RESEARCH PAPER

Diagnosis and Assessment of Severity of Pediatric Pneumonia Using the

Respiratory Index of Severity (RISC) Scoring System
KALYANI PILLAI,1 EDWIN ROS SARTHO,1 TP LAKSHMI,2 V K PARVATHY1
From Departments of 1Pediatrics and 2Radiodiagnosis, Amala Institute of Medical Sciences, Thrissur, Kerala.

Correspondence to: Objectives: To evaluate the correlation between the Respiratory Index of Severity (RISC)
Dr Kalyani Pillai, scoring system and the World Health Organization (WHO) interpretation of chest
Professor of Pediatrics, radiographs, and to evaluate the RISC scoring system in the diagnosis and assessment of
severity of pneumonia in children against chest X-ray. Methods: 241 children presenting to
Amala Institute of Medical Sciences,
a tertiary care center with a clinical diagnosis of pneumonia that necessitated a chest
Amala Nagar, Thrissur 680 555, radiograph to be taken, were enrolled. The RISC scoring was done for all participants by a
Kerala. [email protected] single pediatrician and chest X-ray graded by a single radiologist as per the WHO
Received: August 29, 2019; interpretation of chest radiographs. Results: There was a statistically significant positive
Initial review; February 01, 2020; correlation (P=0.02) between the two scores. RISC score of >1 had a sensitivity of 80.3%,
Accepted: August 09, 2020. and score ≥3 had a specificity of 88.3%, positive predictive value of 61.3% and negative
predictive value of 76.8% for diagnosis and to predict severity of pneumonia. Conclusion:
In a resource-limited setting, RISC scoring can be used to diagnose and predict the severity
of childhood pneumonia.
Keywords: Community-acquired pneumonia, Severity, X-ray chest.

T
he Respiratory Index of Severity (RISC) scoring 2018. The study was approved by the institutional ethics
was formulated in South Africa to provide a committee and informed consent from parents was taken
standardized means of quantifying severity of prior to enrollment. Consecutive patients aged 1 month to 12
respiratory illness among children during years, who visited the pediatric OPD and/or got admitted in
epidemiologic studies [1]. This score has a negative score the pediatric ward with a clinical diagnosis of pneumonia by
for wheezing, which potentially excludes children with the treating doctor, and which necessitated a chest
wheezing, a probable drawback of the Integrated radiograph to be taken, were enrolled for the study. Exclusion
Management of Childhood Illness (IMCI) developed by criteria were subjects with documented congenital heart
the World Health Organization (WHO). disease, chronic respiratory illness, congenital lung
problems, immunosuppressive conditions or known to have
The gold standard for diagnosis of pneumonia is the
neuromuscular disorders with respiratory system
chest X-ray (CXR). In an attempt to improve the consistency
involvement.
of CXR interpretation between epidemio-logical studies, the
WHO standardized the interpretation of CXRs for the For calculation of sample size, prevalence of lower acute
diagnosis of pneumonia in children [2,3]. The CXR respiratory infection was taken as 28% [5]. At significance
interpretation exercise was done and validated under WHO level of 5%, with 20% relative precision, and power of the
[4]. If the RISC scoring system correlates well with CXR, this study 80%, sample size was calculated as 240. To allow for
scoring system could be of use in resource-poor settings potential recruitment failure rate, 285 children were assessed
where CXR is not routinely available. The aim of the study for eligibility for enrollment in the study.
was to evaluate the correlation between the RISC scoring
A detailed history was taken and respiratory system
system and the WHO interpretation of chest radiographs. It
examination done for all subjects, and RISC scoring for HIV
also aims to evaluate the RISC scoring system in the
non-infected children [1] was done by the principal
diagnosis and assessment of severity of pneumonia in
investigator. Variables in the the RISC score represent known
children by validating it against the CXR.
risk factors for severe outcomes of pneumonia in children,
METHODS with a maximum score of 8 points.
This descriptive study was undertaken at a semi-urban CXR scoring was done by a single radiologist, as per the
tertiary care hospital in India from 1 February, 2018 to 31 July, WHO interpretation of chest radiographs [2,4] .

