Infectious Diseases

A systematic review of tools for predicting complications in patients with acute infectious diarrhea

a b s t r a c t

Objective: To identify tools that predict the risk of complications in patients presenting to outpatient clinics or emergency departments (ED) with acute infectious diarrhea.

Methods: Medline, Embase, Cochrane Library, Web of Science and CINAHL were searched from inception to July 2021. Articles reporting on the derivation or validation of a score to stratify the risk of intraVenous rehydration or hospitalization among patients with acute infectious diarrhea in the ED or outpatient clinic were retained for analysis.

Results: Five articles reporting on two different tools were identified. Developed to assess the risk of hospitaliza- tion of children, the EsVida scale has not been externally validated. Developed originally to assess the level of de- hydration in children, the Clinical Dehydration Scale (CDS) was evaluated as a risk stratification tool. For predicting intravenous rehydration, a CDS score >= 1 showed a sensitivity between 0.73 and 0.88 and specificity between 0.38 and 0.69, whereas a CDS score >= 5 showed a sensitivity between 0.06 and 0.32 and specificity be- tween 0.94 and 0.99. For predicting hospitalization, a CDS score >= 1 showed a sensitivity between 0.74 and 1.00 and specificity between 0.34 and 0.38, whereas a CDS score >= 5 showed a sensitivity between 0.26 and 0.62 and specificity between 0.66 and 0.96. High heterogeneity among studies and unclear risk of bias precluded meta-analysis.

Conclusion: As a Risk-stratification tool, the CDS has been validated only for children. Further research is needed to develop and validate a tool suitable for adults in the ED.

(C) 2022

  1. Introduction

Acute infectious diarrhea, defined as lasting <14 days and being of infectious etiology, is usually self-limiting and rarely fatal [1]. A frequent cause of outpatient and emergency department (ED) visits, it is having a significant impact on healthcare resource use around the world. The Centers for Disease Control and Prevention estimate there are >178 mil- lion cases of acute infectious diarrhea in the United States annually. Based on self-reported data, the estimate rates of acute diarrhea per

* Corresponding author at: Centre hospitalier universitaire – Universite de Laval, 2705 Boulevard Laurier, Quebec G1V4G2, Canada.

E-mail address: [email protected] (T. Marx).

person per year has almost doubled in the first decade of the 21st cen- tury [2]. The annual cost of this disease is estimated to be approximately

$77 billion to the United States economy [3]. Acute diarrheal illness is a major public health issue against which control efforts are needed [1].

Except in specific cases (bloody stools, prolonged or severe symp- toms, patient immunocompromised or with significant comorbidities), diagnostic tests to identify a specific pathogen are rarely indicated given the benign and self-resolutive course of most presentations. Moreover, oral rehydration is usually sufficient to prevent dehydration [4]. Most patients at low risk of complications do not need to see a phy- sician and could refrain from unnecessary use of Healthcare resources. However, estimating the risk of a patient is a complex task.

Risk stratification tools may help predict the need for investigation or treatment and thereby optimize the use of healthcare resources.

https://doi.org/10.1016/j.ajem.2022.11.024

0735-6757/(C) 2022

A large number of these tools have been shown to improve cost- effectiveness while minimizing unnecessary diagnostic imaging, treat- ment complications, and unnecessary admissions [5-9]. Most of these tools have been developed as decision aids for physicians, and few are intended to help patients and caregivers identify the most appropriate care pathway for the condition [10,11]. The aim of this systematic re- view was to identify risk stratification tools for use by clinicians or pa- tients to predict the risk of complications arising from acute infectious diarrhea in patients presenting at an outpatient clinic or ED.

  1. Methods

This systematic review followed the PRISMA (referred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [12]. The PICOS framework (Population, Intervention, Comparator, Outcome and Study design) for the review was formulated as follows:

Population: adult or pediatric patients with acute diarrhea visiting an ED or outpatient clinic;

Intervention: Prognostic scores, scales or algorithms for risk assessment;

Comparator: none or no risk-stratification tool applied; Outcome: intravenous rehydration, hospitalization, or mortality;

Study design: observational or interventional trials with data on tool Predictive performance.

The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO – CRD 42020166242).

    1. Search strategy and study selection

A systematic database search was conducted using Medline, Embase, Cochrane Library, Web of Science and CINAHL from inception to July 22, 2021. The search strategy is provided as a supplementary file and was limited to publications in English and French (Table S1). Additional pub- lications were identified by reviewing the studies cited as references and systematic reviews identified through the search.

