ir«* OO14-A-O\
Evaluation of the
NutrienL||p|idard
Menu PB demonstration
An.UhJlTjsun.
r** ■•itttf Stttti
•aaartaaat af
aarlcaltart
Faatf aa< la'rHiaa
$tr»lc< .X
tf IfalaatU* ^fhk.
-
The Unrtrt SWR Dfpirtnwnt of Agncuhut (USDA) prohibits Ascnrrawnon «its programs on the bM of wrcolor. unntulon9n,sn.iv^m.^.(Mi^poli(iultwMsin(l
nunul<xbntfclsuw(Not*lrjrohiij^b*sft
pnm.juriourx.flc.) should conua USD* lAHGf I (tm« [202) 720-2(0) (vow mlTOO).
bMt jcgmpUm«(*D> Sm««yiif Ap^culhR US.r^^
fmploymfni opporTunrty pmployPT
#
Evaluation of the
Nutrient Standard
Menu Planning Demonstration
Summary of Findings
August 1998
United States Submitted by: Submitted to:
OtiirtMtat af
Afriealtar*
Faaf <■< Natritiaa
Abt Associates Inc. Office of Analysis and Evaluation
Strtlct 55 Wheeler Street USDA Food and Nutrition Service
Offict tf Analysis » Cambridge. MA 02138 3101 Park Center Drive, Room 208
IwataMtoi Alexandria, VA 22302
Cantraet Na.
S1-I1M-4-M1
Project Director Project Officer:
Mary Kay Fox John Endahl
USDA
&
Contents
Introduction
Implementation ofNSMP in Demonstration SFAs
Foods Offered Before and After NSMP 6
Nutrient Content of Lunches 8
Nutrient Content of Breakfasts 12
Weighted and Unweighted Nutrient Analyses 16
Program Operations Before and After NSMP 20
Program Costs Before and After NSMP 22
-It
Introduction
The National School Lunch Program (NSLP) ami the School Breakfast Program (SBP) are
administered by the Food and Nunition Service (FNS) of the U.S. Department >>t Agriculture
(USDA). The NSLP operates in over 94,000 schools and institutions. Mure than 20 million children
receive meals through the program on any given day; ahoui halt ot these meals .ire proi ided tree of
charge. The SBP operates in approximately two-thiidsol the schools and institutions ili.it offer the
NSLP. most commonly in schools th.it serve large number* «>t economically disadvantaged children.
v\i ,m average day, roughly seven million children receive breakfast through the SBP The vast
majority of these meab are provided free of charge. School Food Authorities (SFAs) participating in
the NSLP and SBP receive two types of federal assistance: donated surplus commodities and cash
reimbursements.
Meals served in the NSLP and SBP meet defined nutrition standards in order t>> he eligible l«n Federal
subsidies. Program regulations lot the NSLP stipulate th.it lunches must provide, on average,
approximately tine-third ofstudents' Recommended Dieton AUnuunces (RDAs). ToeiiMireth.it these
standards .ire met, program regulations have historically included food-based menu planning
guidelines which define specific types «'i tii\l to he offered a* well a* minimum portion sizes. The
authoncing legislation tor the SBP did not include .1 specific RI\A goal, but .1 meal pattern was
developed to ensure th.it breakfasts would provide approximately one-ttxnth ol the RPA.
While historically the NSLP and SBP have been successful in meeting these nutrition standard*. .1
l*wi USDA study found thai schixil me.il- were not consistent with goals tor total t.it and saturated
tat intake specified in the Dieton Hudc/mc* for Americans. At the time, programs were not required to
meet these guidelines.
Since tin* problem was identified, FNS has worked toenhare e tin* aspect ot the nutritkmal quality .>t
meab offered in the NSLP and SBP. A* part ot tin* ongoing initiative, FNS sponsored a demonstration
project to evaluate the acceptability and impact of an alternative system tor planning lunch and
breakfast menu*. Thi* system, known .1* Nutrient Standard Menu Planning (NSMP), uses
computerised nutrient analysis to assess the nutrient content ot planned menu*.
Abi Associates Inc. ol Cambridge, Massachusetts was awarded .1 contract to conduct an independent
evaluation at the three-year demonstration pniject. Thi* publication summarises major findings from
th.it evaluation While regulatory changes th.it have taken place since the inception ot the NSMP
demonstration have already incorporated NSMP into NSLr and SBP operations, findings h\»m the
evaluation are important in providing Information about requirements ot the cuireni NSMP system
and the potential need tor training, technical assistance, and monitoring.
