The Currently Available Literature on Inpatient Foodservices: Systematic Review and Critical Appraisal

RESEARCH

Review

The Currently Available Literature on Inpatient Foodservices: Systematic Review and Critical Appraisal Dorian N. Dijxhoorn, MD, PhD; Merwin J. M. J. Mortier; Manon G. A. van den Berg, PhD; Geert J. A. Wanten, MD, PhD, MSc ARTICLE INFORMATION Article history: Submitted 20 March 2018 Accepted 25 January 2019

Keywords: Foodservice Hospital Protein Satisfaction Systematic review

Supplementary materials: Table 1 and Table 3 are available at www. jandonline.org 2212-2672/Copyright ª 2019 by the Academy of Nutrition and Dietetics. https://doi.org/10.1016/j.jand.2019.01.018

ABSTRACT Background An adequate hospital foodservice is important to optimize protein and energy intake and to maintain or improve a patient’s nutritional status. Key elements that define an optimal foodservice have yet to be identified. Objectives To systematically describe the effects of published foodservice interventions on nutrition and clinical outcomes and determine which elements should be considered essential. Secondly, to describe the outcome measures used in these studies and evaluate their relevance and validity to guide future research. Methods PubMed, Embase, the Cochrane Library, and the Web of Science databases were searched. Studies that included assessment of nutrition and/or clinical outcomes of hospital foodservice up to December 2017 were eligible. The details of the subject population, the type of intervention, and the effects on reported outcomes were extracted from each study. Results In total, 33 studies that met inclusion criteria were identified, but only nine (27%) were rated as having sufficient methodologic quality. These nine studies concluded that various elements of a foodservice can be considered essential, including using volunteers to provide mealtime assistance, encouraging patients to choose protein-rich foods, adding protein-enriched items to the menu, replacing existing items with protein-enriched items, giving patients the ability to order food by telephone from a printed menu (room service concept), or a combination of these interventions. The interstudy heterogeneity was high for both outcome measures and methods. Conclusions Various foodservice interventions have the potential to improve outcome measures. Recommendations are made to facilitate future research. J Acad Nutr Diet. 2019;-:---.

A

N ADEQUATE HOSPITAL FOODSERVICE IS IMPORtant for maintaining or improving patient nutritional status. Optimal nutrient—especially protein— intake is considered crucial for patient health and outcomes.1 The relevance of adequate hospital foodservice is underscored by the fact that as many as 40% of inpatients experience malnutrition, and this rate increases with the duration of the hospital stay.2 Malnutrition is associated with numerous complications, including loss of functional tissue (ie, muscle), increased infection rates, delayed wound healing, and a prolonged hospital stay. These complications have a substantial influence on health care resources.3-6 Strategies used to provide foodservices for hospital patients differ widely nationally and internationally, and the development of each service seems to be mainly driven by logistic and financial considerations. According to the literature, a wide range of hospital foodservice interventions is available and includes changes in food preparation or selection, mealtime assistance programs, and protection of mealtimes. Yet the key elements that define an optimal foodservice have not yet been identified.7,8 In addition, currently available systematic reviews

ª 2019 by the Academy of Nutrition and Dietetics.

on this topic tend to focus narrowly on one specific intervention. Outcome measures reported in these studies also range widely from food intake to nutritional status, and the way of measuring each outcome differs across the studies.9 To improve nutrient intake during hospitalization and address the high prevalence of malnutrition and associated complications, essential elements of hospital foodservices should be determined. Therefore, the aims of this study were to systematically describe the effects of available foodservice interventions on nutrition, clinical, and patient-reported outcomes and determine those elements that should be considered essential. Secondly, we aimed to discuss the outcome measures used in these studies and evaluate their relevance and validity in guiding future research.

MATERIALS AND METHODS Data Sources and Search Strategy A systematic literature search was performed in PubMed, Embase, the Cochrane Library, and the Web of Science through December 2, 2017. The following medical subject

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

1

RESEARCH heading terms were used: food service, hospital, menu planning, and hospitalization. The search was expanded as described in Figure 1 to identify additional relevant studies using the snowball search method. The search was restricted to randomized controlled trials, cohort studies, and crosssectional studies. No restrictions were imposed on language or publication date.

Eligibility Criteria The articles included in the systematic review met the following criteria: adult hospitalized patients, nutrition and/ or clinical outcomes, and intervention in a hospital foodservice. For the intervention criterion, an intervention in one of the following subcriteria was utilized: preparation and composition of meals, menu, meal delivery, mealtime assistance, or mealtime environment. Studies that met any of the following criteria were excluded: patients on parenteral and/ or tube feeding, financial or health worker-centered outcome measures, nonhospital facilities and services (eg, nursing homes, rest homes, assisted-living facilities, and homedelivered meal services), or limited food preparation changes (eg, reducing the use of salt in meal preparation). When questions were raised regarding the article or in case the full text was not available online, the authors were contacted. The article was excluded if efforts to contact the author(s) were unsuccessful.

Study Selection Two reviewers independently screened the title, abstract, and full text. Any discrepancies were resolved through discussion until consensus or by bringing in a third reviewer. Any duplicate studies were excluded.

Data Extraction The following data were extracted by the two reviewers and reported in Table 1 (available at www.jandonline.org): first author; year of publication; study design; number of patients; type of patient group; mean age; study duration; type of control and intervention; outcome measurement(s); how outcomes were measured; results; and discussion, including reviewers’ judgment of study’s limitations and strengths. In all the included studies, the statistical significance was described using a significance level of a¼.05. The reporting followed the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.10

Quality Assessment Two authors independently performed a quality assessment for each study using the Quality Criteria Checklist for Primary Research (QCCPR), which includes criteria to determine the validity and bias of the study.11 This checklist has also been used in other systematic reviews regarding hospital foodservice interventions.12,13 The assessment of the overall quality of evidence per outcome using the Grades of Recommendation Assessment, Development and Evaluation system was not possible given heterogeneity of outcomes. Blinding (validity question 5) was rated not applicable when patients were aware of their group allocation due to the nature of the intervention, such as in volunteer assistance interventions or when outcomes were measured via survey, 2

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

RESEARCH SNAPSHOT Research Question: Which elements are considered essential to a hospital foodservice to improve nutrition, clinical, and patient-reported outcomes? Key Elements: A variety of foodservice programs and strategies have been used to improve patient outcomes. These include offering mealtime assistance, encouraging patients to choose protein-rich foods, implementing room service, adding protein-enriched menu items or replacing existing items with protein-enriched foods, or a combination of these interventions. questionnaire, or interview. To accurately measure and judge food intake (validity question 7), daily protein and/or energy intakes were either weighed after each meal or estimated by a validated measurement tool and compared with individual requirements. Validated questionnaires were used to measure satisfaction. Each article was rated as positive (strong quality, generalizability, data collection and analysis, and limited bias), neutral (neither exceptionally strong nor weak in quality), or negative (weak quality, data collection and analysis, and likely bias) according to the checklist guidelines. The missing details within each validity question yielded the rating “no.” Any disagreements over ratings were resolved by discussion until consensus or by bringing in a third reviewer.

Data Synthesis and Analysis To provide a concise overview of the included nutritional intervention(s), the types of intervention(s) were separately reported (Table 2). Before mealtime interventions included preparation and composition, menu, and meal delivery, and during mealtime interventions included mealtime assistance and protected mealtime. All term definitions were available from cited sources and listed in Figure 2. The effect on each outcome measure was stated as significantly positive, not different or significantly negative compared with the control group. A meta-analysis was impossible to perform because of the heterogeneity of the interventions and outcomes.

RESULTS Study Selection Based on the inclusion criteria, 8,486 titles and abstracts of potential interest were identified. After the exclusion of 2,902 duplicate articles, 5,584 records were screened, and 42 fulltext articles were assessed for eligibility. Of these eligible articles, 33 were included in the qualitative synthesis according to the flowchart in Figure 3.

Study Characteristics Four randomized controlled trials were included and most of the other studies included investigations that were prospective cohort studies, including seven with a crossover design. The studies were conducted in various countries and continents and between 1996 and 2017. The number of enrolled patients ranged from 8 to 1,012. --

2019 Volume

-

Number

-

RESEARCH Database

Search

Embase

((Hospital food service) OR (menu planning*.ti,ab. OR foodservice*.ti,ab. OR food service*.ti,ab. OR mealservice*.ti,ab. OR meal service*.ti,ab. OR food deliver*.ti,ab. OR meal provision*.ti,ab. OR meal deliver*.ti,ab. OR mealtime*.ti,ab. OR meal time*.ti,ab.)) AND ((hospitalization/) OR (hospital patient/ OR aged hospital patient/ OR hospitalized adolescent/) OR ((hospital* OR inpatient* OR patient OR patients).ti,ab.))

Pubmed

(((("Food Service, Hospital"[Mesh]) OR "Menu Planning"[Mesh])) OR ((menu planning*[tiab] OR foodservice* [tiab] OR food service*[tiab] OR mealservice*[tiab] OR meal service* [tiab] OR food deliver*[tiab] OR meal provision*[tiab] OR meal deliver* [tiab] OR mealtime*[tiab] OR meal time*[tiab]))) AND (("Hospitalization"[Mesh:noexp]) OR ((("Patients"[Mesh:noexp]) OR "Adolescent, Hospitalized"[Mesh]) OR "Inpatients"[Mesh]) OR (Hospital* [tiab] OR Inpatient*[tiab] OR Patient [tiab] OR patients[tiab]))

Web of science

“Food service” and (“Hospitalization” OR “Inpatient”)

Cochrane

("Food service" OR "menu planning" OR foodservice* OR "hospital food service" OR mealservice* OR "meal service" OR "meal deliver" OR "food deliver" OR mealtime* OR "meal time") AND (hospitalization OR hospital OR inpatient OR *patient)

Figure 1. Expanded search of a systematic review of inpatient foodservices (December 2, 2017).

Data Extraction Table 2 shows the type of foodservice intervention(s) in each study, the various outcome measures used, and the quality assessment of the study. Table 1 (available at www. jandonline.org) includes all the extracted data.

rating is shown in Table 2]). Nine studies were rated positive, 19 were rated neutral, and five were rated negative. The interrater reliability between both reviewers was high (96%); all but 14 of 330 validity questions were rated similarly.

Foodservice Interventions Table 2 summarizes the effects of the implemented foodservice interventions per study, including the definition of each intervention and details on outcome measures. The positively rated studies are listed at the top of Table 2.

Meal Preparation and Composition Interventions Eight of 11 studies (73%) reported a food fortification intervention.14-21 In these studies, a positive effect was established on one or more outcome measures (ratings: three negative, three neutral, and two positive). Of three studies that aimed to alter the method of hospital food preparation by serving food in a cart with integrated refrigeration and heating compartments or using a Steamplicity system (Compass Group Healthcare Services), two showed improvement in food intake and satisfaction,22,23 and one reported a negative influence on food and energy intake (ratings: neutral).24

Menu Interventions A total of 13 studies (3 rated positive, 8 rated neutral, and 2 rated negative) reported a menu intervention. In 12 of 13 studies (92%), a positive effect was established on one or more outcome measures by implementing interventions such as an à la carte system (n¼3) or by choosing menu items at the bedside (n¼6; quality rating ranges from negative [n¼2] to positive [n¼3]). Only one study, which was rated neutral, found notable negative effects on food- and energy intake.24

Meal Delivery Interventions Six studies (5 rated neutral and 1 rated negative) reported a meal delivery intervention. In these six studies, the interventions mostly pertained to different ways of transporting the meals from the kitchen to patients. Four of six studies (67%) implemented a bulk trolley system and reported positive effects on outcome measures.23,25-27 However, one study found no significant effects28 and one reported a negative effect.24 These meal delivery intervention studies were all rated neutral, except for one study that was rated negative.27

Mealtime Assistance Across the mealtime assistance studies (n¼10), volunteers assisted at various mealtimes during the day. Four studies were rated positive, five studies were rated neutral, and one study was rated negative. Four of 10 studies (40%) showed a positive effect on one or more outcome measures, such as daily protein intake (ratings: three were positive and one was negative).29-32 In six of 10 studies (60%), no significant improvements were observed (5 rated neutral, 1 rated positive).33-38

Risk of Bias Assessment

Mealtime Environment

All the included studies met the relevance criteria of the QCCPR, enabling the completion of the validity questions (Table 3, available at http://www.jandonline.org [overall

In total, seven studies (2 rated positive and 5 rated neutral) reported an intervention in the mealtime environment. In five of seven studies (71%), protected mealtimes were implemented,

--

2019 Volume

-

Number

-

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

3

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Beelen and colleagues14 2017

147

Dijxhoorn and colleagues31 2017

637

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

Positive þProtein intake The consumed portion was þMeeting determined by protein asking the requirements patient. Body þEnergy intake weight measured Carbohydrate using a scale intake Fat intake

Food fortification: Regular food items replaced by various protein-enriched intervention products Meal choice at bedside: At bedside, meal choice of 3 small protein-rich meals, 6 times a day

Nutritional assistants: Nutritional assistants recommended protein-rich choices and assisted in choosing most optimal menu item

Positive þEnergy intake Plates measured before and after þProtein intake consumption. þMeeting Satisfaction energy measured with a requirements validated þMeeting questionnaire protein requirements Satisfaction

