Incidence of nutritional risk and causes of inadequate nutritional care in hospitals

Transcription

1 Clinical Nutrition (2002) 21(6): r 2002 Elsevier Science Ltd. All rights reserved. doi: /clnu , available online at on ORIGINAL ARTICLE Incidence of nutritional risk and causes of inadequate nutritional care in hospitals J. KONDRUP,* N. JOHANSEN,* L. M. PLUM,* L. BAK,* I. HÒJLUND LARSEN, w A. MARTINSEN, w J. R. ANDERSEN w, H. BÓRNTHSEN, z E. BUNCH, z N. LAUESEN z *Nutrition Unit, Rigshospitalet, w Department of Internal Medicine, NykÖbing Falster Central Hospital, z Department of Surgery, Hobro Hospital, Copenhagen, Denmark (Correspondence to: JK, Nutrition Unit 5601, Rigshospitalet, 9 Blegdamsvej, 2100, Copenhagen Ò, Denmark) AbstractFBackground and aims: Many patients in hospitals are undernourished and nutritional care is inadequate in most hospitals. The aim of this investigation was to gain insight into how this situation could be improved. Methods: Seven hundred and fty randomly selected patients were screened at admission in three hospitals and surveyed during their entire hospitalization. Each time a patient was not treated according to a clearly de ned nutritional standard, the nurse responsible for the patient was interviewed about possible reasons according to preformed questionnaires. Results: The investigators found that 22% of the patients were nutritionally at-risk, and that only 25% of these patients received an adequate amount of energy and protein.the departments had only screened for nutritional problems in 60% of the cases. Only 47% of the patients, who the departments judged to be at-risk patients, had a nutrition plan worked out, and only about 30% of the at-risk patients were monitored by the departments by recording of dietary intake and/or body weight. The main causes for inadequate nutritional care were lack of instructions to deal with these problems, and lack of basic knowledge with respect to dietary requirements and practical aspects of the hospital s food provision. Patient-related aspects and the system of food provision also contributed, but only to a small degree. Conclusions:These ndings form the basis of the strategy to improve nutritional care in these hospitals. r 2002 Elsevier Science Ltd. All rights reserved. Key words: nutrition; malnutrition; undernutrition; screening; patients, hospital Introduction Several studies have shown that a large proportion of patients in hospital is malnourished (1 4). McWhirter and Pennington found that 40% of the patients were malnourished on admission to hospital, and that 75% of the malnourished patients, who remained in hospital for more than 1 week, lost further weight (2). According to a questionnaire study among doctors and nurses (5), most of them agreed that nutritional support would prevent complications during hospitalization. However, only about 20% of them performed nutritional screening and/or assessment, including recording of dietary intake and body weight. The most common reasons given for this lack of clinical practice were insufficient knowledge, low priority, unclear assignment of responsibility and lack of procedures or guidelines. In this study, the primary objective was to define the relative importance of various potential causes of inadequate nutritional care by analyzing a number of randomly selected patient admissions. Each time a disagreement between the actual clinical activity and a well-defined standard was observed, the nurse taking care of the patient was interviewed according to a preformed questionnaire. In addition, the nurses taking care of the patients at admission were asked a number of questions in order to test their basic knowledge of theoretical and practical aspects of nutritional care. Further, we wanted to describe the incidence of nutritional risk among patients in an unselected population of patients. The study was performed among newly admitted patients in three hospitals: a university hospital with 1200 beds, a regional hospital with 300 beds and a local hospital with 150 beds. Patients and methods The following hospitals participated: Rigshospitalet Copenhagen (RH, university), Nykbing Falster Central Hospital (NF, regional), Hobro-Terndrup Hospital (HS, local). At each hospital, a nurse and a clinical dietitian formed a team of investigators. On the day after admission, the investigators selected three patients among all patients admitted in the preceding 24 h, by a 461

