Sepsis is a life-threatening clinical state that is caused by an exaggerated host response to an infection that causes organ dysfunction with a significant risk of death.[1] Despite the progress made in antimicrobial treatment, critical care, and organ support, sepsis is still a major global health concern.[2] Sepsis is estimated to impact about 49 million people annually worldwide and contributes to around 11 million deaths each year, which is about 18% of all deaths worldwide.[3] This burden is especially high in LMICs where the lack of recognition, critical-care capacity, antimicrobial resistance, and late presentation can all be factors in poor outcomes.[4] Therefore, early identification of patients who are likely to deteriorate is a key part of good sepsis management.[5].

This is especially critical given the fact that patients with sepsis can have relatively mild physiological abnormalities early on, which then quickly deteriorate into shock, multiorgan dysfunction, and mortality.[6] Several clinical scoring systems have thus been developed for early recognition and for the assessment of prognosis. The quick Sequential Organ Failure Assessment (qSOFA) was added to the Sepsis-3 consensus and consists of three easy-to-obtain clinical parameters: respiratory rate ≥22/min, altered mentation, and systolic blood pressure ≤100 mmHg.[7] A qSOFA ≥2 is linked to an increased risk for poor outcomes such as prolonged intensive care unit use and mortality.[8] However, although qSOFA is simple and highly specific, its relatively low sensitivity limits its ability to identify all patients at risk of deterioration.[9] A major meta-analysis of 62,338 patients showed a pooled sensitivity of only 46% with a specificity of 82% for qSOFA predicting mortality.[10] .

A more comprehensive assessment of physiological deterioration is available using the National Early Warning Score 2 (NEWS2). It includes the oxygen saturation, respiratory rate, level of consciousness, pulse rate, systolic blood pressure, temperature, and supplemental oxygen requirement. NEWS2 was not created for sepsis specifically, but was developed as a general deteriorating and escalating scale. However, there is growing evidence that it may offer more sensitivity to identify patients with sepsis who are likely to have adverse outcomes. The study in the emergency departments showed that NEWS2 had an area under the receiver operating characteristic curve (AUROC) of 0.80 for recognising sepsis with organ dysfunction compared with 0.71 for qSOFA.[11] Likewise, a recent multi-centre study of 3274 patients with suspected infection from 20 emergency departments across the Middle East identified similar discrimination between NEWS2 and qSOFA for 30-day mortality; however, NEWS2 ≥5 had significantly higher sensitivity (73.5% vs. 42.2%) than qSOFA ≥2.[12] .

Importantly, the relative effectiveness of these scores varies, not completely, across all populations and clinical settings. A recent prospective cohort study of 874 patients with sepsis indicated that NEWS2 was more sensitive for predicting 7-day and 28-day mortality than was qSOFA.[13] Other recent data have shown similar or even better discrimination using qSOFA in some populations, indicating that the discrimination of each of these scores may vary depending on disease severity, patient characteristics, timing of the assessment, and clinical setting.[14] The 2026 guidelines for Surviving Sepsis Campaign therefore highlight that a wider, more comprehensive early warning score for sepsis, like NEWS/NEWS2, is preferred for early recognition of sepsis, and that no single score has optimal sensitivity and specificity.[5]

Although both scores are widely used, a key clinical question remains: when a patient with suspected or established sepsis presents to the bedside, which score will give the more reliable indication of impending death, NEWS2, which takes a broader view of physiology, or qSOFA, which focuses on organ dysfunction.  This separation is particularly important in resource-limited healthcare systems, where quick, cost-effective and repeatable predictive tools can aid in early escalation and prioritisation of patient treatment. Thus, the current study aimed to comparatively analyze and compare the sensitivity, specificity, positive and negative predictive values, and area under the receiver operating characteristic curve (AUROC) of NEWS2 and qSOFA in predicting sepsis mortality. The results could be used to decide which bedside score provides the best mortality risk stratification and which score can be more readily used in the early clinical process of decision-making in patients with sepsis.