(continued on next page)

RESEARCH

4

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies

--

2019 Volume -

Number -

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

-

337

Lindman and colleagues32 2013

86

Methods of measurement

QoRc

Positive Energy intake Intake calculated using food lists Protein intake (recorded twice þSatisfaction per week by Body weight nurse). Satisfaction measured with a nutrition-related quality of life questionnaire. Body weight and handgrip strength measured using scale and dynamometer

À la carte: Foods/ drinks ordered by telephone from a printed menu between 7:00 am and 7:00 pm. Delivered within 45 min

Food caregivers: Kitchen assistants trained as food caregivers serve snacks, guide patients, and motivate eating

þMeeting energy requirements Meeting protein requirements

Positive Food weighed before serving and leftovers registered by nurses. Daily intake calculated as the average of a 3-d assessment (continued on next page)

5

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Doorduijn and colleagues44 2015

Outcomesb

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year Manning and colleagues30 2012

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11) 23

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7) Volunteers at lunchtime: Volunteers assisted at lunch on weekdays trained to feed, encourage higher energy and protein portions of meals

Outcomesb

Methods of measurement

QoRc

Intake calculated by Positive Daily energy using standard intake average weight þDaily protein for serving sizes intake and weighing þLunchtime leftovers. Intake of energy intake in-between þLunchtime snacks noted protein intake through visual Meeting estimates and energy questioning. requirements Requirements þMeeting estimated using protein the Schofield requirements equation (continued on next page)

RESEARCH

6

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

--

2019 Volume -

Number -

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

-

84

Porter and colleagues39 2017

149

Methods of measurement

QoRc

Positive Energy intake Intake visually estimated and Meeting recorded in energy quartiles. requirements Requirements þMeeting calculated using protein basic metabolic requirements rate and þProtein intake estimated activity Handgrip factor. Protein strength requirement set Length of stay at 18% of energy requirement. Handgrip strength measured with dynamometer

Food fortification: Small, energyenriched menu items, supplemented with protein powder

Protected mealtimes: Doors were closed during mealtimes and unnecessary disruptions were minimized

Energy intake Intake measured Protein intake through visual þEnergy deficitd estimation  Protein deficit

Positive

(continued on next page)

7

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Munk and colleagues43 2014

Outcomesb

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Van der Zanden and colleagues45 2015

208

Young and colleagues35 2013

254

Encouragement to choose fortified foods: Verbal prompts given by telephone operators to encourage ordering of fortified foods

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

þNutrient-dense Food orders noted Positive foods orders and protein and caloric contents of food items were retrieved from the hospital database. Patients expressed to what extent they consumed the product in a questionnaire

--

2019 Volume

Positive Energy intake Intake through Volunteers during Protected visual estimation. þMeeting mealtimes: protected Estimated energy Nonurgent mealtimes: requirements activities and requirements Assistant-incalculated based interruptions Protein intake nursing staff or on body mass were limited, þMeeting members of index and body protein staff multidisciplinary weight requirements members teams assisted assisted with meals and during completing mealtimes orders, encouraging high-energy and protein options

-

(continued on next page)

RESEARCH

8

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

Number -

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

-

122

Food fortification: Visual menu and encouragement Fortified menu from staff: Visual that could menu provide 3,680 implemented for kJ/d and 24 g/d protein more midmeals to display than the Foodservice staff standard menu encouraged patients to choose an item at midmeals

þEnergy intake Weight measured Neutral using calibrated þProtein intake seated scales. Nutritional Handgrip status strength Satisfaction measured with dynamometer. Intake visually estimated using a validated 6-point scale. Satisfaction measured with verbal validated questionnaire

Edwards and colleagues22 2006

52

Meal choice at Change in food bedside: Meals preparation: ordered 2 h Ready-plated before meal chilled meals service from an transported to the ward, heated extended choice menu in microwave oven

þFood intake þSatisfaction

Neutral Plates weighed before and after consumption in the CG. The IG used average portion sizes from production records and measured plates after consumption. Satisfaction measured with interviews and a validated questionnaire

9

(continued on next page)

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Collins and colleagues15 2017

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Intervention During Mealtimea

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

þEnergy deficit Intake measured by Neutral Protein deficit patients, relatives and nursing staff using intake record sheets. Requirements estimated using the Schofield equation

Gall and colleagues17 1998

143

Food fortification: Nutrients with high energy density were added to the standard hospital menu

Goeminne and colleagues23 2012

189

Meal choice at Change in food bedside: Meal preparation: choice and Food served in portioning at cart with refrigerating and bedside during each mealtime heating compartment

Hartwell and colleagues25 2007

180

Food cart: Bulk in food cart with refrigerating and heating compartment, portioned at bedside

--

2019 Volume

Bulk trolley: Meal choice at Meals bedside: Meal transported choice at point of consumption, by bulk in extent of choice trolley, food stayed is limited heated and compared with standard system choice at point of service

þFood intake þSatisfaction

Neutral Plates weighed before and after consumption. Satisfaction measured with a validated questionnaire

þSatisfaction

Neutral Satisfaction measured with a consumer opinion card

-

Number

(continued on next page)

RESEARCH

10

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

-

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

-

592

Hickson and colleagues24 2007

57

Volunteers during the day: Additional care from health care assistants to reduce malnutrition and encourage/ enable patients to eat

Meals ordered 2 h Meals were Change in food plated and before meal: preparation: Meals ordered 2 sealed before Ready-plated delivery and h before meal chilled meals heated in service from an transported to the ward, heated extended choice microwave at wards menu in microwave oven

Methods of measurement

QoRc

Neutral Food intakee Plates weighed before and after Barthel score consumption. Infection rate Bodyweight Length of stay measured with Mid-arm electronic seat muscle scale. circumference Circumference of Mortality rate arm measured Weight gain and mid-arm muscle circumference calculated using a formula. Handgrip strength measured with dynamometer Neutral eEnergy intake Average meal eFood intake weight Protein intake determined by weighing three samples of every meal. Plate waste weighed after consumption (continued on next page)

11

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Hickson and colleagues37 2004

Outcomesb

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Hickson and colleagues40 2011

490

Holst and colleagues42 2017

67

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

Neutral Energy intake Average meal Protected eProtein intake weight mealtimes: determined by Mealtimes weighing 3 protected samples of every from meal. Plate waste avoidable weighed after interruptions, consumption providing environment to facilitate eating

--

Improved environment: Environment improved with decorations and music. Patients received a welcome-tray upon admission

Energy intake Intake noted by nursing staff and Protein intake patients. Method þMeeting of measurements energy not specified requirements Meeting protein requirements

2019 Volume

(continued on next page)

RESEARCH

12

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

-

Number -

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

-

Huang and colleagues33 2015

8

Methods of measurement

QoRc

Energy intake Intake determined Neutral by visually Meeting estimating the nutritional consumed requirements portion and Protein intake noting the percentage of the meal that was consumed. Requirements estimated using the Schofield equation

Food intake Volunteers during Protected mealtimes: protected Posters show mealtimes: Volunteers, staff, protected mealtimes, and visitors assist with meals priority for during protected meal trolleys, tables kept mealtimes clear and within patients’ reach

Neutral Intake visually estimated and noted in quartiles

(continued on next page)

13

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Huxtable and 1,012 colleagues34 2013

Volunteers at lunchtime: Trained volunteers assisted lunch on weekdays (feeding, position meal trays, cutting food, and encourage socialization)

Outcomesb

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Intervention During Mealtimea

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Lambert and colleagues28 1999

395

Larsen and colleagues46 2007

113

Trained foodservice personnel (3 IGs): Nursetrained foodservice employee, or cross-trained hospital employee delivered centrally prepared meals directly to patients

--

2019 Volume

À la carte: Meal ordering by telephone up to 24 h in advance from a colorcoded menu and served within 45 min

Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

Satisfaction

Neutral Satisfaction measured using a validated questionnaire rating the quality of service and food

eCarbohydrate intake Energy intake þFat intake þMeeting energy requirements

Intake registered by Neutral patients on a predesigned paper form

(continued on next page)

RESEARCH

14

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

-

Number -

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

-

10

Markovski and colleagues41 2017

34

McCray and colleagues47 2017

128

Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

þCarbohydrate Intake was visually Neutral estimated by intake nurse and the first þEnergy intake author þFat intake þProtein intake

Food fortification: Standard hospital diet enriched with cream, butter, unsaturated oils, and gruel of maize

Dining room: þEnergy intake Intake determined Neutral by measuring the On 3 days of þProtein intake consumed the week, portion of midday patients were meals through a encouraged 5-point scale. to attend a subanalysis dining room performed for at mealtime, patients with instead of malnutrition, eating at cognitive bedside impairment, and low appetite À la carte: Meals ordered à la carte style by phoning anytime between 6:3 0am and 7:00 pm, served within 45 min

þEnergy intake Intake estimated by students of þProtein intake dietetics and þMeeting nutrition using a energy 5-point visual requirements scale. Satisfaction þMeeting measured using a protein survey requirements þSatisfaction

15

(continued on next page)

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Lörefalt and colleagues18 2005

Intervention During Mealtimea

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Intervention During Mealtimea

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Palmer and colleagues36 2015

798

Pietersma and colleagues26 2003

27

Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

Neutral Energy intake Intake visually Volunteers during Protected Protein intake estimated and mealtimes: protected noted in quartiles Posters show mealtimes: using data Volunteers, staff, protected collection sheets. mealtimes, and visitors Intake of nutrientassist with meals priority for dense items during protected meal trolleys, noted as more or tables kept mealtimes. less than 50% clear and consumed within patients’ reach Meal choice at bedside: Meal choice and portioning at bedside during each mealtime

Food cart: Food served in bulk using an electric food cart with a heated surface and a lamp

þSatisfaction

Neutral Satisfaction measured with questionnaire regarding food, appearance, temperature, portion size, variety, and time of service

--

(continued on next page)

RESEARCH

16

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

2019 Volume -

Number -

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

-

407

Barton and colleagues20 2000

35

Volunteers at lunchtime: Volunteers attended a halfday training and assisted nursing staff on intervention wards during lunchtimes Food fortification: Reduced portion sizes with increased energy density

Methods of measurement

QoRc

Neutral Energy intake All leftovers were Protein intake weighed during the day. Subgroup analysis performed for patients with a risk of malnutrition, confusion and use of soft diets and sip feeds þEnergy intake Each plate of lunch Negative þProtein intake and dinner was weighed before and after consumption. Intake of breakfast and snacks was estimated. Daily intake compared to the recommended intake (continued on next page)

17

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Roberts and colleagues38 2017

Outcomesb

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Intervention Before Mealtimea Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

--

Freil and colleagues19 2006

969

Olin and colleagues21 1996

36

Robinson and colleagues29 2002

68

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

2019 Volume

Food fortification: Additional menu options: Meals Meals with chosen from a increased menu offering at energy density least 1 first by adding fat course, 2 main courses, and 2 desserts

þEnergy intake Plates weighed þProtein intake before and after þSatisfaction consumption. Satisfaction measured with questionnaire

Food fortification: Energy-enriched lunch and dinner of standard hospital diet to increase energy density by 50%

þEnergy intake þBodyweight Norton scale

Negative Intake visually estimated and recorded in quartiles. Body weight measured every 3 wk and 6 wk after completion. Norton scale includes variables of functional condition

þFood intake

Negative Intakes measured by nurses and Meal Mates who recorded the percentage of the tray consumed

-

Number

Volunteers during the day: Volunteers were trained to improve appetite and intake, record patients’ food intake

Negative

-

(continued on next page)

RESEARCH

18

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued)

--

2019 Volume

Table 2. Overview of foodservice intervention(s), outcome measures, and quality assessment of 33 studies included in a systematic review of inpatient foodservices intervention studies (continued) Intervention Before Mealtimea

-

Number

Author (s), year

Meal preparation and composition Menu intervention Meal delivery No. of (n[13) (n[6) patients (n[11)

-

Wilson and colleagues27 2000

108

Bulk trolley: Meal choice at bedside: Printed Meals menu filled in at transported in containers wards. Menu and choice was adapted at point regenerated in bulk to of service wards, plated from a hostess trolley

Intervention During Mealtimea Mealtime environment Mealtime assistance (n[10) (n[7)

Outcomesb

Methods of measurement

QoRc

Negative þCarbohydrate Food items for lunch and supper intake weighed þEnergy intake separately before þFat intake and after þProtein intake consumption

a

19

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Specification on the type of interventions indicates before mealtime (preparation and composition, menu, and meal delivery) and during mealtime (mealtime assistance, and protected mealtime) is reported. Outcomes are described as þ[Positive effect of intervention on outcome measure (compared with control), ¼No significant difference between intervention and control, or e¼Negative effect of intervention on outcome measure (compared with control). c QoR¼Quality of Research (positive, neutral, or negative) based on the Quality Criteria Checklist for Primary Research.11 d Deficit calculated as the difference between estimated intake and estimated requirements. e Food intake measured by the weight (in grams) of the consumed food items (not specific macronutrients). b

RESEARCH Term

Definition

Foodservices

Functions, equipment and facilities concerned with the preparation and distribution of readyto-eat food (medical subject headings term, Pubmed).

