Establishment of the Screening Scale on the Aspiration Risk of Patients Receiving Enteral Nutrition

Purpose To establish a set of screening scale on the aspiration risk applicable to patients receiving enteral nutrition, determine the threshold of the scale and evaluate the scale’s reliability-validity. Methods Firstly, the screening scale on the aspiration risk of patients receiving enteral nutrition was mainly formed through literature search, expert brainstorming and consultation. And then a retrospective analysis was undertaken by investigating 484 patients in the intensive care unit of 19 tertiary hospitals around China from May 2019 to June 2019. Reliability-validity test on the screening scale concerning the potential aspiration risk among the vulnerable patients was performed. The items in the scale were assigned with scores through expert evaluation method. At last ROC curve was used to evaluate the diagnosis threshold of such scale.


Introduction
It was recommended that enteral nutrition should be and implemented as early as possible within 24-48 hours in the patients with critically ill, if the patients' digestive system is normal [1][2][3]. As an important way of nutritional support, enteral nutrition is proved to relieve the severity of diseases, reduce the comorbidities, shorter the hospitalization time and improve the prognosis [4]. However, it was also reported that during the enteral nutrition period, 59.1% patients were complicated with at least one kind of complications [5], among which, the most serious complication is aspiration. It is reported that the rate of enteral nutrition relevant aspiration is up to 40% [6]. Aspiration can lead pulmonary complications, including serious lung injury and acute respiratory distress syndrome. It can prolong the hospitalization time and increase the mortality [7]. Therefore, how to identify the high-risk factors of aspiration in patients with enteral nutrition and screen out the high-risk patients is an urgent problem to be solved.
A previous study had evaluated dysphagia and aspiration pneumonia in elderly hospitalization stroke patients [8], however, it was lack of speci city and sensitivity on the aspiration associated with enteral nutrition. Therefore, in this study, based on literature research, expert opinion, and multi-center retrospective analysis, we established the screening scale on the aspiration risk of patients receiving enteral nutrition, determine the threshold of the scale and measure the reliability-validity.

Aspiration Scale
Firstly, PubMed, Cochrane databases, and Web of Science databases were searched from Oct. 2009 to Oct. 2019, using the keywords "enteral nutrition", "aspiration", "aspirate", "aspirating", "factor", "evaluation" without language restrictions. We compiled 12 items for aspiration risk assessment based on physical condition, underlying disease, and treatment intervention in patients who aspirate. Secondly, we invited experts to develop the preliminary aspiration scale. Expert inclusion criteria: engaged in intensive care, medical treatment, teaching, management and scienti c research for more than 10 years.
Finally, 15 experts were selected, with an average age of 47.2 ± 7.4 and experience of 22.8 ± 4.3 years.
There were two rounds of consultation, each lasting for two weeks. The rst round is used to screen items that meet the assignment > 4.0 and the coe cient of variation < 0.2. The second round determined the weight of each item according to experts' experience. Finally, according to the experts and the clinical feasibility of the scale, we established the preliminary aspiration scale.

Aspiration Scale Pre-Detecting
The language expressions of items in the scale were modi ed through pre-detecting, and the expression was matched for the habits of clinical nurses. Random sampling was used to select 20 patients, who met the inclusion criteria in the intensive medicine department of a 3A hospital in Shanghai. Aspiration scale for risks of enteral nutrition aspiration was determined (Table 1). ventilation support. Clinical data of the enrollers were reviewed, including physical condition, underlying diseases and treatment, and the aspiration scale was used to assess the aspiration risk of patients who met the inclusion criteria.
Diagnostic criteria for aspiration [9]: 1) irritant choking cough or blood oxygen saturation decreases; 2) obvious shortness of breath, lung moist rale; 3) intratracheal aspiration of gastric contents; 4) in ammatory reactions such as bronchospasm and thickening of lung texture can be seen by imaging examination. In the process of this clinical research, matron and clinicians had provided training and answered questions to ensure scienti c and effective screening methods.

Evaluation of the aspiration scale
The reliability and validity of the aspiration scale was investigated. The internal consistency reliability of the scale was evaluated by Cronbach's a coe cient. The Inter-rater reliability was evaluated by intragroup correlation coe cient (ICC). Twenty patients were randomly selected, and the researchers and nurse screened patients for aspiration risk. ICC was used to calculate and compare the results of the two evaluators. The content validity index is used to re ect the content validity of the scale.

Statistical analysis
All statistical analyses were performed using SPSS version 20.0 software (SPSS Inc., Chicago, IL, USA.) Continuous variables were shown as mean ± standard deviation (SD), when normally distributed and with equal variances. Categorical data were summarized as frequency counts and percentages and measured with Chi-square or Fisher exact statistic test. The expert's positive coe cient was described by the recovery rate of the consultation table. The concentration of expert opinions was showed by mean ± standard. The coordination degree of expert opinions was described by coe cient of variation and coe cient of coordination (Kendall's W); ROC curve and Youden index (YI) value were used to nd the diagnostic threshold of the scale. P value < 0.05 is considered to be signi cant.