A Prospective Observational Study of Diagnostic Accuracy was done for 6 months at King Edward Medical University, Mayo hospital Lahore, from 1st December 2025 to 31st May 2026. The patients who met the inclusion criteria were consecutively enrolled throughout the study period until hospital discharge or death. In-hospital mortality was considered the primary outcome. The sample size was determined by using OpenEpi version 3.01, based on the previously reported sensitivity of 83.21% of NEWS2 in predicting mortality in patients with sepsis in a prospective observational study from India. In that study, the overall mortality rate was 35.12%, and NEWS2 was found to have an area under the receiver operating characteristic curve (AUC) of 0.781, while qSOFA had an AUC of 0.729.[15] A 95% confidence level, a 10% relative precision, and an expected sensitivity of 83.21% were used to calculate the minimum sample size as about 108 participants. A consecutive sampling technique was used. Patients defined as sepsis-3 (adult ≥18 years) with suspected or documented infection and a significant rise in SOFA score ≥2 points were included. Patients who presented at the Emergency Department and/or were admitted to the Medical Intensive Care Unit were eligible. Informed consent of the patient or their legally authorized representative was obtained. Patients under 18 years, pregnant women, those with incomplete clinical data so that NEWS2 or qSOFA could not be calculated, and those with a history of prolonged resuscitation or critical care treatment before the initial score could be calculated were excluded. Major trauma, burns, and other acute pathologies potentially impacting physiological parameters in isolation from sepsis were also excluded. Patients who did not give consent and those whose legally authorized representatives did not give consent were not enrolled. Data were collected on a structured data collection proforma regarding demographic data such as age and sex, relevant comorbidities, suspected/confirmed source of infection, presenting clinical features, vital signs, oxygen requirement, mental status, and relevant laboratory investigations. The clinical parameters used to compute NEWS2 and qSOFA were gathered prior to the outcome. NEWS2 score was derived from RR, O2 sat, need for supplemental oxygen, systolic blood pressure, pulse rate, temperature, and level of consciousness. The qSOFA score was derived by assigning 1 point for respiratory rate ≥22 breaths per minute, systolic blood pressure ≤100 mmHg and altered mental status. All scores were computed separately at clinical evaluation. Patients were classified as being at elevated risk for adverse outcome using a NEWS2 score ≥5 and a qSOFA score ≥2. Patients were thereafter followed during the period of their hospital stay. In-hospital mortality was the primary outcome. Patients who survived were followed until discharge and patients who died were recorded as mortality events. To minimize incorporation and observer bias, the investigators who computed the scores recorded them prior to knowing the eventual outcome. IBM SPSS Statistics version 26.0 was used to enter and analyze data. Continuous variables were tested for normality by the Shapiro- Wilk test and were shown as mean ± SD for normally distributed variables and median (IQR) for non-normally distributed variables. Categorical variables were reported as frequencies and percentages. The independent-samples t-test or Mann–Whitney U test was used to compare the baseline characteristics of survivors and non-survivors. The chi-square test was used for categorical variables. Diagnostic performance of NEWS2 and qSOFA in predicting IHM was calculated by using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic accuracy with 95% confidence intervals (CIs). Receiver Operating Characteristic (ROC) curves were generated for both scores, and the Area Under the ROC curve (AUROC) was determined to assess overall discriminatory ability. DeLong test for paired ROC curves was used to compare the AUROCs of NEWS2 and qSOFA. The optimal cut-off values were determined using the Youden index. A two-sided p-value <0.05 was considered statistically significant.

A total of 108 patients with sepsis were analyzed. The mean age was 57.8 ± 15.6 years, with 65 (60.2%) males and 43 (39.8%) females. The most common comorbidities were diabetes mellitus and hypertension, and respiratory infection was the most common source of infection (38.9%), followed by urinary tract infection (25.9%). The mean NEWS2 score was 6.1 ± 3.2, whereas the median qSOFA score was 1.4 ± 1.0. (Table 1)

Non-survivors were significantly older, had higher respiratory and pulse rates, lower SBP, saturation, and a higher proportion requiring supplemental oxygen and a change in mental status when compared to survivors. The non-survivors had significantly higher NEWS2 scores (p<0.001) and qSOFA scores (p<0.001). There were no significant differences found for sex, major comorbidities, source of infection, temperature, or hospital stay. (Table 2)

In non-survivors, NEWS2 ≥5 was found in 31 cases (81.6%), while in survivors, the number of cases was 25 (35.7%) (p<0.001). Likewise, the prevalence of qSOFA ≥2 for survivors compared to non-survivors was significantly higher (p<0.001). (Table 3)

