Journal of Hospital Infection (2005) 59, 131–137

www.elsevierhealth.com/journals/jhin

Evaluation of the quality of hospital food from the

kitchen to the patient
H. Réglier-Poupeta, C. Paraina, R. Beauvaisa, P. Descampsa, H. Gilletb,
J.Y. Le Peronb, P. Berchea, A. Ferronia,*
a
Laboratoire de Microbiologie, Hôpital Necker-Enfants Malades, 149 rue de Sèvres, 75743 Paris cedex 15,
France
b
Cuisine Centrale, Hôpital Necker-Enfants Malades, 149 rue de Sèvres, 75743 Paris cedex 15, France

Received 10 October 2003; accepted 16 July 2004

KEYWORDS Summary Food-borne pathogens can multiply if food is not maintained at

Hospital food; Cold
chain; Microbiology
an appropriate temperature and if there are delays between food
preparation and distribution. The aim of this study was to evaluate the
quality of meals during transport from the kitchens to the patients in three
departments of a university hospital. Meals were transported inside
insulated, cooled food carts. We analysed the delays at each step of the
transport process, and measured the temperature inside the food cart and
inside the meals. The total duration of the transport (meanZ85.3 min;
range 44–123 min) conformed to the official recommendations (!2 h at a
temperature !10 8C before consumption). The internal temperature of
73.6% of the 30 food carts followed was below 10 8C. The internal
temperature of the meals was below 10 8C in 91.7% of cases when the
food cart was first opened, but in only 12% of cases by the time the last
patient was served. No pathogens were isolated from any of the samples.
However, 10% of meals, all of which were salads, had total viable counts of
bacteria above the recommended limits. This study confirms that it is
essential to control time and temperature to ensure food quality and safety
in hospitals.
Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights
reserved.