Results
Contents of the screening scale on the aspiration risk of patients receiving enteral nutrition The expert activity was presented by effective recovery rate of expert consultation questionnaire. Fifty questionnaires were distributed respectively in each round of the expert consultation and all 30 questionnaires were recovered with an effective recovery rate of 100%. The coordination index of expert indicators was principally re ected by the coe cient of variation and coordination coe cient (Kendall's w). The coordination index of expert indicators in the two rounds of consultation was shown in Table 1.
In combination of the screening results, expert modi ed suggestions in the 2 rounds of consultation and the opinions by our group members, 4 items in the 2 rounds of consultation were modi ed and 2 items were deleted (one of which was combined with another item). The ultimate 10 items in the primary scale were nally obtained. For the 10 items from the 3 dimensions, the mean values of the degree assignment by the experts were all more than 4.0, and the coe cient of variation was less than 0.2. The screening scale on the aspiration risk of patients receiving enteral nutrition was ultimately formed after modi cation of expression and wording according to the pretest, which was shown in Table 2. The weight and assigned scores of the dimensions and items of the aspiration risk screening scale The weight of the indexes in the scale was presented in Table 2. Considering the clinical utility of the scale, the assigned values were determined by the research group as following. The primary indexes were each assigned as 4 points. The secondary indicators corresponding to the basic condition, including age, body position, consciousness and dysphagia, were assigned as 1 point for each index. The scores of the secondary indicators corresponding to the dimensions of disease and accompanying symptoms (primary disease or associated abnormal condition and symptoms) assigned as 2 points for each index. The scores of the secondary indexes corresponding to the treatment intervention dimension, including drugs, supportive treatment, infusion route and infusion method, were assigned as 1 point for each index.
The diagnosis threshold of the aspiration risk screening scale The area under the ROC curve was 0.749 (P = 0.003) (Fig. 1). The area > 0.7 indicated that the diagnosis criteria possessed high accuracy. From Table 3, the maximum value of YI, sensitivity and speci city were 0.427, 0.800 and 0.627 respectively. Accordingly, the diagnosis threshold of the scale was determined by the group members as 8.5 points. Considering that the diagnostic purpose of this scale was risk screening which put more emphasis on the sensitivity, the ultimate diagnosis threshold was speci ed as 8 points. Reliability-validity analysis of the aspiration risk screening scale Internal consistency of reliability The Cronbach's α coe cients of the 3 dimensions of the scale including general conditions of patients, primary disease and the accompanying symptoms, and treatment, were respectively 0.892, 0.912 and 0.896. While the Cronbach's α coe cient of the summary scale was 0.922, indicating the high internal consistency of the results and ne reliability.

Interrater reliability
The ICC values of the interrater reliability among the 2 evaluators from the 3 dimensions were respectively 0.901, 0.960 and 0.961. The total ICC of the scale was 0.920.

Validity analysis
According to the results of the second round of expert consultation, the content validity indicated that the content validity index of the scale was 0.944 and the range of content validity index of the items was 0.812-1.000. All the items in this scale were factors in uencing aspiration during enteral nutrition, and came from the present studies and the experience of clinician, which could well re ect the conception of aspiration and possessed ne surface validity.

Discussion
Aspiration is considered as one of the main reasons leading to prolonged hospitalization time and increased mortality in patients receiving enteral nutrition, as it could cause aspiration pneumonia, asphyxia or death [1]. It is the consequence resulted from the participation and exertion of multiple elements. Although many studies have widely investigated the risk factors of aspiration [9][10][11], one previous study has indicated that the ICU nurses possessed a low understanding of aspiration [12]. To identify the occurrence of aspiration in patients with enteral nutrition and to take effective measures as early as possible is of great importance to improve the quality of nursing and ensure the safety of patients. Wu et al. has established the risk assessment system of aspiration in ICU patients based on the expert consultation [13], however, the threshold of aspiration was not de ned.
In this study, with literature research of the risk factors leading to aspiration during enteral nutrition, expert brainstorm and expert consultation, the screening scale on the aspiration risk of patients receiving enteral nutrition was established and the diagnostic threshold was de ned. This scale has good clinical pertinence, makes up for the blank of screening tool for intestinal nutrition aspiration, and can effectively promote the improvement of clinical quality.
The reliability of the scale refers to the content of internal consistency of the measuring results. The Cronbach's α coe cient of the screening scale on the aspiration risk of patients receiving enteral nutrition is 0.922, which indicates the high internal consistency of the measuring results and ne reliability. The items were evaluated and modi ed according to expert consultation. The results showed that the content validity and the range of the content validity index of the items were 0.944 and 0.812-1.000 respectively, which were higher than the recommended values [14], and indicated the good content validity of the scale. Simultaneously, with the large-scale literature research, expert brainstorm and the clinical experience, the risk factors in uencing aspiration during enteral nutrition were induced and summarized and the primary item tank was ultimately obtained. These items were capable of re ecting the conception of aspiration and provided ne surface validity.
According to the ROC curve analysis, the area under the ROC curve was 0.749 with P = 0.003, which indicated a high diagnostic value of the scale. Based on the theory of the best critical point, the diagnosis threshold of aspiration during enteral nutrition was de ned as 8.5 points. The diagnostic sensitivity and speci city at this critical point are respectively 0.800 and 0.627. Considering that the assignments of the scale are all integers while the moderate decreasing of the diagnosis threshold can add sensitivity of the scale with the quality speci city being ensured, the diagnosis threshold of the scale was ultimately adjusted to 8 points, which is conducive to maintaining the evaluation authenticity of the scale. When the assessment results ≥ 8 points, there is a risk of aspiration for the patient receiving enteral nutrition.

Conclusion
The established "Screening scale on the aspiration risk of patients receiving enteral nutrition" is capable of identifying the risk factors in uencing aspiration during enteral nutrition directly and conveniently. In this study, the diagnosis threshold as a total of 8 points is de ned of the scale, and a series of relevant factors in uencing aspiration during enteral nutrition are also clari ed. It is of great clinical signi cance to decrease the incidence of the aspiration during enteral nutrition and improve the implementation quality and level of clinical enteral nutrition. Figure 1 ROC curve of the aspiration risk screening scale