INDIAN PEDIATRICS 1052 VOLUME 58__NOVEMBER 15, 2021

PILLAI, ET AL. DIAGNOSIS AND ASSESSMENT OF SEVERITY OF PEDIATRIC PNEUMONIA

Statistical analysis: Analysis was done using Spearman rho Table I Clinical Profile, Respiratory Index of Severity
(ñ) correlation to assess the correlation between RISC score (RISC) score, and Chest X-ray Score of Children With
and WHO CXR score. We divided the children into two Physician-Diagnosed Pneumonia (N=241)
groups by using the WHO CXR score of 3 (lobar pneumonia Characteristics No (%)
on X-ray) as a cut off. Group A with CXR score <3 were
Male 142 (58.9
clinically suspected by physician but CXR was not
suggestive of pneumonia, and group B with CXR score ≥3, Age
clinically suspected by physician and also CXR suggestive 1mo-1 y 96 (39.8)
of pneumonia. By plotting the receiver operating 1y-5y 112 (46.5)
characteristic (ROC) curve, we calculated specificity and 5-12 y 33 (13.7)
sensitivity to determine the appropriate RISC score to Symptoms
diagnose and predict the severity of pneumonia. Analysis Fever 205 (85)
was done by Statistical Software Package for Social Sciences URI symptoms 223 (92)
Version 23 (SPSS 23) Breathlessness 40 (16.5)
Feed refusal 219 (90.8)
RESULTS Immunization as per NIS 239 (99.2)
A total of 285 children with physician-diagnosed pneumonia Disease severity based on treatment
were assessed for eligibility to be enrolled in the study (Fig. Outpatient treatment 90 (37.3)
1). Out of these, 20 were excluded based on exclusion criteria Admission as inpatient 140 (58.1)
and 18 were excluded as a CXR was not taken/available. A Pediatric intensive care unit admission 11 (4.6)
total of 241 children were finally included in the study. CXR score
1 141(58.5)
Baseline clinical characteristics are presented in Table I. 2 39 (16.1)
Of the study subjects 34 (14.1%) had a RISC score ≥3 and 61 3 42 (17.1)
(25.3%) had a CXR score ≥3. 4 16 (6.5)
5 3 (1.2)
There was a statistically significant positive correlation
between RISC score and WHO CXR score (r=0.144, RISC score
P=0.025). The ROC curve was plotted (Fig. 2). Taking WHO 1 59 (24.6)
CXR score of ≥3 (lobar pneumonia score 3) as severe 2 148 (61.2)
pneumonia, a RISC score of >1 had a sensitivity of 80.3% and 3 6 (2.5)
4 26 (10.9)
specificity of 26.1%, suggesting it is a moderate screening
≥5 2 (0.8)
test for pneumonia. A RISC score of ≥3 had a sensitivity of
21.3% and specificity of 88.3% for diagnosis and to predict URI-Upper respiratory tract infection.
severity of pneumonia (Table II).
DISCUSSION
There are hardly any scoring systems which are validated for
Physician diagnosed pneumonia community-acquired pneumonia (CAP) in Indian children
assessed for eligibility
that would have high utility in resource-limited settings. We
↓ demonstrated a statistically signi-ficant correlation between
Excluded, n= 20 n=285 RISC scoring system [1] and the WHO interpretation of
* More than 1 episode of wheeze, n=15 chest radiographs [2].
* Nephrotic syndrome on steroids, n=3
* Congenital heart disease, n=2 Limitations of the study were that RISC score is not
↓ useful to differentiate between types of pneumonia based on
n=265
etiology. RISC score is not useful to monitor resolution of
No chest X-ray symptoms. A negative score for wheeze (-2) may encourage
done, n=18 ↓ false negatives. When developing detection tests, a balance
n=247 must be chosen between risks of false negatives and false
RISC scoring not fully done ↓ positives.
as child not co-operative, n=6 n=241
The major scoring systems designed to predict mortality
risk and serve as guides for admission in community-
Fig. 1 Study flow chart. acquired pneumonia (CAP) are for adults [6], and such scores

INDIAN PEDIATRICS 1053 VOLUME 58__NOVEMBER 15, 2021

PILLAI, ET AL. DIAGNOSIS AND ASSESSMENT OF SEVERITY OF PEDIATRIC PNEUMONIA

WHAT THIS STUDY ADDS?

• In a resource-limited setting, where chest X-ray is unavailable, RISC scoring can be used for diagnosing and
predicting the severity of community-acquired pneumonia.

Table II Test Characteristics of the Respiratory Index of

Severity (RISC) Score (N=241)
Characteristicsa RISC threshold RISC threshold
score >1 score ≥3
Sensitivity (%) 80.3 21.3
Specificity (%) 26.11 88.3
PPV (%) 25.9 61.7
NPV (%) 79.6 76.8
Accuracy (%) 39.8 71.3
aUsing CXR WHO score 3 as diagnosis of lobar pneumonia.