The NSMP Demonstration
In 1,11111.m 1994, FNS selected M volunteer SFAs to participate in the NSMP demonstration. SFA*
were purposefully selected to provide diversity in geographk location. di*tnct *i:e. ttudenl partu ipation
rates, tisid service program characteristics, ami *tatt experience With computerised nutrient analysis.
Evaluation of the Nutrient Standard Menu Planning Demonstration
The demonstration spanned three school years (SY), from SY 1994-95 through SY 1996-97.
Participating SFAs were expected to implement NSMP for both lunch and breakfast in all schools.
Menus were to be planned using newly-developed software systems that evaluated the nutrient
content of planned menus using a weighted nutrient analysis. A weighted analysis incorporates
information about students' food selection patterns and gives more weight to foods that are selected
more often. SFAs were expected to base weighted analyses on district-wide menu production data,
i.e., information on the actual food selection patterns of all students in the district.
Menus planned under NSMP were expected to meet a defined set of nutrient standards. Lunches
were expected to provide one-third of the RDA for calories and key nutrients and breakfasts were
expected to provide one-fourth of the RDA. Both meals were expected to be consistent with Dietary
Guidelines goals for the percentage of calories from total fat and saturated fat.
The sequence and timing of major milestones in both the demonstration and its associated evaluation
are summarized in Exhibit 1. Baseline data were collected in Spring 1994 while traditional meal
patterns were still being used to plan menus and before SFA staff received training on NSMP. Key
staff from participating SFAs attended a three-day training session in June, 1994. Staff were expected
to begin serving meals planned using NSMP the following fall, with the goal of achieving full
implementation by Fall 1995.
Implementation was substantially delayed in most SFAs because of a significant lag in the approval of
NSMP software systems. Software vendors were slow to submit products for USDA review and
approval, and many of the software systems submitted did not satisfy the functional criteria defined by
USDA. By January 1995, only two software systems had been approved. Participating SFAs were
asked to select one of the two available systems and to continue to wotk toward full implementation
by Fall 1995.
Evaluation Design
I FNS 5etects
ffnpipfncwiunfon | SFAS
■ ■ ■ •; ■ • i 95
Data
foWrfton
A 9V96 SY 96-97
Note: Process evaluation data ware ongmaKy scheduled to be collected m Spring 1996
Summary of Findings/August 1998
Nutrient Standards Used in the Evaluation
HUTMENT
Stano^ Defined for NSMP
Nutrients with established RDAs:
Calories, protein, vitamin A, vitamin C,
calcium, and iron
Nutrients included in the Dietary
Guiatwnt&TOfAniencons:
Total fat
Saturated fat
Breakfast One-fourth of the RDA
Lunch: One-third of the RDA
< 30% oftotal calories
< 10% oftotal calories
NRC Recommendations
Carbohydrate
Protein
Cholesterol
Sodium
> 55%of total calories
< 15% of total calories
Breakfast < 75 mg
Lunch: < 100 mg
Breakfast <600mg
Lunch: <800mg
Note fiber content ofplanned menus was also examined but was not compared to a standard There is no scientific
consensus about an appropriate benchmark for dairy consumption of fiber
The evaluation examined the process of NSMP implementation in demonstration SFAs and assessed
how well menus planned under NSMP complied with defined nutrient standards (Exhibit 2). For
purposes of the evaluation, menus were also compared to National Research Council (NRC)
recommendations for carbohydrate, protein, cholesterol, and sodium content. (SFAs were not
expected to meet the NRC recommendations.) In addition, the evaluation assessed differences in
program operations and costs between Spring 1994 (pre-NSMP or before NSMP) and Spring 1997
(NSMP or after NSMP).
The evaluation did not include a control group and sample sizes were small. For these reasons, the
reader u cautioned that differences between pre-NSMP and NSMP measures can not necessarily
be ascribed to NSMP and are not necessarily representative of what might happen in programs
nationwide.
EMMsaoon orinc rtuuMiii ManaaraMcnu nmnuM] Demonstration
Implementation of NSMP in Demonstration SFAs
Lew dan Ml (16) of the original 34 SFAs fully intimated NSMP. Ten of these SFAs implemented
NSMP in full accordance with expectations. The other six SFAs used a modified approach to
weighted nutrient analysis. Modifications included basing data used in the weighted analysis on
production data from a subset of schools rather than all schools in the district or on staff predictions
rather than actual menu production data.