* Corresponding author. Tel.: C33-1-44-49-49-61; fax: C33-1-

44-49-49-60.
E-mail address: agnes.ferroni@nck.ap-hop-paris.fr

0195-6701/$ - see front matter Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jhin.2004.07.023
132
Introduction
In 1999–2000, 1267 outbreaks of food-borne disease
were reported to the French health authorities.
This number has been stable since 1998. The
surveillance data showed that Salmonella (64%),
Staphylococcus aureus (16%) and Clostridium per-
fringens (5.1%) were the most commonly identified
aetiological agents.1 As clinical manifestations are
common in hospitalized patients, the true inci-
dence of outbreaks of food-borne disease in
hospitals and extended-care facilities is not
known. It is unlikely that all events are reported.
Hospitalized people are more likely than non-
hospitalized people to become ill when exposed to
food-borne agents. Small numbers of enteric
pathogens that may be innocuous to most healthy
people can cause disease and even death in highly
susceptible patients, especially in immunocompro-
mised subjects. The main factors that contribute to
the occurrence of food-borne disease are: keeping
food at temperatures outside the recommended
range; inadequate cooking; poor personal hygiene
among food handlers; use of food from unsafe
sources; and use of contaminated equipment. As
hospitalized patients are at increased risk of
becoming ill when exposed to potential food-
borne pathogens, and as hospital food services
need to provide a wide variety of dietary items, it is
critical that appropriate food-handling practices
are maintained. The aim of this study was to
analyse the quality of hospital meals from prep-
aration to distribution to patients. We studied the
transport conditions and their possible effects on
the microbial quality of different food items.
Methods
Organization of the central kitchen
In the Necker-Enfants Malades Hospital (a 795-bed
university hospital), the central kitchen prepares an
average of 813 meals each day using the cold chain
principle. The process is described in Figure 1. The
quality of food is checked at several stages during
delivery of the food products. When appropriate,
food is stored in cold rooms. Foods are cooked,
rapidly chilled (internal temperature !10 8C in less
than 2 h) and refrigerated. On the day of consump-
tion, the food is divided into portions and plated out
as individual meals. Meals are delivered to the
clinical department in insulated, cooled carts. Food
is then reheated before being given to the patients.
H. Réglier-Poupet et al.
Study protocol
This prospective study was carried out over a three-
month period. Every lunch time, an investigator
followed the meals from their preparation in the
central kitchen to their delivery to the patients in
three clinical departments. These wards were
located in different areas of the hospital. A form
was completed for each food cart, stating the type
of food delivered and the critical times in the
distribution of the meals: T1Zentry of the food
carts into the cold room; T2Zexit of the food carts
from the cold room; T3Zarrival in the clinical
department; T4Zfirst opening of the food carts;
T5Zfirst food tray served; and T6Zlast food tray
served. The number of food trays delivered and
errors were noted.
Temperature monitoring
Three temperature recorders (Testor 175, Testo,
Forbach, France) were used to measure the
temperatures inside the food carts at the sites
indicated in Figure 2. These recorded the tempera-
ture every 5 min and plotted temperature as a
function of time.
Two control meals per day were prepared in the
same way as dishes intended for the patients in
order to monitor temperature. A TESTO106-T1
thermometer was used to measure the internal
temperature of the food carts when they were first
opened and when the last food tray was served (just
before microwave reheating).
The temperature of the kitchen in each clinical
department was also recorded.
Microbiological analysis
Two control meals, identical to those used for
checking the temperature, were used for micro-
biological analysis. The first control was taken at
the end of refrigerated storage in the central
kitchen, and the second control was taken when
the last patient was served. All samples were kept
at 4 8C in the laboratory and examined on the same
day. Specific pathogens were routinely sought
according to standard French procedures: meso-
philic organisms (NF V08-051), total and faecal
coliforms (NF V08-050), S. aureus (NF V08-057),
anaerobes (NF V08-019) Salmonella (NF V08-052),
Yersinia and Listeria.2
Briefly, 25 g of food was placed in a sterile plastic
bag. Samples were homogenized in a stomacher
(CML, Nemours, France) in peptone water (total
volume 50 mL) and incubated at 37 8C for 30 min.
The suspension was then centrifuged and 10-fold
dilutions of the supernatant from 10K1 to 10K4
Evaluation of the quality of hospital food from the kitchen to the patient
Figure 1 The cold chain in the central kitchen of the hospital. The recommended time and temperature are indicated
for each stage from the end of cooking until consumption.
were made. Baird-Parker agar was used to isolate
S. aureus (Biorad, Marnes la Coquette, France),
desoxycholate agar for coliforms (BD Bioscience, Le
Pont de Claix, France), CIN agar for Yersinia (Oxoid
Dardilly, France), Oxford agar for Listeria mono-
cytogenes (Oxoid), plate count agar (bioMérieux,
Marcy l’Etoile, France) for mesophilic bacteria and
sulphite iron agar for anaerobes (Biorad).
The results were interpreted according to the
French recommended limits shown in Table I. Foods
were automatically deemed to be unsatisfactory if
Salmonella, Yersinia or Listeria were found.
Statistical analysis
Data were compared by use of Fisher’s test and
Figure 2 Example of the way in which a food tray is
loaded. X, sites at which temperature was measured.
133
ANOVA (Statview software). P values of !0.05 were
considered to be significant.
Results
Transport of the meals
Thirty food carts containing a mean of 20 meals
were followed between 17 May 2002 and 31 July
2002. The delays observed between the critical
times described above are summarized in Figure 3.
Twenty-six carts were stored in the cold room. The
other four carts were delivered directly to the
clinical departments. The length of Delay 2
(between leaving the cold room and arriving in the
clinical department) depended on the distance
between the hospital kitchen and the department,
and on whether a lift was used. For example, the
delay was longest for the department on the fourth
floor.
The total duration of transport (meanZ
85.3 min; range 44–123 min) conformed to the
recommendations of the decree of 29/9/1997,
which stipulated that food should be kept for a
maximum of 2 h at a temperature !10 8C before
consumption.
If the patients were not present when the meals
were served, the trays were placed in the refriger-
ator in the satellite kitchen.
134 H. Réglier-Poupet et al.

