Article Text

Download PDFPDF

Menu selection, glycaemic control and satisfaction with standard and patient-controlled consistent carbohydrate meal plans in hospitalised patients with diabetes
  1. Michelle Curll1,
  2. Monica DiNardo1,
  3. Michelle Noschese1,
  4. Mary T Korytkowski2
  1. 1University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, USA
  2. 2Department of Medicine, Division of Endocrinology, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  1. Correspondence to Dr Mary T Korytkowski, Falk, Room 580, University of Pittsburgh, 3601 Fifth Avenue, Pittsburgh, PA 15213, USA; korytkowski{at}


Problem Medical nutrition therapy is an important component of glycaemic management in hospitalised patients with diabetes; however, there is a lack of information guiding the ordering of specific meal plans in this setting.

Setting University-affiliated academic medical centre.

Methods An administrative decision to gradually replace standard consistent-carbohydrate (CCMP) (standard group) with patient-controlled meal plans (patient-controlled group) presented the opportunity to compare menu selection, adherence to CCMP, glycaemic control and satisfaction as a quality-improvement initiative. Information was obtained from consecutive inpatients with diabetes admitted to units receiving standard (n=30) or patient-controlled meal plans (n=43). Patients received the meal plan according to unit location.

Results No group differences were observed in adherence to CCMP (70% vs 64%, p=0.1), mean capillary blood glucose (CBG) or hyperglycaemia frequency (CBG>180 mg/dl). Hypoglycaemia (CBG<70 mg/dl) occurred more frequently in the patient-controlled group (0.39 vs 3.23%, p=0.04). There were no episodes of severe hypoglycaemia (CBG<40 mg/dl) in either group. The patient-controlled group reported a greater satisfaction and had more opportunities for nutrition education, with a demonstrated improvement in adherence to CCMP following targeted education in six of nine patients with available menu data.

Conclusions The standard group experienced less hypoglycaemia and required less clinician oversight. The patient-controlled group allowed for identification of patients who would benefit from education, required more oversight by nutrition services and reported greater satisfaction with their meal plan. Both meal plans may be appropriate for inpatients with diabetes, provided that a sufficient review is available for patients who make inappropriate selections with the patient-controlled meal plan.

  • Diabetes mellitus
  • diabetic diet
  • nutrition therapy
  • inpatients
  • hospital
  • healthcare quality
  • quality of care
  • patient education
  • patient outcomes
  • healthcare quality improvement
  • hospital
  • inpatients
View Full Text

Statistics from


Medical nutrition therapy (MNT) is an important component of glycaemic management in patients with diabetes in the inpatient as well as outpatient clinical settings.1–3 MNT includes nutrition assessment, establishment of individualised goals, assistance with meal planning and evaluation by nutrition professionals. A major goal of MNT for hospitalised patients is to maintain optimal glycaemic control as a way of enhancing recovery while also providing adequate calories to meet the metabolic demands associated with the underlying illness. In addition, hospitalisation can provide an opportunity to reinforce nutrition principles and dietary choices as a way of preparing patients for continuing self-management after discharge.1 2

There is no single meal plan that meets the needs of all hospitalised patients with diabetes. Procedures, surgeries and variability in appetite present unique challenges to maintaining usual or recommended meal patterns. In addition, food selections and meal times may differ from a patient's usual choices, making it more difficult to predict the efficacy of a home regimen in the hospital.

Consistent carbohydrate meal planning has been advocated for use by hospitals as a practical method of providing nutrition to inpatients with diabetes while also improving metabolic control.4 Because carbohydrate is the nutrient that has the greatest impact on postprandial blood glucose levels, nutritional insulin doses can be better predicted when total carbohydrate intake is the same each day. Alternatively, meals that allow more individual choice are advocated as a way of identifying patients who may benefit from dietary education and of optimising nutritional status among patients at high nutritional risk. The lack of studies comparing the effect of different meal plans on glycaemic control in inpatients with diabetes has contributed to a lack of consistency among hospitals and even among services within one hospital on the type of meal plan offered. This can contribute to confusion on the part of patients who may receive mixed messages regarding appropriate food choices.