After removing duplicates, all abstracts were independently

screened by two reviewers (K.B., S.L., S.B., T.M., G.L) for eligibility using the Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia). Full texts of selected abstracts were also assessed by two independent reviewers (S.L., G.L., C.C., K.B., T.M., S.B.). Included were articles that i) reported on the management of acute diarrhea, ii) concerned the Outpatient setting.(e.g., ED, clinics, primary care prac- tices, telehealth), and iii) presented a scale, score, or algorithm to strat- ify patients based on their risk of complications (e.g., intravenous rehydration, hospitalization, mortality). Conflict assessments were re- solved through discussion. When needed, a third author (S.B.) was solicited to reach a consensus. Excluded were studies that i) presented untested algorithms based on expert consensus, or ii) investigated out- comes other than intravenous rehydration, hospitalization, or mortality. Editorials, case reports, conference abstracts, reviews and practice guidelines on the management of acute infectious diarrhea were also excluded.

    1. Data extraction and quality assessment

Data were extracted from the selected articles by two independent reviewers (C.V-B., L.M-G.) using a standardized form based on the CHARMS (CHecklist for critical Appraisal and data extraction for sys- tematic Reviews of prediction Modelling Studies) [13]. A third opinion was sought (co-author T.M.) when two reviewer opinions were in opposition. The following categories were extracted: author names, journal, year, region where the study was conducted, data collection period, study design, participant description, outcome(s) to be pre- dicted, predictor type and measurement method, missing data treat- ment, model development or validation, and model performance. Study quality was assessed using PROBAST (Prediction model Risk Of

Bias assessment Tool) which rates the methodology and applicability to the review question as “high”, “low” or “unclear” risk of bias, based on a set of questions and a scoring guide [14].

    1. Statistical analysis

A descriptive synthesis was conducted according to the CHARMS checklist. For each study, 2 x 2 tables containing the number of true pos- itives, true negatives, false positives, and false negatives were made ac- cording to the thresholds specified for the scales found. Forest plots (with 95% confidence intervals) of these thresholds, obtained using Review Manager (RevMan) Version 5.4.1, The Cochrane Collaboration, 2020, provided visual representation of the sensitivities and specificities reported in each study. These forest plots, ordered by sensitivity of the studies [15], were visually examined to assess heterogeneity between studies and to determine if a meta-analysis would be performed [16-18].

  1. Results

The literature search retrieved 15,876 articles. After removing 1802 duplicates, 14,074 articles were screened for relevance, of which 218 were retained for full-text assessment. Of these, 212 were excluded, in most cases (70%) because they did not address risk stratification tools or assess the complications of interest. Twelve other articles were found by searching the bibliographies of the included research articles and relevant reviews. All 12 were excluded after assessing the full text. Five studies were included in this review. Fig. 1 shows the flow di- agram for the process of exclusion and inclusion of the publications based on relevance.

    1. Characteristics of the included studies

The characteristics of the five studies included in the present review are summarized in Table 1. All were published between 2008 and 2021. Four were conducted in pediatric tertiary care EDs in North America [19-22], in one case in partnership with a Swiss group [21], whereas the fifth study was conducted in Mexico [23]. All were prospective ob- servational studies, and one was multicentric [21]. Two scales assessing the risk of complications in patients with acute infectious diarrhea were found. Four studies were external validations of the clinical dehydration scale (CDS) whereas one reported the development and internal valida- tion of the EsVida scale and its simplified version [19-23]. These two scales were validated for use by healthcare providers. No tools or scales intended to be self-administering by patients was found.

    1. Participants

The populations examined in the five studies included children only. The age range were one month to five years in three studies [19-21], three months to five years in one study [22] and one to 13 years in the remaining study [23]. All studies were conducted in the ED using the American Academy of Pediatrics definition of acute gastroenteritis as the main inclusion criteria, that is “diarrheal disease of rapid onset, with or without accompanying symptoms and signs such as nausea, vomiting, fever, or abdominal pain” [24]. Two studies also included chil- dren who presented in the ED only with vomiting [20,21], two involved symptoms that lasted fewer than seven days [21,23], and one focused on CDS >= 3 or signs of dehydration lasting fewer than six days with in- travenous rehydration therapy [22]. Cases of diarrhea lasting >10 days, intravenous therapy within the previous 24 h, or an ED visit for the same illness within the previous seven days were all excluded [19,20], as was chronic disease, malnutrition, or any suspected cause of dehydration other than gastroenteritis [19,20,22,23]. History of abdominal surgery, presence of a surgical abdomen, and history of

Image of Fig. 1

Fig. 1. PRISMA flow diagram.

significant head, chest or abdominal trauma within seven days were also excluded from one study [22].