Hospital

Institutions with an organized medical staff that provides medical care to patients (medical subject headings term).

Preparation and composition intervention

Changes in the preparation and composition of meals, ranging from an increase in protein or caloric content; for example, food fortification,20 to a different way of preparing meals (eg, steaming).22,24

Meal delivery intervention

Changes in meal delivery from kitchen to patient: change in the transferring unit; for example, use of a food cart or choice at the point of consumption,25,26 or a change in meal delivery staff.28

Mealtime assistance

Assistance of patients during meal consumption; for example, by volunteers.59

Mealtime environment

Changes in the surroundings during mealtime, such as protected mealtimes35 or modifications in the physical environment during mealtime (eg, eating in a communal dining room, improving environment aesthetics, or listening to music during mealtime).41,42

Menu intervention

Changes in food selection and/or in the way of ordering and/or adjustments in layout of printed menu.12

Protected mealtimes

Mealtimes protected from avoidable interruptions, in a way that patients could eat without being disturbed.60

Steamplicity

A foodservice system (Compass Group Healthcare Services) that operates by preparing food encapsulated in plastic film with a special steam valve. The meal is heated by microwave, which allows pressurized steam to be trapped in the sealed covers, maintaining a constant steam pressure.22,24

Trolley

A cart designed to transport prepared foods from point of preparation or meal assembly to point of service.

Figure 2. Clarification of terms and concepts used in a systematic review of inpatient foodservices intervention studies. and one study showed some improved outcome measures (ratings: positive).39 Three of five studies (60%) showed no significant changes (ratings: two neutral and one positive), and one study showed a negative effect on outcome measures (rating: neutral).40 In the other two neutral-rated studies, all outcome measures improved when patients were encouraged to eat in a dining room,41 and positive effects on some outcomes were reported when the environmental aesthetics improved and written information on nutrition was given.42

Studies with a Positive Quality Rating Based on the critical appraisal of the reviewers, nine studies were rated as positive, meaning relevant outcomes were used. Two studies on preparation and composition showed a positive effect on meeting the protein requirements. The interventions were serving protein-enriched products14 and offering small energy-enriched menu items that were supplemented with protein powder.43 Two studies on mealtime assistance measured protein and energy intake in different ways and reported either improved energy requirements32 or improved protein requirements.30 Another study that evaluated the mealtime environment with protected mealtimes showed no improvement in protein and energy intake, but the energy deficit was reduced.39 One study on menu

20

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

intervention that allowed foods and drinks to be ordered by telephone from a printed menu found no effect on nutrient intake but reported a positive effect on satisfaction.44 Another study in which verbal prompts to encourage ordering fortified foods were given by telephone operators showed an improvement in the amount of ordered target products.45 Two studies implemented more than one intervention. One of these studies reported improved protein and energy requirements by serving protein-rich menu items six times a day at the bedside with proactive advice from nutritional assistants.31 The other study reported an improvement in the number of patients who met their energy and protein requirements by implementing protected mealtimes, additional mealtime assistance, or a combination of both.35 Studies that included a meal delivery intervention were not rated positively. All other studies were rated negative (n¼18) or neutral (n¼6).

Outcome Measures Nutrient Intake. Nutrient intake was measured in 30 of 33 studies (91%). It was defined as the intake of energy and/or specific macronutrients in 77% of studies and was mostly reported in grams per day or energy intake in kilocalories per day or kilojoules per day. Nine of the 30 studies

--

2019 Volume

-

Number

-

Identification

RESEARCH

Records identi ied in PubMed and other databases (n=8,486)

Additional records identi ied through snowball search method (n=4)

Screening

Records after removal of duplicates (n=5,584)

Eligibility

Records screened (n=5,584)

Included

Full-text articles assessed for eligibility (n=42)

Studies included in qualitative synthesis (n=33)

Records removed based on exclusion criteria (n=5,542)

Full-text articles excluded (n=9) - Non-English (n=1) - Abstract only (n=2) - Pilot study (n=1) - No nutrition/clinical outcomes (n=5)

Figure 3. Flow diagram of the literature search for a systematic review to determine which elements are essential in an optimal hospital foodservice. (30%) mentioned intake relative to nutrition requirements.16,17,30-33,42,46,47 Various time frames and methods were used to measure nutrient intake among the studies: 14 studies (43%) measured just the main meals, and 16 studies (53%) also included between-meal snacks. Twelve of 30 studies (40%) used a scale to weigh meals before and after mealtime, whereas 16 studies (53%) visually estimated what portion of the meal was consumed. The portion of food consumed was reported in exact percentages in eight studies, on a 6-point scale in one study, and in quartiles (<25%, 25% to 50%, and so on) in seven studies.

Nutritional Status. The change in nutritional status was assessed in five of the included studies (15%). There were various evaluations of nutritional status: change in bodyweight was measured on seated scales or the method of measuring was not mentioned (four studies).15,21,37,44 Patient handgrip strength was measured using hand dynamometers in two studies,16,37 but the method of measuring midarm muscle circumference was not reported in one study.37

Patient Satisfaction. Patient satisfaction was measured in 10

DISCUSSION

of 33 studies (30%). Questionnaires were validated in five of these studies (50%). The questionnaires that were used were the Nutrition RelatedeQuality of Life Questionnaire,44 which rates the various elements of the meal such as choice, delivery, and quality22; a questionnaire evaluating the effect of treatment and disease on the amount of food consumed23,31; and a consumer opinion card to assess patient and staff acceptance of food items served at military hospitals.25 Interviews were conducted in one study, in addition to a questionnaire.22 --

2019 Volume

-

Number

-

Clinical Outcomes. One of 33 studies (3%) reported clinical outcomes, such as mortality, the number of prescribed antibiotics in case of an infection, and length of stay.37

The primary aim of this study was to provide evidence from the literature on the key factors of an optimal hospital foodservice. This systematic review found only nine studies that have sufficient methodologic quality to meet evidence-based scientific standards.11 Various foodservice interventions were identified (such as using volunteers to provide mealtime assistance, encouraging patients to choose protein-rich foods, adding protein-enriched items to the menu, replacing existing items with protein-enriched foods, or ordering of food by telephone JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

21

RESEARCH from a printed menu) that might improve relevant outcome measures, such as meeting nutritional requirements and increasing patient satisfaction scores. Another recent systematic review also reported a positive effect of mealtime assistance on protein and energy intake in patients older than age 65 years.7 Only two studies showed that a combination of interventions—mealtime assistance and choice at bedside or mealtime assistance and protected mealtimes—seemed to improve protein and energy intake. Strong evidence to support the implementation of only meal delivery interventions was lacking, and the implementation of protected mealtimes showed inconsistent results. These findings agree with a previous systematic review that concluded that there was insufficient evidence for the widespread implementation of protected mealtimes in hospitals.13 Health care institutions that seek to change their foodservice should take these results into account. To facilitate the implementation of an effective foodservice intervention, a pilot study should also be considered to explore needs and barriers in the specific settings of a hospital.48 Recently, a helpful process framework was developed to guide clinicians and researchers before, during, and after the implementation of a mealtime intervention.49 The creators of the framework emphasized the importance of first communicating with other health care institutions that have recently implemented similar changes to learn from their experiences. This review aimed to establish the effect of foodservice interventions only, and diet intervention studies were therefore not included. However, considerable evidence exists in support of the idea that diet measures do influence patient outcomes.50 The second aim was to focus on reported outcome measures in detail and evaluate their relevance and validity to facilitate future research. Nutrient intake was by far the most frequently used outcome measure, which is relevant in light of the increased risk of malnutrition during admission. Yet, this parameter was often measured inaccurately. Many studies merely use visual estimation as a proxy for nutrient intake.51 Furthermore, although a complete daily record of nutrient intake, including between-

meal intake, is recognized as the optimal method, numerous studies measured only one or two main meals during the day. The use of actual nutrient (energy and protein) intake based on the requirements of individual patients is strongly advocated because this measure is a more relevant outcome than the absolute amount of proteins and energy consumed.52 An accurate method to analyze nutrient and meal consumption is by subtracting the weight of each food item at the end of each mealtime from the weight at serving time.31 The effect on patient satisfaction was also measured in various studies using a range of questionnaires of nonspecified validity. Because patient satisfaction is often related to nutrient intake, and food quality is associated with low food intake, an evaluation of patient satisfaction seems to be relevant in all foodservice interventions.53,54 Therefore, the validation of such questionnaires should be a requirement in future research. The nutritional status and clinical outcomes of patients did not significantly improve in any of the studies, most likely due to the relatively short length of hospital stay.55 Extending the intervention period to the outpatient setting, before and/or after hospitalization, would seem crucial to realize a relevant improvement in these outcomes. Researchers should incorporate practical repeated values, such as the Patient Generated Subjective Global Assessment and/or handgrip strength into their studies to provide nutrition-related markers and functional outcomes. Other relevant measures and standard operating procedures are provided on the website of the Dietetic Pocket Guide.56 There are other recommendations that can be taken into consideration to guide future research. Use of research reporting guidelines such as the Strengthening the Reporting of Observational Studies in Epidemiology statement can improve the quality of reporting in observational studies and minimize bias. Peer-reviewed journals should therefore require authors to indicate the guidelines they used.57 Experts in this field should collaborate and develop strict objective criteria, such as the inclusion of a regression analysis that accounts for any

Research question

Recommendation

Which elements might be essential in an optimal hospital foodservice?

Based on studies with high quality of evidence:

Which outcomes and measurements seem to be relevant and valid to use in future nutrition intervention trials?

Nutrient intake outcome: Daily protein and energy intake relative to requirements Measurement: Weight on calibrated scale before and after consumption

      

Mealtime (volunteer) assistance (n¼2)30,32 Encourage patients to choose protein-rich food (n¼1)31 Adding protein-enriched dishes to the menu (n¼1)16 Existing dishes replaced by protein-enriched products (n¼1)14 Ordering of food by telephone from a menu card (n¼1)44 Mealtime assistance plus meal choice at bedside (n¼1)31 Mealtime (volunteer) assistance plus protected mealtimes (n¼1)35

Satisfaction outcome: Satisfaction regarding quality of meals and foodservice Measurement: Validated questionnaire, depending on research question, review literature Out-of-hospital: Nutritional status outcomes such as patient-generated subjective global assessment. Functional status outcomes such as handgrip strength Measurement: By using standard operating procedures56

Figure 4. Recommendations to guide future research of inpatient foodservices.

22

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

--

2019 Volume

-

Number

-

RESEARCH confounding factors that can influence the outcome in nutritional research. Subsequent studies that take recommendations into account (Figure 4) might result in an improvement in the quality of studies, and ultimately, the design of an optimal foodservice in clinical settings. This systematic review has strengths and limitations. The heterogeneity across the included studies was high for both the type of interventions and outcome measures. An adequate overview of this heterogeneous information was presented by categorizing each study based on the type of intervention and by reporting the effects on outcomes. This approach is novel because previous systematic reviews focused primarily on single specific interventions or populations, such as mealtime assistance or elderly patients.7,8,12,13 The QCCPR grading tool is designed to assess the quality of research in the field of nutrition and dietetics, with consideration on aspects of dietary measurement and related errors. Although this grading tool contains instruction guidelines, some aspects of grading are open for interpretation. Therefore, the interobserver variability was minimized as described in the Materials and Methods section. The Grades of Recommendation Assessment, Development and Evaluation system is the preferred tool to assess the overall body of evidence based on the outcome measures across studies. Due to the heterogeneity of outcomes, the grading of the quality of evidence per outcome was not possible.

CONCLUSIONS A concise overview of evidence-based hospital foodservice interventions was created. Based on nine available high-quality studies, we conclude that several types of interventions have the potential to improve outcome measures. These interventions include the use of volunteers to provide mealtime assistance, encouraging patients to choose protein-rich foods, adding protein-enriched items to the menu, replacing existing items with protein-enriched items, ordering food by telephone from a printed menu, or a combination of the above. Health care institutions that wish to improve their foodservice might consider one or more of these interventions.

References 1.

Wall BT, van Loon LJC. Nutritional strategies to attenuate muscle disuse atrophy. Nutr Rev. 2013;71(4):195-208.

2.

Kruizenga H. Nederlandse Prevalentiemeting Ondervoeding in Ziekenhuizen (NPOZ) stuurgroep ondervoeding. http://www. fightmalnutrition.eu/. Published 2016. Updated 2016. Accessed February 9, 2018.

3.

Felder S, Lechtenboehmer C, Bally M, et al. Association of nutritional risk and adverse medical outcomes across different medical inpatient populations. Nutrition. 2015;31(11-12):1385-1393.

4.

Agarwal E, Ferguson M, Banks M, et al. Malnutrition and poor food intake are associated with prolonged hospital stay, frequent readmissions, and greater in-hospital mortality: Results from the Nutrition Care Day Survey 2010. Clin Nutr. 2013;32(5):737-745.

5.

6.

7.

--

Kok L, Berden C, Sadiraj K. Costs and benefits of home care for the elderly versus residential care: A comparison using propensity scores. Eur J Health Econ. 2015;16(2):119-131. Nielsen MA, Biltz C, Freil M, Gut R, Almdal TP. Menu composition of evening meals increases energy and protein intake in patients with low nutritional intake and reduces waste. Ugeskr Laeger. 2004;166(4):267-270. Tassone EC, Tovey JA, Paciepnik JE, et al. Should we implement mealtime assistance in the hospital setting? A systematic literature review with meta-analyses. J Clin Nurs. 2015;24(19-20):2710-2721.