There has been increasing attention to the benefits of glycaemic control in the hospital; however, there are no studies that compare one type of dietary approach with another in hospitalised patients with diabetes.1 5–13 In fact, the contribution of a meal plan with inpatient nutrition education to glycaemic control in the hospital has not been previously investigated. The increasing prevalence of diabetes and hyperglycaemia among hospitalised patients warrants a more systematic investigation of this important area of inpatient glycaemic management.14–16

A recent phased change from a standard consistent carbohydrate meal plan to a patient-controlled meal plan at a 600-bed tertiary care institution provided the opportunity to compare two nutrition approaches in similar populations of patients with diabetes on separate nursing units as part of a quality-improvement (QI) Initiative. The purpose of this project was to compare menu selection and adherence to a consistent carbohydrate meal plan, glycaemic control and patient satisfaction between those on units receiving the standard (standard group) or patient controlled (patient-controlled group) meal plan during this transition period. In addition, we investigated changes in adherence to dietary recommendations following individualised nutrition education in those patients whose menus were identified as varying from recommended selections.


This project was approved by the QI Review Committee of the University of Pittsburgh Medical Center.

Description of meal plans

The standard meal plan menu has limited choices that guide patients towards heart healthy foods with fewer concentrated sweets. Menu selections are entered into a dietary computer system which provides automatic computerised menu corrections with substitutions when meal selections do not fit into the recommended meal plan (CBORD Version 5.0.18) (table 1).

Table 1

Comparison of two inpatient meal plans for patients with diabetes

The patient controlled meal plan incorporates an unrestricted menu with ongoing monitoring of patient's daily meal selections by clinical nutrition staff. The menu includes a broad variety of food choices similar to that available outside the hospital environment. Unlike the standard meal plan, computerised edits to ensure that the carbohydrate content of a meal falls within the recommended range do not occur. Individualised nutrition education is provided to those patients identified as lacking appropriate meal planning skills. This patient controlled meal plan was piloted on selected inpatient units, with plans to replace the current standard consistent carbohydrate meal plan on all units.


Four separate but similar patient care units were selected for patient identification. These units were selected by hospital administration for either continuation of the standard meal plan (standard group) or implementation of the patient controlled meal plan (patient-controlled group). This study was conducted as a QI initiative as a way of evaluating this change in meal plans.

Two units, one general medicine unit and one cardiac service, continued to provide the standard menu to all patients for whom a diabetes meal plan was ordered. Two similar units were selected by the hospital for the patient-controlled meal plan. Each of the four nursing units averages 1600–1700 annual patient admissions with an average length of stay of 3.8–4.0 and days.

Consecutive patients on the four units with an order for a diabetes meal plan by their primary service over a 4-month period were included. Menu and glycaemic data were collected for each patient for up to 5 days, from the time the meal plan was ordered until discontinuation of the meal plan or hospital discharge. Those in the patient-controlled group who were not discharged had three additional days of data collection to allow re-evaluation of menu selections following an education visit by a member of the clinical nutrition staff.

All patients were given a brief overview of the CCMP by a member of the clinical nutrition staff upon admission and were encouraged to select meals with consistent carbohydrate content each day. Instructions to limit the intake of saturated fat were also provided. Those receiving the patient-controlled meal plan were also informed that they would receive all food items as ordered from the menu, but their menus would be reviewed daily and nutrition education provided if persistent non-compliance was observed (table 1).

Patient menus were collected on a daily basis and reviewed. Menu selections were considered adherent when there were greater than or equal to three but less than or equal to five requested carbohydrate servings per meal and less than or equal to two added fat servings per meal. Patients in the patient-controlled group with three or more non-adherent menu choices per day were scheduled to receive a nutrition education visit. Criteria for targeted education visits were based on compliance with carbohydrate selections only. Information regarding the frequency of the education visits (patient-controlled group only) and the effect of this visit on subsequent menu choices was obtained.