Four SFAs achieved onhpartial implementation and tiiree SFAs never implemented NSMP. Directors in
the four partially implemented SFAs reported that they did not achieve full implementation because
they were unahlr to dedicate the number of staff labor hours required. One of the three non-implemented
SFAs attempted to implement NSMP according to the protocol, but never got beyond
the process of entering data and analyzing initial menus. The other two non-implemented SFAs
deviated substantially from the NSMP protocol. Both of these SFAs performed unweighted analyses
and one SFA used unapproved software.
A majority of eleven SFAs that withdrew from the demonstration did so because of concern about
the use of weighted nutrient analysis. Most often, the concern centered around the amount of staff
labor required to obtain district-wide menu production data. Some SFA directors were also
concerned that use of weighted nutrient analysis would limit flexibility in menu planning and
decrease the ability of individual schools to cater to students' preferences. A few directors were
worried that a weighted analysis could only meet nutrient standards if populai high-fat food items
were eliminated or offered much less frequently. These directors believed that such changes would
have a negative impact on lunch participation in middle and high schools.
Other factors contributing to SFA withdrawal included problems with NSMP software
(unavailability, incompatibility, frequent problems with initial releases); the short implementation
time line; a change in SFA directors; and SFA directors' responses to pending program regulations
not directly related to the demonstration.
MOM SFAS required substantially more time to implement NSMP than anticipated. Among the 16
SFAs that fully implemented NSMP, an average of 19 months elapsed between the time SFA staff
started working on NSMP implementation and the time NSMP menus were served in all schools.
The range was very broad, with a minimum of three months and a maximum of 33 months.
Very large SFAs and, to a lesser extent, small SFAs were less likely to implement NSMP than
medium or large SFAs. Most of the very large SFAs (enrollment of 25,000 or more) that did not
implement NSMP withdrew from the demonstration. The primary reason for termination was
concern about the use of weighted analysis, specifically the level of effort involved in collecting and
organizing district-wide menu production data. The burden associated with complying with this
requirement clearly escalates as district size increases, unless a district is fully computerized or has
highly centralized food production.
Three of four small SFAs (enrollment of less than 2,500) that did not implement NSMP also dropped
out of the demonstration. The major issue for small SFAs was the amount of work involved in NSMP
implementation giver. Tiited staffing.
SFA» that had some centralized food production prior to NSMP were more likely to achieve full
implementation, without modification, than SFAs with little or no centralized production. The
presence of centralized food production may benefit NSMP implementation by reducing the amount of
work required to document district-wide food usage. In addition, a more centralized program is likely to
face fewer problems in identifying standardized recipes and ensuring that these recipes are used in all
food production locations.
SFAs that had fully centralized menu planning prior to NSMP were more likely to achieve full
implementation, without modification, than SFAs with decentralized menu planning. In order ro
implement NSMP in situations where schools have limited or total autonomy in menu planning, SFAs
must either reduce the level of local control—i.e., move to centralized menu planning—or analyze
menus for individual schools.
The number of staff hours devoted to NSMP implementation varied widely. The total number of
hours devoted to NSMP implementation ranged from 334 hours, for a small district that had centralized
menu planning prior to the initiation ofNSMP and some centralized food production, to 3,830 hours,
for a very large district that maintained partially decentralized menu planning for middle and high
schools even after NSMP implementation. The median was 1,139 hours. Based on reported staff
salaries, exclusive of fringe benefits, total costs for start-up labor ranged from $4,272 to $41,434, in 1997
dollars, with a median of $16,139.
Data entry and staff training together accounted for more than half of all labor hours spent on
NSMP implementation (Exhibit 3). Less than ten percent of all reported labor was associated with
planning and analyzing menus. Thus, nine out of every ten hours spent on NSMP implementation
Mott of the Time Devoted to NSMP Implementation
Wat Spent on Data Entry and Staff Training
Contribution of NSMP Tasks to Total Implementation Hours
Miscellaneous
Review Analyses
Plan Menus
Recipe
StarxJardization J A °** Entry
and Testing
Staff Training
Sourc* Dau from 16 fully imptamanud ifAt
"•uauon wait iwuifiri xenoetu menu running utfnoniuiuofi
was devoted, essentially, to building the system infrastructure. Only one of every ten hours was spent
using the system to assess status of current and interim menus.
A majority of SFAi reported that one or more NSMP tasks imposed a significant burden on SFA
staff (Exhibit 4). Only six of the 23 directors who continued in the demonstration through Spring
1997 reported that NSMP implementation did not impose an undue burden on SFA staff. The specific
task most often cited as a significant burden was obtaining menu production data to support use of
weighted nutrient analysis.
Tasks that were least often perceived as imposing a significant burden were those required for any
menu planning option: planning menus, marketing healthy food choices, developing specifications for
purchased foods, and monitoring purchased foods.