2 462 INADEQUATE NUTRITIONAL CARE IN HOSPITALS random numbers system. The following patient-groups were excluded: one-day s admissions, children (o15 years), healthy women admitted for child delivery and patients with psychiatric disorders. Altogether, 750 patients were selected, 250 in each hospital. After the random selection, the investigators visited the patients and the nurses involved in their care. As a standard for evaluation, we used the official Danish guidelines (6), as described in (7). These guidelines follow the standards that meet the requirements given by the Joint Commission on Accreditation of Healthcare Organizations (8). The nurse was asked whether the patient had a nutritional screening performed (as a minimum including a recording of body weight and an evaluation of this body weight in relation to the patient s medical problems and status at admission). If they had, the nurse was asked whether the staff considered the patient to be nutritionally at risk. If they did, she was asked for documentation of a nutrition plan (estimation of energy and protein requirement, a decision about food/feeding regime and a plan for monitoring intake and body weight). Whenever the nurse, on behalf of the staff, failed according to this procedure, she was questioned immediately about the reasons, according to a preformed questionnaire. Finally, the nurse was given 20 short questions in order to test her knowledge with respect to basic theoretical and practical aspects of performing nutrition care. The same nurse was given these 20 questions once only, and the questioning was stopped after 268 nurses had been asked. The total interview of patient and nurse lasted 5 10 min in most cases. After this, the investigators evaluated the nutritional risk of the patient as described in (7). The degree of malnutrition was categorized as none, slight, moderate or severe, employing information on height (according to the patient s information), body weight (weighed if possible, otherwise according to the patient s information or estimated), recent weight loss and period of weight loss (0 3, 3 6, 6 9, 9 12 months; if within 0 3 months, supplemented with information about whether it was within the last 1, 2 or 3 months), dietary intake during the last week before admission (categorized as intakes of 0 25%, 25 50%, 50 75%, % of estimated energy requirement). Energy requirement was based on the factorial method (BMR calculated by the Harris Benedict equation multiplied by an activity factor in most cases, or by a stress factor in a few cases), as explained in more detail in (7). The concept of categorizing dietary intake into quartiles was taken from (9) which showed an acceptable degree of correlation between actual intake and the categorization approach. A score of 3 was given for severe malnutrition (BMI o18.5, recent weight loss 45% in the last month, or an intake of 0 25% of requirement). A score of 2 was given for moderate malnutrition (18.54 BMIo20.5, recent weight loss 45% in the last 2 months, or an intake of 25 50% of requirement). A score of 1 was given for slight malnutrition (recent weight loss 45% in the last 3 months, or an intake of 50 75% of requirement). When the score was dependent on BMI, it was a further condition that the low BMI was associated with an impairment of the patient s general condition. Finally, the severity of disease, as an indicator of stress metabolism, was categorized as none, slight (score = 1), moderate (score = 2) or severe (score = 3). Then, the score for nutritional status and for severity of disease were added to give a total score which could range from 0 to 6. Any patient with a total score 3 was considered to be nutritionally at risk. This scoring system is explained in more detail in (7). If the patient was defined to be nutritionally at-risk by the investigators, the patient s dietary intake was assessed daily during the admission. This was based on dietary recordings, if available, and otherwise the information collected from the patients and nurse was categorized as intakes of 0 25%, 25 50%, 50 75%, %, % or % of estimated requirement for energy and protein for maintenance of body weight. The weekly average of dietary intake was estimated as one of the six categories mentioned. The categorization was done by energy or protein, whichever was the lowest. It was decided not to include patients in this analysis if they stayed in hospital for less than 4 days (defined as o1 week). At discharge from hospital, the mean category of intake for the patient s stay was calculated by using dummy values of 1 6 for the categories mentioned. The investigators also collected daily information about the patients dependency on other persons (assistance to toilet visits, confined to bed or chair), surgical procedures, complications to surgery, days spent in the intensive