NEWS2 ≥5 had a sensitivity of 81.6% and a negative predictive value of 86.5%, while qSOFA ≥2 had a lower sensitivity (57.9%) but higher specificity (82.9%) and positive predictive value (64.7%) for predicting in-hospital mortality. NEWS2 and qSOFA had a diagnostic accuracy of 70.4% and 74.1%, respectively. (Table 4)

ROC curve analysis showed that NEWS2 had good discriminatory performance for mortality, with an AUROC of 0.823 (95% CI: 0.743–0.903), while qSOFA demonstrated an AUROC of 0.746 (95% CI: 0.649–0.843). The best cut-off point for NEWS2 was 5.5 (sensitivity 81.6%) and for qSOFA was 1.5 (sensitivity 71.1%). (Table 5)

The AUROC values for predicting in-hospital mortality were 0.823 for NEWS2 and 0.746 for qSOFA, with a significant difference (DeLong p=0.029). (Table 6)

Table 1. Demographic and clinical characteristics of study participants (n=108)

Table 2. Comparison of demographic and clinical characteristics between survivors and non-survivors

Table 3. Distribution of NEWS2 and qSOFA categories according to in-hospital mortality

Table 4. Diagnostic performance of NEWS2 and qSOFA for predicting in-hospital mortality

Table 5. Receiver operating characteristic analysis of NEWS2 and qSOFA for prediction of in-hospital mortality