All SFAs made errors in their nutrient analyses. The most common errors were omission of a
planned menu item and inaccurate serving projections. In a weekly analysis, about half of the SFAs
omitted five or more menu items. Condiments, including salad dressing, tartar sauce, mayonnaise, and
other high-fat items, were the most frequent omissions. Major menu items omitted from an analysis
were generally alternative entree choices (e.g., salad bars, other specialty bars, or sandwiches) or one or
more types of milk. In general, such omissions were made consistently in each dairy analysis. Errors in
serving projections generally resulted from failure to separate out a la carte and adult servings from
menu production figures.
Foods Offered Before and After NSMP
Changes in the types of food offered in NSMP menus were consistent with recommendations made in
NSMP training. Compared to pre-NSMP menus, NSMP lunch menus included more of the following
foods: skim milk and flavored lowfat milk; fresh fruit; raw vegetables and salads; extra bread and grain
choices; pasta-based entrees; rice; and desserts. NSMP lunch menus also included leu of the following
foods: whole milk; french fries and similar potato products; entrees that tend to be high in fat, such as
breaded meat, poultry and fish, burgers, nachos, and hot dogs; and snack chips. Secondary school
menus showed more change than elementary school menus.
NSMP breakfast menus included more skim milk and flavored lowfat milk; canned and fresh fruit; fruit
juice; breakfast potatoes; yogurt; cold cereals; and pancakes, and less whole milk; sausages; muffins; and
French toast.
NSMP menus included approximate the same number of dairy options as pre-NSMP menus.
There were no significant differences between pre-NSMP and NSMP menus in the average number of
choices offered to students, within menu item category, each day.
The variety of foods offered over the course of a week was comparable before and after NSMP. There
were no significant differences between pre-NSMP and NSMP menus in the number of different types of
food offered over the course of a week within major food categories, e.g., fruits, vegetables, entrees.
Summary of Rndsngs/August 1990
SFA Directors Found Obtaining Mmu Production Data to bo tho Most
Burdantomo NSMP Task
Obtaining menu
production data
Entering/analyzing
recipes
Developing
standardized recipes
Entering/analyzing
menus
Obtaining data from
food manufacturers
Training staff
Entering nutrient data
into the database
Monitoring purchased
foods
Developing purchasing
specifications
Marketing/Promotions
Planning menus
1
Number of SfA Directors Reporting That Task Posed a Significant Burden
Source OMa from 23 Sf A directors involved In the demonstration through Spring 1997
f4 \\%m w ww ^M*,#^w1w1 ■v »■ .»a OjuMrWmISiMuI yHunu r■w#.nnvny fu*lt■n ■i■o InWs»i»r^iiJiWon1
Nutrient Content of Lunches
This analysis is based on data from 19 SFAs that implemented NSMP, either district wide or for
elementary lunches only. All findings are based on weighted nutrient analyses.
Lunches served both before and after implementation of NSMP provided one-third or more of the
RDA for calories and key nutrients (Exhibit 5). Lunches served at both points in time provided
more than one-third of the RDA, with the exception of calories in secondary school lunches.
Elementary school lunches served after NSMP provided significantly fewer calories and less protein
than lunches served prior to NSMP Secondary school lunches served after NSMP provided
significantly more vitamin A than lunches served before NSMP. These differences have little
substantive importance, however, because lunches served at both points in time exceeded the NSMP
standard of one-third of the RDA.
Lunches served after NSMP were significantly lower in fat than pre-NSMP lunches, but continued
to exceed the Dietary Quidelines goal of no more than 30 percent of calories from fat (Exhibit 6).
NSMP lunches derived significantly fewer calories from fat than lunches served before NSMP (31%
for elementary NSMP lunches and 33% for secondary NSMP lunches versus 36% for both elementary
and secondary pre-NSMP lunches). The five percentage point drop in the percent of calories from fat
in elementary school lunches represents an overall decrease in actual fat calories of 19 percent. For
secondary school lunches, the three percentage point drop in percent of calories from fat represents a
12 percent decrease in actual fat calories.
Lunches served after NSMP provided significantly fewer calories from saturated fat than pre-
NSMP lunches, but continued to exceed the Dietary Quutelme* goal of less than 10 percent of
calories from saturated fat (Exhibit 6). The percentage of calories from saturated fat in elementary
school lunches dropped three percentage points over the course of the demonstration (14% versus
11%). This difference represents a decline of 25 percent in actual saturated fat calories. The
percentage of calories from saturated fat in secondary school lunches dropped one percentage point
(13% versus 12%), equivalent to a 16 percent decrease in actual saturated fat calories.