care unit and the use of antibiotics. At discharge, the length of stay was noted in weeks (all patients) and number of days (at-risk patients). Once a week, the investigators collected the departments information about body weight, or, if the department had not weighed the patients, the investigators took care of this. If the patient had not been weighed at least twice weekly, or if the patient had not had a dietary recording performed by the staff at least 5 days per week, the nurse was questioned about possible reasons. If the patient had not eaten sufficiently according to the investigator s estimate, the nurse was questioned about possible reasons for insufficient intake. When calculating the number of causes given (see Results), any single cause for a particular patient was allowed to appear only once. In a few cases, the same patient appeared with different causes in different weeks, and in these cases all causes were counted. If the patient was not an at-risk patient at admission, the investigators examined the treatment plans for the patients in order to evaluate whether the patient would become an at-risk patient within a week, e.g. because of a scheduled major operation. These patients were

3 CLINICAL NUTRITION 463 considered risk-prevention patients and they were followed up as above. If the patient was neither an atrisk patient nor a risk-prevention patient, the investigators collected data once a week regarding nutritional status and severity of disease, similar to the information acquired at admission, in order to evaluate whether the patient would become an at-risk patient during the hospital stay. Results are presented as means7sem. Statistical difference was tested by Students t-test for continuous data and Fisher s exact test for categorical data. In a few cases, data were obviously not normally distributed and in these cases the Mann Whitney test was applied. Logistic regression, generalized linear model analysis and survival analysis (for analyzing length of stay) was performed by Systat 10, SPSS Inc., Chicago, USA. Results Incidence of nutritional risk Table 1 shows the diagnoses of the patients included in the study, divided into at-risk and non-risk patients. The ranges of scores for severity of disease allocated to each diagnostic group are also shown. Table 2 shows the data for at-risk and non-risk patients at admission. For 10 of the 750 patients, data were too incomplete to allow a score, in most cases because the patients had been dismissed, or transferred to other hospitals, before the investigators arrived. At-risk patients had a lower mean body weight and BMI. A higher proportion of at-risk patients had a recent weight loss and the magnitude of weight loss was larger. Also, a larger proportion of atrisk patients had insufficient dietary intake during the week preceding admission and among these a larger proportion had a severely impaired intake. The score for nutritional status, for severity of disease and the total score were also significantly higher in the at-risk group. In order to elucidate how the different components of the scoring system influenced the categorization as atrisk or non-risk patients, a logistic regression analysis was performed with the categories for BMI, recent weight loss, recent dietary intake and severity of disease as independent continuous variables and at-risk or nonrisk patient as dependent binary variable. The result was (expressed as coefficients for incremental steps in category for each variable): BMI: 1.7; recent weight loss: 2.4; recent intake: 3.8; severity of disease: 4.3. All coefficients were significantly different from 0 (Po0.0001; McFadden s rho 2 for the equation = 0.82). In addition to the 124 patients, who were judged to be at-risk patients at admission (Table 2), 29 patients were judged to be risk-prevention patients and a further group of 14 patients developed a state of nutritional risk during the hospital stay. Altogether, 167 of the 740 patients who could be evaluated, were at-risk patients, equalling 23% of all admitted patients. Table 1 Diagnoses at admission for patients at-risk and patients not at-risk. The range of scores for severity of disease is also shown Condition At risk Not at risk N Score N Score Internal medicine Cardio vascular medicine Ischemic heart disease Incompensated heart disease Other cardio vascular Respiratory medicine Chronic obstructive pulmonary disease Pneumonia Other acute respiratory illness Neurological disorders Vascular disease Convulsions Other neurological illness Gastrointestinal disorders Upper G.I. disorders Lower G.I. disorders Pancreatic disease, hepatic cirrhosis Various