Studies assessing paediatric outcomes and timing of admission
| Study group | Study type | Outcome | Exposure | Key result | Comment | |
| Adil et al 5 | Patients aged 1–18 years admitted to USA hospitals between 2002 and 2011 due to ischaemic or haemorrhagic stroke. 19 386 patients. Cases found from National Inpatient Sample (NIS). | Retrospective cohort | Mortality Length of stay (LOS) Hospital charges Discharge status | Weekend or weekday admission | Ischaemic stroke Children admitted at weekends had a higher rate of discharge to nursing facilities (vs home) (OR 1.5 (95% CI 1.1 to 1.9), P=0.006). Haemorrhagic stroke Inhospital mortality was higher in children admitted at weekends (OR 1.4 (95% CI 1.1 to 1.9), P=0.04). LOS and charges were also higher in both groups in children admitted at the weekend. | Limited by reliance on correct coding of diagnoses within NIS database and under reporting of comorbidities. Adjustment was made for age, sex and confounding factors significant (P<0.05) in univariate analysis. |
| Arias et al 6 | 20 547 emergency admissions to 15 PICUs in the PICUE database in the USA between 1995 and 2001. | Retrospective cohort | Mortality within 48 hours of admission to PICU. | Weekend (19: 00 Friday to 07:00 Monday) or weekday admission Day (07:00–17:00) or evening (17:00–07:00) | No significant difference between weekend and weekday admission regarding severity of illness. Death at the weekend (OR 1.0 (95% CI 0.78 to 1.28)) Death in the evening (OR 1.28 (95% CI 1 to 1.62)) | Used PRISM scoring as method for adjusting for illness severity. |
| Arslankoylu et al 7 | 210 consecutive admissions of children aged 1 month–18 years to PICU in a Turkish hospital between 2005 and 2006. | Retrospective cohort | Mortality within 24 hours, 48 hours, 72 hours and overall. Length of stay. Duration of ventilation. | Weekday or weekend Daytime (08:00–17:00) or evening (17:00–08:00) | No statistically significant difference in mortality rates between weekend and weekday or daytime and evening. Patients admitted during the daytime had greater PIM2 mortality scores (P=0.01) and greater need for mechanical ventilation (P=0.03) | Used PIM2 scoring as a way of adjusting for illness severity. |
| Attenello et al 8 | 99 472 patients aged less than 20 years between 2000 and 2010 with hydrocephalus identified from NIS and KID databases in the USA. | Retrospective cohort | Mortality Routine discharge rate | Weekday or weekend admission Time to shunt insertion | Weekend admission was not associated with increased mortality (P=0.46) Weekend admission was not associated with a significant difference in the routine discharge rate (RR 1.00 (95% CI 0.97 to 1.03), P=0.98) | Multivariate analysis to account for confounding factors that included time to shunt insertion. Adjustment for disease severity, time to procedure and admission acuity. |
| Berkley et al 9 | 12 893 children aged ≥90 days admitted to a rural Kenyan hospital between 1998 and 2001. | Prospective cohort | Mortality within 4 hours, 4–48 hours and after 48 hours | Weekend or weekday admission | Adjusted for illness severity, weekend admission; death within 4 hours was increased (OR 2.05 (95% CI 1.26 to 3.31)) as was death within 4–48 hours (OR 1.54 (95% CI 1.17 to 2.03)) Death beyond 48 hours was not (OR 1.09 (95% CI 0.8 to 1.47)) | Created their own prognostic indicator score and used this to adjust for illness severity. |
| Desai et al 10 | 580 patients less than 18 years undergoing 710 emergency neurosurgical procedures between 2011 and 2014 in a Texas Children’s Hospital. | Retrospective cohort | Mortality Surgical complications (including infection, CSF leakage, postoperative haemorrhage, and new neurological deficit) | Weekday hours (07:20–19:00) or weekday out-of-hours or weekend (Saturday 12:00–Sunday 24:00) procedure | Combined morbidity and mortality at nights and weekends OR 1.79 (95% CI 1.083 to 2.961), P=0.0227 No significant difference in mortality (P=0.058), but only three deaths in total. No significant differences in complications between groups. | Accounted for baseline health status using anaesthetic scoring system (ASA). |
| Goldacre et al 11 | All patients with a discharge diagnosis of meningococcal disease as per HES and mortality records from 1999 to 2010. 12 697 patients aged less than 15 years. | Retrospective cohort | Mortality | Weekend or weekday admission | 30-day case fatality rates showed no evidence of increased mortality for patients admitted at weekend (P=0.92) No significant change when adjusted for confounding factors | Adjusted for sex, age, year of admission, IMD score and day of the week of admission |