    1. Outcomes

Among the four CDS studies, intravenous rehydration [20] and hospitalization [22] were used as the primary outcome. In two of these, intravenous rehydration and hospitalization were assessed as secondary outcomes [19,21]. In the only study of the EsVida scale, the primary outcome was hospitalization [23]. In all five studies, outcome assessors were blind to the scale results. In none of these studies was the performance of the score as a predictor of mortality evaluated.

    1. The clinical dehydration scale (CDS)

The CDS was originally developed to assess the level of dehydration in children aged one month to three years. It is based on four clinical characteristics: general appearance, appearance of the eyes, hydration of the mucous membranes (e.g., tongue) and the presence of tears (Table 2). Each characteristic is scored as 0, 1, or 2 for a total ranging from 0 to 8 whereby 0 indicates no dehydration, 1 to 4 indicates some dehydration and 5 to 8 indicates moderate to Severe dehydration [25]. Three studies performed external validation of the CDS with pro- spective cohorts of children aged one month to five years evaluated upon arrival in the ED by a triage nurse formally trained in the use of the CDS [19-21]. The hypothesis that the three CDS categories of dehy- dration would be positively associated with the likelihood of requiring intravenous rehydration or hospitalization was tested using univariate

analysis.

In the other study, CDS dehydration level was validated externally as a predictor of hospitalization of children aged three months to five years and having a score of at least 3 or clinical signs of dehydration [22]. The CDS was recorded by a research nurse after the physician had deter- mined that intravenous rehydration was required. CDS performance was analyzed with a cutoff score value of at least 5, which corresponds to a moderate or severe dehydration. The area under the curve (AUC)

for the outcome of hospitalization was 0.65 (95% CI 0.57-0.73) [22]. The results are reported in Table 1.

The performance of the CDS based on different cutoffs is presented as forest plots of sensitivity and specificity for predicting intravenous re- hydration or hospitalization (Fig. 2). A CDS cutoff of 5 or above showed better predictive specificity but lower sensitivity compared to a cutoff of 1 or above.

Substantial heterogeneity was observed among these studies. The sensitivities and specificities of the different CDS cutoffs vary consider- ably, especially for predicting hospitalization. Heterogeneity appeared less pronounced for the intravenous rehydration outcome and a CDS cutoff of 5 or above. Except for the Kinlin and Freedman study, in which children were included if they already had clinical signs of dehy- dration and intravenous therapy was underway [22], no other factor explaining this heterogeneity was identified and no subgroup analysis could be considered. Meta-analysis was therefore deemed inappropri- ate and was not performed [16-18].

    1. The EsVida scale

EsVida is the other prognostic scale used in the studies examined in this systematic review. Originally developed by Fernandez-Garrido et al. without statistical modelling, this scale is based on 13 questions derived from a systematic review and selected by a panel of pediatricians [23]. The characteristics are scored as present (1 point) or absent (0 point) for a total score ranging from 0 to 13. For its initial validation, its perfor- mance at predicting the Need for hospitalization was evaluated for pa- tients aged under 13 in a single pediatric ED. Since this gave inconclusive results, the scale was simplified and refined to focus on fac- tors associated more significantly with hospitalization according to “a stepwise logistic regression analysis” without further specification. Composed of five factors (comorbidities, malnutrition, more than five bowel movements per day, pallor and capillary refill time more than two seconds), the modified EsVida scale (Table 3) ranges from 0 to

42.5 points based on a weighting that favors the factors most signifi- cantly associated with the need for hospitalization [23]. With a cutoff

Table 1

T. Marx, C. Vincent-Boulay, L. Marquis-Gendron et al.

American Journal of Emergency Medicine 64 (2023) 7885

81

Characteristics of the included studies

Authors, year, journal

Setting, country

Data collection period

Study design

Participants

Outcomes of interest

Candidate predictors

Method for measurement

Handling of missing data

Sample size (no. of participants, no. with outcomes)

Model development or model evaluation

Model performance?

Goldman et al.,

Tertiary

January

Prospective

Children, 1

IV rehydration

CDS (physical

Nurse

5,

IV rehydration:

External validation

CDS >= 1 CDS >= 5

(2008)

care

2005 to

observational

month to 5

examination, 4 criteria)

completed the

excluded

206, 62

IV rehydration: IV rehydration:

Pediatrics

pediatric

May 2005

study

years of age

CDS

Univariate analysis

sens: 0.73 (0.61 sens: 0.06 (0.00

ED

Ontario,

assessment after triage

p < 0.001

to 0.84) to 0.13)

Sp: 0.69 (0.62 Sp: 0.99 (0.98

to 0.77) to 1.00)

Canada

NPV: 0.85 (0.79 NPV: 0.71 (0.65

Bailey et al., (2010)

Tertiary

April 2008

Prospective

Children, 1

IV rehydration

CDS (physical

Nurse

No

IV rehydration:

External validation

to 0.92) to 0.77)

PPV: 0.51 (0.40 PPV: 0.80 (0.45

to 0.61) to 1.15)

CDS >= 1 CDS >= 5

Acad Emerg Med

care

to March

observational

month to 5

Hospitalization

examination, 4 criteria)

completed the

missing

150, 41

IV rehydration: IV rehydration:

pediatric

2009

study

years of age

CDS

data

Univariate analysis

sens: 0.88 (0.78 sens: 0.32 (0.17

ED

assessment

after triage

Hospitalization:

150, 4

IV rehydration: p <

to 0.98) to 0.46)

Sp: 0.46 (0.37 Sp: 0.94 (0.89

Canada

0.001

Hospitalization: p =

to 0.56) to 0.98)

NPV: 0.91 (0.84 NPV: 0.78 (0.71

to 0.98) to 0.85)

NA

PPV: 0.38 (0.28 PPV: 0.65 (0.44

to 0.48) to 0.86)

Hospitalization: Hospitalization: Sens: 1 (1.00 to Sens: 0.50 (0.01

Gravel et al.,

Tertiary

June 2009

Prospective

Children 1

IV rehydration

CDS (physical

Nurse

2,

IV rehydration:

External validation

1.00) to 0.99)

Sp: 0.38 (0.30 Sp: 0.88 (0.82

to 0.46) to 0.93)

NPV: 1 (1.00 to NPV: 0.98 (0.96

1.00) to 1.00)

PPV: 0.04 (0.44 PPV: 0.10

to 0.86) (-0.03 to 0.23)

CDS >= 1 CDS >= 5

(2010)

care

to March

observational

month to 5

Hospitalization

examination, 4 criteria)

completed the

excluded

262, 41

IV rehydration: IV rehydration:

Arch Pediatr

pediatric

ED

2010

study

years of age

CDS

assessment

Hospitalization:

Univariate analysis

IV rehydration: p <

sens: 0.85 (0.75 sens: 0.19 (0.08

to 0.95) to 0.30)

Montreal

after triage

262, 14

0.001

Hospitalization:

Sp: 0.38 (0.31 Sp: 0.97 (0.95

to 0.44) to 0.99)

and

Quebec,

p-value = 0.005

NPV: 0.92 (0.86 NPV: 0.84 (0.80

to 0.98) to 0.89)

Canada

Geneva,

PPV: 0.24 (0.17 PPV: 0.60 (0.35

to 0.30) to 0.85)

Switzerland

Hospitalization: Hospitalization: Sens: 0.74 (0.54 Sens: 0.26 (0.06

to 0.93) to 0.46)

Sp: 0.34 (0.28 Sp: 0.96 (0.93

to 0.40) to 0.98)

Kinlin et Freedman

Tertiary

December

Prospective

Children >3

Hospitalization

CDS (physical

Post-treatment

No

Hospitalization:

External validation

NPV: 0.94 (0.89 NPV: 0.94 (0.91

to 0.99) to 0.97)

PPV: 0.08 (0.04 PPV: 0.33 (0.09

to 0.12) to 0.57)

CDS >= 5

(2012)

care

2006 to

observational

months of

examination, 4 criteria)

assessment of

missing

226, 52

AUC: 0.65 (0.57-0.73)

sens: 0.62 (0.48 to 0.75)

Pediatrics

pediatric

ED

April 2010

study

age, with IV

rehydratation

the CDS by

nurse and

data

Sp: 0.66 (0.58 to 0.73)

NPV: 0.85 (0.78 to 0.81)

Toronto,

required

physician

PPV: 0.35 (0.25 to 0.45)

Canada

(continued on next page)

score of 12 or more, the AUC is 0.84 (95% CI 0.75-0.90), compared to

0.69 (95% CI 0.60-0.79) for the EsVida scale with cutoff score of 6 or more [23]. The results are reported in Table 1.

EsVida scale >=5:

sens: 0.75 (0.62 to 0.85)

Sp: 0.57 (0.42 to 0.70)

NPV: 0.66 (0.51 to 0.81)

PPV: 0.68 (0.56 to 0.80)

Modified EsVida scale >=12: sens: 0.70 (0.54 to 0.80)

Sp: 0.70 (0.42 to 0.83)

NPV: 0.66 (0.52 to 0.79)

PPV: 0.74 (0.61 to 0.86)

    1. Risk of bias assessment

Model performance?