internal medicine Infection Intoxication Diabetes mellitus Observation Nephrology Other internal medicine Surgery Neurological surgery Chest surgery Major abdominal surgery Minor abdominal surgery, including appendicitis Abdominal pain, not specified Major orthopedic surgery Fractured neck of the femur Trauma Vascular surgery Other surgery Malignant disease (oncology/ hematology) Ophthalmology Table 2 Nutritional risk status of newly admitted patients (mean 7 SEM) Risk patients Non risk patients N Age (years) Body weight (kg) * Height (cm) BMI (kg/m 2 ) * Weight loss, all, 0 3, 3 6, 43, 5, 2, 13 49, 15, 3, or 9 12 months, N Recent weight loss (0 3 months), N 43 49* Recent weight loss (kg) * Inadequate intake (o75% 98 77* of requirement), N Category of intake 0 25%, 16, 40, 42 0, 3, % or 50 75% of requirement, N Nutritional status, score * Severity of disease, score * Total score, * *Po either by Mann Whitney t-test or by Fisher s exact test. At the local, county and university hospitals 28 (11%), 48 (19%) and 48 (19%) patients were judged to be at-risk patients at admission (P = 0.02 for HS vs

4 464 INADEQUATE NUTRITIONAL CARE IN HOSPITALS Relative cumulative frequency (1 83 patients) Fig Intake (% of requirement) Adequacy of intake among at-risk patients with 1 week in hospital. NF or RH). At the local, county and university hospitals, the average scores for nutritional status were similar, but the scores for severity of disease were different: , and , respectively (P = for test of linear trend as post test) and the total risk scores were , and , respectively (P = for test of linear trend as post test). A total of 12 patients died (eight at-risk patients and four non-risk patients, P = 0.001) and 8 were judged to be terminally ill (seven at-risk patients and one non-risk patient, P = 0.002). These patients were excluded from the following analyses. Figure 1 shows the distribution of average dietary intakes during admission. Of the 167 at-risk patients, the 77 patients who stayed 1 week and who were not terminally ill or died, are included (see Methods). Among these patients, only about 25% had 75 99% of their needs covered. Intake of 75% of requirement seems crucial, since all patients with an intake below 75% of their estimated requirement experienced a weight loss (7). A further analysis (data not shown) indicated that the average intake was not better for the patients with a long stay in hospital (3 weeks), compared to those with a short stay (r 2 weeks). The average length of stay of the patients in Figure 1 was weeks (mean7sem). Table 3 shows the distribution of weight gains and losses from admission to discharge. The table includes 52 of the 83 at-risk patients with 1 week stay, who were weighed at admission, who had no edema and who were weighed weekly during the stay. The average length of stay for these patients was weeks (mean7sem). Sixteen of the patients had some weight loss (2%, or an average of kg, mean7sem) while nine of these had a considerable weight loss (5%, or an average of kg, mean7sem). Figure 2 shows that the length of stay (in weeks) was significantly longer for at-risk patients, compared to non-risk patients, as analyzed by survival rate statistics. Patients who died or who developed a terminal stage of disease (N = 20) were not included in the analysis. About 95% of the non-risk patients had been discharged after 2 weeks while this was the case for at-risk patients only after 5 6 weeks. In order to elucidate a possible association between the components of nutritional risk assessment and length of stay, an analysis by the generalized linear model with stepwise backward elimination (P40.15) was performed with length of stay (weeks) as independent variable and scores for nutritional status and severity of disease as independent variables. In this analysis, recent intake (P = 0.002) and severity of disease (Po0.0005) were significantly associated with length of stay, while BMI and recent weight loss were not. For the equation, multiple r 2 was 0.14, indicating that other factors had a major influence on length of stay. Expressed as median values, the length of stay (weeks) for non-risk patients (median; interquartile range) was 1 (1 1) and for at-risk patients 1 (1 3; Po by the Mann Whitney test). Table 3 Weight gains or losses. Number of patients of a total of 52 a Gained 5% of initial body weight 4 Gained 2% and o5% of initial body weight 7 Weight stable 25 Lost 2% and o5% of initial body weight 7 Lost 5% of initial body weight 9 a Patients at nutritional risk at admission, who stayed in hospital for more than 1 week and who had no edema and who could be weighed.