Glycaemic control was measured as the mean of all blood glucose levels, the mean nadir and peak blood glucoses within each group. Hypoglycaemia was determined as the number and percentage of blood glucoses less than 70 mg/dl (3.9 mmol/l); hyperglycaemia as the number and percentage of blood glucose between 180 mg/dl and 300 mg/dl (10 and 16.6 mmol/l); and severe hyperglycaemia as greater than 300 mg/dl (16.6 mmol/l). All blood glucose measurements ordered by the primary service were obtained from a computer data repository (MARS V 3.2). Medications known to influence blood glucose control, such as insulin, oral diabetes medications and steroids, were recorded.17

Patient survey

A registered dietitian interviewed all available patients between days 2 and 5 of their hospitalisation to obtain the following survey information: length of time since diagnosis of diabetes, previous nutrition education, home nutrition management, satisfaction with meals in the hospital, estimated percentage of each meal that was consumed in the hospital, and the frequency with which food was received from sources other than the delivered meal trays and snacks during hospitalisation. All patients were offered a list of outpatient nutrition education centres during the interview.

Statistical analysis

Patient demographic information and glycaemia are described using measures of central tendency (means, medians) and spread (SD, range) for continuous data and frequencies and percentages for categorical data. The number of meal plan days was calculated as the total number of days that menus were received from the patients using the prescribed meal plan. To examine differences in the meal satisfaction, χ2 tests were performed. Exact tests were applied when appropriate (SAS 8.0; SAS Institute, Cary, North Carolina).

The frequencies of hypoglycaemia and hyperglycaemia were calculated as the number of blood glucoses meeting these definitions divided by the total number of measured blood glucoses. Differences in glycaemic measures between the two groups were analysed using linear regression adjusted for sex (due to gender differences) and steroid use.17


A total of 73 consecutive patients for whom a diabetes meal plan was ordered were identified. There were 30 patients in the Standard Group who received their meal plan for a total of 108 meal plan days and 43 in the patient-controlled group who received their meal plan for a total of 169 meal plan days. Patient demographic information is presented in table 2.

Table 2

Patient demographic information

Menu selections

Based on review of menu choices between the groups, no difference was observed in adherence to a consistent carbohydrate meal plan (standard group 70% (159/227) meals vs patient-controlled group 64% (278/437) meals, p=0.1). Only one patient made persistent menu errors despite nutrition education and was changed from the patient-controlled to the standard meal plan.

In the patient-controlled group, education visits were indicated in 23 (54%) patients due to carbohydrate selections consistently falling outside the recommended range. An education visit was completed in 13 (54%) of these 24 patients. Menu selections before and following the education visit were analysed in nine patients who remained in the hospital for up to 3 days following the educational session. Of these, six patients demonstrated improved meal planning capabilities following the education visit.

Glycaemic control (table 3)

No differences in mean blood glucose or rates of hyperglycaemia were observed between the groups. One patient in the standard group and seven patients in the patient-controlled group received corticosteroids (greater than 5 mg/day of prednisone or equivalent dose). After adjusting for sex and meal plan, those patients who received steroids had a significantly more severe hyperglycaemia (p=0.003) than those not receiving steroids. However, after adjusting for sex and steroid use, there was a significantly higher percentage of hypoglycaemia and a lower mean nadir blood glucose in the patient-controlled group (p=0.03).

Table 3

Glycaemic control of inpatients receiving standard meal plan versus patient-controlled meal plan

Patient survey

The survey was completed by 17 of 30 patients in the standard group and 30 of 43 patients in the patient-controlled group. The majority of those in both groups reported a duration of diabetes of greater than 5 years. Despite this, 45% of all participants reported that they had not received education from a registered dietitian ever or in the last 5 years. Reported meal plans at home included: no restrictions (17%), ‘no added sugar’ without regard to total carbohydrate intake (38%), consistent carbohydrate meal plan (32%) and exchange lists (13%) (table 4).

Table 4

Survey results addressing diabetes and nutrition management prior to admission

Patient satisfaction with their inpatient meal plan (88% vs 97%, p=0.03) and portion size (47% vs 67%, p=0.01) was higher in the patient-controlled group than the standard group. A small percentage of subject in each group reported that the foods provided in the hospital were similar to what they consumed at home (24% vs 20%, p=0.61).

The majority of patients in each group (84 vs 82%) reported that they consumed more than 50% of the food received on their meal trays. Hypoglycaemia did not occur more frequently in those patients who reported consuming <50% of meals. A small percentage of patients in each group (17% vs 13%) reported that they received additional food from outside sources while they were in the hospital. A trend towards more hyperglycaemia was observed among these patients (p=0.07).