The risk of bias (Table 4) was assessed using PROBAST, which as- sesses both the risk of bias and the applicability of prediction model studies, based on four aspects: participants, predictors, outcome, and analysis. All five studies were rated as applicable (relevant). In contrast, the risk of bias was unclear or high for three studies because of the type of analyses performed [19-21], and high in one case since the prediction model was not validated externally [23]. As recommended in the PROBAST scoring method, the risk of bias should be downgraded to high in such cases [14].

Model development or model evaluation

EsVida scale: development (no modelling method) and internal validation

AUC: 0.69

(0.60-0.79)

Person’s test: r =

-0.28

Modified EsVida scale: development (stepwise logistic regression analysis) and internal validation

AUC: 0.84

(0.75-0.90)

Person’s test: r =

-0.21 (-0.4, 0.01)

Abbreviations: ED = emergency department; IV = intravenous; sens = sensitivity; spe = specificity; NPV = negative predictive value; PPV = positive predictive value; AUC = area under the curve.

* Results are reported with 95% confidence intervals.

  1. Discussion

Sample size (no. of participants, no. with outcomes)

Hospitalization: 97, 53

The aim of this systematic review was to identify tools that would predict the care needs (intravenous rehydration, hospitalization) of pa- tients with acute infectious diarrhea, in contrast to other similar reviews that have identified tools that estimate the level of dehydration. Two tools were found: the CDS and the EsVida scale. Of these two, only the CDS has undergone formal external validation for children aged one months to five years. At a cutoff score >= 1, the sensitivity of the CDS ranged from 0.73 to 0.88 for predicting intravenous rehydration and from 0.74 to 1.00 for predicting hospitalization. A cutoff score >= 5 de- creased the sensitivity and increased the false-negative complication rate. However, these results were highly heterogeneous across the stud- ies and did not allow meta-analysis. The literature search strategy did not retrieve any document describing tools that predicted mortality or were intended for use in adults. Both the CDS and the EsVida scale are designed to be completed by healthcare providers.

Candidate predictors

Method for measurement

Handling of missing data

No missing data

EsVida scale (personal history, social management, severe gastroenteritis criteria, physical examination, 13 criteria)

Evaluation by nurse and physician after triage

Modified EsVida scale (history, physical examination, 5 criteria)

Various clinical tools have been developed to estimate the level of dehydration in children. To our knowledge, only the EsVida scale was derived specifically to predict the risk of the complication of interest in this study, namely hospitalization. However, we have found no exter- nal validation of this tool. In addition, the EsVida scale is composed of more questions than the CDS, which may hinder its use in the ED where decision making must be simple and quick. In contrast, tools such as the Gorelick scale, the WHO (World Health Organization) scale and the CDS were developed to assess the dehydration but not the risk of complications per se [25-27]. A tool called the modified Vesikari scale has been developed to predict patient prognosis and healthcare needs (consultation with healthcare providers, fluid rehy- dration or hospitalization) based on a gastroenteritis illness severity score, but for research purposes only, not in the clinic [28]. In a study of the accuracy of the three scales widely use in the ED (the WHO, the Gorelick scale and the CDS) at evaluating children with diarrhea or vomiting, the CDS and the Gorelick scale were found to be better diag- nostic tools for dehydration, with AUCs of respectively 0.72 (0.60, 0.84) and 0.71 (0.57, 0.85) compared to 0.61 (0.45, 0.77) for the WHO

Study design

Participants

Outcomes of interest

Prospective observational study

Children 1

year to 13 years of age

Hospitalization

scale [29].

Data collection period

Uncertain

Guidelines and reviews based on clinical evidence or expert consen- sus recommend the use of a tool to evaluate dehydration of children with acute infectious diarrhea. To be useful in the ED, these tools must be short and quick to complete so as not to add to the workload of the care providers. The CDS is evidence-based, easy to use to assess dehy- dration status and stratify patients according to their risk of complica- tions, and has a moderate to high inter-reliability when used by healthcare providers [30-33]. Based on available evidence, mild dehy- dration appears to be at low risk of complications and may therefore be treated safely at home since oral rehydration fails in only 3.6% of cases in this Patient category [34]. In addition, no significant difference was found between oral and intravenous rehydration in this category in terms of hospitalization or return to the ED [35]. A CDS with a cutoff

Table 1 (continued)

Authors, year, journal

Setting, country

Fernandez-Garrido et al.,

(2021)

Rev. Gastroenterol Mex

Pediatric ED

Mexico City, Mexico

Clinical deshydration scale [25]

Characteristic score

0

1

2

General appearance

Normal

Thirsty, restless or lethargic but irritable when touched

Drowsy, limp, cold, or sweaty, + / – comatose

Eyes

Normal

Slightly sunken

Very sunken

Mucous membranes (tongue)

Moist

Sticky

Dry

Tears

Tears

Decreased tears

Absent tears

A score of 0 represents no dehydration; a score of 1 to 4, some dehydration; and a score of 5 to 8, moderate to severe dehydration.

of 1 therefore may be used to select low-risk patients for whom oral re- hydration would be safe and ancillary tests unnecessary. In the fives

Table 3

Modified EsVida scale [23]

studies examined in the present review, more than one third of the chil-

dren with acute infectious diarrhea managed in the ED had a CDS lower than 1.