5 CLINICAL NUTRITION 465 Fig. 2 % 100% all patients in group (At-risk: 152; non-risk: 578) Length of stay At-risk Not at-risk P < Mantel-Haenszel test week number Among at-risk patients, the median length of stay was 9 days with an interquartile range of An analysis similar to the one above was performed within the group of patients at-risk with a length of stay allowing at least one period of assessment of adequacy of dietary intake (6 days; N = 70). The association between length of stay (days) and the components of screening at admission and events during the admission was evaluated. Length of stay (days) was the dependent variable and the admission categories for BMI, recent weight loss, recent dietary intake and severity of disease, together with dietary intake during the admission and number of days with dependency on others, were independent continuous variables and use of antibiotics, subjection to surgery and transfer to intensive therapy were binary variables. By this stepwise analysis, recent weight loss (P = 0.096), intake during the admission (P = 0.057), severity of disease (P = 0.009) and number of days with dependency (Po0.0005) were selected. All variables, including intake during the admission, were positively associated with length of stay. For the equation with these variables, adjusted multiple r 2 was 0.44 (Po0.0005). On its side, number of days with dependency on others was related to recent weight loss (P = 0.014) and performance of surgery (P = 0.011), and not to the other independent variables mentioned. The care provided Table 4 shows how the patients were treated by the departments. Only 59% of the patients were screened at admission to hospital. Among those screened, the departments had judged 21% to be at-risk patients, either at admission (12%), or as risk-prevention patients or as becoming at-risk patients during the stay (9%). Only 47% of these patients had a nutrition care plan worked out and only 33% and 39 %, respectively, was monitored by recording dietary intake or body weight. Table 5 shows the results of the questionnaires dealing with the causes of inadequate nutritional care at admission. The main reason for not screening the patients was the lack of instructions to do so. The main reasons for not classifying a screened patient as an atrisk patient, at variance with our judgement, was the lack of guidelines to perform this screening and the expected short duration of the stay. Outright disagreement with the investigators was the cause in a few cases only. Table 6 deals with the patients that the departments had judged to be at-risk, but for whom they had not worked out a nutrition care plan. In this case, they could give one answer in each of several categories. The most common category was the usual routine of care, in most cases stating that they usually just observed the patient. The next category was education, in which the most common answer was lack of education in estimating the needs of the patient. In the category dealing with food, the most common cause was difficulties arising from problems with chewing, swallowing, nausea or vomiting. It is noticeable that the kitchen s food did not seem to be a problem. Finally, in the category of support, the most common reason was that the patient was not interested. In case the departments had not monitored their plan at the investigators weekly status, i.e. by recording dietary intake and/or weighing the patient (71 cases), the dominant answer was that it was not necessary (data not shown). Table 7 deals with the reasons why patients had an insufficient intake, as apparent from the investiga-