Lunches served after NSMP provided significant' nore calories from carbohydrate than pre-
NSMP lunches (Exhibit 6). NSMP lunches derived significantly more calories from carbohydrate
than lunches served before NSMP (54% for elementary NSMP lunches and 52% for secondary NSMP
lunches versus 49% for both elementary and secondary pre-NSMP lunches). Despite these increases,
lunches served in both types of schools continued to fall short of the NRC recommendation of more
than 55 percent of calories from carbohydrate.
Lunches served after NSMP provided significantly less cholesterol then pre-NSMP lunches.
Amounts of sodium and fiber were similar (Exhibit 7). NSMP lunches served in both elementary
schools and secondary schools were significantly lower m cholesterol than lunches served prior to
NSMP. These differences are of limited importance, however, because lunches served at both points in
time met the NRC recommendation for cholesterol. Sodium content was essentially the same in pre-
NSMP and NSMP lunches and, at both pomts in time, exceeded the NRC recommendation by a
substantial margin. Fiber content of pre-NSMP and NSMP lunches was not significantly different.
JWrwHWy ^W rwr*sW^Bfrtl^W»l I*"
With the Exception of Calories In Secondary Schools, Lunches Served
Before and After NSMP Provided More than One-third of the RDA
Elementary School Lunches
110%
1
100%
80%
60%
40%
20%
0%
Catorlaa ProMn Vitamin A Vitamin C Calcium Iron
Secondary School Lunches
100%
Catorlaa Protaln Vitamin A Vitamin C Calcium Iron
Before NSMP After NSMP
• Difference is statistically significant at the 5 percent level
' * DUforence n statisticaey significant at the 1 percent level
r. ■ »L ■■*!!! Ml aillu fcl.^-t.-.a f|.- ■■■ ,,1 »* mt. ■ ■ I ■ - ft.-. „.|..||.,
■ . ■ hit 6 Lunches Served After NSMP Provided Fewer Calories from
Fet and Saturated Fat than Lunches Served Before NSMP
Elementary School Lunches
a
1
60%
>55% 54% • •
50% 49%
40% 36%
*
30%
<30% 31% ee
20%
10%
0%
14%
<io%| "&•*
■15%^^
Total Fat Satur «J Fat Carbohydrata Protaln
Secondary School Lunches
<3
1
60%
50%
40%
30%
20%
10%
0%
Total Fat Saturated Fat Carbohydrata Protaln
| Dietary Guidelines or NRC Recommendation
■ Before NSMP ■ After NSMP
* DUtmtnce s statistically sigrnhcant at the S percent level
* * Difference is statistically significant at the 1 percent level
Summary of FtrnMngs/Auaust 199S
LunchM Served Before and After NSMP Met the NRC Recommendation
for Cholesterol But Not for Sodium
Elementary School Lunches
100
<100
I
1500
§
Chowtttrol Sodium
Secondary School Lunches
100
!
1500 1466 1436
Cnotottorof Sodium Floor
| NRC Recommendation
Before NSMP ■ After NSMP
•• OMfrwnct
a stMBtKsKy ugnfictnl M the 5 percent level
a statenc»!ty agnMcam at the 1 percent level
EvoluMion of the Nutritm Standard MwwPtMMi^ 11
Nutrient Content of Breakfasts
This analysis is based on data from 12 SFAs that implemented NSMP for breakfast menus. All findings
are based on weighted nutrient analyses.
With the exception of calories, breakfasts served both before and after NSMP provided one-fourth
of the RDA (Exhibit 8). This finding is consistent with SFA director reports that the calorie
standard for breakfast was one of the most difficult to meet.
NSMP breakfasts served in secondary schools provided significantly greater amounts of vitamin A
and iron than pre-NSMP breakfasts (Exhibit 8). Although, on average, pre-NSMP breakfasts served
in secondary schools provided more than one-fourth of the RDA for vitamin A and iron, the average
amount of both nutrients was significantly higher in NSMP breakfasts. Both of these increases are
attributable to an increase in the use of cold cereals, most of which were fortified with vitamin A and
iron, in NSMP breakfasts.
Breakfasts served before and after NSMP satisfied the Dietary Quidelines goal for calories
from fat (Exhibit 9). While both pre-NSMP and NSMP breakfasts were consistent with the
Dietary Guidelines goal of no more than 30 percent of calories from fat, the percentage of calories
from fat in NSMP breakfasts *rved in secondary schools was significantly lower than pre-NSMP
breakfasts (24% versus 29%).