Integrating the CDS into ED triage of children with acute infectious diarrhea may represent a research opportunity. The scale could help ex- perienced Triage nurses identify low-risk patients and provide Care pathways that do not necessarily involve an emergency physician. However, since the CDS has been validated only in children under 5 years of age and has not been studied in ED Triage protocols, the safety and effectiveness of such an approach are not supported by the litera- ture at this time. Although the CDS has been shown to be useful in clin- ical practice, it still needs to be used in combination with other criteria such as capillary refill more than two seconds, prolonged skinfold or altered neurologic status [25,26,36] to determine the need for labora- tory tests and intravenous therapy [30]. Further research is therefore needed to improve the performance of the CDS and to evaluate new uses that could improve Healthcare efficiency.

With our search limited to Prognostic tools for predicting complications following acute infectious diarrhea, namely the need for

Factor Score,

if present

Comorbidity (chronic underlying disease reported by the parents with 15 evidence of its medical management; a history of prematurity is noted)

Malnutrition (weight for age and sex in the third percentile) 13

>5 bowel movements within 24 h 6

Pallor 5

Capillary refill time > 2 s 3.5

Total score from 0 to 42.5.

intravenous rehydration and hospitalization, and mortality, the number of documents retrieved exceeded 15,000. Including diagnostic tools de- veloped to estimate patient dehydration would have made the number unmanageable and increased the likelihood of overlooking articles from which we could have extracted data relevant to assessing their prognos- tic performance for our study outcomes. Based on our review of the lit- erature cited by three recent systematic reviews [35,37,38] on this topic,

Image of Fig. 2

Fig. 2. Forest plot showing the sensitivity and specificity of the Clinical Dehydration Scale (CDS) for intravenous dehydration and hospitalization. Definition: CDS = clinical dehydration scale; IV = intravenous; TP = true positive; FP = false positive; FN = false negative; TN = true negative.

Table 4

Prediction model Risk Of Bias ASsessment Tool (PROBAST) results

Study

ROB

Applicability

Overall

Participants

Predictors

Outcome

Analysis

Participants

Predictors

Outcome

ROB

Applicability

Goldman 2008

+

+

+

+

+

+

+

Bailey 2010

+

+

+

?

+

+

+

?

+

Gravel 2010

?

+

+

?

+

+

+

?

+

Kinlin 2012

+

+

+

+

+

+

+

+

+

Fernandez-Garrido 2021

+

+

+

?

+

+

+

+

Definition: ROB = risk of bias.

+ indicates low ROB or low concern regarding applicability;

– indicates high ROB or high concern regarding applicability;

? indicates unclear ROB or unclear concern regarding applicability.

it appears unlikely that we missed such articles. Despite a rigorous and extensive literature search (14,074 abstracts screened), only five arti- cles were included in the review. Targeted outcomes (intravenous rehy- dration and hospitalization) have been little studied in the literature in contrast to the dehydration. Our aim was to identify scales that could di- rectly guide clinical interventions rather than simply assess dehydra- tion. Therefore, very few articles met our inclusion criteria. Moreover, PROBAST indicated a high or unclear risk of bias for more than three quarters of the articles cited by these reviews. Even the results of the five studies examined in the present review should therefore be interpreted with caution. The heterogeneity observed among these studies, due probably to the participant inclusion and exclusion criteria, precluded a valid meta-analysis.

In conclusion, this systematic review identified two scales that have been used to predict the need for intravenous rehydration or hospitaliza- tion in patient presenting to an outpatient clinic or an ED with acute infec- tious diarrhea. The CDS was derived and validated for children aged one month to five years at multiple sites, whereas the EsVida scale was devel- oped and validated for patients younger than 13 years of age in a single small population. Although the CDS is the more studied of these two scales, the quality of evidence supporting its widespread use is low and the risk of bias is high. It may still be used in the ED to support the assess- ment by emergency care providers without replacing clinical judgement, however a better validated scale to predict the risk of complications is needed to assist in ED triage for patients with acute infectious diarrhea. The findings of this review will be useful for guiding the derivation and validation of new decision support tools tailored to the ED setting to en- able clinicians to propose more efficient care pathways to patients, both adults and children, at low risk of complications following acute infectious diarrhea. Further research is needed to develop and validate a risk- stratification tool suitable for adults in the ED.