The decision on the part of the hospital to gradually replace a standard meal plan with a patient-controlled meal plan provided the unique opportunity to conduct a QI project comparing menu selection, glycaemic data and patient satisfaction in hospitalised patients with diabetes. Despite similarities in the menu selections and adherence to a consistent carbohydrate pattern between the two groups, subtle but clinically important differences in glycaemic control were observed. The patient-controlled group experienced more hypoglycaemia and lower mean nadir blood glucose than those on the standard meal plan. There were no episodes of severe symptomatic hypoglycaemia in either group. Inadequate carbohydrate consumption may have contributed to hypoglycaemia, based on the finding that 50% of the patients who experienced hypoglycaemia in the patient-controlled group had chosen fewer than three carbohydrate servings at a meal.18 Hyperglycaemia occurred with a similar frequency in both groups despite the availability of a wider variety of food selections in the patient-controlled group.

An important finding of the survey was the high percentage of patients identified who have either never received nutrition education from a registered dietitian as part of their diabetes management or who have not seen a registered dietitian for over 5 years. Hospital admission may serve as an opportunity to identify these patients, provide survival skills education and direct them to the outpatient setting for more extensive MNT following discharge. In fact, 22 of 47 patients surveyed (47% standard group, 30% patient-controlled group) expressed an interest in obtaining more nutrition education after they leave the hospital.

Although the numbers were small, more than 60% of those in the patient-controlled group who were evaluated following targeted nutrition education demonstrated improved meal-planning capabilities. This information helps underscore the opportunity to provide education for hospitalised patients with diabetes, but also the importance of monitoring by clinical nutrition staff in a timely manner following hospital admission if the patient-controlled meal plan is used by an institution. This monitoring allows rapid determination of a patient's ability to make informed meal choices and need for nutrition education. More frequent selection of higher fat food items was observed in the patient-controlled group, suggesting that menu monitoring and nutrition education include information regarding both fat and carbohydrate meal content.18 19

Although there were no differences in portion sizes between the two meal plans, the patient-controlled group reported greater overall satisfaction with their meals and portion sizes. Whether increased satisfaction was related to a larger variety of menu items or to feelings of greater control over personal choices while in the hospital is not known but warrants further study.

There are several limitations to this QI project. One is that data collection was limited to four hospital units which restricted the number of patients with available menu and glycaemic data. The small sample size may have obscured important differences in glycaemic control between the two groups and limits the generalisability of the results. The expansion of the patient-controlled meal plan on all hospital nursing units limited our ability to study even more patients on the standard meal plan. Variability in staffing of nutrition personnel and lack of control over the timing of patient discharge contributed to an inability to both complete all indicated education visits as well as to evaluate the effectiveness of these visits.

An interesting aspect of this QI project is the insight that it provides into patient experiences in the hospital. This project was conducted in a hospital setting without any interruption of usual routines, allowing a glimpse into hospital patterns that create potential limitations to the use of a meal plan that requires significant oversight. These were non-critically ill patients on general medicine and cardiac services. Although the numbers were small, more than 10% of patients in each group reported receiving food from outside sources. While information regarding the types and timing of food from outside sources that was consumed is not known, it is possible that this extra nutrition can confound attempts at determining the impact of any therapeutic intervention on glycaemic control in the hospital.


Both standard and patient controlled approaches to consistent carbohydrate meal planning are reasonable choices for hospitalised patients with diabetes; however, each offers unique advantages and disadvantages. The standard meal plan requires fewer clinical resources, but also provides less opportunity for inpatient nutrition education. The patient-controlled meal plan provides a method of identifying patients who may be in need of diabetes education with the opportunity for patient-specific diabetes nutrition education which may improve the care of patients both in the hospital and following discharge.18 Hospitalisation can thus provide an environment for reinforcing sound nutrition principles and appropriate dietary choices that are carried to the home setting. In addition, it is associated with a higher degree of patient satisfaction. Education visits are particularly important when a patient-controlled meal plan is used. It is essential that adequate clinical nutrition staff be available to review menu data and provide nutrition education as needed. The higher frequency of hypoglycaemia observed in the patient-controlled group in this report may have been prevented by closer monitoring of menu choices.


View Abstract


  • Funding This work was sponsored by funding from the United States Air Force administered by the US Army Medical Research Acquisition Activity, Fort Detrick, Maryland, Award Number W81XWH-04-2-0030.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.