Table I French recommended microbiological limits for different items of food

Bacteria Number cfu/g
Good Acceptable Unsatisfactory
Salads
Mesophilic organisms !105–1.5.106 1.5.106–5.106 O5.106
Faecal coliforms !10–30 30–100 O100
Meat
Mesophilic organisms !5.105–1.5.106 1.5.106–5.106 O5.106
Faecal coliforms !102–3.102 3.102–103 O103
S. aureus !102–3.102 3.102–103 O103
Anaerobes !30–90 90–300 O300
Cooked plates and plates with eggs
Mesophilic organisms !3.105–9.105 9.105–3.106 O3.106
Faecal coliforms !10–30 30–100 O100
Total coliforms !103–3.103 3.103–104 O104
S. aureus !102–3.102 3.102–103 O103
Anaerobes !30–90 90–300 O300
Vegetables
Mesophilic organisms !5.105–1.5.106 1.5.106–5.106 O5.106
Faecal coliforms !10–30 30–100 O100
Total coliforms !103 3.103–104 O104
S. aureus !102–3.102 3.102–103 O103
Anaerobes !30–90 90–300 O300

Temperature monitoring had an internal temperature of below 10 8C. The

During transportation, 22 of the 30 food carts
(73.6%) had an internal temperature below 10 8C.
There was no difference between clinical depart-
ments. The only parts of the carts that were not
below 10 8C were the upper and lower levels. There
were reasons for these exceptions: the freezing pad
of the food cart was not completely frozen; one
food cart was washed with warm water just before
transport because of a maintenance problem; and
the door of the food cart was left open.
We tested the internal temperature of 60 meals.
The initial measurement was taken when the food
cart was first opened, and a second measurement
was taken when the last patient was served. When
the food cart was first opened, 55 plates (91.7%)
other five dishes were slightly too warm, all
between 10 8C and 12 8C. In two of these cases,
the temperature inside the food cart was O10 8C
during transportation.
The internal temperature of the last patient’s
meal was between K1.4 and 23 8C (mean: 14.9 8C).
The temperature increased by an average of 6.9 8C
between the first opening of the food cart and the
last patient being served (0.3–15.6 8C). When the
last patient was served, 88% (53/60) of the meals
were O10 8C (Figure 4). Eight of these meals were
at room temperature (20–25 8C). The temperature
was not found to be dependent on the food type
(PZ0.853; ANOVA).
The trays were prepared in a satellite kitchen in
the clinical department. The temperature in these

