Table 1.

Characteristics of studies of pediatric early warning systems in resource limited settings.

References	Country	Setting	No. of subjects	Age range	System	Study design	Purpose	Trigger & response	Key results
Olson et al. (12)	Malawi	Large referral hospital with high dependency unit	54 Cases; 161 Controls	< 15 years old	Inpatient Triage, Assessment and Treatment (ITAT)	Prospective nested case-control study	Develop ITAT score to identify clinical deterioration in inpatients and prompt physician evaluation	ITAT score >3 triggers physician assessment	ITAT score >3: OR for death 4.8, predicts mortality with AUROC 0.76
Olson et al. (14)	Malawi	Large referral hospital with high dependency unit	1,642 Assessments	< 15 years old	Inpatient Triage, Assessment and Treatment (ITAT)	Prospective quality improvement project	Improve pediatric inpatient surveillance by provision of vital sign equipment implementation of vital sign assistant program	ITAT score >3 triggers physician assessment	Vital signs assistants associated with increased frequency and accuracy of ITAT scores & increased clinician notiifcations; Mortality reduced from 9.3% to 5.7%
Chaiyakusil and Pandee (15)	Thailand	Emergency Department in tertiary university hospital with ICU	1,136	< 15 years old	Pediatric Early Warning Score (PEWS)	Prospective descriptive study	Validation of PEWS on arrival in Emergency Department to predict inpatient admission disposition	None	Predicts Admission: AUROC 0.73, PEWS 1 or above, Sensitivity 78%, Specificity 59.6%; Predicts ICU admission: AUC 0.98, PEWS 3 or above, Sensitivity 100%; Specificity 90.5%
George et al. (16)	Uganda, Tanzania & Kenya	Mix of 6 large and small hospitals	3,125*; 1,053**; 5,098**	2 months - 12 years	Bedside PEWS	Retrospective data used to develop and validate PET score	Develop an easy to use, bedside score for risk of mortality based on clinical signs	None	PET score associated with mortality w/in 48 hrs w AUROC 0.77–0.86; Predicts mortality better than PEWS, PRISMIII & Aquamat
Miranda et al. (17)	Brazil	Pediatric referral hospital without ICU	271	0–10 years	Brighton PEWS (BPEWS)	Integrative Review	Literature review, translation and adaptation of BPEWS for Brazilian Portuguese and Pilot test	None	BPEWS was successfully translated and adapted for Brazilian Portuguese and in the pilot study, 26.6% of children were considered at risk of clinical deterioration based on PEWS score >2
Miranda et al. (18)	Brazil	Pediatric referral hospital without ICU	271	0–10 years	Brighton PEWS adapted for Brazilian Portuguese (BPEWS-Br)	Prospective evaluation of diagnostic acuracy	Evaluation of diagnostic accuracy of BPEWS-Br compared to provder assessment of clinical deterioration	None	BPEWS-Br associated with clinical signs of deterioration, AUROC 0.919; BPEWS-Br score >2, sensitivity 73.9% and specificity 95.5%
Agulnik et al. (13)	Guatemala	Pediatric oncology hospital with ICU	129 Cases; 129 Controls	< 18 years	Pediatric Early Warning Score (PEWS)	Retrospective matched case-control study	Validation of the ability of PEWS to predict the need for unplanned ICU transfer	PEWS algorithm followed, high PEWS prompted physician and/or ICU evaluation	PEWS correlated with unplanned ICU transfer: AUROC 0.94; PEWS >3, sensitivity 88% and specificity 93%; Higher PEWS prior to PICU transfer associated with increased morbidity & mortality, higher PIM2 and increased support requirements in ICU
Agulnik et al. (19)	Guatemala	Pediatric oncology hospital with ICU	287	< 18 years	Pediatric Early Warning Score (PEWS)	Retrospective cohort study the year before and year after implementation of PEWS.	Describe the effect of implementation of a PEWS system on the frequency of clinical deterioration events	PEWS algorithm followed, high PEWS prompted physician and/or ICU evaluation	Abnormal PEWS in 93% of unplanned PICU transfer; Reduced frequency of clinical deterioration events, severe sepsis, septic shock and organ dysfunction; Reduced ICU utilization; No change in mortality

^*Development data set;

^**Validation data set.