6 466 INADEQUATE NUTRITIONAL CARE IN HOSPITALS Table 4 Screening and nutritional care as performed by the departments Y N Y N Y N Y N Y N Y N Y N Y N Screening at admission Screened upon request Screening, total Risk patient at admission Risk prevention or during stay Nutrition plan Recording of daily dietary intake Recording of body weight % After request by the study team. 2 According to the opinoin of the department, or receiving nurse. 3 According to the opinion of the department, or receiving nurse. See text for definition of the term risk prevention. 4 Required documented estimate of energy and protein requirement, prescription of dietary regime and a plan for monitoring intake and body weight. 5 Required written documentation of dietary intake for at least 3 of 5 week days in at least 3/4 of the weeks of stay. If the patient was discharged within 4 days after admission (N = 39), it was decided that the activity could not be evaluated. In addition, data from patients who died, or who were judged to be terminally ill, were not evaluated. Therefore data were available for only 93 of 137 at-risk patients. 6 Required recording of body weight at least twice per week. If the patient was discharged within 4 days after admission (N = 39), it was decided that the activity could not be evaluated. In addition, data from patients who could not be weighed, who died, or who were judged to be terminally ill, were not evaluated. Therefore data were available for only 72 of 137 risk patients. 7 Percent of 750 admitted patients who were screened spontaneously by the department. 8 Percent of 678 patients screened. 9 Percent of all 137 risk patients. 10 Percent of 93 patients who could be evaluated (see foot note 5). 11 Percent of 72 patients who could be evaluated (see foot note 6). tors s recordings. Again, the nurses could give one answer in each of several categories. The most common category was the one related to patient problems, most commonly that the patient had a decreased appetite. After this, the education category was the most common, in that nurses lacked tools to estimate the patients needs and also to estimate the actual content of energy and protein in the hospital s menus. The category related to food and catering also had a few answers in several options, most commonly that the menus of the hospital were not suitable for the patient. Not shown is a separate category dealing with the routine of care, in which the nurses had the options of answering, that they had too little time for feeding the patient, had too little time for urging and motivating the patient, or that the patient deliberately had been kept starving due to examinations, operations, etc., but there were actually only four answers in this category. Table 5 Reasons why patients were not screened, or were not classified as risk patients Why was this patient not screened for nutritional risk at admission? We do not know how to do it 5 There is no instruction to do it 232 We just forgot in this case 78 Total 315 Why was this patient not classified as a risk patient? We do not have guide-lines to define a risk patient 35 The patient is going to stay here for a short time only 21 We disagree: nutrition has no importance for the clinical course 6 of this patient Total 62 Table 8 shows the results with regard to the basic knowledge among nurses. Questions 1 3 dealt with the screening system used by the investigators, and most nurses gave correct answers to these questions. This is Table 6 Reasons for not making a plan for nutritional care (N = 84) Category Answer N Education We are not trained in estimating requirement 29 We are not trained in composing a diet 4 We are not trained in monitoring nutritional 8 status and dietary intake Total 41 Care We plan to do it do it tomorrow 2 Sorry, we forgot 2 In this situation, we usually just observe the patient 45 Total 49 Food It is impossible due to difficulties in chewing, 12 swallowing, nausea or vomiting The kitchen s food is not suitable for this patient 0 The taste of liquid supplements is not acceptable 1 to the patient Tube feeding would be too stressful for this patient 5 Parenteral nutrition is too expensive or too 1 complicated Total 19 Support We had too little support from the doctor 5 We had too little support from the dietitian 0 The patient is not motivated 10 Total 15 Terminal The patient is terminally ill 9 The sum of totals = 133, since for 41 patients there were reasons in more than one category. The most frequent combination was Education and Care.