Table 6. Comparison of AUROCs of NEWS2 and qSOFA

The present study showed that the overall prognostic discrimination of NEWS2 was superior in predicting in-hospital mortality in septic patients when compared to qSOFA. NEWS2 achieved an AUROC of 0.823 compared with 0.746 for qSOFA (p=0.029). The specificity of NEWS2 was also higher compared to qSOFA (82.9% vs. 64.3%), while the negative predictive value of NEWS2 was also higher (86.5% vs. 78.4%). The results indicate that NEWS2 might be better suited as a tool for initial mortality-risk stratification, and qSOFA as a tool with more specificity. Our results were similar to the prospective study by Verma et al (2023) which enrolled 373 patients attending an Indian emergency department with suspected sepsis. NEWS2 had a higher AUROC than qSOFA (0.781 vs. 0.729; p<0.001), with NEWS2 demonstrating 83.21% sensitivity compared with 77.10% for qSOFA. The investigators found that NEWS2 was better than qSOFA at predicting in-hospital mortality. Their AUROC results compared closely with our results, further supporting the evidence that NEWS2 is superior in discriminating between emergency patients with sepsis.[15] A similar prospective study conducted in Turkey in 2021 included 463 patients in the ED and compared NEWS and qSOFA with SIRS. NEWS had an AUROC of 0.772 for mortality prediction, while qSOFA had an AUROC of 0.758. The sensitivity of NEWS was superior to that of qSOFA (77% vs. 39%), and the specificity of qSOFA was superior to that of NEWS (91% vs. 64%), but the differences were not great. This is similar to our findings, where greater sensitivity was seen with NEWS-based scoring compared to greater specificity with qSOFA.[16] This is supported by a 2022 meta-analysis by Wang et al. of 26 studies with 62,338 patients who were suspected of having sepsis. The pooled analysis revealed that qSOFA had a high specificity (82%) but moderate sensitivity (46%), while NEWS had a moderate specificity (52%) and high sensitivity (73%). The results of the meta-analysis were slightly more accurate overall for the prediction of death with qSOFA, but they do confirm that the important clinical characteristic seen in our study was the relative specificity of qSOFA, which may underestimate the number of patients at risk of death, while NEWS can detect a larger number of high-risk patients.[10] Another Italian retrospective study from 2022 showed that NEWS and qSOFA better predicted in-hospital mortality in a cohort of 556 patients with sepsis and septic shock than qSOFA alone. NEWS comparison with other evaluated tools showed better performance for all type of infectious etiology, including Respiratory, Urinary and abdominal infections. These findings are significant for our study as respiratory and urinary infections were the two most common sites of infection in our population, and the physiological assessment which formed part of NEWS2 may have prognostic significance for a wide range of infection sources.[17] Our findings also matched those of another general-ward study conducted in 2021 that compared several early warning scores for identifying sepsis. NEWS2 achieved the best discrimination (AUROC 0.803) and the best recall, while qSOFA achieved the best specificity. Therefore, although this study investigated primarily for early susceptibility to sepsis, not death, the results that NEWS2 was more sensitive and qSOFA more specific are very concordant with the diagnosis pattern shown in our study.[18] In a 2023 prospective multicenter Spanish study, 535 patients with suspected infection who were evaluated in the prehospital setting were evaluated for NEWS2, qSOFA, and modified SOFA. NEWS2 had an AUROC of 0.761 vs 0.731 for qSOFA, while modified SOFA performed best of all. The results suggest that NEWS2 discriminated better than qSOFA, even prior to admission to hospital. These differences in AUROC could be due to differences in patient selection, when the patients were assessed for scoring, disease severity, and our study specifically included patients with established sepsis rather than all patients with suspected infection.[19] An additional systematic review and meta-analysis of 57 studies in 2023 revealed significant variations between the scores. Both qSOFA and NEWS had high specificity and sensitivity for sepsis prediction.qSOFA had high specificity and low sensitivity; NEWS had high sensitivity and good specificity. However, for mortality outcomes, there was some variation in performance based on the definition and timing of mortality. This variability accounts for the fact that scores have not consistently been found to be superior in one study over another, and highlights the importance of the clinical context and the clinical outcome being predicted. However, our findings corroborate the increasing data that suggest NEWS assessment is more valid when the goal of the assessment is to identify patients at risk of deterioration and death.[20] This has been further supported by more recent evidence, which has shown the benefit of NEWS2. NEWS2 performed significantly better than qSOFA and the other scores in the 2024 emergency-department study that compared 11 early warning scores and three shock indices, showing the highest point estimate for predicting subsequent sepsis, with an AUROC of 0.75. The sensitivity of NEWS2 ≥5, however, was only 45%, meaning that NEWS2 should not be used independently as a screening test. This observation is significant as our study showed significantly higher sensitivity for predicting mortality, which may indicate that NEWS2 performance is better in patients already diagnosed with sepsis than in an undifferentiated emergency population.[21] An observational study conducted in Colombia in 2025 specifically focused on mortality prediction for sepsis or septic shock patients with cardiovascular disease (CVD) or neurovascular disease (NVD). NEWS2 performed significantly better in terms of AUROC (0.80 vs 0.5868 for qSOFA), and NEWS2 ≥8 was a significant predictor of mortality. The magnitude of the superiority was larger than seen in our cohort but in the same direction. The strong ability of NEWS2 to predict outcome is maintained in clinically more challenging patients with significant comorbidity, as these analyses suggest.[22] These findings have clinical significance. Most screening tools used at the bedside should not require a complex calculation or laboratory investigations to determine if an individual is at high risk. Our results indicate that NEWS2 might be the better option for initial risk stratification and escalation of monitoring, while qSOFA might offer additional information about the risk for severe physiological compromise. It is important to note that neither score is to be used alone. There is growing evidence in the modern era that a structured early warning score should be used in addition to clinical assessment, infection evaluation, organ-function assessment, and appropriate laboratory investigations rather than a single 'bedside' score. In summary, the present study contributes to the growing evidence that NEWS2 is more sensitive and discriminates better than qSOFA for predicting mortality in adult patients with sepsis, while qSOFA has an advantage in terms of specificity. The AUROC of 0.823 for NEWS2 suggests good discrimination and implies that the use of NEWS2 as a broader physiological assessment could be beneficial in identifying patients at higher risk of death. However, the variability found in published studies makes it essential to validate the choice of threshold or score locally prior to using a specific threshold or score as a standalone decision-making tool. LIMITATIONS This study had several limitations. First, it was carried out in a single hospital and a fairly small sample; this may not help to generalize the results to other hospitals and populations. Second, the use of consecutive sampling might have resulted in selection bias. Third, the serially calculated scores were not assessed, meaning the value of changing scores in the hospital was not assessed; thus, the prognostic value of serial scoring could not be evaluated. Fourth, there was inadequate adjustment for potential confounding factors (infection source, comorbidities, treatment, and time of antimicrobial therapy). Lastly, only in-hospital mortality was reviewed and longer-term mortality following discharge (28- or 90-day) could not be validated.

The study demonstrated that NEWS2 had superior overall discrimination and sensitivity compared with qSOFA for predicting in-hospital mortality in patients with sepsis. The results indicate that NEWS2 could effectively be used as a preliminary bedside risk stratification tool to alert clinicians to patients who need closer surveillance and prompt escalation of care. These scores, however, should not be relied upon solely in conjunction with clinical evaluation, and additional multicenter trials with multiple scores and longitudinal measurements are warranted.