2019 Volume

-

Number

-

8.

Mahoney S, Zulli A, Walton K. Patient satisfaction and energy intakes are enhanced by point of service meal provision. Nutr Diet. 2009;66: 212-220.

9.

Whitelock G, Aromataris E. Effectiveness of mealtime interventions to improve nutritional intake of adult patients in the acute care setting: A systematic review. JBI Database System Rev Implement Rep. 2013;11:263305.

10.

Moher D, Liberati A, Tetzlaff J, Altman DG; the PRISMA Group. Preferred reporting items for systematic reviews and metaanalyses: The PRISMA statement. Ann Intern Med. 2009;151(4): 264-269.

11.

Academy of Nutrition and Dietetics. Evidence Analysis Manual: Steps in the Academy Evidence Analysis Process. https://www.andeal.org/ vault/2440/web/files/2016_April_EA_Manual.pdf. Published 2012. Updated 2016. Accessed February 14, 2018.

12.

Ottrey E, Porter J. Hospital menu interventions: A systematic review of research. Int J Health Care Qual Assur. 2016;29(1):62-74.

13.

Porter J, Ottrey E, Huggins CE. Protected Mealtimes in hospitals and nutritional intake: Systematic review and meta-analyses. Int J Nurs Stud. 2016;65:62-69.

14.

Beelen J, Vasse E, Janssen N, Janse A, de Roos NM, de Groot L. Proteinenriched familiar foods and drinks improve protein intake of hospitalized older patients: A randomized controlled trial. Clin Nutr. 2018;37(4):1186-1192.

15.

Collins J, Porter J, Truby H, Huggins C. A foodservice approach to enhance energy intake of elderly subacute patients: A pilot study to assess impact on patient outcomes and cost. Age Ageing. 2017;46(3):486-493.

16.

Munk T, Beck AM, Holst M, et al. Positive effect of proteinsupplemented hospital food on protein intake in patients at nutritional risk: A randomised controlled trial. J Hum Nutr Diet. 2014;27(2):122-132.

17.

Gall MJ, Grimble GK, Reeve NJ, Thomas SJ. Effect of providing fortified meals and between-meal snacks on energy and protein intake of hospital patients. Clin Nutr. 1998;17(6):259-264.

18.

Lorefalt B, Wissing U, Unosson M. Smaller but energy and proteinenriched meals improve energy and nutrient intakes in elderly patients. J Nutr Health Aging. 2005;9(4):243-247.

19.

Freil M, Nielsen MA, Biltz C, Gut R, Mikkelsen BE, Almdal T. Reorganization of a hospital catering system increases food intake in patients with inadequate intake. Food Nutr Res. 2006;50:6.

20.

Barton AD, Beigg CL, Macdonald IA, Allison SP. A recipe for improving food intakes in elderly hospitalized patients. Clin Nutr. 2000;19(6):451-454.

21.

Olin AO, Osterberg P, Hadell K, Armyr I, Jerstrom S, Ljungqvist O. Energy-enriched hospital food to improve energy intake in elderly patients. JPEN J Parenter Enteral Nutr. 1996;20(2):93-97.

22.

Edwards JS, Hartwell HJ. Hospital food service: A comparative analysis of systems and introducing the ’Steamplicity’ concept. J Hum Nutr Diet. 2006;19(6):421-430.

23.

Goeminne PC, De Wit EH, Burtin C, Valcke Y. Higher food intake and appreciation with a new food delivery system in a Belgian hospital. Meals on Wheels, a bedside meal approach: A prospective cohort trial. Appetite. 2012;59(1):108-116.

24.

Hickson M, Fearnley L, Thomas J, Evans S. Does a new steam meal catering system meet patient requirements in hospital? J Hum Nutr Diet. 2007;20(5):476-485.

25.

Hartwell HJ, Edwards JS, Beavis J. Plate versus bulk trolley food service in a hospital: Comparison of patients’ satisfaction. Nutrition. 2007;23(3):211-218.

26.

Pietersma P, Follett-Bick S, Wilkinson B, Guebert N, Fisher K, Pereira J. A bedside food cart as an alternate food service for acute and palliative oncological patients. Support Care Cancer. 2003;11(9):611-614.

27.

Wilson A, Evans S, Frost G. A comparison of the amount of food served and consumed according to meal service system. J Hum Nutr Diet. 2000;13:271-275.

28.

Lambert LG, Boudreaux J, Conklin M, Yadrick K. Are new meal distribution systems worth the effort for improving patient satisfaction with foodservice? J Am Diet Assoc. 1999;99(9):1112-1114.

29.

Robinson S, Clump D, Weitzel T, et al. The Memorial Meal Mates: A program to improve nutrition in hospitalized older adults. Geriatr Nurs. 2002;23(6):332-335.

30.

Manning F, Harris K, Duncan R, et al. Additional feeding assistance improves the energy and protein intakes of hospitalised elderly patients. A health services evaluation. Appetite. 2012;59(2):471-477.

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

23

RESEARCH 31.

Dijxhoorn DN, van den Berg MGA, Kievit W, Korzilius J, Drenth JPH, Wanten GJA. A novel in-hospital meal service improves protein and energy intake. Clin Nutr. 2018;37(6 pt A):2238-2245.

32.

Lindman A, Rasmussen HB, Andersen NF. Food caregivers influence on nutritional intake among admitted haematological cancer patients—a prospective study. Eur J Oncol Nurs. 2013;17(6):827-834.

33.

Huang CS, Dutkowski K, Fuller A, Walton K. Evaluation of a pilot volunteer feeding assistance program: Influences on the dietary intakes of elderly hospitalised patients and lessons learnt. J Nutr Health Aging. 2015;19(2):206-210.

34.

Huxtable S, Palmer M. The efficacy of protected mealtimes in reducing mealtime interruptions and improving mealtime assistance in adult inpatients in an Australian hospital. Eur J Clin Nutr. 2013;67(9):904-910.

35.

Young AM, Mudge AM, Banks MD, Ross LJ, Daniels L. Encouraging, assisting and time to EAT: Improved nutritional intake for older medical patients receiving Protected Mealtimes and/or additional nursing feeding assistance. Clin Nutr. 2013;32(4):543-549.

36.

Palmer M, Huxtable S. Aspects of protected mealtimes are associated with improved mealtime energy and protein intakes in hospitalized adult patients on medical and surgical wards over 2 years. Eur J Clin Nutr. 2015;69(8):961-965.

37.

Hickson M, Bulpitt C, Nunes M, et al. Does additional feeding support provided by health care assistants improve nutritional status and outcome in acutely ill older in-patients? A randomised control trial. Clin Nutr. 2004;23(1):69-77.

nutritional status, and offers opportunities to improve intake. Clin Nutr. 2016;35(5):1174-1180. 45.

van der Zanden LDT, van Essen H, van Kleef E, de Wijk RA, van Trijp HCM. Using a verbal prompt to increase protein consumption in a hospital setting: A field study. Int J Behav Nutr Phys Act. 2015;12:110.

46.

Larsen CS, Toubro S. The effect of conventional v. a la carte menu on energy and macronutrient intake among hospitalized cardiology patients. Br J Nutr. 2007;98(2):351-357.

47.

McCray S, Maunder K, Krikowa R, MacKenzie-Shalders K. Room service improves nutritional intake and increases patient satisfaction while decreasing food waste and cost. J Acad Nutr Diet. 2018;118(2):284-293.

48.

Dijxhoorn DN, Wanten GJA, van den Berg MGA. Developing mealtime interventions: Considerations based on a single center experience. Clin Nutr. 2018;37(4):1437-1438.

49.

Conchin S, Carey S. The expert’s guide to mealtime interventions - A Delphi method survey. Clin Nutr. 2018;37(6 pt A):1992-2000.

50.

Baldwin C, Weekes CE. Dietary advice with or without oral nutritional supplements for disease-related malnutrition in adults. Cochrane Database Syst Rev. 2011 Sep 7;(9):CD002008.

51.

Castellanos VH, Andrews YN. Inherent flaws in a method of estimating meal intake commonly used in long-term-care facilities. J Am Diet Assoc. 2002;102(6):826-830.

52.

Kruizenga H, Beijer S, Huisman-de Waal G, et al. Guideline on malnutrition. http://www.fightmalnutrition.eu/. Published 2017. Accessed February 14, 2018.

38.

Roberts HC, Pilgrim AL, Jameson KA, Cooper C, Sayer AA, Robinson S. The impact of trained volunteer vealtime assistants on the dietary Intake of older female in-patients: The Southampton Mealtime Assistance Study. J Nutr Health Aging. 2017;21(3):320-328.

53.

Johns N, Hartwell H, Morgan M. Improving the provision of meals in hospital. The patients’ viewpoint. Appetite. 2010;54(1):181-185.

54.

39.

Porter J, Haines TP, Truby H. The efficacy of Protected Mealtimes in hospitalised patients: A stepped wedge cluster randomised controlled trial. BMC Med. 2017;15(1):25.

Keller H, Allard J, Laporte M, et al. Food access is associated with poor appetite and low food intake in acute care patients: The Canadian Malnutrition Task Force (CMTF) study. Clin Nutr. 2013;32:73.

55.

40.

Hickson M, Connolly A, Whelan K. Impact of protected mealtimes on ward mealtime environment, patient experience and nutrient intake in hospitalised patients. J Hum Nutr Diet. 2011;24(4):370-374.

Health at a glance 2017: OECD indicators. https://doi.org/10.1787/ health_glance-2017-en. Accessed February 22, 2018.

56.

Wierdsma N, Kruizenga H, Stratton R. Dietetic pocket guide. http:// www.dieteticpocketguide.com. Accessed February 1, 2018.

57.

Lachat C, Hawwash D, Ocke MC, et al. Strengthening the Reporting of Observational Studies in Epidemiology-Nutritional Epidemiology (STROBE-nut): An Extension of the STROBE Statement. PLoS Med. 2016;13:e1002036.

41.

Markovski K, Nenov A, Ottaway A, Skinner E. Does eating environment have an impact on the protein and energy intake in the hospitalised elderly? Nutr Diet. 2017;74(3):224-228.

42.

Holst M, Beermann T, Mortensen MN, et al. Optimizing protein and energy intake in hospitals by improving individualized meal serving, hosting and the eating environment. Nutrition. 2017;34:14-20.

58.

Munk T, Beck A, Holst M, et al. Positive effect of protein-supplemented hospital food on protein intake in patients at nutritional risk: A randomised controlled trial. J Hum Nutr Diet. 2014;27(2):122-132.

Naithani S, Thomas JE, Whelan K, Morgan M, Gulliford MC. Experiences of food access in hospital. A new questionnaire measure. Clin Nutr. 2009;28(6):625-630.

59.

Roberts HC, Pilgrim AL, Elia M, et al. Southampton Mealtime Assistance Study: Design and methods. BMC Geriatr. 2013;13:5.

60.

Murray C. Improving nutrition for older people. Nurs Older People. 2006;18(6):18-22.

43.

44.

Doorduijn AS, van Gameren Y, Vasse E, de Roos NM. At Your Request room service dining improves patient satisfaction, maintains

AUTHOR INFORMATION D. N. Dijxhoorn is a resident, Internal Medicine Department, Radboud University Medical Centre, Nijmegen, the Netherlands; at the time of the study, she was a PhD student, Department of Gastroenterology and Hepatology, Radboud University Medical Centre, Nijmegen, the Netherlands. M. J. M. J. Mortier is a medical student and G. J. A. Wanten is a gastroenterologist, Department of Gastroenterology and Hepatology, and M. G. A. van den Berg is a clinical dietitian researcher, Department of Gastroenterology and Hepatology - Dietetics and Intestinal Failure, Radboud University Medical Centre, Nijmegen, the Netherlands. Address correspondence to: Dorian N. Dijxhoorn, MD, PhD, Department of Gastroenterology and Hepatology, Radboud University Medical Center, PO Box 9101, 6500 HB Nijmegen, the Netherlands. E-mail: [email protected]

STATEMENT OF POTENTIAL CONFLICT OF INTEREST No potential conflict of interest was reported by the authors.

FUNDING/SUPPORT There is no funding to report.

AUTHOR CONTRIBUTIONS All authors designed the research; D. N. Dijxhoorn and M. J. M. J. Mortier conducted the research; D. N. Dijxhoorn had primary responsibility for final content; and all authors wrote, read, and approved the final manuscript.