Breakfasts served after NSMP provided significantly fewer calories from saturated fat than
breakfasts served before NSMP, and satisfied the Dietary guidelines goal for saturated fat (Exhibit
9). NSMP breakfasts served in both elementary and secondary schools provided less than ten percent
ofcalories from saturated fat (actual values of 9.5% and 9.7%, rounded to 10% in Exhibit 9),
compared to 13% and 14%, respectively, for pre-NSMP breakfasts.
Breakfasts served before and after NSMP satisfied the NRC recommendation for calories from
carbohydrate (Exhibit 9). While both pre-NSMP and NSMP breakfasts were consistent with the
NRC recommendation of more dun 55 percent of calories from carbohydrate, NSMP breakfasts
served in both elementary and secondary schools provided significantly more calories from
carbohydrate than pre-NSMP breakfasts.
Breakfasts served before and after NSMP provided comparable amounts of cholesterol and sodium
(Exhibit 10). Breakfasts served both before and after NSMP satisfied the NRC recommendation for
cholesterol. The same is true for sodium in elementary school breakfasts, however, secondary school
breakfasts exceeded the recommendation for sodium at both points in time. Fiber content was higher
in breakfasts served after NSMP and this difference was statistically significant for secondary school
breakfasts.
Summary of FMngs/AuguM 1900
With the Exception of Calories, Breakfasts Served Before and
After NSMP Provided More than One-fourth of the RDA
Elementary School Breakfasts
100%
80%
60%
40%
20%
0%
98% 98%
CalortM Protein Vitamin A Vitamin C Calcium Iron
Secondary School Breakfasts
100%
80%
60%
40%
20%
0%
Catorfaa Protaln Vitamin A Vitamin C Calcium Iron
Before NSMP After NSMP
Difference B statistically significant at the S percent level
Difference s statistically significant at the 1 percent level
Evaluation ofthaNutrtant Standard Manu Wanning PainonsUatton 13
Breakfasts Served After NSMP Provided Fewer Calories from Fat
end Saturated Fat than Breakfasts Served Before NSMP
Elementary School Breakfasts
65% •
70%
60%
50%
40%
30%
20%
10%
0%
Total Fat Saturated Fat Carbohydrate Protein
Secondary School Breakfasts
64%**
Total Fat Saturated Fat Carbohydrate Protein
| Dietary Guidelines or NRC Recommendation
■ Before NSMP ■ After NSMP
•• Omtnna
is iMBEWy signifkent at the 5 percent ftnvE
o rtatBtxart, ognaXant at the 1 percent lev*
With the Exception of Sodium in Secondary Schools, Breakfasts Served
Before and After NSMP Met NRC Recommendations for Cholesterol
and Sodium
Elementary School Breakfasts
1500
H
10
<600
537 545 II
Sodium Fiber
Secondary School Breakfasts
100 1500 10
713
■I Sodium
| NRC Recommendation
Before NSMP BJ After NSMP
* Ofhrtnct o Mjfoacjrfy itgnhcant * ttw 5 ptrant m*
of the Nutrient Standard Menu Planning Demonstration 15
Weighted and Unweighted Nutrient Analyses
A key component of NSMP is use of a weighted nutrient analysis to determine nutrient content of
planned menu*. A weighted analysis incorporates information about student selection patterns and
does not assume that every student take* one serving of every type of food offered (e.g., milk, entree,
fruits, vegetable*). This approach provides a picture of the average lunch served to students. In
contrast, an unweighted nutrient analysis represents a simple average of all foods offered, providing a
picture of the average toioS offered to students.
Menus planned before and after NSMP were analyzed using both weighted and unweighted analysis
techniques and the results of the two analyses were compared. Findings for weighted and unweighted
analyses of Spring 1997 lunch menus are summarized below. The reader is cautioned that the patterns
reported here are not necessarily representative of what can be expected for programs nationwide.
The sample of SFAs included in this analysis was very small and was self-selected. Moreover, some of
the SFAs did not complete NSMP implementation and those that did planned Spring 1997 menus
using weighted analysis. It is possible that the results presented here may have been different if all
SFAs had been included or if SFAs had planned Spring 1997 menus using unweighted analysis.
Unweighted analysis of NSMP lunch menus resulted in greater estimated contributions to RDAs
for calories and ke> nutrients (Exhibit 11). The disparity between weighted and unweighted
estimates is greatest for vitamins A and C, and tends to be larger for secondary school lunches.