Figure 3 Description of the different delays observed during the meal transport process. Delay 1, time between
arrival of food carts in the cold room and leaving the cold room; Delay 2, time between the food cart leaving the cold
room and arriving in the clinical department; Delay 3, time that the cart spent in the department with its doors closed;
Delay 4, time taken to make up the individual trays; and Delay 5, time between the first and the last food tray served to
the patients. Each number indicates the mean (range) of the delays observed for the 30 food carts followed in the three
wards (in minutes).
Evaluation of the quality of hospital food from the kitchen to the patient
Figure 4 Internal food temperature. Internal food
temperature was noted for 60 plates when the food cart
was first opened (open bars) and when the last patient
was served (solid bars).
kitchens varied from 17 8C to 27.3 8C. This tem-
perature and the time spent at room temperature
affected the internal food temperature (P!0.05;
ANOVA).
Bacteriological analysis
The microbiological contents of 51 pairs of food
samples (one taken on leaving the hospital kitchen
and the other taken from the same meal when it
Figure 5 Bacteriological interpretation of the different
meals. The culture results were interpreted according to
the published norm (see Table I). The percentages of
unsatisfactory (solid bars), acceptable (stippled bars) and
good (open bars) meals were calculated for each type of
food.
135
was served to the last patient) were compared.
Different items were cultured: 40 salads, 22 cooked
meals with sauce, 22 meat dishes, 12 vegetable and
six preparations containing eggs (Figure 5).
The only unsatisfactory meals were the salads:
10% of the total meals and 25% of the salads. We
detected coliforms (O10 cfu/g) in one dish
(Chinese salad).
We detected a high rate of mesophilic bacteria in
salads (55% of salads) and in preparations contain-
ing eggs (65%). No S. aureus, Salmonella spp.,
Yersinia spp., Listeria spp. or anaerobes were
isolated from any of the samples.
Seven meals changed category between leaving
the central kitchen and being served (Table II). One
meal was ‘good’ when it left the central kitchen but
was only ‘acceptable’ when it was served. Six meals
were ‘acceptable’ when they left the central
kitchen but were ‘unsatisfactory’ when they were
served. Salads were involved in all cases. The
multiplication factor of bacterial growth varied
from 1 to 25 (data not shown).
Discussion
This study made it possible to locate the potential
problems concerning the time, temperature and
microbiological quality of hospital food during
distribution. This is the first evaluation of this
type in France. We followed each step of the
transport of meals from the central hospital kitchen
until consumption by the patients. In most cases,
the cold chain was maintained during the transport
of food in our hospital. The cold chain system makes
it possible to offer flexible menus, and ensures the
efficiency of food production and microbial quality
even if certain psychotropic organisms such as
Listeria and Yersinia can multiply at 4 8C. However,
as this system is sensitive to even the smallest
dysfunction, repeated controls and continuous staff
training are essential.
By studying the whole process, especially the
transport of meals, we were able to detect some
points that could be improved, even if the overall
timing of the transport process conformed to the
recommendations. For example, we noted that a
long delay between the arrival of the food cart in
the clinical department and the first opening of the
food cart (up to 50 min) can increase the internal
temperature of the food. Meals should not arrive in
the clinical department too early as the availability
of staff to serve the meals changes from day to day.
Moreover, after opening the cart, the meals were
often left at ambient temperature for a long time
136
Table II Interpretation of the microbiological results
Interpretation of the microbiological Good
results on leaving the central kitchen Acceptable
Unsatisfac-
tory
The same meal was analysed when it left the hospital central kitchen and when the last patient was served. A total of 51
microbiological comparisons were made.
before being consumed, thus making it possible for
bacteria to grow.
The number of trays and the organization of the
staff who prepare them for serving influence the
preparation time. Two systems were observed. In
one case, all the trays were prepared and then
reheated after their preparation by one person. In
the other case, one person prepared the trays and
another person reheated them one by one. This
latter situation was more efficient, but required
two people. It was noted that when the patient was
not present for the meal, the tray was stored in the
refrigerator of the local kitchen, but only after the
distribution of all the other meals.
Holding food at temperatures outside the rec-
ommended range is an important and common error
responsible for food-borne disease outbreaks.3,4
Our study allowed us to follow up, for the first time