24

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

--

2019 Volume

-

Number

-

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices

2019 Volume -

Number

Author(s), year, design

-

Barton and colleagues20 2000 RCTd with crossover

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Strengths: Design: Part of population crossed over. Patients randomly allocated. Outcome measures: Intake analyzed on group basis and individual basis. Food weighed before and after consumption Limitations: Outcome measures: Nutritional requirements not taken into account

(continued on next page)

24.e1

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Energy and protein intake: Energy intake at lunch and CG: Normal menu 35 supper: IG1 and IG2 higher Each plate of lunch and Elderly rehabilitation ward (n¼13): Standard intakes than CG: 1,111195 kcal dinner weighed before cereal breakfast (19/35 with (IG1) vs 916176 kcal (IG2) vs and after consumption. cardiovascular accident) IG1: Fortified menu: 825136 kcal (CG). Intake calculated by Reduced portion sizes CG: 7511 Protein intake at lunch and subtracting waste from with increased energy IG1: 778 supper: CG had higher intakes food provided. Intake of density (n¼14) IG2: 789 than IG2: 347 g (CG) vs 316 g breakfast and snacks IG2: Cooked breakfast: 56 d (14-d cycles) (IG2). No difference between CG estimated. Daily intake Normal menu with IG2 group does not cross and IG1. was noted compared to cooked breakfast over patients’ recommended Mean daily energy intake: Close to (varying daily from recommended levels on level of intake tomatoes on toast to average. Higher intake in the IGs eggs and bacon) to than in the CG: 1,711195 (IG1) replace the standard vs 1,744176 (IG2) vs 1,425136 cereal breakfast (n¼8) (CG). Mean daily protein intake: Below recommended levels on all menus. Higher intake in IG2 than in CG: 576 (IG2) vs 477 (CG). No difference between CG and IG1

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Beelen and colleagues14 2017 RCT

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb 147 Medical wards CG: 79.27.0 IG: 77.77.8 6 mo. Both groups observed simultaneously

CG: Standard menu Standard energy- and protein-rich hospital menu for patients aged 65 y (n¼80) IG: New menu Same standard menu, but regular products were replaced by intervention products and some additional options (n¼67)

Outcome measures

Resultsc

Nutrient intake: Nutritional intake: Protein intake on Day 4 Protein intake was higher in IG adjusted for bodyweight. (105.7 g; 1.51 g/kg BW) than CG Percentage of reached (88.2 g; 1.22 g/kg BW). protein requirements (1,2 Fulfillment of protein g/kg/d). requirements was higher in IG Intake of energy and other than CG (79% vs 48%). macronutrients. Patients Energy intake was higher in IG were asked how much of than CG (31.1 vs 28.6 kcal/kg). the portion was No difference in intake of consumed. BWe measured carbohydrates and fat on SECA scalef

Discussion Strengths: Outcome: Protein requirements were taken into account Limitations: Outcome: Energy requirements were not taken into account

(continued on next page)

RESEARCH

24.e2

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number -

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Collins and colleagues15 2017 Nonrandomized, nonblinded controlled pilot study

122 Geriatric wards CG: 80 (75-87) IG: 84 (75-88) 4 mo. Both groups observed simultaneously

Resultsc

Discussion

Nutritional status: Change in weight (calibrated seated scales) and handgrip strength (hand dynamometer) between Day 1 and 14 Energy and protein intake: Daily energy and protein intake, based on all mealtimes, at Day 1 and 14 visually estimated using a validated 6-point scale. Standard serving sizes were known Satisfaction: Verbal validated questionnaire at Day 14. 4 Domains: Food quality, meal service, staffing and service, and physical environment

Nutritional status: No difference between groups in change in weight or HGSg Energy and protein intake: Total daily energy and protein intake, at Day 14 was significantly higher in IG (132 kJ/ kg/d and 1.4 g/kg/d) than CG (105 kJ/kg/d and 1.1 g/kg/d) Satisfaction: Higher rating in the domain of physical environment in IG

Strengths: Outcome measure: Nutrient intake was adjusted for weight. Intake, nutritional status, and satisfaction were all analyzed Limitations: Patients’ requirements were not taken into account

(continued on next page)

24.e3

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

CG: Standard menu Paper menu prior to 3 main meals. Midmeals offered by foodservice staff at bedside. No encouragement or visual menu for midmeals (n¼38) IG: New menu Could provide up to 3,680 kJ/d and 24 g/ d protein more than standard menu. More options. Less energy dense items were removed. Visual menu implemented for midmeals to show all available options. Foodservice staff encouraged patients to choose an item at midmeals (n¼30)

Outcome measures

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Dijxhoorn and colleagues31 2017 Preepost prospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Energy and protein intake: Energy and protein intake: CG: Traditional meal 637 Percentage of met service Higher energy intake relative to Surgical and medical requirements and 3 meals per day served requirements in IG on Day 1 wards percentage of patients by nutritional (88%34% vs 70%39%) and CG: 5917 that met requirements at assistants. Dinner Day 4 (84%40% vs 73%31%). IG: 6016 preference indicated in Day 1 and 4 of oral intake Higher % fulfilled energy 17 mo: CG for 10 mo and were calculated. morning from menu requirements in IG (37% vs 14%). IG for 11 mo. Subtracting weight at end Higher protein intake relative to list (n¼326) of meals from weight at IG: FoodforCare meal requirements in IG on Day 1 serving time. service (7933 vs 5928) and Day 4 (73 Food appreciation: 6 meals per day, small 38 vs 5929). Higher % Validated questionnaire protein-rich menu fulfilled protein requirements in items. Special attention scored various domains IG on Day 1 (24% vs 8%) and on regarding food paid to meal Day 4 (23% vs 8%) appreciation and presentation. Trained Food appreciation: FfC group accessibility.58 The other nutritional assistants more satisfied with appearance recommend proteinand smell of the meals. Ratings questionnaire included 2 rich choices and assist questions that rated meal were similar. in choosing most service and food quality optimal menu item from 1 to 10 (n¼311)

Discussion Strengths: Design: Large population Outcome measures: Nutritional requirements were calculated Limitations: Patients and investigators were not blinded due to the nature of the intervention

(continued on next page)

RESEARCH

24.e4

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Doorduijn and colleagues44 2015 Preepost prospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Resultsc

Discussion

Patient satisfaction: -Rating meal service (scale 110) -Questionnaire: Assess the nutrition-related quality of life (scale: 27-162) Nutritional status: Body weight and handgrip strength measured by nurse on day of admission and day before discharge Food intake: Energy intake and % of patients reaching protein requirements calculated with food lists (recorded twice per week by nurse) for patients requiring an energy and protein rich menu (n¼72)

Patient satisfaction: CG: 7.5 vs IG: 8.1 Questionnaire: Increase from 124.5 points (CG) to 132.9 points (IG) Nutritional status: No difference in BW and handgrip change between both groups. Malnutrition Universal Screening Tool (MUST) score increased in 4 (CG) and 6 (IG) patients, and decreased in 18 patients in both groups (significance unknown) Food intake: No significant differences between groups

Strengths: Outcome measures: Multiple outcome measures used Limitations: Outcome measures: Food intake in the control group was not measured, except for the subgroup requiring an energy- and protein-rich diet

(continued on next page)

24.e5

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

CG: Traditional meal 337 service: 3 meals per Surgery and general day. Drinks served medical wards between meals. 6515 8 mo. CG for 5 mo, IG for 3 Decided before 10:00 am what to eat the mo next day (n¼168) IG: At Your Request concept: Foods and drinks ordered by telephone from a printed menu between 7:00 am and 7:00 pm. Delivered within 45 min (n¼169)

Outcome measures

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

52 Edwards and Postoperative surgical colleagues22 ward 2006 Prospective cohort 25-68 (standard deviation NSh) 4 wk. Each system observed during 2 wk

CG: Cook-chill system: Cyclical menu, food ordered the day before. Cold bulk food was loaded into trolley and transported to ward. Before meal, food was regenerated, plated and taken to patients’ beds (n¼28) IG: Steamplicity systemi: Meals ordered 2 h before meal service, from an extended choice menu. Individual, readyplated chilled meals were transported to the ward, held chilled, and heated/cooked in microwave oven (n¼24)

Outcome measures

Resultsc

Discussion

Stakeholders satisfaction: Semistructured interviews with randomly selected patients (n¼5), staff members, and visitors, exploring main issues in patient satisfaction and meal experience Patient acceptability: Validated questionnaire distributed at beginning and end of stay Nutrient intake: In CG, plates weighed before and after consumption on a digital scale. In IG, average portion sizes taken from production records. Leftover food weighed after lunch and dinner

Stakeholders satisfaction: Steamplicity was rated higher in 5 out of 15 questions of the questionnaire, namely in the subjects of food texture, food presentation, overall meal satisfaction, knowing the available meal options, and sufficiency of portion size. Patient acceptability: Patients were positive about the Steamplicity system overall, in terms of food choice, ordering, delivery, and food quality Nutrient intake: Lunch: Mean¼202 g for the CG and mean¼282 g for the IG (unknown significance). Dinner: Mean¼226 g for the CG and mean¼310 g for the IG (unknown significance)

Strengths: Outcome measures: Satisfaction measured using questionnaires and interviews. Intake measured with digital scale Limitations: Outcome measures: Food intake not measured per patient, only whole group. Intake measured in grams per day, without details of protein and energy intake. Requirements not calculated. Intake only recorded during lunch and dinner

(continued on next page)

RESEARCH

24.e6

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Freil and colleagues19 2006 Pre post prospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

24.e7

Discussion

Energy intake: 1st quartile: Increase from 128 kJ (95% CI 79-178) in the CG, to 560 kJ (95% CI 489-631) in IG1, to 1,021 kJ (95% CI 939-1,104) in IG2 2nd quartile: No change from CG to IG1. Increase of 400 kJ from IG1 to IG2 3rd quartile: Intake approximately 2,000 kJ in all groups, no significant difference between groups 4th quartile: Intake of 2,400-2,500 kJ in all groups, no significant difference between groups Protein intake: 1st quartile: Increase from 0.7 g per patient in CG to 4.1 g in IG1 to 8.1 g in IG2 2nd quartile: Increase from 10.3 g in IG1 to 20.1 g in IG2 3rd quartile: 18-25 g in all groups, no significant change between groups 4th quartile: 25-35 g, no significant change between groups Patient experience and satisfaction: Patients were fairly positive about both systems. Increase in satisfaction regarding meal appearance, taste, and general satisfaction

Strengths: Population: Large population. Design: Patients were divided into groups according to energy and protein intakes Outcome measures: Served and wasted foods were weighed individually Limitations: Outcome measures: Validity of questionnaire unclear Requirements were not taken into account

(continued on next page)

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Energy and protein intake: CG: Old fixed menu: 969 Intake calculated by Breakfast, lunch, and Gynecological, breast subtracting wasted from dinner covered 20%surgery, and orthopedic served evening meals. 25% of total energy surgery ward Meals weighed by scale needs. Snacks cover NS before and after 42 d. Each system observed 30%. No possibility for consumption. Results patients to for 14 d grouped in 4 quartiles individualize their with respect to calculated menu (period 1 energy and protein intake n¼376). per patient IG1: Individualized Patient experience and system pilot: Meals satisfaction about with increased energy density (by adding fat). evening meals: Choosing from a menu Questionnaire regarding meal appearance, cart offering at least 1 quantity, taste, and first course, 2 main courses, and 2 desserts general satisfaction. Either a positive or a negative (period 2 n¼328) opinion about each of IG2: Individualized these subjects could be system: 2 y after selected (n¼70). implementation (period 3 n¼265)

Resultsc

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Energy and protein deficits Energy deficit and intakes: CG had CG: Standard hospital 143 Gall and a higher energy deficit than IG: and intakes: diet. (n¼81) Medical and orthopedic colleagues17 e7 kcal/d vs e341 kcal/d. ward and female elderly IG: Intervention group: Protein and energy 1998 Energy intake 17.5% higher in IG requirements estimated Standard hospital diet, Prospective cohort wards and increased from 1,404 kcal/ using the Elia and with in addition Male: 60.472.36 d to 1,650 kcal/d fortified foods at lunch Schofield formulae. Female: 74.01.50 Energy and protein deficit Protein deficit and intakes: and supper (n¼62) 3 d. Both groups were Protein intakes and below calculated by subtracting observed estimated requirements in both mean protein and energy simultaneously groups. Mean intake of 51.2 g/ intakes from estimated requirements per patients. d (CG), 43.5 g/d (IG: female Intake measured by patient, orthopedic ward) and 66.7 g/ relatives, and nursing staff d (IG: male medical ward) using food-intake record sheets. Food charts checked at least every 48 h

Discussion Strengths: Outcome measures: Energy and protein intakes were compared to the respective requirements. Limitations: Outcome measures: Food intakes were only observed over a 3d period. Oral intake might be noted inaccurately when done by (ill) patients and family/ friends

(continued on next page)

RESEARCH

24.e8

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

-

Strengths: Outcome measures: Meals weighed before and after serving time and snacks were included Limitations: Population: Implementation at only 1 ward, which hampers generalizability Outcome measures: Protein and energy intake not measured

(continued on next page)

24.e9

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Nutritional intake: Weight Nutrient intake: Increased for CG: Standard system: 189 Goeminne and at end of the 3 meals Meals ordered 1 day breakfast, lunch, and dinner. Department of respiratory colleagues23 (breakfast, lunch, and beforehand. Total daily intake increased with diseases 2012 dinner), was subtracted Individually portioned 236 g (95% CI 163-308). Prospective cohort CG: 70 (59-75) from weight at serving beforehand and Food appreciation and IG: 71 (62-76) time. All data measured transported with a accessibility: Patients preferred 2 mo. Each system and collected by the heated food cart Meals on Wheels to the current observed for 1 mo nutritional assistant. Extra (n¼83) system in the subdomains snacks included in data. IG: Meals on Wheels: choice, sensation of hunger, and Patients were asked at Food appreciation and food quality accessibility: Validated mealtime how much questionnaires with 5 and what they wished domains58 completed at to eat. Portioning at bedside. Food cart with discharge a refrigerator and a heating compartment (n¼106)