Regardless of which analysis is used, however, NSMP lunch menus in both elementary and secondary
schools provided one-third or more of the RDA for calories and key nutrients with the exception of
calories in secondary schools. Using weighted analysis, secondary school lunch menus, on average.fell
just short of the one-third RDA target for calories.
Weighted and unweighted estimates of the percentage of calories from fat, saturated fat, carbohydrate,
and protein in NSMP lunch menus were very similar (Exhibit 12). Regardless of which analysis is
used, NSMP lunch menus did not meet Dietary Gutdeimes goals for calories from fat or saturated fat.
Unweighted analysis of NSMP lunch menus resulted in somewhat greater estimates of cholesterol,
sodium, and fiber content (Exhibit 13). Although some of these differences are statistically
significant, they do not affect conclusions about whether NSMP lunch menus met NRC
recoinmendarions.
Weighted and Unweighted Analyses Produced Different Mean
Values for Calories and RDA Nutrients, But Conclusions
About Whether Lunches Met NSMP Standards Are Similar
Spring 1997 Elementary School Lunches
103%
106%*
v^e-"wfT1W rBr-o«iArWalf«l A Vitamin C Calcium Iron
Spring 1997 Secondary School Lunches
100%
Calorie* ProMn Vitamin A Vitamin C Calcium Iran
Weighted Unweighted
a suasocatf significant it me 5 percent
a subsocttf MpvAcanr m me ' pmtmt
ofaw Watnawt Standard Mami Hanntnq DamwmiaUun 17
Weighted and Unweighted Analyses Produced Similar Results
for Calories from Fat, Saturated Fat, Carbohydrate, and Protein
60%
Spring 1997 Elementary School Lunches
>"* 54% 53%
20%
10%
Total Fat Saturated Fat Carbohydrate Protoin
Spring 1997 Secondary School Lunches
60%
52% 52%
rCm^4t M^aarlDi iOll..n..yI.n-mj , I—TOT■ M ■T I
I Dietary GukMnes or NRC Recommendation
Wetghted ■ Unweighted
o jcaittfcafy ugnHcmt m ttm 5 ptran mm.
Weighted and Unweighted Analyse* Produced Different Mean
Value* for Cholesterol, Sodium, and Fiber, But Conclusions
About Whether Lunch** Met NRC Recommendations Are Similar
Spring 1997 Elementary School Lunches
100 1800
Sodium Fiber
Spring 1997 Secondary School Lunches
100 1800 10
Sj NRC Recommendation
Weighted SJ Unweighted
a XMntKaKy ugnficarf m ttm 5 pmcmtt Intl.
it tumksff stgnmcmt M tht 1 pmant m«
cvMUamHMi ovine rvMintmManOafu Mtmi Planning Uvmonscnmon 19
Program Operations Before and After MSMP
UM at NSMP did not influence NSLP participation rate* (Exhibit 14). The overall rate of
student participation in the NSLP was essentially constant over the course of the demonstration.
SBP participation rates increased significantly over time, however, the sizeable increase that was
observed (a 26% increase in overall participation, from 6.5% to 8.2%) probably cannot he
attributed to NSMP. Each of the SFAs that showed a substantial increase in breakfast
participation over time had made a concerted effort to increase SBP participation, particularly
among middle and high school students, as part of an ongoing national initiative.
The vast majority of kitchen managers did not believe that NSMP had an impact on levels of
plate waste. More than two-thitds of kitchen managers reported that the level < 4 plate waste
observed after NSMP, for a variety of different food items, was no different than what was
observed prior to NSMP.
Moreover, with the exception of cooked vegetables and, to » lesser extent, enttees in elementary
and middle schools, changes that were reported we ; largely positive, i.e., managers indicated that
students were wasting less food since the implementation of NSMP. Fifteen to 25 percent of
kitchen managers reported that students wasted more cooked vegetables aftet implementation of
NSMP than before NSMP. Fourteen to 18 percent of managers in elementary and middle schools
reported an increase in waste of lunch entrees.
Average Dally NSLP Participation Remained the Same Over
the Course of the Demonstration
100%
80%
60%
40%
20%
0%
Uudma Bj/m for
Bsfort NSMP After NSMP
1
SfAv
More SFA» used centralized Menu planning and cycle menus after implementing NSMP
than before NSMP. Four of six fully implemented SFAs that entered the demonstration with a
decentralized menu planning system switched to a fully centralized system. Eight of eleven fully
implemented SFAs that did not use a cycle menu prior to NSMP had adopted one by the end of
the demonstration.