Financial support

This work was supported by Fonds d’acceleration des collaborations en sante Ministere de l’Economie, de la Science et de l’Innovation du Quebec.

Author contributions

SB, MB and MGB conceived the study and obtained research funding. EPR realized the search strategy and its update. TM, KB, SL, GL, CC, MM and SB screened the abstract and the full text to select the studies included. CVB, LM and TM realized the extraction of the data and the statistical analysis. TM and SB drafted the manuscript. All authors contributed substantially to its revision, approved the final version to be published and written acknowledgment to be accountable for all aspects of the work.

CRediT authorship contribution statement

Tania Marx: Writing – review & editing, Writing – original draft, Validation, Supervision, Methodology, Data curation. Claudia

Vincent-Boulay: Writing – review & editing, Visualization, Validation, Data curation. Laurance Marquis-Gendron: Writing – review & editing, Visualization, Validation, Data curation. Kathryn Bareil: Writing – review & editing, Visualization, Validation, Data curation. Samuel Leduc: Writing – review & editing, Visualization, Validation, Data curation. Gabrielle Lefebvre: Writing – review & editing, Project admin- istration, Data curation. Catherine Cote: Writing – review & editing, Visualization, Validation, Data curation. Myriam Mallet: Writing – review & editing, Project administration, Funding acquisition, Concep- tualization. Emmanuelle Paquette-Raynard: Writing – review & editing, Methodology, Data curation. Maurice Boissinot: Writing – review & editing, Funding acquisition, Conceptualization. Michel G. Bergeron: Writing – review & editing, Funding acquisition, Conceptual- ization. Simon Berthelot: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Methodology, Funding acquisition, Conceptualization.

Declaration of Competing Interest

None.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2022.11.024.