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Hartwell and colleagues25 2007 Cross-sectional cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb 180 Orthopedic wards Median age range: 60-79 2 mo. 1 mo each system

Outcome measures

Resultsc

Satisfaction: CG: Plated system: Menu Satisfaction: During mealtimes, patients filled Temperature: Improvement in selection 24 h before potatoes, poached fish, and in a consumer opinion diner. Catering staff card with a 7-point rating minced beef. and domestic workers function as foodservice scale for 5 attributes of 5 Flavor: Rated higher in broccoli, carrots, and poached fish. food items were used. personnel (n¼NS) Portion size: Rated higher in Reliability and validity IG: Bulk trolley system: broccoli and potato. have been evaluated Menu choice at point Texture: Improvement in all foods: of consumption. Broccoli, carrots, potato, Nursing staff was poached fish, and minced beef. appointed as Overall opinion/satisfaction: foodservice personnel. Improved in broccoli, potato, Range of menu items is and poached fish limited compared to CG (n¼NS)

Discussion Strengths: Outcome measures: Different components of the meal were rated in the opinion card. Limitations: Population: Implementation at only 1 ward, which hampers generalizability

(continued on next page)

RESEARCH

24.e10

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Hickson and colleagues37 2004 RCT

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Discussion

(continued on next page)

24.e11

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Strengths: No differences in any results BW: Using electronic seat CG: Usual care group: 592 Outcome measures: Multiple except for number of scales. Mid-arm muscle 3 acute medicine wards for Standard ward care patient-related outcome circumference: Calculated intravenous antibiotics (n¼300) the elderly measures were used prescribed for infections: 6 d in using a formula and IG: Feeding support CG: 82 (76-86) Limitations: measuring arm of patient. CG and 4 d in IG (P¼0.02) group: Standard ward IG: 82 (77-87) Population: Only older care plus nutrition care Barthel score: Score for Unknown patients included assessing a person’s ability from a health care Outcome measures: Number to perform activities of assistant, including of intravenous antibiotics daily living. Handgrip identifying and is not an ideal reflection of strength: Measured with reducing malnutrition, infection rate. Food intake dynamometer. Food encouraging and measured in only 6% of intake: Weighing meals enabling patients in the study population. Part beforehand and using feeding, and offering of intake estimated, based food records for breakfast, snacks and drinks on food record book snacks, and drinks. (n¼292) Infection rate: Estimated by number of antibiotics prescribed. Length of stay: Until medically fit for discharge and until actual discharge. Mortality in hospital

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Hickson and colleagues24 2007 Prospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

--

Food intake: Weight of food Food intake: More food was CG: Traditional bulk 57 consumed in CG compared with served at lunch Medical and surgical wards cookechill system. determined by weighing 3 IG (467 g vs 358 g). Energy intake Performed previously 75 (55.8-83.5) samples of every meal and higher in CG than in IG (2,074 k (Wilson and 1 mo for the IG. CG data calculating average. Plate J vs 1,779 kJ). Protein intake did colleagues27). Printed collected from other study (Wilson and menus were filled in at waste weighed after each not differ between groups meal to calculate overall colleagues27; study wards (not mentioned which day). Meals were food intake, protein, and period unknown) energy intake transported to wards and regenerated in bulk. Food plated from a trolley at wards (n¼57) IG: Steamplicity system. Patients ordered meals 2 h before meal service, from an extended choice menu. Individual, ready-plated chilled meals were transported to the ward, held chilled, and heated/cooked using a microwave oven (n¼57)

Discussion Strengths: Population: The study population is diverse, making the study more generalizable, and increasing external validity Limitations: Design: 7 y between IG and CG period (Wilson and colleagues27): Hospital environment could have changed Population: Small population, heterogeneity Outcome measures: Energy deficit not calculated in CG. Served meals not weighed individually. Intake only measured during lunch.

2019 Volume

(continued on next page)

RESEARCH

24.e12

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

-

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

-

CG: Standard system: A Nutrient intake: Way of 490 Nutrient intake: No difference in Hickson and measuring described in ready plated and Patients at high risk of energy intake between groups. colleagues40 sealed catering system Hickson and colleagues24 malnutrition. Median protein intake higher in 2011 (n¼39) the CG (14.0 g [4-26 g] vs 7.5 g Prospective cohort CG: 40 wards IG: Protected mealtimes: IG: 34 wards [1-28 g]; P¼0.04) During mealtime, all Age: NS Study period of 4 mo. Each nonurgent clinical activity was stopped system 2 mo on the ward (n¼66) 67 Patients at risk of malnutrition. Medical and surgical wards CG: 62.915.1 IG: 67.217.9 Study period of 6 wk. 3 wk per group

Nutrient intake: Nutrient intake: CG: Standard system: Energy and protein intake No difference in energy and Energy- and proteindense foods, including was noted by nursing staff protein intake between CG and or by the patients in food IG. supplements. 3 main records. No information >75% fulfillment of energy meals and 3 inrequirements was higher in IG between meals (n¼30) about weighing or than CG (67.6% vs 40.0%). No estimation mentioned. IG: New environment: difference in fulfillment of Environmental aesthetics Noting of midmeals and protein requirements improved with painted snacks not mentioned. Fulfillment of >75% of walls, decoration and music added to dining requirements was also analyzed room. Welcome-tray with special serving and written materials about nutrition upon admission. Welcome interview by nurse, to identify specific individual preferences and challenges (n¼37)

Strengths: Population: Large study population and many different wards Limitations: Outcome measures: Nutrient intake was only measured during lunch Strengths: Outcome measure: Requirements of energy and protein taken into account Limitations: Design: Short study period Outcome measure: No information on way of measuring food intake

(continued on next page) 24.e13

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Holst and colleagues42 2017 Preepost Prospective cohort

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Huang and colleagues33 2015 Prospective cohort (pilot) (patient is own control)

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

--

Lunchtime and daily energy Lunchtime and daily energy and CG: No volunteer 8 macronutrient intakes: No and macronutrient assistance: Standard 2 aged care wards, difference. intakes: Meal trays hospital care without identified as “at-risk” or sighted before and after Daily energy and protein intakes, volunteer feeding malnourished compared with estimated daily meal. Percentage of meal assistance 834.5 requirements: No differences in IG: Volunteer assistance: consumed estimated 4 d for each patient; 2 amount of patients meeting through observation and Trained volunteers d with and 2 d without energy and protein recorded assisted, at lunch on volunteer assistance requirements Daily energy and protein weekdays, with intakes, compared with feeding, correctly positioning meal trays, estimated daily requirements: Dietary cutting up foods, requirements calculated handling cutlery, using Schofield equation opening packaging, and encouraging patients with conversation and socialization. Staff was educated on serving meals and helping patients with eating. Menus and toppings were improved to give more refined impression

Discussion Strengths: Design: Patients are their own controls Outcome measures: Requirements taken into account Limitations: Population: Very small study population Design: Volunteer assistance only during lunch on weekdays

2019 Volume

(continued on next page)

RESEARCH

24.e14

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

-

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

-

Strengths: Population: Large study population. Outcome measures: Only staff filled in data collection sheets. All main meals of the day included. Limitations: Outcome measures: Requirements not taken into account

(continued on next page)

24.e15

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Nutrient intake: Data CG: Standard system: 1,012 Nutrient intake: No difference in Huxtable and collection sheet used by Standard care before (1,632 mealtime intake between groups. colleagues34 the observing dietitians to Proportion of inpatients implementation of the observations) 2013 note intake data at all 3 protected mealtimes consuming >50% of nutrientProspective cohort Adults on medical and main meals. Intake program (PMP) (799 surgical wards. dense items did not differ and estimated visually observations) CG: 6518 they were more likely to (validated) and proportion consume >50% if items were IG: Protected mealtimes IG: 6618 consumed noted as <1/4 , placed in reach 2 y. Each system during 1 y program: Volunteers, 1 staff, and visitors /4 to 1/2, 1/2 to 3/4 , >3/4 assisted with meals during protected mealtimes. Posters on wards displayed times reserved for protected mealtimes, priority was requested for meal trolleys, and tray tables were kept clear and within patients’ reach (833 observations)

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

395 Lambert and Patients from 19 acutecolleagues28 care hospitals from 1999 Prospective cohort medical/surgical departments 50 Study period unknown

Outcome measures

Resultsc

--

2019 Volume

Patient satisfaction: No Patient satisfaction: A: Traditional differences in rating. Group A Questionnaire rated level foodservice rated food quality higher than distribution: Employee of agreement with 16 group B (3.70.6 vs 3.50.7; statements about quality delivers meals directly P0.05). Groups C and D rated of service for meal tray to patients (n¼92) food quality not differently delivery and food quality B: Traditional nursing compared with the other groups on a scale of 1 to 5 service distribution: Nurse delivers meals directly to patients (n¼131) C: Nontraditional foodservice distribution: Specifically trained foodservice employee performs a wait-type service by delivering meals directly to patients (n¼109) D: Nontraditional nursing meal distribution: Hospital employee, crosstrained to provide a full range of patient-care services, delivers meals to patients (n¼63)

Discussion Strengths: Design: Four types of intervention were compared in 1 study. Population: Large study population and different wards Limitations: Year: Old study, which makes it hard to generalize

(continued on next page)

RESEARCH

24.e16

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

-

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Larsen and colleagues46 2007 Prospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Discussion

Energy intake: Patients in IG2 had higher intake in CG than IG2: 6.6 vs 7.9 kJ/d; 1,576 vs 1,887 kcal. Insignificant result, after adjustment for BW. 86% of estimated requirement in CG, 89% in IG1, and 101% in IG2. Fulfillment of carbohydrate and fat requirements: Relative carbohydrate intake was lower and fat intake higher in CG (32% and 52%) compared with both IG1 (43% and 41%) and IG2 (42% and 41%)

Strengths: Outcome measures: Items from hospital shop or visitors included. Energy intake relative to requirements measured. Limitations: Outcome measures: Patients noted intake by themselves, which reduces accuracy of registration

(continued on next page)

24.e17

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

CG: Fixed menu: Meals Food intake: Patients 113 registered their own produced in central Cardiology Department intake on a pre-designed kitchen. Dispatched to CG: 7516 paper form, noting time of departments at fixed IG1: 7313 day, description and hours. Standard IG2: 7116 servings consisted of 3 amount of foods and 10 wk. 3 wk (CG), 3 wk drinks served, consumed meals, planned 14 da (IG1), and 3 mo later: 4 and left over, including before production wk (IG2) items from hospital shop (n¼48) IG: À la carte/Free menu: or visitors. Colored printed menu (almost 100 items) distributed among patients. Ordered by telephone, up to 24 h in advance, whenever they liked. Immediate served within 45 min

Resultsc

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

86 Lindman and Department of colleagues32 hematology 2013 Prospective cohort CG: 49(21-71) IG: 62(23-81) 6 d: 3 d in each group

Outcome measures

Nutrent intake: Food CG: Standard system: weighed separately before Assistants worked in serving. Leftovers ward kitchens, in registered. Intake charge of stock recorded by nurses. Daily management and intake and fulfilled orders to central requirements calculated kitchen (n¼41) as average of 3IG: Food caregivers: Kitchen assistants were d assessment. Types of discomfort and side trained as food effects that influenced caregivers, and served snacks, guided patients dietary intake noted by nurses, taken into account and relatives during in data analysis mealtime, and motivated patients to eat (n¼45)

Resultsc

Discussion

Nutrient intake: Energy requirements fulfilled in 76.2% (95% CI 64.6-87.9) vs 93.3% (95% CI 82.3-104.3) in the CG and IG, respectively. Protein requirements fulfilled in 64.3% (95% CI 53.7-75.0) vs 69.1% (95% CI 59.6-78.5) in the CG and IG, respectively. No difference in number of patients fulfilling requirements with 75%

Strengths: Outcome measures: Food weighed before and after mealtime. Requirements measured. Discomfort and side effects influencing intake included in analysis Limitations: Design: Only 1 ward observed. Short observation period Outcome measures: Tool assessing discomfort and side effects not validated

(continued on next page)

RESEARCH

24.e18

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Lörefalt and colleagues18 2005 Prospective cohort (patient is own control)

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb 10 Geriatric rehabilitation ward 81.73.2 6 d: Each menu 3 d

Outcome measures

Discussion

Nutrient intake: Mean daily energy intake was 1,864513 kcal/d (CG) vs 2,564490 kcal/d (IG). Lunch energy intake was 562133 kcal/d (CG) vs 794171 kcal/ d (IG) and supper intake 391132 kcal/d (CG) vs 822196 kcal/d (IG). Mean protein intake was higher in IG than CG: 72 g vs 88 g

Strengths: Design: Patients were their own controls Outcome measures: Requirements measured Limitations: Population: Small population Design: Short period Outcome measures: Requirements not taken into account measured

(continued on next page)