| Goldstein et al 12 | 439 457 patients less than 18 years old requiring surgery abscess drainage, appendicectomy, hernia repair/VP shunt) on same day as admission using the NIS and KID databases between 1988 and 2010. | Retrospective cohort | Mortality Haemorrhage Accidental puncture or laceration Wound dehiscence Wound infection Blood transfusion | Weekend or weekday admission | Risk of death (OR 1.63 (95% CI 1.21 to 2.2)) Accidental puncture or laceration (OR 1.4 (95% CI 1.12 to 1.74)) Blood transfusion (OR 1.14 (95% CI 1.01 to 1.26)) | Looked at number of ICD9 codes associated with each patient as a way of assessing comorbidities. Adjusted for age, gender, race, insurance status, comorbid diagnoses, geographical region, type of hospital, admission type and surgical procedure |
| Gonzalez et al 13 | 176 patients aged less than 18 years undergoing extracorporeal life support (ECMO) between 2004 and 2015 in a tertiary centre in the USA | Retrospective cohort | Mortality Survival to discharge Complications including haemorrhage cannula repositioning and conversion from venovenous to venoarterial ECMO | Regular hours (07:00–19:00) or out-of-hours (including weekends) (off hours) | The overall complication rate was lower in off-hours group (45.7% versus 61.9% P=0.034) | The most common indication for ECMO were congenital diaphragmatic hernia (CDH) and persistent pulmonary haemorrhage (23%) |
| Goodman et al 14 | Children with newly diagnosed acute lymphocytic leukaemia or acute myeloid leukaemia admitted to hospital between 1999 and 2011 and identified by the Pediatric Health Information System database. 12 043 admissions in 43 USA hospitals. | Retrospective cohort | Mortality during primary admission Length of inpatient stay Time to chemotherapy Organ system failure | Weekend or weekday admission | Weekend mortality rate (OR 1.0 (95% CI 0.8 to 1.6)) If admitted at weekend: admission duration: 1.4 days longer (95% CI 0.7 to 2.1), time to chemotherapy initiation> 0.36 day longer (95% CI 0.3 to 0.5) and risk of respiratory failure: OR 1.5 (95% CI 1.2 to 1.7) | Adjusted for severity of illness at presentation defined as requiring PICU within 48 hours of hospital admission, demographics and hospital level. |
| Hixson et al 15 | 5968 patients aged 0–21 years admitted to a PICU in a single hospital in the USA between 1996 and 2003. | Prospective cohort | Mortality prior to discharge from hospital | Weekday or weekend Daytime (07:00–19:00) or evening (19:00–07:00) | Following adjustment for confounding variables, there was no significant difference in death rate at weekends or in the evenings: weekend: P=0.146 evening: P=0.711 | Used PRISM3 to assess for illness severity. |
| McCrory et al 16 | All patients less than 18 years admitted to a PICU between 2009 and 2012. 23,4192 admissions to 99 USA PICUs | Retrospective cohort | Mortality during PICU admission | Inhours (07:00–18:59 Monday–Friday) or off-hours (1900–0659 or Saturday/Sunday anytime) | Weekend death (OR 1.01 (95% CI 0.94 to 1.09), P=0.79) Night death (OR 0.86 (95% CI 0.81 to 0.92), P<0.0001) Risk of death between hours 06:00 and 11:00 OR 1.27 (95% CI 1.26 to 1.39), P≤0.0001 | Used PRISM scoring as a way of assessing severity of illness. Data adjusted for confounders including illness severity. No evidence of increased out of hours mortality. |
| McShane et al 17 | All patients admitted to 29 PICUs in England and Wales between 2006 and 2011. 86 980 patients and 4087 deaths. | Retrospective cohort | Mortality before discharge from PICU | Out-of-hours (nights, weekends and public holidays) or inhours admission | 47.1% of admissions were out of hours, with 79.2% being emergency admissions. Risk-adjusted mortality for planned admissions was higher in out-of-hours (OR 1.99 (95% CI 1.67 to 2.37), P<0.001) but not for emergency admissions (OR 0.93 (95% CI 0.86 to 1.01), P=0.07) | Adjustment PIM2, year of admission, sex ethnicity, age group and diagnostic group |
| Numa et al 18 | 4456 emergency PICU admissions of children aged 0–14 years in an Australian hospital between 1997 and 2006. | Prospective cohort | PICU mortality Length of stay | Inhours (08:00–18:00 Monday–Friday and 08:00–12:00 Saturday–Sunday) or after-hours (all other times) | Following adjustment for illness severity. After-hours risk of death (OR 0.712 (95% CI 0.518 to 0.980), P=0.037) Shorter median length of stay for after-hours versus inhours admissions. | Used paediatric index of mortality (PIM) scoring to adjust for illness severity. |
| Nwosu et al 19 | All inpatient deaths (3934) at a tertiary referral hospital in Nigeria occurring between 1998 and 2007. | Retrospective cohort | Mortality | Weekend or weekday | Weekend deaths less in SCBU (ratio 0.88), paediatric ward (ratio 0.88) and children’s ED (ratio 0.87). The labour ward had significantly higher weekend to weekday death rates of 1.72:1. | Ages ranged from a few hours to 94 years |