References

  1. Riddle MS, DuPont HL, Connor BA. ACG clinical guideline: diagnosis, treatment, and prevention of acute diarrheal infections in adults. Am J Gastroenterol. 2016;111(5): 602-22.
  2. Scallan E, Griffin PM, Angulo FJ, et al. Foodborne illness acquired in the United States-unspecified agents. Emerg Infect Dis. 2011;17(1):16-22.
  3. Scharff RL. Economic burden from health losses due to foodborne illness in the United States. J Food Prot. 2012;75(1):123-31.
  4. Graves NS. Acute gastroenteritis. Prim Care. 2013;40(3):727-41.
  5. Katz DA, Lynch ME, Littenberg B. Clinical prediction rules to optimize cytotoxin test- ing for Clostridium Difficile in hospitalized patients with diarrhea. Am J Med. 1996; 100(5):487-95.
  6. Depinet H, von Allmen D, Towbin A, et al. Risk stratification to decrease unnecessary diagnostic imaging for acute appendicitis. Pediatrics. 2016;138(3):e20154031.
  7. Govani SM, Guentner AS, Waljee AK, et al. Risk stratification of emergency depart- ment patients with Crohn’s disease could reduce computed tomography use by nearly half. Clin Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Assoc. 2014;12(10):1702-1707.e3.
  8. Dhatt S, Sabhaney V, Bray H, et al. Improving the diagnostic accuracy of appendicitis using a multidisciplinary pathway. J Pediatr Surg. 2020;55(5):889-92.
  9. Shen EX, Lord A, Doecke JD, et al. A validated risk stratification tool for detecting high-risk small bowel Crohn’s disease. Aliment Pharmacol Ther. 2020;51(2): 281-90.
  10. Rosenbloom ST, Daniels TL, Talbot TR, et al. triaging patients at risk of influenza using a patient portal. J Am Med Inform Assoc JAMIA. 2012;19(4):549-54.
  11. Judson TJ, Odisho AY, Neinstein AB, et al. Rapid design and implementation of an in- tegrated patient self-triage and self-scheduling tool for COVID-19. J Am Med Inform Assoc JAMIA. 2020;27(6):860-6.
  12. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.
  13. Moons KGM, de Groot JAH, Bouwmeester W, et al. Critical appraisal and data extrac- tion for systematic reviews of prediction modelling studies: the CHARMS checklist. PLoS Med. 2014;11(10):e1001744.
  14. Wolff RF, Moons KGM, Riley RD, et al. PROBAST: a tool to assess the risk of bias and applicability of prediction model studies. Ann Intern Med. 2019;170(1):51-8.
  15. Schriger DL, Altman DG, Vetter JA, et al. Forest plots in reports of systematic reviews: a cross-sectional study reviewing current practice. Int J Epidemiol. 2010;39(2): 421-9.
  16. Kim KW, Lee J, Choi SH, et al. Systematic review and meta-analysis of studies evaluating Diagnostic test accuracy: a practical review for clinical researchers-part I. general guidance and tips. Korean J Radiol. 2015;16(6):1175-87.
  17. Campbell J, Klugar M, Ding S, et al. The systematic review of studies of diagnostic test accuracy. Joanna Briggs Inst Rev Man. 2015:1-46.
  18. Macaskill P, Gatsonis C, Deeks J, et al. Chapter 10: Analysing and presenting results. Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy; 2022 Version 10. 210apr. J.-C. (The Cochrane Collaboration).
  19. Bailey B, Gravel J, Goldman RD, et al. External validation of the clinical dehydration scale for children with acute gastroenteritis. Acad Emerg Med Off J Soc Acad Emerg Med. 2010;17(6):583-8.
  20. Goldman RD, Friedman JN, Parkin PC. Validation of the clinical dehydration scale for children with acute gastroenteritis. Pediatrics. 2008;122(3):545-9.
  21. Gravel J, Manzano S, Guimont C, et al. Multicenter validation of the clinical dehydra- tion scale for children. Arch Pediatr Organe Off Soc Fr Pediatr. 2010;17(12):1645-51.
  22. Kinlin LM, Freedman SB. Evaluation of a clinical dehydration scale in children requir- ing intravenous rehydration. Pediatrics. 2012;129(5):e1211-1219.
  23. Fernandez-Garrido CL, Bernardez-Zapata I, Iglesias-Leboreiro J, et al. EsVida scale de- velopment and validation to assess the requirement of emergency management in cases of infectious gastroenteritis. Rev Gastroenterol Mex Engl. 2021;86(2):145-52.
  24. American Academy of Pediatrics, Provisional Committee on Quality Improvement, Subcommittee on Acute Gastroenteritis. Practice parameter: the management of acute gastroenteritis in young children. Pediatrics. 1996;97(3):424-35.
  25. Friedman JN, Goldman RD, Srivastava R, et al. Development of a clinical dehydration scale for use in children between 1 and 36 months of age. J Pediatr. 2004;145(2): 201-7.
  26. Gorelick MH, Shaw KN, Murphy KO. Validity and reliability of clinical signs in the diagnosis of dehydration in children. Pediatrics. 1997;99(5):E6.
  27. World Health Organization. The Treatment of Diarrhoea: A Manual for Physicians and Other Senior Health Workers. 4th ed.. Geneva Switz World Health Organ; 2005.
  28. Freedman SB, Eltorky M, Gorelick M, Group PERCGS. Evaluation of a gastroenteritis severity score for use in outpatient settings. Pediatrics. 2010;125(6):e1278-85.
  29. Jauregui J, Nelson D, Choo E, et al. External validation and comparison of three pedi- atric clinical dehydration scales. PLoS One. 2014;9(5):e95739.
  30. Guarino A, Ashkenazi S, Gendrel D, et al. European Society for Pediatric Gastroenter- ology, Hepatology, and Nutrition/European Society for Pediatric Infectious Diseases Evidence-based guidelines for the management of acute gastroenteritis in children in Europe: update 2014. J Pediatr Gastroenterol Nutr. 2014;59(1):132-52.
  31. Carson RA, Mudd SS, Madati PJ. Clinical practice guideline for the treatment of pedi- atric acute gastroenteritis in the outpatient setting. J Pediatr Health Care Off Publ Natl Assoc Pediatr Nurse Assoc Pract. 2016;30(6):610-6.
  32. Lo Vecchio A, Vandenplas Y, Benninga M, et al. An international consensus report on a new algorithm for the management of infant diarrhoea. Acta Paediatr Oslo Nor 1992. 2016;105(8):e384-389.
  33. Hartman S, Brown E, Loomis E, et al. Gastroenteritis in Children. Am Fam Physician. 2019;99(3):159-65.
  34. Gavin N, Merrick N, Davidson B. Efficacy of glucose-based oral rehydration therapy. Pediatrics. 1996;98(1):45-51.
  35. Freedman SB, Pasichnyk D, Black KJL, et al. Gastroenteritis therapies in developed countries: systematic review and meta-analysis. PLoS One. 2015;10(6):e0128754.
  36. Duggan C, Refat M, Hashem M, et al. How valid are clinical signs of dehydration in infants? J Pediatr Gastroenterol Nutr. 1996;22(1):56-61.
  37. Falszewska A, Szajewska H, Dziechciarz P. Diagnostic accuracy of three clinical dehy- dration scales: a systematic review. Arch Dis Child. 2018;103(4):383-8.
  38. Steiner MJ, DeWalt DA, Byerley JS. Is this child dehydrated? JAMA. 2004;291(22): 2746-54.