7 CLINICAL NUTRITION 467 Table 7 Reasons for insufficient dietary intake (N = 57) Category Reason N Education We lack tools to estimate energy and protein 18 requirement We lack tools to estimate the energy and protein 11 content of the menus Total 27 Patient related The patient was not motivated after all 8 The patient had a poor appetite 22 The patient was too weak to eat 8 The patient had nausea and vomiting 10 Total 42 Food The food we ordered did not arrive 0 The food was ill prepared 1 The food was ill served 3 The menus were not suitable for the patient 9 There are too few varieties in snacks 2 The diet with modified consistency was too 1 monotonous The enriched diet for low-intake patients was 0 not suitable We have insufficient experience with tube feeding 3 We have insufficient experience with parenteral 1 nutrition Total 19 Terminal The patient became terminally ill 7 In each category the sum of answers can be larger than the total, since various answers could apply to an individual patient in different weeks of his stay, but in a particular week only one answer could be given per category. The sum of totals = 99, since 30 patients had answers in more than one category. The most frequent combination was Education and Patient related. noticeable since very few nurses were actually acquainted with the screening system. Questions 4 10 dealt with recent developments in clinical nutrition, and the knowledge of these subjects was rather scarce. Questions dealt with nutrition care including practical aspects of the hospital s catering and most nurses had insufficient knowledge about most of these items. The answers to the last question validate the content of the other 19 questions. Discussion This study agrees with most earlier studies that a large proportion of hospitalized patients is malnourished. However, several aspects distinguish this investigation from most of the previous studies. First, instead of only estimating nutritional status, we estimated nutritional risk involving the combination of nutritional status and severity of disease. As discussed in more detail elsewhere (7), our screening system is based on an analysis of randomized controlled trials, and it is clear that most of these studies dealt with patients who at the same time suffered from some degree of malnutrition and some degree of disease activity, or severity, as an indicator of increased nutritional requirements. We therefore believe that indications for nutritional support is not only a matter of nutritional status. Secondly, we have studied incidence rather than prevalence, in contrast to many previous studies that studied patients already inhospital, rather than those arriving at the hospital. Thirdly, we have monitored the patients during their entire hospital stay, rather than merely doing a time fixed survey. As far as we know, only the study of McWirther and Pennington (2) shares most of these features with our study. However, in contrast to their study, we dealt with all departments in the hospitals, we involved hospitals of three levels (university, county, local), and we examined the possible causes of inadequate care in a direct relation to events during the hospital stay. We found that about 20% of the patients admitted to hospital were nutritionally at-risk. According to the logistic regression analysis, all components of the scoring system contributed to defining the patients as at-risk patients. In combination with the data in Table 2, this means that inclusion of severity of disease in the definition of nutritional risk does not lead to selection of patients who simply are severely ill. Only 25% of the atrisk patients received an amount of food that could be considered adequate, even by a very permissive criterion (75% of requirement for weight maintenance covered). Accordingly, a large fraction of these patients lost weight during their stay. However, we did not see severe weight losses as frequently as reported in (2). This could be due to the fact that they focused on five departments, which would be expected to have a high proportion of nutritionally complicated patients. We also found that the length of stay was significantly longer for the at-risk patients. The median length of stay was not increased but a large fraction of the at-risk patients had a prolonged stay, so that the time elapsing before 95% of the patients were discharged was 2 weeks for non-risk patients and 5 6 weeks for at-risk patients. When considering all at-risk and non-risk patients, recent dietary intake and severity of disease were independently related to length of stay (weeks). Within the group of atrisk patients, the length of stay (days) was independently related to recent weight loss and intake during admission, in addition to severity of disease and number of days with dependency on others. Both analyses suggest that nutritional variables are related to length of stay. Intake during admission was also positively related to length of stay, suggesting that targeted intake was reached more commonly in patients who stayed in hospital for a prolonged period. The number of deaths was significantly higher among the at-risk patients, but the data do not allow any further speculations as to how this was associated with nutritional status and severity of disease. The incidence of at-risk patients was higher in the university and county hospitals, compared to the local hospital. Further, the average risk scores showed a significant trend of linearity according to type of hospital which was related to severity of disease, rather than severity of nutritional status. These findings are

