professional resources,uch as journals, for guidance (3). A structured approachor menu review avoids overreliance on anecdotal evi-ence and supplies objective information for decisionaking. One method for evaluating menu performance isplate waste study. Previously reported studies have

. L. Connors is an assistant professor, Hospitalityanagement, University of North Texas, and S. B. Roz-

ll is chief clinical dietitian, Nutritional Services, Den-on Community Hospital, Denton, TX.

Address correspondence to: Priscilla L. Connors, PhD,D, Hospitality Management, University of Northexas, PO Box 311100, Denton, TX 76203-1100.E-mail: pconnors@unt.eduCopyright © 2004 by the American Dietetic

ssociation.0002-8223/04/10401-0005$30.00/0

wdoi: 10.1016/j.jada.2003.10.012

4 Journal of THE AMERICAN DIETETIC ASSOCIATION

ndicated that patients waste an average of 16% of calo-ies served and 21% of foods served (4,5).Plate waste refers to the volume or percentage of food

iscarded and can be used to measure menu effectiveness6). Three approaches have been reported: weighed orhysical measurement, 24-hour recall, and visual. Al-hough the weighed method optimizes accuracy, it is la-or intensive, requires space for holding soiled trays andeasuring waste, and becomes increasingly unmanage-

ble as sample size grows (6,7). In addition when thebjective is simply to compare meal consumption with aredetermined standard, the high level of accuracy pro-ided by a physical measurement may not be necessary.he recall method asks patients to rate the amount of

ood they discarded over the past 24 hours (8,9). Thisethod is inexpensive and avoids tray handling, how-

ver, it depends on retrospective self-reported data anday not provide enough information for specific foods.he third method is a visual plate waste study. For theurpose of taking a benchmark measurement of foodtems wasted on a cycle menu, a visual plate waste studyrovides adequate detail yet does not significantly alteroodservice activities.

Although Kirks and colleagues caution that for aggre-ated food groups a visual plate waste study may provideisleading results, if the purpose is to highlight those

tems generating the greatest amount of waste, this typef study provides useful information (10). In addition, itoes not require additional space for holding soiled trays,dds little to labor costs, collects data on actual ratherhan intended behavior, does not require interaction withatients, and, in keeping with the precepts of universalrecautions, minimizes handling of soiled trays. The pur-ose of this article is to describe how a visual plate wastetudy can be used to take benchmark measurements ofood items and guide menu modifications.

ETHODSive nutrition professionals reviewed procedures, includ-

ng a tick sheet with a six-point scale that was based onne designed by Comstock and colleagues (7). The tickheet was preprinted with food items listed in a left-handolumn adjacent to horizontal rows of circles designed toeflect an ascending scale of none consumed (1) to allonsumed (6) (Figure). Plate waste data were recordedirectly on the tick sheet, with a separate form used forach tray. The same investigator was responsible for dataollection in both phases of the study. Before collectingata the investigator tested plate waste estimation accu-acy by scoring waste on 20 soiled lunch trays and com-aring results to those obtained by using weighed plate

aste. Scoring strategy was adjusted until an accuracy

© 2004 by the American Dietetic Association

lo1lcpceiq

votmctssfhaplwupssbpsw

RAs

ddswAHsncmccmh

adwatbsorptd

cmrvbasf

F

evel of 80% was achieved. For each meal in both phasesf the study the investigator monitored tray assembly for5 minutes to observe a minimum of 10 trays and estab-ish an accurate image of a standard portion for eventualomparison with waste on soiled patient trays. Althoughortion sizes were standardized as 3 oz meat and one-halfup starch, fruit, or vegetable, the shape and texture ofach food item influenced portion appearance. Familiar-ty with the plated portion enhanced accuracy of subse-uent waste estimation.At the conclusion of each meal, soiled trays were indi-

idually collected at bedside by nursing staff and placedn patient carts. Carts were transported for visual inspec-ion in the dish washing room. The investigator usedenu slips on returned trays as a reference to original

ontents and inspected each tray before removing it fromhe cart and placing it on a tray rack. Trays without menulips were included only if the remaining foods clearlyuggested the original contents. The extra step of trans-erring trays to a rack before washing did not change theour and a half normally required to clean meal trays,nd data collection was completed in this time. For eachhase of the study, plate waste data were collected for allunches and dinners over 7 consecutive days. Breakfastas excluded because of its uncomplicated menu and pop-larity with patients. Trays for patients receiving liquid orureed diets as well as trays that were delivered but noterved were excluded from the study. For each meal a meancore was tabulated for the entree, starch, vegetable, fruit,read or roll, and milk that was served to the majority ofatients. It was determined that any item with a meancore below a benchmark of 4.0 was producing excessiveaste and would be scrutinized for possible replacement.

ESULTS AND DISCUSSIONNorth Texas acute care hospital was the site for the

igure. Tick sheet for plate waste study.