in our hospital, the temperatures inside the food
carts and inside the meals. Foods were maintained
at appropriate temperatures in most cases. In the
other cases, the errors responsible for this tem-
perature increase were analysed and corrected.
These abnormalities may alter the food quality.
Thus, it was important to compare the bacterio-
logical status of the meals when they left the
central kitchen and when the last patient was
served. The increase in the total number of bacteria
was probably due to the delay between opening the
cart in the ward and the distribution of meals to
patients, particularly in the summer when the
ambient temperature can reach 30–35 8C. The
poor bacteriological results obtained for salads,
even when they left the kitchen, may lead us to
carry out a more thorough study of the distribution
of this type of food.
The absence of pathogenic bacteria, such as S.
aureus, Listeria, Yersinia, Salmonella and anae-
robes, was not surprising because these bacteria
are generally found to be causes of outbreaks.4–7
Over the past five years, we have analysed 795
items of food distributed in our hospital. A control
H. Réglier-Poupet et al.
Interpretation of the microbiological results at
the time the last patient was served
Good Acceptable Unsatisfac-
tory
36 1 0
– 6 6
– – 2
tray containing three different plates is tested
weekly. We found C. perfringens in one case,
coliforms in 20 dishes and S. aureus in seven
cases. The salad plates were most likely to be
contaminated: 11% had total viable counts above
the recommended limits (personal data). These
data are in accordance with the literature. In 1988,
Sandys and Wilkinson detected no pathogens in
3393 food items, but showed that food such as
salads often contained total viable counts greater
than 105 cfu/g.8,9
This study stresses that staff need to be aware of
each stage where system failure can occur. The
time/temperature control appeared to be the most
critical issue throughout the whole process. A
Hazard Analysis Critical Control Point (HACCP)
approach should be adopted to ensure quality and
safety in food production services. Few studies have
looked at the implementation of the HACCP system
in hospitals.10–13 In France, the decree of 1997
proposed that this approach should be adopted by
Assistance Publique des Hôpitaux de Paris (AP-HP)
hospitals. A HACCP team responsible for co-operat-
ing with the food services in developing written
policies and procedures was constituted. Moreover,
personnel should be trained in good food-handling
practices. However, each hospital has to define the
periodicity of the different time/temperature con-
trol and the microbiological analyses. The HACCP
method is currently applied in the satellite kitchens
in the wards.
References
1. Haeghebaert S, Le Querrec F, Gallay A, Bouvet P, Gomez M,
Vaillant V. Les toxi-infections alimentaires collectives en
France en 1999 et 2000. B.E.H. 2002;23:105—109.
2. Association Française de Normalisation. Microbiologie ali-
mentaire, Tome 1: Máthodes horizontales, 2nd edn. Paris:
AFNOR; 1999.
3. Cookson ST, Hughes JM, Jarvis WR. Nosocomial
Evaluation of the quality of hospital food from the kitchen to the patient
gastrointestinal infections. In: Wenzel RP, editor. Preven-
tion and control of nosocomial infections. 3rd edn.
Baltimore: Williams and Wilkins; 1997. p. 925—975.
4. Slutsker L, Villarino ME, Jarvis WR, Goulding J. Foodborne
disease prevention in healthcare facilities. In: Bennet JV,
Brachman PS, editors. Hospital infections. 4th edn. Phila-
delphia: Lippincott Raven; 1998. p. 333—341.
5. Dryden MS, Keyworth N, Gabb R, Stein K. Asymptomatic
foodhandlers as the source of nosocomial salmonellosis.
J Hosp Infect 1994;28:195—208.
6. Graham JC, Lanser S, Bignardi G, Pedler S, Hollyoak V.
Hospital-acquired listeriosis. J Hosp Infect 2002;51:
136—139.
7. Mason BW, Williams N, Salmon RL, et al. Outbreak of
Salmonella indiana associated with egg mayonnaise sand-
wiches at an acute NHS hospital. Commun Dis Public Health
2001;4:300—304.
8. Sandys GH, Wilkinson PJ. Microbiological evaluation of a
hospital delivered meal service using pre-cooked chilled
foods. J Hosp Infect 1988;11:209—219.
137
9. Remington Jack S, Stephen C, Schimpf F. Please do not eat
the salad. N Engl J Med 1981;304:433—434.
10. Angellilo IF, Viggiani NM, Greco RM, Rito D. HACCP and food
hygiene in hospitals: knowledge, attitudes, and practices of
food-services staff in Calabria, Italy. Collaborative group.
Infect Control Hosp Epidemiol 2001;22:363—369.
11. Boreng BJ, David BD. Development of Hazard Analysis
Critical Control Points model. HACCP models for Q.C. of
entree production in hospital food service systems. J Am
Dietetic Assoc 1978;73:524—529.
12. Shanaghy N, Murphy F, Kennedy K. Improvements in the
microbiological quality of food samples from a hospital cook-
chill system since the introduction of HACCP. J Hosp Infect
1993;23:305—314.
13. Richards J, Parr E, Riseborough P. Hospital food hygiene: the
application of hazard analysis critical control points to
conventional hospital catering. J Hosp Infect 1993;24:
273—282.