8 468 INADEQUATE NUTRITIONAL CARE IN HOSPITALS Table 8 Basic knowledge about nutrition problems among nurses (N = 268; results are given as % of total answers for each question) % Yes No U 1 A patient with a hip fracture and a BMI , is she considered a risk patient according to official recommendations? A stroke patient with an intake of 25 50% of estimated requirement during the last week, is he considered a risk patient according to official recommendations? 3 A patient with a chronic lung disease who has had a weight loss of 5 10% during the last 6 weeks, is he considered a risk patient according to official recommendations? 4 Will two weeks preoperative nutrition therapy reduce postoperative infections? Can early enteral nutrition after major gastrosurgery reduce the need for analgesics? Can tube feeding be started after major gastrosurgery, even if there is 150 ml in the stomach? Do most undernourished patients lose further weight during a hospital stay? If a patient loses 10% of his body weight, is the loss of muscle strength also 10%? Is 30% of the patients in hospitals considered to be at nutritional risk? Do most stroke patients develop signs of malnutrition during their hospital stay? Does the standard meal plan in our hospital require that the patients gets about 25% of the total intake from snacks? Is g protein/day the average protein requirement for a patient? Is kj the average energy requirement for a patient? Is bouillon a good meal for a patient with a low intake? Does junket contain about 3 g protein/100 ml? Does this hospital s primary menu offer about 9000 kj in the 3 main meals? Is this hospital s primary menu a better choice for a patient with low intake than the secondary menu? If you wish to change the order of a main menu to a liquid menu, can this be done until 4 p.m.? (without extra cost) If you ordered and stored in the freezer: is there a choice of 5 10 frozen dishes that can be used as alternatives to the main menus? 20 Were these questions relevant? U = respondent does not know; Bold = correct answer. explained by the different functions of the hospitals, with an increasing degree of specialization. A few major areas causing the lack of focus on nutritional problems emerge from our questionnaires. First, departments ought to have guidelines and instructions regarding nutritional screening and therapy, including precise rules for how long time patients are just observed without intervention. Secondly, the nurses have insufficient knowledge about theoretical and practical aspects of nutrition. Especially, it seems that the nurses are grossly unaware of the importance of snacks between the main meals. However, since the nurses found the 20 basic questions relevant, it is very likely that the nurses would find information and seek education, provided that the departments requested this care through instructions and guidelines. Thirdly, the area of patient-oriented aspects such as lack of appetite and suitability of the hospital s food is also an issue, although it seems quantitatively to be of much less importance than the first two areas. This last issue should not receive the main attention in the beginning, but once the areas of guidelines and knowledge of the nurses are covered, this issue might become more significant. At present, the managements of the three hospitals involved in this study have sent out instructions to the departments including guidelines for nutritional screening and care. It is an important feature of the managements instructions that they are not simply issued as a policy, but rather as detailed instructions which the single departments may modify, provided there is an adequate documentation for the modification. Following this, the study investigators are now teaching and supporting the departments in implementing improved nutritional care routines, and in a final study period the effects of these interventions will be measured. Acknowledgements The study was financed by the Danish Ministry of Health and the hospitals participating in conjunction with the county of Nordjylland, the county of Storstrom and the Copenhagen Hospital Corporation Organization (H:S). References 1. Naber T H, Schermer T, de Bree A et al. Prevalence of malnutrition in nonsurgical hospitalized patients and its association with disease complications. Am J Clin Nutr 1997; 66: McWhirter J P, Pennington C R. Incidence and recognition of malnutrition in hospital. BMJ 1994; 308: Edington J, Boorman J, Durrant E R et al. Prevalence of malnutrition on admission to four hospitals in England. The Malnutrition Prevalence Group. Clin Nutr 2000; 19: Kelly I E, Tessier S, Cahill A et al. Still hungry in hospital: identifying malnutrition in acute hospital admissions. QJM 2000; 93: Rasmussen H H, Kondrup J, Ladefoged K et al. Clinical nutrition in danish hospitals: a questionnaire-based investigation among doctors and nurses. Clin Nutr 1999; 18: Pedersen A N, Ovesen L F (eds). Recommendations for Food in Public Institutions in Denmark. Copenhagen: Danish Ministry of Food and Agriculture, Kondrup J. Can food intake in hospitals be improved? Clin Nutr 2001; 20 (Suppl 1): Anonymous. ASPEN. Standards for nutrition support: hospitalized patients. Nutr Clin Pract 1995; 10: Olin A O, Osterberg P, Ha dell K et al. Energy-enriched hospital food to improve energy intake in elderly patients. JPEN 1996; 20: Submission date: 12 March 2002 Accepted: 5 July 2002