Table 2 Post emergency care outcomes

 Mortality (n = 2)

Studies of ED trigger systems had conflicting results:

  • In-hospital deaths (n = 1) [19]

• no difference in 30-day mortality after implementation of ED triggers [49], and

  • 30 days (n = 1) [49]

• patients who fulfilled ED CIC during ED care were more likely to die in hospital than those who did not fulfil trigger criteria (6.3% vs 1.4%, p = 0.044) [19].

 ICU admission (n = 1)

Implementation of ED triggers made no significant difference to days spent in special care units (ICU or intermediate care) [49].

  • ICU cohort with no comparator [49]

 Hospital admission (n = 10)

Patients who fulfilled ED CIC during ED care were more likely to be admitted to hospital (72.7% vs 40.8%, p < 0.001) [19].

  • Hospital admission vs discharge from ED [19]

 Length of stay (n = 1)

The introduction of ED triggers had no significant impact on hospital stay [49].

  • Hospital length of stay [49]

 Mortality (n = 22)

The majority of studies of aggregate scoring systems showed that patients who died had higher scores than those who survived and this was the case for MEWS [34, 35, 37, 39, 42, 45, 48, 53], NEWS [32, 36, 40, 56, 61], NEWS2 [61], NEWS-C [61], qSOFA [33, 61], REMS [42, 61], GAP [39], BEWS [54], RTS [60] and Novara-COVID scores [58]. One study showed that MEWS progression (increasing scores) was associated with significantly higher 7- and 30-day mortality [59]. One study of a single trigger system applied at repeated intervals during ED care but with no specific response to deterioration reported showed that more patients who died in hospital fulfilled hospital MET criteria (applied in ED) compared to those who survived [55]. MEWS and RTS were good predictors of 24-h mortality, but MEWS had better predictive efficacy than RTS [60]. Most aggregate scoring systems were good or excellent predictors of 2-day in-hospital mortality but less predictive of in-hospital mortality at 7 and 28-days [57]. One study showed that REMS had the highest overall accuracy for 48-h and 7-day mortality compared to MEWs, NEWS, NEWS2, NEWS-C, and qSOFA [61].

  • In-hospital deaths (n = 9) [32, 34, 35, 42, 45, 55, 56, 58]

  • 24 h (n = 1) [60]

  • 48 h (n = 3) [54, 57, 61]

  • 5 days (n = 3) [33, 36, 37], 7 days (n = 3) [57, 59, 61]

  • 14 days (n = 1) [33]

  • 28 days (n = 3) [33, 39, 57], 30 days (n = 7) [32, 36, 40, 48, 49, 53, 59]

  • 90 days (n = 1) [36]

  • Not reported (n = 1) [47]

 ICU admission (n = 14)

Patients admitted to ICU had higher aggregate scores than those not admitted to ICU and this was the case for MEWS [34, 35, 37, 42, 45], NEWS [32, 40], PEWS on ED arrival [43] and on ED discharge [41, 43], BEWS [54] and REMS [42] on ED arrival; NEWS at 1-h of ED care [40] and NEWS at ED discharge [40]. The majority (52.5%) of patients admitted to ICU had MEWS = 2–3 and no patient with MEWS ≥7 was admitted to ICU [52]. There was no significant difference in the proportion of patients admitted to ICU with MEWS≥4 versus MEWS < 4 [53]. Most aggregate scoring systems were poor predictors of need for ICU admission from wards within 2-days of hospital admission [57]. One study showed that NEWS had the highest overall accuracy for predicting ICU admission at both 48 h and 3 days, compared to MEWs, NEWS2, NEWS-C, qSOFa, and REMS [61].

  • ICU admission compared to no ICU admission with not (n = 6) [34, 35, 40, 45, 52–54]

  • ICU admission compared to ward admission (n = 5) [32, 37, 41–43]

  • ICU admission compared to discharge from ED (n = 2) [37, 42]

  • ICU cohort with no comparator (n = 2) [52]

  • ICU admission from wards < 2 days of admission (n = 1) [57]

  • ICU admission within 48 h and 7 days of ED arrival [61]

 Hospital admission (n = 10)

Patients requiring hospital admission tended to have higher aggregate scores than those not admitted and this was the case for MEWS [35, 37, 48], NEWS [36, 40], VIEWS [37], PEWS [41, 51], on ED arrival; and EWS at 15 min and 30 min of ED care [38]. One study reported that patients admitted to hospital had significantly lower MEWS and REMS scores than patients discharged from ED or admitted to ICU [42]. One study reported no difference in the odds of hospital admission between patients with MEWS ≥4 versus MEWS < 4 [45].

  • Hospital admission vs discharge from ED (n = 9) [35–38, 40–42, 45, 48, 51]

  • Hospital admission vs died in ED (n = 2) [37, 42]

  • Hospital admission vs ICU admission (n = 2) [37, 42]

 Clinical stability (n = 3)

Significantly lower levels of clinical stability occurred in patients with Novara-COVID scores of 3 (OR = 0.28, 95%CI 0.13–0.59) or 4–5 (OR = 0.03, 95%CI 0.006–0.12) [58]. One study showed that MEWS progression (increasing scores) was associated with significantly higher 24-h APACHE-II scores [59]. One study showed that severely injured patients had significantly higher MEWS and MEWS-A scores on ED arrival than less severely injured patients [62]. MEWS-A had greater predictive value than MEWS in identifying severely injured patients [62].

  • No transfer to higher intensity of care (low to intermediate) and no in-hospital death (n = 1) [58]

  • Initial and 24-h ICU APACHE-II scores (n = 1) [59]

  • Injury severity (n = 1) [62]

 Length of stay (n = 2)

The hospital length of stay studies had conflicting results:

  • Hospital length of stay (n = 2) [40, 55]

• NEWS ≥7 on ED arrival was associated with longer hospital length of stay [40], and

  • ICU length of stay (n = 1) [55]

• no significant association between MET call criteria in the ED and hospital length of stay [55].

There was no correlation between MET call criteria in the ED and ICU length of stay [55].