| Peeters et al 20 | All emergency PICU admissions for children aged less than 18 years in two hospitals in the Netherlands between 2003 and 2007. 3212 admissions | Prospective cohort | PICU mortality Duration of ventilation Length of PICU admission | Office hours (08:00–18:00 Monday–Friday) or off-hours (18:00–08:00 Monday–Friday and all weekends and bank holidays) | 66% patients admitted during off-hours. Adjusting for severity of illness, there was no significant effect of off-hours admission on mortality. PIM1 score (OR 0.95 (95% CI 0.71 to 1.27), P=0.73) PRISM2 score (OR 1.03 (95% CI 0.76 to 1.39), P=0.82) | Used PIM1 and PRISM2 scoring systems to adjust for severity of illness. |
| Sayari et al 21 | Patients admitted between 1997 and 2009 aged less than 8 days and undergoing repair of tracheo-oesophageal fistula and oesophageal atresia underwent repair. Cases found from KID US 2913 cases. | Retrospective cohort | Hospital mortality Complications Resource use Total charges | Weekend or weekday procedure | Weekend versus weekday surgical procedure had no significant effect on mortality or resource utilisation (P≥0.05). Surgical complication rates were higher in those undergoing a weekend procedure (OR 2.2 (95% CI 1.01 to 4.8), P=0.048). | Additional adjustment made for comorbid risk factors. |
| Auger et al 22 | 55 383 paediatric patients from a US hospital between 2006 and 2012. Newborns and children who died were excluded. | Retrospective cohort | Unplanned readmission within 30 days of discharge | Weekend admission or weekend discharge | Weekend admission had significantly higher odds of unplanned readmission adjusted (OR=1.09 (95% CI 1.004 to 1.18), P<0.05). Being discharged on the weekend was not associated with unplanned readmission adjusted (OR=0.97 (95% CI 0.91 to 1.03)). | ICD9 codes used to identify children with chronic complex conditions. |
| Yeung et al 23 | 28 patients between 2005 and 2010 undergoing oesophageal atresia with tracheoesophageal fistula repair at a tertiary children’s hospital Canada. | Retrospective cohort | Mortality Intraoperative complications Oesophageal complications Pneumothorax | Inhours (08:00–15:30 Monday–Friday) or after-hours (remaining time period including weekends and bank holidays) | Significant increase in oesophageal leaks in after-hours group (P=0.014). Postoperative ventilation time significantly longer in after-hours group (16.1 days vs 9.3 days), P<0.001. No significant difference in mortality. | Small sample size. |
| Miller et al 24 | Review of 140 reported paediatric medication errors occurring in a tertiary USA paediatric hospital in 2008. | Retrospective cohort | Medication dispensing errors (pharmacy and nursing) | Day or night or weekday (Monday 07:00–Friday 18:59) or weekend | Weekday day nursing shift 1.17 errors/1000 doses versus weekend 2.55 errors/1000 doses (P=0.0004) or versus night 2.12 errors/1000 doses dispensed (P=0.005). No significant association between time and severity of the error. | Administration errors were most common. |
| Thompson et al 25 | Children less than 2 years of age with a primary admission diagnosis of failure to thrive between 2003 and 2011 in 42 US hospitals. 23 332 cases from the PHIS database. | Retrospective cohort | Length of hospital stay Cost of admission | Weekend or weekday admission | Patients admitted at the weekend had longer LOS incident rate ratio (IRR) (1.2 (95% CI 1.18 to 1.22), P<0.002). Weekend admissions cost more and had an increased LOS of 1.93 days | Results adjusted for the number of discharge diagnoses. |
| Fendler et al 26 | 10 042 patients admitted to a ward treating haemato-oncology, diabetic and nephrology patients in a children’s hospital in Poland between 2000 and 2010. | Retrospective cohort | Length of hospital stay | Weekend or weekday admission | Weekend admission was associated with a longer duration of hospital stay by 1.86 days (95% CI 1.6 to 2.13). | Controlled for other clinical factors. |
| Johnson et al 27 | Patients with a diagnosis of hypoplastic left heart syndrome or other single right ventricle anomalies and a planned Norwood procedure. | Retrospective cohort | Mortality Hospital length of stay Transplant-free survival Intensive care unit length of stay Days of mechanical ventilation | Weekend or weekday admission and day of procedure | No difference in death (P=0.9), hospital (P=0.7) or ICU (P=0.5) length of stay, days ventilated (P=0.3) between weekend and weekday groups. | Known risk factors for poor post-Norwood procedure outcomes were controlled for. |
CSF, cerebrospinal fluid; HES, hospital episode statistics; ICD9, international classification of diseases; IMD, index of multiple deprivation score; KID, Kids’ Inpatient Database; PHIS, paediatric health information system; PICU, paediatric intensive care unit; PRISM, paediatric risk of mortality score; RR, relative risk; SCBU, special care baby unit; VP, ventriculoperitoneal shunt .