24.e19

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Nutrient intake: First author CG: Standard hospital menu: Energy contents and nurse visually estimated food intake and for portions were noted in household standardized as measures (cups, pieces, 2150kcal/day. Menus pre-defined by hospital etc.) in food record book kitchen, but patients Energy requirements were determined by calculating could choose quantity basal metabolic rate of foods. IG: Energy and proteinenriched menu: Quantity of lunch and supper half of standard hospital portion. Density of energy and nutrients increased corresponding to whole portion of standard menu

Resultsc

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Discussion

--

2019 Volume -

Manning and colleagues30 2012 Prospective cohort (patient is own control)

23 Aged care wards 83.28.9 20 d. Each system 10 d

Lunchtime protein and energy CG: Standard system. No Lunchtime energy and protein intakes: Standard intakes: More protein and volunteer assistance average weight for serving energy intake in IG vs CG: 1,730 (n¼23) vs 1,334 kcal and 21.8 vs 17.5 g. sizes was used. Leftover IG: Volunteer feeding Mean daily protein intake of 43.0 food weighed after each assistance. Volunteers (IG) vs 51.7 g (CG). No difference meal. In-between snacks assisted at lunchtime in energy intake on weekdays and were noted through visual trained to feed patients estimates and questioning Meeting daily nutritional requirements: More people met and encourage them Meeting daily energy and daily energy requirements in IG protein requirements: to eat the products than CG: 59% (CG) compared Estimated requirements with high energy and calculated using Schofield with 71% IG). protein density first There was a trend of more people equation (n¼23) meeting daily energy requirements (NS): 58% (CG) vs 64% (IG)

Strengths: Design: Patients were their own controls Outcome measures: Requirements measured Limitations: Population: Small population Design: Volunteers only at lunchtime

Markovski and colleagues41 2017 Prospective observational pilot study

34 Geriatric rehabilitation wards 79.111.8 Study period of 3 mo Participants crossed over

Nutrient intake: Proportion CG: Bedside. Meals consumed at midday consumption at meals determined bedside. through 5-point IG: Dining room. On 3 consumption scale: 0, 1/4 , days of the week, 1 patients were /2, 3/4 , 1. Energy and encouraged to attend protein intake were a dining room, calculated using nutrient supervised by 1-2 staff analysis of the food. members Subanalysis for patients with malnutrition (MSTj2), cognitive impairment (MMSEk), and low appetite

Strengths: Design: Participants were their own controls Outcome measure: Intake measured in different groups of patients using subanalyses Limitations: Design: Small population Outcome measures: Leftovers not weighed Only midday meals measured. Requirements not taken into account

Nutrient intake: Energy and protein intake higher in IG than CG (2,158.3 vs 1,723 kJ and 28.2 vs 22.5 g) MST2: intake higher in IG than CG (2,295.0 vs 1,331.0 kJ and 27.3 vs 19.9 g) MMSE 25: intake higher in IG than CG (2,136.6 vs 1,479.4 kJ and 27.2 vs 19.0 g) Low appetite: Protein intake higher in IG than CG (24.0 vs 16.6 g)

Number

(continued on next page)

RESEARCH

24.e20

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

-

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

McCray and colleagues47 2017 Retrospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Resultsc

Discussion

Nutrient intake: Intake of all meals and snacks estimated by nutrition and dietetics students with a 5-point visual scale: 0, 1/4 , 1/2, 3/4 , 1. BW used to estimate energy and protein requirements Satisfaction: Analyzed using survey data results from the organization’s routine "Press Ganey" survey. 7 Food domains regarding food quality and foodservice

Nutrient intake: Energy and protein intake was higher in IG than CG (1,588 vs 1,306 kcal/d and 65.9 vs 52.3 g/d). The percentage of estimated energy and protein requirements was also higher in IG (75.1% vs 63.0% and 84.7% vs 65.0%) Satisfaction: Improvements in all domains were seen

Strengths: Outcome measures: Requirements taken into account. Both intake and satisfaction were tested Limitations: Outcome measures: Intake measurement not validated

(continued on next page)

24.e21

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

CG: Traditional 148 foodservice model. General medical, surgical, Order meals with paper and oncology wards menu, up to 24 h CG: 62.919.5 before meal. Meals are IG: 66.315.1 Study period of 3 y (9 mo delivered at set times (n¼85) of observation) IG: Room service. Meals ordered by à la carte style, by phoning anytime between 6:30 am and 7:00 pm. Meals are prepared on demand and delivered within 45 min (n¼63)

Outcome measures

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

--

2019 Volume

CG: Standard hospital 84 Munk and menu. 3 main meals Departments of oncology, colleagues16 (served from a buffet) orthopedics, and 2014 provided 50%-75% of Single-blinded RCT urology nutritional CG: 74 requirements. 3 inIG: 75 between meals, served Each group 18 wk by nursing staff or buffet staff IG: Proteinsupplemented foodservice. Supplemental to CG. À la carte menu of small menu items enriched with energy-dense ingredients, supplemented with protein powder. Patients, ward staff, or research assistants could order by telephone and menu items arriving within 20 min of ordering. Nursing staff assisted patients if needed

Outcome measures

Resultsc

Number of patients fulfilling Number of patients fulfilling 75% of energy 75% of energy and protein requirements: No difference in and protein fulfilling energy requirements. requirements: More patients in IG met protein Requirements calculated using basic metabolic rate requirements: 30% vs 66% Mean energy and protein intake: and estimated activity Higher energy intake in IG than factor. Protein CG (þ693 kJ). Mean protein requirements set at 18% intake higher in IG (þ9.6 g/d), of energy requirement. also in relation to BW (þ0.2 g/kg) Mean energy and protein intake: Mean intake over Handgrip strength, length of hospital stay: No differences 3- 7 d, according to BW. Visually assessed and recorded in quartiles by nursing staff or patients (validated). Handgrip strength: Measured in right hand using dynamometer. Length of hospital stay

Discussion Strengths: Population: Diverse patient group Design: Patients randomly allocated Outcome measures: Requirements taken into account. Nutritional status also measured Limitations: Outcome measures: Intake noted in quartiles instead of exact amount consumed

(continued on next page)

RESEARCH

24.e22

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

-

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Olin and colleagues21 1996 Prospective double-blinded cohort (patient is own control)

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Discussion

(continued on next page)

24.e23

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Strengths: Nutrient intake: Mean energy Nutrient intake: Visually CG: Standard system. 36 measured and recorded in intake increased with 40%, from Design: Double-blinded Foodservice system 2 wards in a long-term study 25 (CG) to 35 kcal/kg/d (IG) quartiles of full amount unknown geriatric care hospital BW: In IG, BW increased 3.4%, from Outcome measures: Use of served, done by ward IG: Energy-enriched CG: 793 54.4 kg to 55.7 kg after 6 wk. No the Norton Scale hospital food. Lunches staff. IG: 847 Limitations: weight change in CG Body weight: Weighed 12 wk. Both systems 6 wk and dinners of before start study, every 3 Modified Norton scale: No change Year: Old study. May not be standard system generalizable to the in overall activity measures wk, and 6 wk after enriched (average current system completion increase of 50%) with Design: Standard system not Modified Norton scale: natural ingredients described Scoring system includes Outcome measures: variables of functional requirements not taken condition: mobility, into account activity and well-being

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

798 Palmer and Surgical and medical colleagues36 wards 2015 Prospective cohort 6319 2 y. Each system 1 y

Outcome measures

Aspects of protected CG: Standard system. mealtimes associated Standard care before with intake: Intake implementation of visually estimated PMP. (n¼122) (validated) and noted in IG: Protected mealtimes quartiles (eg, 1/2-3/4 ) at program: Volunteers, staff, and visitors data collection sheets at assisted with meals all 3 main meals. during mealtimes. Proportion of nutrientPosters on wards dense food items displayed times classified as <50% or reserved for mealtimes >50%. Barriers to and requests for consumption and priority for meal interruptions collected trolleys. Tray tables were kept clear and within patients’ reach (n¼210)

Resultsc

Discussion

Aspects of protected mealtimes associated with intake: PMP not associated with intake. Appropriate positioning and presence of volunteers during mealtime; having more time to eat, no interruptions and meals within reach, positively associated with intake In patients that required mealtime assistance, protein intake was positively associated with volunteers in wards at mealtime and patients appropriately positioned

Strengths: Population: Large study population Outcome measures: Only staff filled in data collection sheets. All main meals of the day included Limitations: Outcome measures: Requirements not taken into account

(continued on next page)

RESEARCH

24.e24

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Pietersma and colleagues26 2003 Prospective cohort (patient is own control)

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Discussion

(continued on next page)

24.e25

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Patient satisfaction: All items rated Strengths: Patient satisfaction: CG: Traditional food 27 Design: Patients were their higher in IG, except for meal Questionnaire with 11 tray. Meals prepared Acute oncology and own controls temperature. centrally and delivered questions (scale of 1-5) palliative care unit about food appearance, 95% of patients preferred the food Outcome measures: Various to patients on thermal NS aspects of meal rated in cart to the food tray. 90% trays. Ordering 24 h in temperature, portion size, 10 d. Lunch served preferred to choose portion size questionnaire variety and time of advance. Supper according to CG and themselves and 94% preferred Limitations: service. Patients delivered by tray supper according to IG to choose food type themselves Population: Age not completed 1 copy after (n¼27) specified. Small lunch and 1 after supper IG: Electric food cart. population Same food served as in (n¼22) Outcome measures: Survey CG, but in bulk. Cart not tested for validity or with heated surface reliability and lamp. Deciding what and how much to eat at bedside. Lunch delivered by cart (n¼27)

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Porter and colleagues39 2017 Randomized controlled trial

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Nutrient intake: Energy and Nutrient intake: CG: Standard hospital 149 (38 crossed over) protein intake measured No change in energy and protein Rehabilitation and geriatric system. Mealtime intake. Energy deficit decreased by nutrition and dietetics processes that were wards in IG compared with CG after students through already in place. No CG: 80.510.7 correction for baseline observations of protected mealtimes IG: 78.612.9 determinants consumptions at main policy (n¼82) 4 wk IG: Protected mealtimes. meals and midmeals. Estimations (in quarters) During mealtimes doors were closed and were converted to energy and proteins. Energy and unnecessary protein deficits were disruptions minimized calculated by subtracting (n¼105). Ward and mean protein and energy foodservice staff intakes from estimated trained before requirements. More than 1 intervention observation per participant was made (n¼416)

Discussion Strengths: Outcomes: In addition to intake, energy and protein deficit also analyzed Limitations: Outcomes: Design: Way of handling withdrawals not described

(continued on next page)

RESEARCH

24.e26

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Roberts and colleagues38 2017 Quasiexperimental study

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb 407 Acute medical female ward Observation year: CG: 87.85.8 IG: 87.15.3 Intervention year: CG: 87.95.1 IG: 87.15.3 2 y: 1 y pre- and 1 y postintervention

Outcome measures

Resultsc

Strengths: Design: Control and intervention wards compared pre- and postintervention. Outcome measure: Leftover items weighed Limitations: Design: Specific population: all female. Outcome measure: No requirements taken into account

(continued on next page)

24.e27

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Nutritional intake: Nutritional intake: CG: Standard hospital system. No volunteer Daily and lunchtime energy No differences in daily or lunchtime energy and protein and protein intake. All assistance (n¼104 intakes between intervention food and drinks were preintervention year; and control wards in either n¼82 postintervention recorded and leftover observational or intervention items were weighed. year) year Standard portion sizes IG: Southampton Improvement in daily energy and were known. mealtime assistance. protein intake in patients with Subgroup analysis in Trained volunteers confusion, after adjusting for BW patients with risk of assisted nursing staff malnutrition, current on intervention wards confusion, and use of soft during lunchtimes diets and sip feeds (n¼117 preintervention year; n¼104 postintervention year) The control wards and intervention wards were compared before and after implementation of the intervention

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design Robinson and colleagues29 2002 Non-RCT

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb 68 Elderly (>65 y) hospitalized patients requiring assistance with feeding CG: 78.2 IG: 77.8 Each group 2 mo

Outcome measures

Resultsc

CG: Standard system. No Mean intake: Nurses in the Mean intake: 32.45% vs 58.88% of CG group and Meal Mates the tray consumed in CG and IG, assistance during respectively in the IG recorded mealtimes (n¼34) percentage of tray IG: Memorial Meal consumed for each Mates. Volunteers patient, including foods trained to improve and fluid. (Method of appetite and intake, recording, amount of and how to record meals recorded [per day], patients’ food intake and exact volunteer roles (n¼34) unknown)

Discussion Strengths: Design: Concurrent controls were used. Limitations: Design: No mention of frequency and time of volunteer assistance. Populations: Only elderly patients that require assistance with feeding. Outcome measures: Only mean intake measured, no energy or protein intakes. Variation in observers (nurses vs volunteers) (continued on next page)

RESEARCH

24.e28

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Van der Zanden and colleagues45 2015 Prospective cohort (patient is own control)

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Strengths: Outcome measures: Both nutritional value and questionnaires used to measure intake Limitations: Design: Patients who asked family or nurses to order were not included (ie, possible selection bias). Outcome measures: Only intake at lunch reported. Ordered food was reported, but not consumed amount

(continued on next page)