tudy. A daily average of 66 inpatients were served meals u

uring the first phase (August 2001), and 68 were serveduring the second phase (August 2002). Average length oftay was 4.5 and 4.0 days, respectively. The populationas evenly divided between male and female patients.pproximately 82% of patients were white, 10% wereispanic, and 8% were Asian or African American. No

ingle diagnostic related group or age range predomi-ated. The nutritional services department operated aonventional food production system with a 7-day cycleenu. Although patients had the option to make menu

hoices, most received nonselect trays. A hazard analysisritical control point system was in place, and all foodset minimum temperature standards as outlined in the

ospital policies and procedures manual.Fourteen entrees, seven starches, seven vegetables,

nd six fruits were observed during the first phase, con-ucted for 1 week in August 2001. A total of 383 traysere observed (190 lunch and 193 dinner), with an aver-ge of 27 trays per meal. All fruits served met or exceededhe benchmark value. Twelve entrees exceeded theenchmark value, including beef and macaroni (5.0),outhwest chicken (4.7), and turkey with gravy (4.7). Thenly entrees to score below the benchmark value were potoast (3.6) and Salisbury steak (2.8). Nursing staff re-orted that patients who had difficulty cutting dislikedhe large chunks in the pot roast, and patients with poorentition found the Salisbury steak difficult to masticate.Three starches, mashed potatoes (4.5), baked potato

hips (4.3), and dressing (4.3), scored above the bench-ark value, although baked potato (3.8), noodles (3.7),

ice (3.5), and au gratin potatoes (3.2) scored below. Allegetables, including mixed vegetables (3.6), peas (3.5),roccoli (3.4), carrots (3.4), green beans (3.4), corn (2.8),nd spinach (2.8), scored below the benchmark, as didliced whole wheat bread (3.6) and skim milk (3.6). Theollowing changes were made based on the evidence of

npopularity that was shown by the first phase of the

Journal of THE AMERICAN DIETETIC ASSOCIATION 95

prrtb

cwFsgiaM(ctpmsbp

wofcabsesat

wvtmlt

ATowd

lfMvtte

R

1

9

late waste study. Pot roast and Salisbury steak wereeplaced with steak fingers and chopped steak. Noodles,ice, and au gratin potatoes were replaced with red pota-oes and macaroni and cheese. A dinner roll that wasaked on site replaced sliced whole wheat bread.In August 2002 the second phase of the study was

onducted. A total of 467 trays (231 lunch, 236 dinner)ere observed, with an average of 33 trays per meal.ourteen entrees, seven starches, eight vegetables, andix fruits were observed. All entrees, including steak fin-ers (5.4) and chopped steak (4.0), as well as all starches,ncluding baked potato (4.4), macaroni and cheese (4.6),nd red potatoes (4.1), cleared the benchmark value.ixed vegetables (4.4), peas (4.3), broccoli (4.2), carrots

4.1), corn (4.1), and peas and carrots (4.0) met or ex-eeded the benchmark, although green beans (3.9) andossed salad (3.8) scored below it. Dinner rolls (4.6) out-erformed the previously offered sliced bread, but skimilk (3.8) continued to score low. The findings of this

tudy are consistent with previous reports that slicedread and milk are among the least popular food items foratient meals (4,5).To determine the impact of menu changes on plateaste, a series of t tests were conducted to compare phase

ne and two by meal category (Table). Plate waste scoresor entree, starch, vegetable, and bread/roll were signifi-antly higher after the changes, although those for fruitnd skim milk were not. This outcome is not surprisingecause menu changes (replacement of two entrees, threetarches, two vegetables, and sliced bread) did notqually impact all categories. More importantly, resultsuggest that menu changes based on plate waste evidencere an effective way to decrease waste and improve pa-ient consumption of meals.

To better understand those factors that influenced foodaste, a visual plate waste study was conducted to pro-ide objective information for decision making. Actionaken based on initial results improved menu appeal andade food production activities more effective. A fol-

ow-up study confirmed that patients were eating more ofhe foods provided.

PPLICATIONShis article summarizes how the nutritional services inne hospital successfully carried out a two-phase plateaste study. Exploration of causative factors behind un-

Table. Effect of changes on plate waste by menu category

Category

Phase 1

Observations Mean SDa

Entree 328 4.3 1.8Starch 229 3.9 1.9Vegetable 357 3.5 2.2Fruit 210 4.5 2.0Bread/roll 262 3.4 2.3Skim milk 295 3.6 2.3

aSD�standard deviation.bTwo-tailed significance.cSignificant at .001.dSignificant at .005.

esirable waste and poor patient consumption high-

6 January 2004 Volume 104 Number 1

ighted correctable problems and provided justificationor expenditures targeted at improving performance.

easuring menu performance is an important task. Aisual plate waste study can provide valuable informa-ion for determining the desirability of each food item onhe menu. A popular menu improves consumption andnhances the nutritional status of patients.

eferences1. Folio D, O’Sullivan-Maillet J, Touger-Decker R. The

spoken menu concept of patient foodservice deliverysystems increases overall patient satisfaction, thera-peutic and tray accuracy, and is cost neutral for foodand labor. J Am Diet Assoc. 2001;102:546-548.

2. Griffin B, Dunn JM, Irvin J, Speranza IF. Standardsof professional practice for dietetics professionals inmanagement 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 wastein a hospital cook-freeze production system. J AmDiet Assoc. 1986;86:941-942.

5. Hirsch KM, Hassanein RS, Nelson SJ. Factors influ-encing plate waste by the hospitalized patient. J AmDiet Assoc. 1979;75:270-273.

6. Buzby JC, Guthrie JF. US Department of Agricul-ture, Economic Research Service. Plate waste inschool nutrition programs: Final report to congressE-FAN-02.009. Washington, DC: US GovernmentPrinting Office; 2002.

7. Comstock EM, Pierre RG, Mackiernan YD. Measur-ing individual plate waste in school lunches. J AmDiet Assoc. 1981;79:290-296.

8. Weisberg K. A remedy for patient meals: MemorialSloan-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.

0. Kirks B, Wolff H. A comparison of methods for platewaste determination. J Am Diet Assoc. 1985;85:328-

Phase 2

t P brvations Mean SD

4.9 1.5 �04.60 .000c

4.6 1.6 �04.51 .000c

4.0 2.0 �02.91 .004d

4.6 2.1 �0.702 .4834.6 2.1 �06.50 .000c

3.8 2.3 �0.788 .431

Obse

373374371304309304

331.