Most SFA directors spend the same amount of time on NSMP maintenance as they spent on
traditional menu planning activities prior to NSMP. Directors in nine of the 16 SFAs that
achieved full implementation reported that NSMP maintenance requires about the same
amount of their time as routine menu planning activities required before NSMP. Three
directors reported that NSMP maintenarv e requires less of their time than previous menu
planning. Only four directors reported that NSMP maintenance takes more of their time than
previous menu planning practices.
About half of the SFA directors reported that other staff members spend more time on
NSMP maintenance than on previous menu planning practices. Required NSMP activities
that reportedly contribute to an increased level of staff effort include: reviewing and entering
data for purchased products; collecting and synthesizing menu production data; updating
nutrient analyses with new production data or early substitutions; and continuing to modify and
test recipes.
State agencies plan to use a variety of approaches in scheduling SMI and CRE reviews Of
12 State directors who had made decisions about how to incorporate SMI reviews, five States
were conducting combined SMI/CRE reviews. Four States were conducting SMI and CRE
reviews on separate schedules, requiring two visits to each SFA every five years. The remaining
three States planned to use both scheduling approaches, depending on the circumstances in
each SFA.
Almost all State directors using combined SMI/CRE reviews reported that these reviews took
more time than traditional CRE reviews. Three State directors estimated the increase to be
about four hours and two estimated it to be six to eight hours.
21
Program Casts Before and After NSMP
The co»t of producing reimbursable meal* under NSMP wa» comparable to or lower than the cost
before NSMP (Exhibit 15). Pre-NSMP and NSMP meal cow* were similar for middle and high
school lunches. NSMP costs were significantly lower than pre-NSMP costs for lunches in elementary
schools ($1.23 per lunch versus $1.36. in 1997 dollars). While it is possible that NSMP contributed to
this decrease (three SFA directors reported that costs declined after implementing NSMP), another
possible explanation is an increase in the number of meals produced over time, with a stable labor pool
and fixed other costs. (Although participation remained constant, enrollment — and the number of
lunches produced — increased in most districts.) Average costs for NSMP breakfasts were comparable
to pre-NSMP breakfasts.
Annual a la carte revenues were comparable before and after NSMP. Although average annual a la
cane revenues increased about I 5 percent per year, from $42.75 per student before NSMP (1997
dollars) to $44.74 per student af er NSMP, this difference was not statistically significant.
NSMP did not affect SFAs' utilization of donated commodities. The average level of commodity
assistance received after implementation of NSMP, expressed on a per-meal basis, was lower than the
level of assistance received before NSMP ($0,149 versus $0,163, in 1997 dollars). This difference was
not statistically significant and is attributable to both a decline in the entitlement over time and a
On Average, the Coot of Producing a Reimbursable Lunch Under
NSMP Wat Comparable to or Lower than the Cost Before NSMP
I
$2 00
$150
3 $100 I $0 50
$000
Harare NSMP After NSMP
** DOfnnet H «SfMej«y agntksn m rtm l puctn mm:
Souc* On tan 1« My nvtamnM SfAt
large disparity in the amount of bonus commodities available in the two .school years. On a dollar-for-dollar
basis, the availability of bonus commodities in SY 1993-94 (before NSMP) was more than three
times as great as the availability in SY 1996-97.
Although all SFAs entered the demonstration with computer capabilities, about half of the SFAs
had to purchase computer hardware in order to implement NSMP. The total cost of hardware
purchases ranged from $500 to $6,500, with a median of $2,867. The most common purchase was a
computer with increased processing power. Other purchases included computer work stations, needed
in order to provide staff working on NSMP with more immediate access to computer hardware.
Eight SFAs had to purchase software. SFAs that did not have to purchase software either owned
nutrient analysis software prior to the demonstration or received a complimentary copy because they
were participating in the demonstration. Costs for software purchases ranged from $475 to $4,600,
with a median of $555. The $4,600 expenditure was incurred for a nutrient analysis module of a
multi-faceted software package for food service program management.
Seven SFAs had to purchase other goods or services in order to implement NSMP. The most
common purchase was gram scales needed for accurate measurement of ingredients and portions. The
largest equipment purchase reported by an SFA was $53,750 for 35 milk coolers needed to store an
expanded array of fruit juices offered in NSMP lunches.
Total costs for hardware, software, and other goods and services ranged from zero to more than
$56,000, with a median of $210. Excluding the site with the milk cooler purchase, the maximum
cost for NSMP-related purchases was $11,100. Among SFAs that had at least one hardware, software,
or other NSMP-related purchase, the minimum cost was $200 and the median cost was $3,900.
of tn# Nirtnant scandaro
Home Economics and Household Collection