24.e29

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Ordering target product for Ordering target product for CG: Control: Patients 208 lunch: 6.5%, 45.2%, and 45.3% of lunch: Whether or not ordered by telephone. Hospitalized patients orders, the target product patients ordered the Calls were answered 60.617.8 showed up for CG, IG1, and IG2, target product. according to the 14 d in cycles. CG on Day Protein and caloric content respectively. Target product was standard system 1, IG1 on Day 2, IG2 on of ordered lunch: All data ordered significantly less in CG (n¼93) Day 3 and CG again on Protein and caloric content of retrieved from hospital IG1: Prompt: Telephone Day 4 ordered lunch: No difference database operators asked if between groups. patients would like the Patient rating: Using a Patient rating: 65% of patients questionnaire, patients target product, dairy who ordered target product said to what extent they product with 7.1 g reported eating “most” or “all” of consumed the target proteins per portion, it with their order (n¼62) products (5-point Likert IG2: Praise-then-prompt: scale) Telephone operators praised patient about placing the order, “Good that you ordered [food product]” and ended with a prompt (n¼53)

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), year, design

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

108 patient meals Wilson and General medical and colleagues27 orthopedic ward 2000 Prospective cohort NS Study period unknown

Outcome measures

Resultsc

Nutrient intake: Food items Nutrient intake: Intakes greater in CG: Plated system. the IG for energy (319 vs 414 Printed menus filled in for lunch and supper at wards. Meals plated- menu items were weighed kcal), protein (14 vs 18 g), fat (11 vs 16 g) and carbohydrates (41 separately before put on up on a belt-run and vs 51 g) plate and after transported to the consumption ward in trolleys, regenerated and served directly (n¼51) IG: Bulk service. Printed menus filled in at wards, bulk supply is estimated accordingly. Containers transported to the ward and regenerated in bulk. Food then plated from hostess trolley at ward. Patients can change menu choice at point of service (n¼57)

Discussion Strengths: Outcome measures: Each food item weighed before and after consumption. Fat, protein, and carbohydrate intake individually analyzed. Limitations: Design: Study period not mentioned Population: Age NS

(continued on next page)

RESEARCH

24.e30

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -

--

Table 1. Data extraction of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume -

Number

Author(s), year, design

-

Young and colleagues35 2013 Preepost prospective cohort

n Sample Mean–standard deviation or median (range) age (y) Study period CGa and IGb

Outcome measures

Resultsc

Strengths: Design: Multiple types of interventions compared with each other Outcome measures: Requirements measured Limitations: Design: Way of handling withdrawals not described

24.e31

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Nutrient intake: Measured Nutrient intake: No differences in CG: Preintervention 254 energy intakes between groups. by dietitian and dietetic data. Standard 3 internal medicine wards More adequate energy intake in assistants through visual mealtime procedures. 808 the IGs than in the CG. 8% 1 y. CG observed for 6 mo. Assistance not given in estimation (validated). adequate intake in CG vs 20%Each meal was inspected structured manner PMl, AINm and PMþAIN 31% (range) in the IGs. No on delivery and (n¼115) were simultaneously differences between the completion. Energy and observed during 6 mo PM: Protected different IGs mealtimes. Nonurgent protein content was calculated using known Protein intakes showed a trend activities and toward improvement in AIN and interruptions limited at food compositions. PMþAIN groups. Not significant Estimated requirements mealtimes. Patients when adjusted for BW were calculated based on encouraged and body mass index and BW More participants in IGs had assisted with intake. adequate protein intake than in Lunch breaks of nurses the CG. 12% in CG vs 21%-31% rearranged to in IGs. No differences between maximize patient IGs assistance (n¼39) AIN: Additional assistant-in-nursing. 1 AIN member employed per ward to assist patients with meals and completing orders (encouraging high energy and protein options) and to document intake when requested (n¼58) PM+AIN: Combined intervention. Both interventions implemented together (n¼42)

Discussion

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

CG¼control group. IG¼intervention group. c Significantly different if P<0.05. Only significant results are reported (better or worse compared with the other group). d RCT¼randomized controlled trial. e BW¼body weight. f Seca. g HGS¼handgrip strength. h NS¼not specified. i Steamplicity system (Compass Group Healthcare Services). j MST¼Malnutrition Screening Tool. k MMSE¼Mini-Mental State Examination. l PM¼protected mealtimes. m AIN¼additional assistant-in-nursing. b

RESEARCH

24.e32

a

--

2019 Volume -

Number -

--

Table 3. Quality assessment using the Quality Criteria Checklist for Primary Research of included studies in a systematic review of inpatient foodservices

2019 Volume

Validity Itemsa

-

Number -

Study design

Quality rating

1

2

3

4

5

6

7

8

9

10

Comments

Barton and colleagues20 2000

RCTb with crossover

Negative

þc

ed

e

e

þ

þ

e

e

e

þ

Groups not comparable. Eligibility criteria not specified. Outcomes not reported in intake per requirement.

Beelen and colleagues14 2017

RCT

Positive

þ

þ

þ

þ

e

þ

þ

e

þ

þ

Normality of data and adjustment for confounding not specified.

Collins and colleagues15 2017

Non-randomized nonblinded pilot study

Neutral

þ

þ

þ

þ

NA

þ

-

þ

þ

þ

Patients’ requirements not calculated.

Dijxhoorn and colleagues31 2017

Pre-post prospective cohort

Positive

þ

þ

þ

þ

NA

þ

þ

þ

þ

þ

Blinding not applicable due to nature of the intervention.

Doorduijnand colleagues44 2015

Pre-post prospective cohort

Positive

þ

þ

þ

-

NA

þ

þ

þ

þ

þ

Blinding and withdrawals not specified.

Edwards and colleagues22 2006

Prospective cohort

Neutral

þ

-

-

-

NA

þ

-

þ

-

þ

No baseline table. Intake only measured at main courses. Effect of intervention not clear described in conclusion. Patients’ requirements not calculated.

Freil and colleagues19 2006

Pre-post prospective cohort

Negative

þ

-

-

-

-

þ

-

-

þ

þ

Eligibility criteria and patients characteristics not specified. Patients’ requirements not calculated.

Gall and colleagues17 1998

Prospective cohort

Neutral

þ

þ

-

þ

-

þ

-

-

þ

þ

Baseline demographics not specified. Intake noted (and not weighted) by patients/ family/nurses on record sheets.

Goeminne and colleagues23 2012

Prospective cohort

Neutral

þ

-

þ

-

NA

þ

-

þ

þ

þ

Exclusion criteria, way of handling blinding and withdrawals not specified. Patients’ requirements not calculated.

Hartwell and colleagues25 2007

Cross-sectional cohort

Neutral

þ

-

-

-

NA

þ

þ

-

þ

þ

Concise baseline table and description of statistical analysis

Hickson and colleagues37 2004

RCT

Neutral

þ

þ

þ

þ

NAe

þ

e

þ

þ

þ

Intake measured using food records

(continued on next page)

24.e33

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), y

Validity Itemsa

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS --

2019 Volume

Author(s), y

Study design

Quality rating

1

2

3

4

5

6

7

8

9

10

Comments

Hickson and colleagues24 2007

Prospective cohort

Neutral

þ

þ

e

e

e

þ

e

e

þ

þ

No comparison of baseline characteristics. Only measurements at lunchtime. Energy requirements were measured in IG but there was no comparison with CG. Withdrawals not clearly mentioned

Hickson and colleagues40 2011

Prospective cohort

Neutral

þ

þ

e

e

NA

þ

e

þ

e

þ

Only measurements at lunchtime. Baseline characteristics and limitations not specified

Holst and colleagues42 2017

Preepost prospective cohort

Neutral

þ

þ

e

þ

NA

þ

e

e

þ

þ

No information on way of measuring food intake. No adjustment for confounding

Huang and colleagues33 2015

Prospective cohort (pilot)

Neutral

þ

e

e

e

NA

þ

e

þ

þ

þ

Baseline characteristics and exclusion criteria not specified. Small sample size, short duration and only volunteers present at lunchtime while daily intake is an outcome factor

Huxtable and colleagues34 2013

Prospective cohort

Neutral

þ

þ

þ

þ

NA

þ

e

þ

þ

þ

Patients’ requirements not calculated

Lambert and colleagues28 1999

Prospective cohort

Neutral

þ

e

e

e

e

þ

þ

e

þ

þ

Baseline characteristics, exclusion criteria, normality, and adjustment for confounders not specified

Larsen and colleagues46 2007

Prospective cohort

Neutral

þ

e

þ

e

NA

þ

e

þ

þ

þ

Eligibility criteria, way of blinding, and handling of withdrawals not clearly specified. Intake self-noted by patients

Lindman and colleagues32 2013

Prospective cohort

Positive

þ

þ

þ

þ

NA

þ

þ

þ

þ

þ

Blinding not applicable due to nature of the intervention

Lörefalt and colleagues18 2005

Prospective cohort (patient is own control)

Neutral

þ

þ

NA

þ

e

þ

e

e

þ

þ

No baseline table. Patients’ requirements not calculated

Manning and colleagues30 2012

Prospective cohort (patient is own control)

Positive

þ

þ

NA

e

NA

þ

þ

þ

þ

þ

Methods of handling withdrawals not clearly described

-

Number

(continued on next page)

RESEARCH

24.e34

Table 3. Quality assessment using the Quality Criteria Checklist for Primary Research of included studies in a systematic review of inpatient foodservices (continued)

-

--

Table 3. Quality assessment using the Quality Criteria Checklist for Primary Research of included studies in a systematic review of inpatient foodservices (continued)

2019 Volume

Validity Itemsa

-

Number -

Study design

Quality rating

1

2

3

4

5

6

7

8

9

10

Comments

Markovski and colleagues41 2017

Prospective pilot study (patient is own control)

Neutral

þ

þ

NA

e

NA

e

e

e

þ

þ

Control and intervention not specified. Patients’ requirements not calculated. No adjustment for confounding factors

McCray and colleagues47 2017

Retrospective cohort

Neutral

þ

þ

e

NA

NA

þ

e

e

þ

þ

No adjustment for confounding factors mentioned. Meal intake observation tool not validated

Munk and colleagues16 2014

Single-blinded RCT

Positive

þ

þ

þ

þ

NA

þ

þ

þ

þ

þ

Data analysis was blinded, but participants and data assessors were not due to nature of intervention (risk of performance and detection bias)

Olin and colleagues21 1996

Prospective double blinded cohort (patient is own control)

Negative

e

e

NA

þ

e

e

e

e

e

þ

Eligibility criteria not specified. Body mass index not in baseline table. Standard system not described. Patients’ requirements not calculated. Limitations not clearly mentioned

Palmer and colleagues36 2015

Prospective cohort

Neutral

þ

þ

þ

e

NA

þ

e

þ

þ

þ

Patients’ requirements not calculated. Way of handling withdrawals not specified.

Pietersma and colleagues26 2003

Prospective cohort (patient is own control)

Neutral

þ

þ

NA

þ

NA

þ

e

e

þ

þ

No baseline table. The survey was not tested for validity or reliability. Statistics described too briefly

Porter and colleagues39 2017

RCT

Positive

þ

þ

þ

-

NA

þ

þ

þ

þ

þ

Way of handling withdrawals not mentioned

Roberts and colleagues38 2017

Pre-post prospective cohort

Neutral

þ

þ

þ

þ

NA

þ

e

e

þ

þ

Patients’ requirements not calculated. No adjustment for confounding factors

Robinson and colleagues29 2002

Non-randomized controlled trial

Negative

þ

þ

e

e

NA

e

e

e

þ

Uf

No baseline table. Study duration not specified. Intake measured in percentages but not regarding specific macronutrients or calories. Conflict of interest unclear: Some authors also worked as volunteers

Van der Zanden and colleagues45 2015

Prospective cohort (patient is own control)

Positive

þ

þ

NA

þ

NA

þ

þ

þ

þ

þ

Telephone operators not fully blinded, but patients were. Food orders measured instead of actual food intake

24.e35

(continued on next page)

RESEARCH

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), y

Validity Itemsa

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Author(s), y

Study design

Quality rating

1

2

3

4

5

6

7

8

9

10

Comments

Wilson and colleagues27 2000

Prospective cohort

Negative

þ

e

e

e

NA

þ

e

e

e

þ

Eligibility criteria and baseline table not specified. Nutritional requirements not measured. Limitations not clearly mentioned

Young and colleagues35 2013

Pre-post prospective cohort

Positive

þ

þ

þ

e

NA

þ

þ

þ

þ

þ

Way of handling withdrawals not described

a

Studies were rated on 10 items: 1¼Research question stated, 2¼Subject selection free from bias, 3¼Comparable study groups, 4¼Method for withdrawals described, 5¼Blinding used, 6¼Interventions described, 7¼Outcomes stated and measurements valid and reliable, 8¼Appropriate statistical analysis, 9¼Appropriate conclusions and limitations described, and 10¼Funding and sponsorship free from bias. Shaded areas indicate validity items must be satisfied for a positive quality rating. b RCT¼randomized controlled trial. c þ¼item present. d eitem not present. e NA¼not applicable. f U¼unclear.

RESEARCH

24.e36

Table 3. Quality assessment using the Quality Criteria Checklist for Primary Research of included studies in a systematic review of inpatient foodservices (continued)

--

2019 Volume -

Number -