sensitivity of 59% and specificity of 78%. With a score

threshold of 4, the sensitivity was 32.6% and specificity was
93.1%. A median RISC score of 1 corresponded with a risk of
mortality of 0% (95%CI: 0%-0.6%) in the study by Reed, et al.
[1]. In our study, using a RISC score of >1 had a higher
sensitivity and was a moderate screening test to predict
pneumonia.
Validation of Bacterial pneumonia score (BPS) score was
Fig. 2 ROC curve for Respiratory Index of Severity (RISC) done with multiplex PCR examinations of blood specimens by
scoring system in pediatric pneumonia. Imilda, et al. [14]. BPS had 69% sensitivity and 60%
specificity, 42% positive predictive value, and 81% negative
predictive value. In our study, which was validated with X-
for children are scarce [7,8]. Only the modified ray chest, a RISC score of ≥3 had a higher specificity and
predisposition, insult, response and organ dysfunction positive predictive value and almost similar negative
(PIRO) score has so far been applied to children with CAP [6]. predictive value for diagnosis and to predict severity of
We used the RISC score for our study as it has six predictors pneumonia.
for assessing the severity of pneumonia (hypoxia, chest
indrawing, feed refusal, wheeze, malnutrition, age) and is To conclude, we propose that RISC score may be used for
easy to administer. Other risk models [9,10] are relatively diagnosing pneumonia in resource-constrained areas to
cumbersome to administer. supplement IMCI protocols. Assessment of a combination
of the two in community-based studies may provide
The WHO introduced the Integrated Management of additional information on this aspect.
Childhood Illness (IMCI) to standardize and improve
treatment and thus prevent major causes of death in children Ethics clearance: Amala Institute of Medical Sciences; No.
AIMSIEC/05/2018 dated January 22, 2018.
<5 years [10]. The first drawback of IMCI diagnostic criteria
Contributors: KP: concept and design of the study; ERS: analyzed
is of over-diagnosing pneumonia by including children with and collected data; drafted the manuscript; TPL: analyzing data;
wheezing. In our study too, 36.5% of patients had a wheeze VKP: supervised cognitive and behavioral assessments. All authors
on presentation. However, the RISC scoring system approved the final version of manuscript, and are accountable for
overcomes this by giving a negative score for wheezing. The all aspects related to the study.
second drawback of IMCI is a missed diagnosis, as it has a Funding: None; Competing interest: None stated.
low sensitivity [11,12]. REFERENCES
Hooli, et al. [13] did an external validation of RISC by 1. Reed C, Madhi SA, Klugman KP, et al. Development of the
calculated classification performance measures at respiratory index of severity in children (RISC) score among
thresholds of 3 and 4. Using a RISC score of 3 had a young children with respiratory infections in South Africa.

INDIAN PEDIATRICS 1054 VOLUME 58__NOVEMBER 15, 2021

PILLAI, ET AL. DIAGNOSIS AND ASSESSMENT OF SEVERITY OF PEDIATRIC PNEUMONIA

PLoS One. 2012;7:e27793. validation of a clinical prediction rule to distinguish bacterial

2. Cherian T, Mulholland EK, Carlin JB, et al. Standardized from viral pneumonia in children. Pediatr Pulmonol.
interpretation of paediatric chest radiographs for the diagnosis 2006;41:331-7. Erratum in: Pediatr Pulmonol. 2006;41:494.
of pneumonia in epidemiological studies. Bull WHO. 9. Williams DJ, Zhu Y, Grijalva CG, et al. Predicting severe
2005;83:353-9. pneumonia outcomes in children. Pediatrics. 2016;138:
3. O’ Grady KA, Torzillo PJ, Ruben AR, et al. Identification of e20161019.
radiological alveolar pneumonia in children with high rates of 10. Emukule GO, McMorrow M, Ulloa C, et al. Predicting
hospitalized respiratory infections: comparison of WHO - mortality among hospitalized children with respiratory illness
defined and pediatric pulmonologist diagnosis in the clinical in Western Kenya, 2009-2012. PLoS ONE. 2014; 9:e92968.
context. Pediatr Pulmonol. 2012;47:386-92. 11. Deodari AK, Chellani H, Carlin JB, et al. Clinico-
4. Taylor E, Haven K, Reed P; SHIVERS Investigation Team. A epidemiological profile and predictors of severe illness in
chest radiograph scoring system in patients with severe acute young infants (<60 days) reporting to a hospital in North India.
respiratory infection: A validation study. BMC Med Imaging. Indian Pediatr. 2007;44:739-48.
2015;15:61. 12. Uwemedimo OT, Lewis TP, Essien EA, et al. Distribution and
5. Prietsch Silvio OM, Fischer Gilberto B, César Juraci A, et al. determinants of pneumonia diagnosis using Integrated
Acute lower respiratory illness in under-five children in Rio Management of Childhood Illness guidelines: A nationally
Grande, Rio Grande do Sul State, Brazil: Prevalence and risk representative study in Malawi. BMJ Glob Health.
factors. Cad Saúde Pública. 2008; 24:1429-438. 2018;3:e000506.
6. Uwaezuoke SN, Ayuk AC. Prognostic scores and bio-markers 13. Hooli S, Colbourn T, Lufesi N, et al. Predicting hospitalised
for pediatric community - acquired pneumonia: How far have paediatric pneumonia mortality risk: An external validation of
we come? Pediatric Health Med Ther. 2017; 8:9-18. risc and mrisc, and local tool development (RISCMalawi) from
7. Arbo A, Lovera D, Martínez – Cuellar C. Mortality predictive Malawi. PLoS One. 2016;11:e0168126.
scores for community – acquired pneumonia in children. Curr 14. Imilda I, Yani FF, Hariyanto D, Basir D. Bacterial pneumonia
Infect Dis Rep. 2019;21:10. score to identify bacterial pneumonia, Paediatr Indones.
8. Moreno L, Krishnan JA, Duran P, Ferrero F. Development and 2015;55:79-82.

INDIAN PEDIATRICS 1055 VOLUME 58__NOVEMBER 15, 2021